摘要:

THE DERBYSHIRE SILICA FIREBRICK CO., LTD.FRIDEN - HARTINGTON - Nr. BUXTONManufacturers of theHIGHEST GRADE REFRACTORIESFOR THECHEMICAL, CARBONIZING AND STEELINDUSTRIESand WATER TUBE BOILERSGas-fired wire patenting Furnace built with D.S.F. Bricks and Insulated with” Dome Insulating BricksU S E” DOME ” BRAND INSULATING BRICKSManufactured by the D.S.F. Co., Ltd.FOR FUEL CONSERVATIONTelegrams : Telephone :“ Silica,” Friden, Hartington Hartington 230Bacteriological Peptone(Evans)A new product developed in the research laboratories o fTHE EVANS BIOLOGICAL INSTITUTE, and released tomeet an urgent demand from bacteriologists and mediamakers for a reliable British-made Peptone.Bacteriological Peptone (Evans) is a distinctive producthaving a variety of applications and exhibiting the followingfeatures:7 It i s free from contaminating organisms and toxic7 It contains no fermentable carbohydrate.7 Unlike many of the older peptones it provides insaline solution ALONE a complete media for awide variety of bacteria.fT With the addition of dextrose it produces anexcellent growth of pneumococci, haemolyticstreptococci, and organisms of the coli-typhoidand welchii groups.7 No serum need be added when it i s used fortyping streptococci by the fermentative reactions.7 It conforms t o the requirements of the lndol test.11 Potent diphtheria, tetanus and other toxins canbe obtained with media containing this peptone.by-products.Bacteriological Peptone (Evans) i s issued in 5 and 8-oz.amber bottleswith bakelite screw caps.Prices and an illustrated brochure will be sent on application toHome Medical Dept., Hanover Street, Liverpool.Made in England atThe Evans Biological Institute byEvans Sons Lescher & Webb Ltd.Liverpool and London1Owing to the large aemana lor uovernrnenr purposes, w e die d C ~ I - G I I L E ; I C Q L I Yrestricted as regards the purpose for which these steels can be supplied11~ SOUTH-WEST ESSEX TECHNICALCOLLEGE AND SCHOOL OF ART(Essex Education Committee)FOREST ROAD, WALTHAMSTOW, E.17H. LOWERY, M.Ed., Ph.D., D.Sc., F.1nst.P.SCIENCE DEPARTMENTR. W. JUKES, BSc., F.C.S.J. BOR, M.Sc.Tech., A.1nst.P.W. B. CROW, Ph.D., D.Sc.Principal :Head of Department :Senior Lecturer in Physics :Senior Lecturer in Biofogy:Full-time Day Courses extending over 3 or 4 yearsare held in preparation for the London UniversityB.Sc.General and Special Degrees in Chemistry,Physics, Botany, Zoology and Mathematics; also forA.I.C. and A.1nst.P.~ Many research investigations are in progress in' the Science Department and students may becomeassociated with any of these with a view to takingeither the M.Sc. or Ph.D. Degree.Part-time Degree and National Certificate Coursesare also available in Science subjects.The College was opened in 1938 and replaces the formerLeyton and Walthamstow Technical Colleges.The building is entirely new and includes an assembly hallseating 1200 people; a full-size swimming bath; two gymnasia;classrooms; lecture theatre and smaller lecture rooms; fullcomplement of science laboratories !biology, chemistry,physics, photometry, mathematics and mechanics) ; engineeringworkshops ; building science laboratory ; architectural studios ;geography, typewriting, bookkeeping, retail trades, com-modities, speech-training, music practice, cookery, dress-making and cookery demonstration rooms; art studios;common rooms for staff and students; refectory; library.Full particulars and Prospectus of classes may be obtainedoit application to the PrincipalHomogeneous Lead Lined VesselsThe Oxley process of homogeneous lead coatingprotects any form of chemical vessel with a coating,inside or out, of chemically pure lead.Operators are specially trained for this work, anda strict system of supervision and inspection giveperfect confidence that the finished work will standboth pressure and vacuum tests of the utmost severity.W e undertake the manufacture of all kinds of steelvessels, welded or rivetted, and their lining by thehomogeneous or ordinary lead lining process. Alsoall chemical lead work, including coils, etc.HUNSLET, LEEDS, 10London Office : Winchester House Old Broad Street E.C.1'Phones: 27468 (3 lines) 'Grams: '' OXBROS," Leeds, 10V rTHE POLYTECHNICRegent Street, W.1DEPARTMENT OF CHEMISTRY & BIOLOGY(Evacuated to the Storey Technical College,Lancas t er )Head of Department: H.LAMBOURNE, M.A., M.Sc., F.I.C.DAY COURSESB.Sc. Degree Special and General (External), LondonAssociateship of the Institute of Chemistry (A.I.C.)First Medical, Pre-Medical and Preliminary ScientificUniversity.Diploma.Courses in Chemistry, Biology and Physics.Prospectus free on a4plication to theDirector of EducationUNIVERSITY OF ST. ANDREWS (SCOTLAND)Chancellor: The Right Hon.the EARL BALDWIN OF BEWDLEYVice-Chancellor and Principal: SIR JAMES COLQUHOUN IRVINERectw: Air Vice-Marshal SIR DAVID MUNRODean of the Faculty of Science: Professor ROBERT JAMES. DOUGLAS GRAHAhlThe University confers the following Degrees, all open to men or women :BSc. (Ord. and Hons.). Ph.D., D.Sc.S@SlON 1941-42 opens 8th October, 1941. The curriculum for PureScience may be taken in St. Andrews or Dundee, the curriculum forEngineering in Dundee.Residential.Entrance Scholarships for Men and Entrance BursarlnCompetitions+une annually. Entries due 2nd May. United College, St.Andrews.-Residential Entrance Scholarships-Nine of f100 pcr annum. tenablefor 3 or 4 years. Bursaries open t o Science Students-Four of €50. one of €40.three of $30. seven of lesser amounts from S.25 to f13. tenablc for 3 or 4 yeors.University College, Dundee.-Bursaries open to Science Students.-Ten of from250 t3 f25. tenable for 3 or 4 years.Preliminary Examinations.-March and September. Entries due 6th Februaryand 6th August.Post Graduate Study and Research in Chemistry. Mathematics, Astronomy.Natural Philosophy, Zoology. Botany, Geology, Anatomy, Physiology, Engin-eering. Students of theUniversity are eligible for Carnegie Trust SchoIarships and Fellowships andfor other Scholarships.Fellowships and Grants.Matheson Chemistry Scholarships and Bursaries.-A Scholarship of LlOO forentrant students for three years and a Bursary of €30 for students entering theHonours Chemistry Class, for one year, tenable a t United College. S t . Andrews.Chemistry Research Students shouId communicate with Professor Read, UnitedCollege, or Professor Wynne-Jones. University College, Dundee.Residence Halls for Men Students a t St. Andrews.-St. Salvator’s Hall, DeansCourt, Swallowgate.Full information may be got from The Secretary of the University,71 North Street. St. Andrews, or the Dean of the Faculty of Science.Botany Department. Bute Medical Buildings, St.Andrews.Provision made for duly qualified Research Workers.v. . . Good “Service” isevidently the aim of thischarming tennis playerand Good Service is thekeynote of the ProdoriteOrganisation-a servicewhich is the result ofyears of research andpractical experience. Abrief idea of our field ofactivity is given belowbut we should bepleased to give youfull particulars of ourproducts on request orour specially trainedTechnical Staff will bepleased to give adviceon any problem youmay put up to us.~Acid-proof Concrete Cement Prodor Acid-proof CementAcid-proof Tank Linings De-scaling Tanks for MetalIndustry Bulk Storage of Acids Acid-proof DrainChannels Acid-proof Mastics, “Prodorkitt ” and Prodor-phalte “ Lithcote Acid-proof Linings for the Chemical,Brewery and Food trades “Coldkitt” Elastic Compound“Prodoraqua” Liquid Hardener and Waterproofer forConcrete ”Prodordur ” Special Cement Floor SurfacesHard wearing and Decorative Floor Tiles “Prodorglaze”Wall Finish.EAGLE WORKS,WEDNESBURYPhone :WEDnesbury 0284(Private Branch Exchange)~~,RTILLERY HOUSEARTILLERY ROWLONDON, S.W.1Phone :Abbey 1547 & 1548viTHE INSTITUTE OF CHEMISTRYOF GREAT BRITAIN AND IRELANDThe Institute of Chemistry was established in 1877 to providethe Government and the public with the means of recognis-ing those who have been properly trained and proved to becompetent to practise chemistry as a profession.In 1885 theInstitute was granted a Royal Charter with authority to grantcertificates of competency, and to register persons qualifiedto practise.The aims of the Institute include the elevationof the profession of chemistry and the maintenance of theefficiency, integrity and usefulness of persons practising thesame, by compelling the observance of strict rules ofmembership, and by setting up a high standard of scientificand practical efficiency.Particulars of the Regulations and Examinations of theInstitute can be obtained (free) on application.All communications to be addressed to the Registrar,-THE INSTITUTE OF CHEMISTRY30 Russell Square, W.C.1CHEMICAL SOCIETYMEMORIAL LECTURESVOLUME 111, 1914-1932Price 6s. Od., postage 7d.CONTENTSTHE EMIL FISCHER MEMORIAL LECTURE.Delivered October 28, 1920THE BAEYER MEMORIAL LECTURE.Deltvered May 10, 1923THE VAN DER WAALS MEMORIAL LECTURE.Delivered November 8, 1923THE KAMERLINGH ONNES MEMORIAL LECTURE.By ERNEST COHEN,Delivsred February 19, rgz7THE ARRHENIUS MEMORIAL LECTURE.Delivered May 10, 1928THE THEODORE WIL.LIAM RICHARDS MEMORIAL LECTURE. By Sir HAROLDDelzvered APrzl 25, 1929THE OTTO WALLACH MEMORIAL LECTURE. By LFOPOLD RUZICKA.Delivsred Marcn 10, 1932By M. 0. FORSTER, F.R.S.By W. H. PERKIN, F.R.S.By J. H. JEANS, F.R.S.For. Mem. R.S.By Sir JAMES WALKER, F.R.S.HARTLEY, C.B.E., F.R.S.Publishers :THE CHEMICAL SOCIETY, BURLINGTON HOUSE,PICCADILLY, LONDON, W.1....VllBISAcetic Acid Acetic Anh dAcetone Amy1 AcetatedrideB u ty 1 Ace t a t e Buy1 AlcoholButyraldehyde CrotonaldetydeDiacetone Alcohol Ethyl AcetateLactates Oxalates ParaldehydePhthalates StearatesT h y are all manufactured in England !yBRITISH INDUSTRIAL SOLVENTS LIMITEDWartime Address :GREAT BURGH, EPSOM, SURREYTelephone : Burgh Heath 3470-6iCHEMICAL PLANTREBUILT & GUARANTEEDBYBARBERS’ENSURESCOMPLETESATISFACTIONOur comprehensive stock at Hayesincludes-STILLS fabricated in various metalsMIXERS & KNEADERS,STEAM JACKETED MIXERS & PANS,CALORIFIERS, CONDENSERS,FRACTIONATING COLUMNS,HYDRO EXTRACTORS,VACUUM OVENS & PUMPSFILTER PRESSES all types & sizes,DEPHLEGMATORS & HEATEXCHANGERS,GLASS LINED EQUIPMENT,AUTOCLAVES all types.KESTNER FILM DRYERS,VACUUM DISTILLATION PLANT,TRIPLE ROLL MILLS,DISINTEGRATORS & BALL MILLS,SCREENING PLANT,CRYSTALLISERS, EVAPORATORS,CONFECTIONERY PLANT,GAS PRODUCERS, ACID PUMPS,DUST EXTRACTORS,MILLING & GRINDING PLANT,ALUMINIUM & COPPER VESSELS,SULPHONATORS, CENTRIFUGES,VULCANISERS.PAINT PLANT, DYEING PLANT,ETC., ETC.WE HOLD COMPREHENSIVE STOCKSOF REBUILT & GUARANTEEDCHEMICAL PLANT BY LEADINGMAKERS.IMMEDIATE DELIVERY OF PLANTREQUIRED BY THE CHEMICAL &ALLIED TRADESC.BARBER LTD.CHEMl CAL E N GIN E ERSandMACHINERY MERCHAN JSSILVERDALE GARDENSHAYES, MIDDX.Tekfihone : Telegrams :Hayes, Middx. “ Barchem,”73516 Hayes, Middx.JAHN ”RAIL END”ENGINEERING WORKSEVAPORATORSIn double and quadruple effect, plusThermo Re-compression, the world’smost economic system.Jahn’s Britishand U.S. Patents. Numerous large in-stallations concentrating various tradeliquors operating in home and overseatropical factories. Steam jet producesVacuum, i t s exhaust heats. Air PumpPower at no cost !FOR BEEF EXTRACT,BONE, GELATINE,GLUCOSEFor Oil Deodorising, Distilling, OilHardening with Hydrogen andCatalyst“FARMA” MILKCONDENSING PLANTMACH I NERYfor starch, from cassava-tapiocaroots, potatoes, etc., for cocoa,sweets, chemicals. Fillingand Packaging.and of~ ~~~~Telegrams: BELLAMY, PHONE, LONDONTelephone: EAST 1892JOHNBELLAMYLTD.43 BYNG ST., MILLWALL, E.14TANKS for OIL,PETROL,&c.MIXING& BLENDING Plant,PRtSScJRE and VACUUMVessels,AIR RECEIVERS, AUTOCLAVES, etc.,STORAGE BUNKERS andHOPPERS,STEEL CHIMNEYS, FLUESand DUCTS‘ LYTE-WATE” ROADWAGON TANKSPURE HYDROGENPure Hydrogen (9995%) for hydrogenationand metallurgical processes, together withoxygen (99’80/0), are produced direct from thecell without the need for further treatment.SIMPLY PRODUCEDThe plant is arranged so that every part ofthe process is simplified and safe, withoutcomplication or the need for specially skilledlabour.AT LOWCOSTAs a result, cost of operation and maintenanceare particularIy low, renewals are at a minimum,installations show a very long life.Plants ofall capacities have been supplied to mostcountries of the world.THE INTERNATIONALELECTROLYTIC PLANT CO. LTD.SANDYCROFT * CHESTERxBLACKIEA Textbook of Organic ChemistryA.BERNTHSEN, Ph.D.J. J. SUDBOROUGH, Ph.D., D.Sc., F.I.C.NEW EDITION REVISED BY - Honorary Fellow of, and formerly Professor ofOrganic Chemistry in the Indian Institute of Science, Bangalore.Large crown ~vo., xii + 1371 pp. 17s. 6d. net.A completely revised edition of this standard textbook for Honours and Researchstudents, with extensive additions on electronics, biochemistry, synthetic processes,and stereochemistry.Electrochemistry andElectrochemical AnalysisBy H. J. S. SAND, Ph.D., D-Sc., F.T.C., lately Head of theDepartment of Inorganic and Physical Chemistry, and Lecturer onElectrolytic Analysis, a t the Sir John Cass Technical Institute,London.Vol. 1. Electrochemical Theory.4s. 6d. net.Vol I I. Gravimetric Electrolytic Analysis and ElectrolyticA Theoretical and Practical Treatise for Students and Analysts, by an ackuowledgedBritish authority on the subject.Marsh Tests. 6s. net.Systematic Inorganic ChemistryFrom the Standpoint of the Periodic LawBy R. M. CAVEN, D.Sc.(Lond.), F.I.C., and G. D. LANDER, D.Sc.Sixth edition, revised and enlarged by A. B. CRAWFORD, B.Sc.,Ph.D.(Glas.), A.R.T.C., A.I.C., Lecturer in Analytical Chemistry inthe Royal Technical College, Glasgow. 10s. net.Principles of Organic ChemistryBy H. P. STARCK, M.A.(Cantab.), Head of the Science Depart-ment, The Technical College, Kingston-on-Thames. 12s. 6d. net.A textbook of Organic Chemistry, both theoretical and practical, for students readingfor medical, pharmaceutioal, and general examinations in chemistry.The treatment isdetailed and a large number of examination questions are included.Quantitative Chemical Analysisand Inorganic PreparationsBy R. M. CAVEN, D.Sc.(Lond.), F.T.C. 9s. net. Also in two parts.Part I.-Preparation of Inorganic Salts, and Simple Exercisesin Gravimetric and Volumetric Analysis. 4s. net.Part I I.-Volumetric Analysis, Gravimetric Separations,Analysis of Minerals and Alloys, Preparation of InorganicCompounds. 5s. 6d. net.-50 OLD BAILEY, LONDON, E.C.-xiDERBYSHIRE STONE LIMITEDControlling 12 large Quarries inDerbyshire and North StaffordshireoffersCarboniferous Limestoneof high CaCO, content forGENERAL CHEMICAL ANDMETALLURGICAL PROCESSES- and -Limestone Powderfor most ‘ I FILLING ” REQUIREMENTSBANK HOUSE, MATLOCK, DERBYSHIRETelephone: Matlock 396 Telegrams: Derbystone, MatlockACID RESISTING CEMENTSPEClA LISTS INFOR ACID AND ALKALI PROCESSESTHE LINING OF HIGH-PRESSURE DIGESTERSOur Linings withstand Direct Acid Proof Floors to with-Steam and Instant Cooling stand abrasion from Heavywithout Fracturing.loads, etc.Cements, Bricks, Tiles, etc., for Our Products can be applied toAcid Storage Tanks, Pickling vessels of Iron, Wood, Brickwork,Vats, Bleaching Cisterns, etc. Concrete or Stone, etc.Used by the leading Chemical and Allied TradesF. HAWORTH LTD.PEEL BROW - RAMSBOTTOM - LANCS.TELEPHONE: RAMSBOTTOM 3242xiiCHEMICAL SOCIETYMEMORIAL LECTURESVOLUME I, 1893-1900 (Reproduced by a photolithographic process)Price 10s.6d., postage 7d.CONTENTSTHE STAS MEMORIAL LECTURE.THE KOPP MEMORIAL LECTURE.By J. W. MALLETT, F.R.S. With an additionalDelivered December r j , 1892Delivered February 20, 1493By the Kt. Hon. Lord PLAYPAIR, G.C.B.,F.R.S.; Sir F. A. ABEL, Bart., K.C.B., F.R.S.; W. H. PERKIN, Ph.D., D.C.L., F.R.S.;H. E. ARMSTRONG. Delivered May 5, 1893THE HELMHOLTZ MEMORIAL LECTURE. By G. A. FITZGERALD, M..4., DSc.,F.R.S. Delicered January 23, 1896THE LOTHAR MEYER MEMORIAL LECTURE.F.J.C. Delivered May 28, 1896THE PASTEUR MEMORIAL LECTURE. By P. FRANKLAND,, Ph.D., B.Sc., F.R.S.Delavered Mnrch 25, I897THE KEKULE MEMORIAL LECTURE. By F. R. JAW, F.R.S. 'Delivered December I 5, 189 7THE VICTOR MEYER MEMORIAL LECTURE.By T. E. THORPE, Ph.D., D.Sc.,Delivered February 8 , 1900THE BUNSEN MEMORIAL LECTURE. By Sir H. E. ROSCOE, B.A., Ph.D., D.C.L.,Delivered March 29, 1900Facsimile Letter of Stas.By T. E. THORPE, D.Sc., F.R.S.THE MARIGNAC MEMORIAL LECTURE;. By P. T. CLEVE. I895THE HOFMANN MEMORIAL LECTURE.By P. P. BEDSON, M.A., D.Sc.,LL.D., F.R.S.LL.D., D.Sc., F.R.S.THE FRIEDEL MEMORIAL LECTURE. By J. M. CRAFTS. 1900THE NILSON MEMORIAL LECTURE.VOLUME 11,1901-1913 (Reproduced by aphotolithographic process)By 0. PETTERSSON.Delivered July 5, 1900Price 8s. Od., postage 7d.CONTENTSTHE RAMMELSBERG MEMORIAL LECTURE.THE RAOULT MEMORIAL LECTURE.THE WISLICENUS MEMORIAL LECTURE.By Sir HENRY A. MIERS, F.R.S.Delivered December 13, 1900Delivered March 26, 1902Delivered January 25, 190jDelzvered June 21, 1906THE WOLCOTT GIBBS MEMORIAL LECTURE.By F. WIGGLESWORTH CLARKE.Delivered June 3 , 1909Delivered October 21, rgogF.R.S. Delivered February 17, igroDelivered November 23, 191 IBy Sir WILLIAM RAMSAY, K.C.B., F.R.S.Delivered February 29, 1g12Delivered June 26, 1912Delivered October 17, 1912Delivered May 22, 1913Delivered October 23, 1913By J. H. VAN'T HOFP, F.R.S.By W. H. PERKIN, Jun., F.R.S.THE CLEVE MEMORIAL LECTURE. By Sir THObfAS EDWARD THORPE, C.B., F.R.S.THE MENDELEEFF MEMORIAL LECTURE.THE THOMSEN MEMORIAL LECTURE.By Sir WILLIAM A.:'rILDEN, F.K.S.By Sir THOMAS EDWARD THORPE, C.B.,THE BERTHELOT MEMORIAL LECTURE. By H. B. DIXON, F.R.S.THE MOISSAN MEMORIAL LECTURE.THE CANNIZZARO MEMORIAL LECTURE.By Sir WILLIAM A. TILDEN, F.R.S.THE BECQUEREL MEMORIAL LECTURE.THE VAN'T HOFF MEMORIAL LECTURE.THE LADENBURG MEMORIAL LECTURE.By Sir OLIVER LODGE, F.R.S.By JAMES WALKER, F.R.S.By F. S. KIPPIXG, F.R.S.Publishers :THE CHEMICAL SOCIETY, BURLINGTON HOUSE,PICCADILLY, LONDON, W.1.XiSCIENTIFIC BOOKSH. K. LEWIS & Co. Ltd.A very large selection of new and standard works in everyThe Department for Scientific Books, English and Foreign, is onthe first floor.Foreign Books not in stock obtained under Licence.Orders and Inquiries by Post promptly attended to.branch of Science always available. (Lift to first floor).Underground : Euston Square, Warren Street. ‘Buses : Euston Road and TottenharnCourt Road.SCIENTIFIC LENDING LIBRARYAnnual Subscription, Town or Countrg, from One GuineaThe LIBRARY is useful to SOCIETIES and INSTITUTIONS, andto those engaged on SPECIAL RESEARCH WORK, etc.TheLibrary includes all Recent and Standard Works in all branches ofMedical and General Science.Reading and Writing Room (first floor) oFen daily.New Books and New Editions are added to the Library and areavailable to Subscribers immediately on publication.Catalogue of the Library, revised to December 1937; containingthe titles of books arranged in alphabetical order under the Authors’names, with size, price and date, and an Index of Subjects with thenames of the Authors who have written on them. Price 16s. net (toSubscribers, 8s.).Bi-monthly List of New Books and New Editions is issued free to allSubscribers and Bookbuyers regularly.Detailed Prospectus on Application.Every work is the latest edition.STATIONERY DEPARTMENT : Scientif;c and General.Loose-Leaf Notebooly, Record Cards, Filing Cabinets,Slide Rules, Graph-papers, etc.SECONDzHAND BOOKS : 140 GOWER STREET.Large and oaried stock.BOO~S not in stock sought for, and reported when found.H. K. LEWIS & Coo Ltd.PUBLISHERS AND BOOKSELLERS136 COWER STREET, LONDON, W.C.1Telegrams : Telephone :‘PUBLICAVIT, WESTCENT, LONDON. EUSton 4282 ( 5 linesPURIFICATIONOFWATERFOR ALL PURPOSESBOILER FEEDPROCESS WORKTEXTILE PURPOSESTOWN SUPPLY, Etc.IS THE SPECIALITY OFJOHN THOMPSON(KENNICOTT WATER SOFTENERS) LTDHEAD OFFICE AND WORKSWO LV E RH A M PTO NEstablished 40 YearsPLACE YOUR WATER PROBLEMSBEFORE US AND AVAIL YOUR-SELF OF OURUnrivalled ExperienceAlso Manufacturers of Domestic Water SoftenersTo face title Page xv

ISSN:0365-6217    

DOI:10.1039/AR94037FP001

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

ERRATA.VOL., 1939, 36.Page Line337 22 for “Ester” read (‘Ether.”337453 14 24} for “ester” read “ether.

ISSN:0365-6217    

DOI:10.1039/AR9403700006

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

ANNUAL REPORTSON THEPROGRESS OF CHEMISTRY.RADIOACTIVITY AND SUBATOMIC PHENOMENA.INTRODUCTORY AND SUMMARY.THE year 1940 has produced no spectacular progress in nuclearphysics. The <‘ boom ” in papers about nuclear fission-thebreak-up into two roughly equal fragments of the nuclei of theheaviest elements-which followed the discovery of this remarkablephenomenon early in 1939 has almost faded out. On the otherhand, investigation of the radioactive isotopes produced by thebombardment of various elements with neutrons and with artificiallyaccelerated projectiles such as protons, deuterons, and a-particlesis being pursued in many places and with ever more powerfulmethods. Long-lived isotopes of common elements such ashydrogen, carbon, chlorine, and calcium were discovered and arelikely to gain great importance as tracers in chemical and biologicalinvestigations.Radioactive isotopes of the elements 85 and 93have been produced and have made possible the study of thechemical properties of these elements. The phenomenon of nuclearisomerismthe occurrence of nuclei with identical charge and massbut different radioactive properties-has been extensively studiedand a number of cases have been elucidated, thanks largely to theuse of the P-ray spectrograph. No less than four cases of nuclearisomerism have been found among the isotopes of indium alone,and about 24 pairs of nuclear isomers can now be regarded as wellestablished, there being some evidence for another 10.The accurate determination of the magnetic moment of freeneutrons is perhaps the most important item in the field of neutronresearch.The scattering of slow neutrons by crystals was studiedby several investigators. They found, in agreement with theory,that diffraction effects play a considerable part and cause the scat-tering from a large crystal to be much less than from the aameamount of a fine powder where diffraction is negligible. Sincesome of the older measurements of nuclear cross-sections wer8 RADIOACTIVITY AND SUBATOMIC PHENOMENA.done with crystals as scatterers, their results are likely t o needrevision.NUCLEAR FISSION AND TRANSURANIC ELEMENTS.The work of this year has brought only confirmation and someslight extensions of the knowledge gained in 1939, which was fullyreported upon last year.The number of ions produced by the fission fragments of uraniumand thorium was again the subject of several investigations.1Kanner and Barschall confirm the earlier result that the distributioncurve shows two maxima corresponding to mean kinetic energiesof 64 and 97 m.v. I n addition, they proved that a particle of onegroup is always associated with a particle of the other group.Theyused a thin layer of uranium, sputtered upon a very thin aluminiumfoil which enabled both fragments to escape into the ionisationchamber so that their total ionisation was measured. I n thisarrangement, only one maximum was observed, corresponding toan energy of 151 m.v., which agrees well with the sum of the twoenergies given above, allowance being made for an energy loss ofabout 8 m.v.in the aluminium foil. Jentschke and Prankl 1 findthat the minimum between the two maxima is less pronounced iffission is produced by fast rather than slow neutrons. This mustmean that with increasing neutron energy more symmetricalfission modes become possible, a result which is confirmed by chemicalevidence (see below).Cloud-chamber photographs of the tracks produced by fissionfragments2 show numerous “ branches ” on each track, resultingfrom close collisions between the fragments and nuclei of the gas.N. Bohr3 showed that, on account of the large mass and chargeof the fission fragments, nuclear collisions are much more importantfor them than for a-particles or protons, which lose their energyalmost exclusively by collisions with electrons.The energy lossof fission fragments was also studied by Haxel (experimentally)and by W. E. Lamb * (theoretically).Chemical analysis of the fission products of uranium and thorium,which led to the identification of more than 30 radioactive periodsin 1939, has been continued. Only a few new periods, however,were discovered, and nearly all the results summarised in lastyear’s table still hold good.1 W. Jentschke and F. Prankl, Physikal. Z . , 1939, 40, 706; 0. Haxel,2. Physik, 1939, 112, 681 ; M. H. Kanner and H. H. Barschall, Physical Rev.,1940, 57, 372.K. J. Brostrram, J. K. Berggild, and T . Lawitsen, Physical Rev., 1940,58, 651.Ibid., p. 654. Ilbid., p. 696FRISCH : NUCLEAR FISSION AND TRANSURANIC ELEMENTS.9G. N. Glasoe and J. Steigman used the method, described lastyear, of bubbling air through a uranium salt solution during orafter irradiation with neutrons, the air bubbles carrying with themall the isotopes of krypton and xenon which result from the fission.By placing a negatively charged wire in the air sample, the decayproducts of these gases were collected and subsequently subjectedto chemical separations. By comparing the initial activities result-ing from successive extractions from the same gas sample, the half-life of the respective parent gas could be determined.Apart from 88Rb, of 18 mins. period, the growth of which froma 3-hours krypton was confirmed, the authors found that a rubidiumisotope of 15-4 mins.grows from a krypton isotope with a periodof a few minutes and decays into an isotope of strontium with ahalf-life of about 51 days. This long-lived strontium has beenknown for some time and is probably 8gSr.In addition, the authors confirmed the growth of a 33-mins.cmium from a 17-mins. xenon, but found no evidence of its decayinto a long-lived barium (see last year's Report). According t otheir evidence, the 300-hours barium grows from a very short-lived caesium isotope, which is formed from an equally short-livedxenon.Two longer-lived xenon isotopes were studied in two differentplaces 6 with substantially identical results. A genetic relationwas established between two known activities : the 5-&days xenonwas found to grow from the 22-hours iodine, and a new xenonactivity, with a period of 9.4 hours, was found to grow from a 6.6-hours iodine.Both these xenon periods result also from thoriumfission, but with different relative intensity. Segrb and Wu alsoshowed that an isotope of element 43, decaying with a half-life of6.6 hours, can be extracted from the 67-hours molybdenum whichresults from uranium fission, thus confirming the identification ofthe latter as 99Mo or lolMo.A zirconium isotope of 17-hours period, decaying into a 75-mins.columbium, and some evidence for a zirconium with more than20-days period, was found by A. V. Grosse and E. T. Booth.' Thisfills the gap in the light group of fission products, so that now activeisotopes of all the elements between bromine and element 43 andbetween antimony and lanthanum have been found among theproducts of the fission of uranium with slow neutrons.The gap between the two groups seems to be real as far as fissionPhysical Rev., 1940, 58, 1.6 E.Segre and C. S. Wu, ibid., 57, 652; R. W. Dodson and R. D. Fowler,7 /bid., p. 664,ibid., p. 96610 RADIOAaTIVITY AND SUBATOMIC PHENOMENA.by slow neutrons is concerned. Active isotopes of silver, cadmium,and indium have, however, been produced by irradiation ofuranium with very fast neutrons, produced by deuteron bombard-ment of lithium. The two silver periods found, of 7.5 days and3 hours, agree with the known periods of 1llAg and 112Ag. Of thethree cadmium periods, of 50 mins., 5.5 hours, and 2.5 days, thelast two decay into active indium isotopes of 2.1 hours and 4.5hours, respectively, which is strong evidence for their identity with117Cd and 115Cd.All these activities are not produced by thermalneutrons, or even by the neutrons from beryllium, bombardedwith deuterons.It seems likely that an increase in the energy of the bombardingneutrons would extend the range of the mass ratio of the fissionfragments, not only towards smaller values, as indicated by theseexperiments (and those of Jentschke and Prankl, see above), butalso towards larger values ; i.e., elements below bromine and abovelanthanum would be produced. It seems, however, that a t presentnot even all the active substances produced in slow-neutron fissionare known.E. Fermi9 estimates that the added activities of allthe known fission products account for only about one-half of thetotal, chemically unseparated, activity of irradiated uranium(after allowance for the presence of active isotopes of uranium andelement 93, see below), and suggests that the surplus may be due inpart to unknown active isotopes of rare-earth elements.Fission of uranium under deuteron bombardment, briefly reportedlast year, was further studied by R. S. Krishnan and T. E. Banks,lowho also obtained positive results with thorium. Photo-fission,i.e., fission caused by y-rays (from fluorine, bombarded with protons),was observed by R. 0. Haxby, W. E. Shoupp, W. E. Stephens, andW. H. Wells,ll both for uranium and for thorium. The weaknessof the effect explains why earlier observers had failed to detect it.Evidence for spontaneous fission of uranium (but not of thorium)was reported by Flerov and Petrjak.12 From the brief note(cabled) it is not clear to what extent the effect-which wouldcorrespond to a mean life of 10l6 to lo1' years-was caused by theneutrons known to accompany the cosmic radiation.It was mentioned in last year's Report that N.Bohr,13 from ananalysis of the energy dependence of the fission phenomena, hadconcluded that the fission observed with neutrons of thermal energyY. Nishina, T. Yasaki, H. Ezoe, K. Kimura, and M. Ikawa, Nature, 1940,Ibid., p. 640.146, 24.l o Ibid., 145, 860.l1 Physical Rev., 1940, 58, 92. l2 Ibid., p. 89.l3 Ibid., 1939, 65, 418; N.Bohr and J. A. Wheeler, ibid., 56, 426FRISCH : NUCLEAR FISSION AND TRANSURBNIa ELEMENTS. 11was to be ascribed to the light isotope of uranium, 236U, whichconstitutes 007% of ordinary uranium. This conclusion has nowbeen confirmed by experiments with separated is0topes.1~ Samplescontaining a few micrograms of the heavy isotope, and correspond-ingly smaller samples of the light one, were prepared by means ofa mass spectrograph. When these samples were subjected to anintense bombardment with slow neutrons, fission was observed inthe light but not in the heavy isotope. It was also found that therare isotope 2MU (U-11) does not contribute appreciably to thefission effects in ordinary uranium.With regard to the emission of neutrons in uranium fission, newmeasurements 15 and a careful analysis l6 of earlier experimentsindicate that the average number of neutrons emitted per fissionis close to 3.This figure is somewhat larger than the estimate(2.3) given in last year's Report, but the situation concerning thepossibility of a nuclear chain reaction has not been materiallychanged thereby.Instead of causing fission, a neutron hitting a uranium nucleusmay be captured. It has been known for some time that theirradiation of uranium with slow neutrons produces a p-active isotopeof uranium, with a period of 24 mins., and there was strong evidencethat this was S9U rather than =W. This, too, has now been con-firmed by experiments with separated isotopes.17 In addition,the growth of an active substance of 2-3-days period from 239U hasbeen observedIs and had to be ascribed t o the formation of anactive isotope of element 93.Preliminary experiments on thechemical behaviour of this new element showed that it is precipitatedquantitatively by hydrofluoric acid in the presence of a reducingagent (sulphur dioxide), cerium being used as a, carrier. In thepresence of an oxidising agent (bromate in strong acid) it is notprecipitated. In the reduced state with a thorium carrier it isprecipitated by iodate, and in the oxidised state with uranium assodium uranyl acetate. It is also precipitated with thorium onthe addition of hydrogen peroxide, and in basic solution if carbonateis excluded. These properties indicate that its two valency statesare similar to those of uranium, the chief difference being in thevalue of the oxidation potential between the two valencies, suchthat the lower state is more stable in the new element.It hasl* A. 0. Nier, E. T. Booth, J. R. Dunning, and A. V. Grosse, PhysicalRev., 1940, 57, 546, 748; K. H. Kingdon, H. C. Pollock, E. T. Booth, andJ. R. Dunning, ibid., p. 749.l6 T. Hagiwara, Mem. Coll. Sci., Kyoto Imp. Univ., 1940, A, 23, 19.l6 L. A. Turner, Physical Rev., 1940, 57, 334.l' E. T. Booth, J. R. Dunning, A. V. Grosse, and A. 0. Nier, ibid., 58,475.E. McMillan and P. H. Abelson, &bid., 57, 118512 RADIOACTIVITY AND SUBATOMIC PHENOMENA.little if any resemblance to its lower homologue, rhenium, for itis not precipitated by hydrogen sulphide in acid solution, is notreduced to the metal by zinc in acid solution, and does not have anoxide volatile a t red heat.The latterprobably decays ‘into 235U by the emission of an a-particle.Noa-particles were, however, observed from a strong, aged sample of23993, and it is concluded that the life of 23994 must exceed 106years. Moreover, no fission fragments were observed, so spontaneousfission is improbable, too.If uranium is bombarded with very fast neutrons (from lithium + deuterons) a 8-active uranium isotope with a half-life of 7 daysis produced.lg Since slow neutrons fail to produce this activity itis attributed to 237U, resulting from 238U by a (n, 2%) (neutron-loss) reaction. In decaying it must form 23793, but no activitydue to this isotope was discovered when element 93 was extracted(see above) from an aged sample of 237U.Since 23993 emits (negative) P-rays, it must form 23994.ARTIFICIAL RADIOACTIVITY.The study of artificial radioactivity has made great strides duringthe last few years, and a comprehensive report by G.T. Seaborg,20which covers the literature up to July 15th, 1940, lists more than350 different periods, of which more than 150 can be attributed todefinite isotopes. This progress is due in part to the ever-increasingnumber of workers in this field, but mainly to the development ofexperimental resources. In the first place, the construction ofnumerous cyclotrons and their development towards greater beamintensity and higher energy must be mentioned.The most powerfulinstrument a t present is the 60-inch cyclotron a t Berkeley, whichproduces deuterons up to 16 m.v. and helium nuclei (a-particles)up to 32 m.v. energy. Particles of these energies can readilypenetrate even into the heaviest nuclei. The strong activities whichthereby become available greatly facilitate the chemical identifica-tion of the various periods and permit us to unravel genetic relation-ships by means of successive chemical separations which necessitatethe measurement of small fractions of the initial activity. Increas-ing use is being made of the P-ray spectrograph to study both thecontinuous energy spectrum of the @-rays (positive or negativeelectrons) emitted from the nucleus itself and the superimposed“lines” of electrons of sharply defined energy, which have beenl9 Y .Nishina, T. Yasaki, H. Ezoe, K. Kimura, and M. Ikawa, Physical20 Chem. Reviews, 1940, 27, 199.Rev,, 1940, 57, 1182; E. McMillan, ibid., 58, 178FRISCH : ARTIFICIAL RADIOACTIVITY. 13ejected from the K- or L-shell of the atom by the internal con-version of nuclear energy. Although two different radioactivesubstances may have the same half-life within the limits of experi-mental error, it is very unlikely that the energy spectra of theparticles which they emit should be identical as well, and the p-rayspectrograph is thus of great help in the identification (or otherwise)of activities of similar half-life.The spacing and intensity of the electron lines give informationabout excited states of the nucleus concerned, and have been par-ticularly useful in the study of isomeric nuclei (see below). I nsome @-ray spectrographs the photographic plate has been replacedby a counting tube, which in this way records single electronswithin a narrow range of energies.This arrangement is moresensitive and more suitable for quantitative intensity measure-ments than the photographic plate. By setting it for any oneelectron line, the decay of the substance emitting this line can bedetermined with little disturbance from the presence of otheractivities.Long-lived Isotopes of Common Elements.-The discovery of 3H,a p-active isotope of hydrogen, was briefly announced in last year’sReport. Evidence for a decay with a few months’ half-life wasreported but soon withdrawn again; the apparent decay had beendue to diffusion through a rubber tube.An indirect determinationof the period gave a value of about 30 years.21 A known numberof 3H nuclei was produced by exposing lithium for a, certain time toa known density of slow neutrons [reaction 6Li (n, a) 3H] and theenumber of nuclei decaying per second was determined by dissolvingthe lithium in water and introducing a known fraction of the hydro-gen evolved into a tube counter. From these data, it is possibleto calculate the time after which all the nuclei would have decayedif decay continued a t its initial rate. By multiplying this time(the mean life) by 0.69, one obtains the half-life. The p-rays of3H are exceedingly soft, their maximum energy being only 0.013m.v.; it is therefore necessary to introduce any materialcontaining 3H into the counter, in gaseous form, in order to observethe activity.Another interesting isotope is 14C, the radioactivity of which wasfirst observed by S.Ruben and M. D. Kamen.22 It had been knownfor some time that 14C is produced in various nuclear reactions suchas 14N(n, p), 13C(d, p ) , and llB(a, p ) , and from the energy balancein these reactions a p-radiation of about 0.3 m.v. energy was expected.Ruben and Kamen found that the P-rays of 14C are much softer,21 R. D. O’Neal and M. Goldhaber, Physical Rev., 1940, 58, 574.22 Ibid., 57, 54914 RADIOACTIVITY AND SUBATOMIC PHENOMENA.only about 0.1 m.v., which explains the earlier failures to defectthis activity.No decay has been observed, but from an estimsteof the number of nuclei produced by the reaction 13C(d, p)14C, andthe number decaying per time unit, the period must be at least1000 years. 14C should become very useful as a tracer in chemicaland biological work, where its long period will permit the study ofslow processes, for which the period (20 mins.) of llC is too short.The softness of the radiation of I4C, however, necessitates the use ofextremely thin-walled counters, or preferably the introduction ofthe sample into the counter) in gaseous form.A calcium isotope with a half-life of 180 days has been produced 23by the action of deuterons on calcium and of fast neutrons onscatndium, and must, in consequence, be regarded as 45Ca.Itsp-rays are not unduly soft and it should therefore be very suitableas a tracer. The 2-5-hours calcium, which had been used for suchpurposes, was found to form a radioactive scandium and musttherefore be considered as less suitable. Incidentally) the formationof the active scandium (57-mins. period, 49Sc) shows that the 2.5-hours activity is due to 49Ca, formed from the recently discovered,rare isotope 48Ca.Chlorine strongly absorbs slow neutrons and becomes radio-active with a period of 37 mins. By means of the thermo-diffusionmethod (see this vol., p. 153), J. W. Kennedy and G. T. Seaborg 24prepared samples of chlorine in which the abundance of the heavyisotope was strongly reduced, and found that the intensity of the37-mins.period was much reduced, too. This shows that the activityis to be assigned to 38Cl. It is €ar too weak, however, to accountfor all the neutrons absorbed by chlorine, and it has been assumedfor some time that most of the neutrons are captured by 35c1,forming an active 36Cl of very long period. According to unpub-lished experiments by D. C. Grahame and H. Walke (quoted byKennedy and Seaborg 24), such an isotope, with a period of morethan a year, has now been found, but no further informationregarding its properties is 8s yet available.bombarded bismuth with energetic '' artificial a-particles, )' Le.,helium which had been accelerated to an energy of 32 m.v. by meansof the Berkeley cyclotron. Bismuth seemed a particularly interest-ing target, since capture of an a-particle with subsequent emissionof one or more neutrons (which, in heavy elements, is much morelikely than the emission of protons) would lead to the formationof an isotope of element 85.The bombarded bismuth emitted 01-Element 85.-D. R. Corson, K. R. MacKenzie, and E.23 H. Walke, F. C. Thompson, and J. Holt, Physical Rev., 1940, 57, 177.24 IbicE., p. 843. 26 I b d . , 58, 672FRIScltE : BTIFICIAL RADIOACTIVITY. 15and y-rays and decayed with a half-life of 7-5 hours. The carrierof this activity was found to evaporate from the bismuth on meltingand could be collected on a, water-cooled plate.Numerous chemical experiments showed the characteristicproperties of this substance and supported the view that it is anisotope of element 85.Although it is a higher homologue of iodine,it is not precipitated by silver nitrate from a slightly nitric acidsolution, with potassium iodide as a carrier. It has quite markedmetallic properties and a closer resemblance to its neighbourpolonium than to iodine; it can, however, be separated frompolonium-for example, by precipitation with sulphur dioxide in3~-hydrochloric acid, with tellurium as a carrier. It resemblesiodine in that it is concentrated in the thyroid of guinea pigs.26Analysis of its radiation supported the assignment to this sub-stance of the atomic number 85. Two groups of a-particles areemitted, of 4.52 and 6.55 cm. range. The different intensity ofthe groups, as well as other evidence, excludes the possibility thatthe two a-particles are emitted in succession by the same nucleus,and suggests a branching decay.The longer of the two groups hasvery accurately the same range as the a-rays of actinium-C', apolonium isotope with the mass 21 1. The authors therefore assumethat the active isotope has the mass 211, formed from 209Bi bycapture of an a-particle and emission of two neutrons.' Some ofits nuclei decay into 207Bi, by emission of an a-particle of 462-cm.range, while the others decay, by capture of a K-electron and emis-sion of the corresponding X-rays (which have been identsed) intoactinium-C', which has a period of the order of lo3 sec. and turnsinto stable 207Pb with the emission of a 6.57-cm. a-ray. If thispicture is correct (which seems highly probable), then m7Bi musthave a life of at least a month, since no radiation attributable to ithas been observed.The chemical properties of element 85 me of particular interestin connexion with attempts to detect the presence of this elementin Nature.It is now clear that the methods used by E. Buch-Anderson 27 in his search for element 85 in the natural radioactivefamilies would not have led to en enrichment of 85 in the fractionsstudied by him, and the negative result of his search is thereforeno proof of the non-occurrence of the element. W. Minder28obtained evidence for a p-radiation from radium-A, which he believes26 J. G. Hamilton and M. H. Soley, Proc. Nut. Acad. Sci., 1940, 26, 483.27 K g l .Danske Vid. Selskab., Math.-Pys. Medd., 1938, 16, 6.2 8 Helu. Physica Acta, 1940, 13, 144.* The capture of an a-particle with 32 m.v. gives such a high excitationenergy to the compound nucleus that even after the " evaporation " of oneneutron enough energy is left to evaporate a second one16 RADIOACTIVITY AND SUBATOMIC PHENOMENA.to be of nuclear origin rather than secondary. p-Decay of radium-A(2 = 84) would indeed result in the formation of element 85, butno actual proof of the presence of this element is given. H. Holubeiand Y. Cauchois29 have also reported experiments which theyinterpret as evidence of the existence of element 85 in the decayproducts of radon. A theoretical discussion concerning thepossible formation and presumable radioactive properties of element85 is given in a paper by L.A. Turnerm30The discovery of a radioactive isotope of element 93 is reported,together with some of its chemical properties, in the section onnuclear fission (p. 11).Nuclear Isomerism.-The number of papers upon this phenomenonand the complexities encountered in its study are so great that acomprehensive account is out of question. Instead, it is proposed,together with a short recapitulation of the essential features, todiscuss a few representative papers and to present a, survey, in theform of a table, of our present knowledge.The term ( ( nuclear isomerism ” was introduced to describe thefact that nuclei with identical mass and charge number (Le., nucleiof the same isotope) may yet have different (radioactive) properties.The difference between isomeric nuclei cannot be assumed to liein a different arrangement of their component particles (as inisomeric molecules), since anything like a rigid structure inside anucleus seems incompatible with quantum theory. Instead, itis now generally assumed that the difference is essentially one inenergy content or, in other words, that we have to do with differentenergy states of the same nucleus.The lower one may be assumedto be the ground state of this nucleus, and the upper one a “ meta-stable ” excited state. A nucleus in the upper state can get ridof its excess energy, either by radiation (emission of a y-ray) or bygiving it to one of the extranuclear electrons (“ internal conver-sion ”), and thereby drop into the ground state.In most casesthis will happen within a very small fraction of a second. If,however, the nucleus in a particular excited state has got a momentof rotation which differs by several units * from that of the groundstate (and of any intermediate state), and if, at the same time, theenergy difference between the two states is not too large, then thetransition may be so highly “ forbidden ” that the life of the upperstate gets long enough to be observed and then we get isomerism.29 Cornpt. rend., 1939, 209, 39.30 Physical Rev., 1940, 57, 950.* One unit = h/Za; h = the quantum of action. According to quantumtheory, the moment of rotation of any system always changes by an integralnumber of these unitsFRISCH ARTIFIUIAL RADIOACTIVITY.17All cases of isomerism so far observed may be classified under oneof the following headings :(A) A metastable state decays into the stable ground state, withthe emission of electrons and/or y-rays and X-rays.(B) A metastable state decays into the ground state, which inturn undergoes p-transformation (we shall take this term as includ-ing any transformation which leads to a neighbouring element,i.e., electron or positron emission from the nucleus, and K-electroncapture).(C) Both metastable state and ground state undergo p-trans-formation.The following kinds of evidence indicate that an activity is dueto an isomeric transition (1.T.) in a stable isotope (case A) :(1) Its production by a process which can give rise to stableisotopes only, of the element in question. The production of the4.5-hours l151n * (the asterisk indicating an excited state) byirradiating indium with X-rays of 1.5 m.v., reported last year,is the classical example.More recently, a 1.6-mins. activity hasbeen excited in lead, with X-rays of the same energy.31(2) Emission of the characteristic X-rays of the element inquestion. (In the case of a P-transition, the X-rays of one of theneighbouring elements would be emitted, if any.) These X-raysmay be identified either by means of a crystal spectrograph or,much more easily, through their absorption in suitably chosenfilters.(3) Evidence from the energy spectrum of the electrons emitted.The absence of a continuous spectrum is not sufficient evidencein itself, since it might be a case of K-capture. If lines due to theejection of both K- and L-electrons are obsepved, their energydifference may be used to ascertain the atomic number of the nucleusafter the transition.From the intensity ratio of the lines it ispossible to calculate the change in rotational moment whichaccompanies the transition.L. W. Alvarez, A. C. Helmholz, and E. Nelson32 studied theelectron spectrum of the 6.7-hours cadmium isotope, which wasknown to emit silver X-rays, indicating its decay to silver. Fromthe unusually high ratio of L- to K-conversion electrons, theyconcluded that the difference in momentum between this excitedstate of silver and the ground state should be 4 units; from thisand from the energy of the state (obtained from the energy of theconversion electrons) they calculated that it should have a life ofabout 30 secs.Experiment showed this prediction to be surprisingly31 B. Waldman and G. B. Collins, Phyaical Rev., 1940, 57, 338.32 Ibid., p. 66018 RADIOACTIVITY AND SUBATOMIC PHENOMENA.accurate, for the authors were able to extract, from the 6-7-hourscadmium, an active silver fraction with a half-life of 40 secs. Inthis way the p-spectrograph not only helps to identify isomerictransitions, but may even help to discover new ones.Isomerism of either type B or type C will have to be assumedif two periods, in view of the way in which they are produced,have to be assigned to the same isotope.If the nucleus in itsground state is @-active, the isomeric state has the choice ofeither going to the ground state or undergoing @-transformation,and both processes will occur side by side. I n all cases so farknown, however, the probability of one of the two processes isfound to be negligible compared with that of the other; we aretherefore left with the two limiting cases B and C, as definedabove.Although in case B the @-rays are emitted from the ground stateonly, their decay curve is complex, since the decay of the nucleiin the ground state is partly compensated by the decay of theexcited nuclei into the ground state. The decay is, in fact, indis-tinguishable from that of a mixture of two radioactive substancesand does not even allow one to tell which of the two periods cor-responds to the I.T.and which to the p-decay. Study of the p-rayspectrum again supplies valuable information: I n case B, thecontinuous spectrum, due to the p-transition, remains the samethroughout the decay ; i€ the two periods were due to two different@-active bodies, such a behaviour would be highly improbable.In addition, one often observes conversion electrons and y-rayswhich decay with a single period, that of the isomeric state.Absorption experiments are, in general, accurate enough to findwhether the P-spectra of the two periods are identical or not, and toobserve the soft conversion electrons and/or the y-rays associatedwith one of them, while a P-spectrograph permits one, in addition,to study the exact location of the electron lines and to apply thesort of argument which we have discussed under A.If X-raysare observable (they are sometimes masked by strong y-rays)their identification as the characteristic X-rays of the element inquestion offers a simple and convincing proof of the I.T.I n several cases it has been possible, as reported last year, toseparate the nuclei in the two isomeric states by a chemical pro-cedure. This possibility exists only in case B, and depends onwhether the active element can be made part of a suitable com-pound which either breaks up or is activated by the isomeric tran-sition, so that the nuclei which have gone to the ground state maybe separated chemically from the others. The two bromine periodsof 18 mins.and 4.5 hours were the first to be separated in thiFRISCI-I : ARTIFICIAL RADIOACTIVITY. 19way.33*34 Don C. DeVault and W. F. Libby 35 more recentlyinvestigated a large number of methods for the separation of thebromine isomers. They find that about 15% of the lower isomerrefuses to be extracted even by the most effective methods, andsuggest that this represents the fraction of isomeric nuclei decayingby emission of a y-ray rather than by internal conversion, the recoilfrom the y-quantum being too weak to disturb the molecule.Methods for the separation of isomers of tellurium and seleniumwere developed by G. T. Seaborg, J. J. Livingood, and J. W.Kennedy,36 and by A. Langsdorf and E. Scgr&,3' respectively.The element indium38 is a veritable show-case of isomerism.Metastable states of both the stable indium isotopes lI3In and1151n have been produced and their radiations have been studied.The p-active l141n has a metastable state which decays into theground state (case B), and 116In shows two independent p-periodswith different energy spectra (case C).In the following table, col.2 indicates the type of isomerism asdefined at the beginning of this section. Col. 3 indicates on whatIs0 tope.49Ca44sc'Ti52Mn6OCO69Zn79 or 81Se80BrS3Krs5Sr8'Sr87YssZr08 *r 10143Table of Isomeric Nuclei.Type-B or CBB or CCB or CBBBAB ?AB ?B ?A *Evidence. Periods.1 30mins. (-)2.5 hours ( -)1, 2 4.1 hours (,!? 452 hours (I.T.)1 2.9 mins.(-)72 days (-)21 mins. (,!? +)6.5 days (/3 + 1, 21 11 mins. (-)5.5 years (6 -)57 mins. (/3 -)13.8 hours (1.T.)19 mins. (,!? -)1 hour (1.T.)18 mins. (#? -)4-5 hours (1.T.)113 mins. (12.)70 mins. (1.2". 1 )66 days (K)2.7 hours (1.T.)14 hours (1.T. ?)80 hours (K)78 hours (#I 3.)4.5 mins. (I.T. ?)6.6 hours (1.T.)Ref.2339404142433733, 34, 3537444444444533 E. Segr6, R. 8. Halford, and G. T. Seaborg, Physical Rew., 1939, 55, 321.34 Don C. DeVault and W. F. Libby, aid., p. 322.35 Ibid., 1940, 58, 688. 36 Ibid., 57, 363.38 J. L. Lawson and J. M. Cork, ibid., p. 982.* Listed as " A " because no #?-transition from ground state h m been3 7 Ibid., p. 105.found; it is, however, almost certainly b-unstable20 RADIOACTIVITY AND SUBATOMIC PHENOMENA.Isotope.lo4RhlosAggg1 0 7 or 1 0 9 ~108 or llOA?Cd131n1141n1151n6InlZ7TelaDTe131Te134cs152 and 154EU176 or liiLuleoTa182 or ls41rTable of Isomeric Nuclei (contd.)Type.BB or CAB or CA ?ABA eBB13CB or C??B or CB or C?AEvidence .Periods.44 secs. ( p -)4.2 mins. (I.T.)25 mins. ( p +)8.2 days (K ?)40 secs. (I.T.)22 secs. (llOAg, /3 -)2.3 mins. (lo8Ag, /3 -)225 days ( p -)50 mins. (I.T. ?)105 mins. (I.T.)72 sec. (p -)50 days (I.T.)13 secs. (/3 -)54 mins. (p -)90 days (I.T.)72 mins. (/3 -)32 days (I.T.)25 mins. (/3 - )1.2 day (I.T.)1.7 year ( p -)12 mins.9.4 hours (/3 -)1 Year (/3 -)4 hours6 days8.2 hours (K)1.5 mins.(,!I - ?)19 hours ( p -)60 days (j3 - ? )4.5 hours (I.T.)9.3 hours (/3 -)3 horns ( p -)105 mins.1 A 2 1 mins.18 hours (/3 - ?)13 hours (/3 - ?)3-3 days ( p - ?)4 5 days (/3 - ? )1.6 mins. (I.T.)Ref.464732484938383838363636505152535454543139 H. Walke, Physical Rev., 1940, 57, 163.40 H. Walke, E. J. Williams, and G. R. Evans, Proc. Roy.Xoc.,1939, A, 171,360.41 J. J. Livingood and G. T. Seaborg, Physical Rev., 1938, 54, 391;F. A. Heyn, Physica, 1937, 4, 160, 1224; J. J. Livingood and G. T. Sea-A. Hemmendinger, ibid., 1940, 58, 929.borg, Physical Rev., 1938, 53, 847.43 Idem, ibid., 1939, 55, 457.4 4 L. A. DuBridge and J. Marshall, ibid., 1940, 58, 7.45 G.T. Seaborg and E. SegrB, ibid., 1939, 55, 808.4 6 B. Pontecorvo, ibid., 1938, 54, 542.48 J. J. Livingood and G. T. Seaborg, ibid., 54, 88.49 M. Dod6 and B. Pontecorvo, Compt. Tend., 1938, 207, 287.50 D. C. Kalbfell and R. A. Cooley, Physical Rev., 1940, 58, 91.51 K. Fajans and D. W. Stewart, ibid., 1939, 56, 625.52 G. Hevesy and H. Levi, Nature, 1936, 137, 185.53 0. Oldenberg, Physical Rev., 1938, 53, 35.54 E. McMillan, M. D. Kamen, and S. Ruben, ibid., 1937, 52,375.4 7 M. L. Pool, ibid., 53, 116FRISCH : NEUTRONS. 21kind of evidence the assignment is based : “ 1 ” on the way in whichthe activity is produced; “ 2 ” on the energy distribution of theemitted particles ; ‘‘ 3 ” on identification of the characteristicX-rays; and “ 4 ” denotes that the two isomeric nuclei have beenseparated by chemical means.p + and p - after the periodmean that the nucleus in question undergoes p-transformation,with the emission of a positron or an electron, respectively; Kindicates K-electron capture, and I . T. means isomeric transition.(-) means that it is uncertain whether the electrons emitted areof nuclear origin or conversion electrons.NEUTRONS.The magnetic moment of free neutrons has now been measuredwith great accuracy 55 and found to be 1.93 0.02 nuclearmagnetons (one nuclear magneton = 111840 Bohr magneton).The result is very nearly what had been expected on the assumptionthat the magnetic moment of the deuteron is the sum of the momentsof the proton and the neutron.The method, which is similar t o the one by which Rabi and hiscollaborators have determined the nuclear moments of manyelements, is capable of very great accuracy.A beam of slowneutrons was “ polarised ” by passage through magnetised iron andthen passed through a homogeneous magnetic field, upon whichwas superimposed a weak field the direction of which alternateda t a high frequency. The homogeneous field causes the neutronsto precess like small gyroscopes. If the frequeney of this pre-cession happens to coincide with the frequency of the superim-posed field, the precession becomes disturbed and the beam isdepolarised. On emerging from this high-frequency treatment,the neutrons had to pass through a second slab of magnetised iron,which served as an analyser. The homogeneous field was graduallyaltered until a sudden change in the beam intensity indicated thatdepolarisation was taking place. In this way the rate of pre-cession in a given field was measured, and from it the magneticmoment of the neutrons can immediately be calculated.Diffraction phenomena affecting the scattering of slow neutronshave been studied by several authors. F. Rasetti 56 found thatprecipitated and carefully dried calcium carbonate had a molecularcross-section which was equal to the sum of the cross-sections ofthe component atoms. A crystal of calcite showed a cross-sectionless than one-third of that, while ground powders of various grainsizes gave intermediate values. The explanation is that a single5 6 L. W. Alvarez and F. Bloch, Physical Rev., 1940, 57, 111.5 6 Ibid., 58, 32122 RADIOACTIVITY AND SUBATOMIC PHENOMENA.crystal scatters only those neutrons whose de Broglie wave-lengthhappens to fulfil the Bragg condition for reflection by a lattice plane,while the other neutrons pass through almost freely. In a powder,however, a neutron passes through many grains of different orient-ations; in addition, the Bragg condition is less selective in smallgrains. H. G. Beyer and M. D. Whitaker 57 found similar anomalieswith crystals of iron, nickel, and quartz. Apart from the funda-mental interest of these experiments, they are of practical importancein so far as they call for a redetermination of all those slow-neutronscattering cross-sections which have been measured with crystallinescatterers. (Capture of neutrons, and scattering of faster neutronsshould not be affected by the crystalline structure.)0. R. FRISCH.5 7 Physical Rev., 1940, 57, 976

ISSN:0365-6217    

DOI:10.1039/AR9403700007

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

GENERAL AND PHYSICAL CHEMISTRY.1. INTRODUCTION.A VARIETY of topics is dealt with in this section of the Reports. Inthe first section there is a discussion of the forces of interactionbetween very simple molecules. Here there are now two lines ofapproach. By the use of wave mechanics it is possible to calculatefairly accurately these interactions, provided that the system issimple enough. On the other hand, valuable information can alsobe obtained from the statistical properties of a gas-in particularthe deviations from the ideal gas laws. W. J. C. Orr has describedand examined the results obtained by the two methods of approach.Coming next to a more complicated problem, namely, the structureof molecules, there is now a great range of methods, each of whichis almost indispensable to the others.In recent Reports, thecontributions made by the study of infra-red and Raman spectraand electronic spectra have been described. This year L. E. Suttonshows how much the study of electron diffraction by vapours andthe measurement of dipole moments have contributed to this fieldof research.Photochemistry can now be regarded as a branch of chemicalkinetics, as its relationship with the latter subject is very close.Accordingly the report on reaction mechanisms is confined ex-clusively to a discussion of recent work in this branch. There havebeen prolonged controversies in photochemistry, not only about theinterpretation of results, but also about the validity of the resultsthemselves. It is in a way unfortunate that the subject has oftenbecome clouded over in this manner.However, clarity is nowbeginning to appear, as it is realised that the behaviour of photo-chemically excited molecules is indeed much more complex than hashitherto been suspected. Long-held dogmas have had to beabandoned-often to the mutual satisfaction of rival theories.M. Ritchie consequently gives an account of the efforts made toprobe further into these reactions and bring harmony even a t theexpense of complexity into the subject.Classical colloid chemistry is a subject often shunned by themore theoretical of physical chemists, probably because of itsrather qualitative nature. There is, however, a vigorous revivaltaking place which must be chronicled. A. S.C. Lawrence describesin some detail the clarification of ideas in many parts of this branchof chemistry. H. W. M24 GENERAL AND PHYSICAL CHEMISTRY.2. INTERMOLECULAR ENERGIES.Explicit determinations of the potential energy of a pair of mole-cules as a function of their distance apart may in principle beobtained either from direct quanta1 calculations or from inductiveanalyses of the thermodynamic and kinetic behaviour of actualphysical systems. Limiting our field in this review to a discussionof the interactions of the rare-gas atoms and chemically saturated,non-polar, spherically symmetrical molecules, we shall summarisethe methods which have had practical success in this aim and describecurrent developments.(a) Direct Calculations.-The calculation of the interatomicenergies of particles, containing more than one electron, involvesin practice approximations concerning (a) the analytic form of thewave function, Y, or related electron distribution p, assumed forthe separate particles, and (b) the analytic behaviour of thesefunctions as the particles are brought together.Among the systemshere considered Y has been accurately calculated in the case ofhelium on1y.l The values of 9’ (or p) for more complicated atomshave been obtained either from the Fermi-Thomas statisticalmodel or from some modification of the Slater central field model.The Fermi-Thomas method represents the atom as a nucleussurrounded by an electron gas in which the number of electrons inany element of phase space h3 is limited, in accordance with thePauli principle, to two of opposite spin.In this way approximate,smoothed-out electron distribution curves are obtained. In theSlater method the electrons of an atom are considered to be movingindependently in the central field of the nucleus. The total wavefunction Y for a system of N electrons may then be expressed asit determinant whose N2 terms are single-electron $, mutuallyorthogonal, and normalised functions of the space and spin co-ordinates. Such a form is antisymmetrical in the electrons, asrequired by the Pauli principle, and includes exchange of theelectrons among the available eigen-functions. Numerical methodsof calculating these satisfying the Schroedinger equation forE. A. Hylleraas, 2.Pl~ysik, 1928, 48, 469; 51, 150; 1929, 54, 347;1930, 65, 209; A. F. Stevenson and M. F. Crawford, Physical Rev., 1938,54, 275.a L. H. Thomas, Proc. Camb. Phil. SOC., 1927, 23, 542 ; E. Fermi, 2. Physik,1928, 48, 73. A description of the use of this method and a bibliographyare given by H. Hellmann, “ Einfuhrung in die Quantenchemie,” Chap. 1.J. C. Slater, Physical Rev., 1929, 34, 1293; also E. U. Condon and G. H.Shortley, “ The Theory of Atomic Spectra,” Chap. 6ORR : INTERMOLECULAR ENERGIES. 25particular atoms have been developed by D. R. Hartree* andV.As two particles are brought together, their wave functions areincreasingly influenced by each other’s presence, and the potentialenergy of the system alters accordingly. On the statistical modelthis energy is composed of the mutual electrostatic attractions ofelectron clouds and nuclei, the exchange energy of the electrons,and the repulsion due t o the kinetic energy of the electrons, thelast term predominating.Calculations have been carried out forsystems having completed electron shells by W. Lenz,6 H. Jensen,‘P. Gombas,8 and others. Since the Fermi-Thomas distributionfalls off too slowly a t large distances, these calculations tend to givetoo large a repulsive potential. H. Hellmann has shown, however,that the statistical method yields good results in the case of twohelium atoms when wave-mechanical distributions are employed.Hence, although this method cannot lead t o high precision calcul-ations of repulsive energy, it can provide useful approximationsfor systems too complicated t o be treated wave-mechanically.Three general approximation methods are available by whichthe quantum-mechanical energy of two interacting neutral atomsmay be calculated, vix., (u) the generalised perturbation method ofHeitler and London, ( b ) the perturbation method neglecting electronexchange between the atoms, and ( c ) the variation method.Inthe first method a wave function, having a determinantal form,containing t,h for the electrons of both atoms is formed, and theSchroedinger equation is solved for the potential energy by theapproximation methods first developed by W. Heitler and F.London lo and Y . Sugiura In the caseof atoms having rare-gas configurations this method yields a re-pulsive potential a t all distances. A calculation accurate to 1 or2% for the case of two helium atoms has been made by J.C. Slater,12for the hydrogen molecule.Proc. Camb. Phil. Soc., 1928, 24, 89, 111.2. Physik, 1930, 61, 126; 62, 795. A list of references on this methodup to 1936 is given by H. Hellmann, op. cit., Chap. 3. Further referencesare : D. R. Hartree and W. Hartree, Proc. Roy. Soc., 1938, 164, A, 167;166, A, 450 ; Proc. Camb. Phil. Soc., 1938, 34,550 ; D. R. Hartree, W. Hartree,and B. Swirles, Phil. Trans., 1939, 238, A, 229; A. F. Stevenson, Proc. Roy.Soc., 1937, 160, A, 588; Physical Rev., 1939, 56, 586; M. F. Manning andL. Goldberg, Physical Rev., 1938, 53, 662; R. L. Mooney, ibid., 1939, 55,557; W.J. Yost, ibid., 1940, 58, 556; A. 0. Williams, ibid., p. 723.6 2. Physik, 1933, 77, 713.Ibid., p. 722; 1936, 101, 141, 164.* P. Gombas and T. Neugebauer, ibid., 1934, 89, 480; P. Gombas, ibid.,92, 796; 93, 378; cf. also R. P. Bell, Proc. Roy. Soc., 1940, 174, A, 504.@ 2. Physik, 1933, 85, 180. lo Ibid., 1927, 44, 455. l1 Ibid., 45, 488.l2 Physical Rev., 1928, 32, 349; cf. also N. Rosen, ibid., 1931, 38, 25526 GENERAL AND PHYSICAL CKEMISTRY.using analytic wave functions. W. E. Bleick and J. E. Mayer 13treat the casc of two neon at,oms, using approximate Hartree wavefunctions. The results of both these calculations are well repre-sented by a simple exponential law, B(R) = Aexp.(- a&). Sucha relation is also suggested as a limiting law by calculations on theexchange forces between hydrogen-like atoms and K , L, M , and Nelectron shells,14* l5 but a t normal equilibrium distances it is foundthat much more complicated expressions are involved.A recentapproximate method of reducing this many-elect ron problem toa number of single-electron problems has been suggested by M. F.Mamotenko and A. A. Schuchowitzky.lG This treatment introducesspecial orthogonalisation principles designed to conserve the Pauliprinciple and to reduce to a minimum the contribution of the ex-change integrals. Using J. C. Slater’s atomic wave functions,l7the repulsive potentials of pairs of hydrogen and helium atoms areobtained in fair agreement with more elaborate methods.ll* l2In the perturbation method the Hamiltonian expression for theinteraction energy, H , is averaged over all the unperturbed statesof the separate systems.For the case where the energy levelsEn are non-degenerate, the expansion, as far as the second approxim-ation, is as follows :I n the variation method, suitable analytic forms of Y are chosencontaining arbitrary parameters, which then are varied to makethe expression 46 =/YHFdT//YqdT a minimum. The energy socalculated will provide an upper limit t o the experimental valueeven when the central field Y is used, since the latter neglectscorrelations in the relative motion of electrons. In the case ofnon-polar, spherically symmetrical molecules the first-order per-turbation energy is zero. For unsymmetrical molecules such ashydrogen,l8 however, the firsborder mutual quadruple interactionsurvives and gives attractive and repulsive potentials dependingI3 J .Chem. Physics, 1934, 2, 252.l4 A. Unsold, 2. Physik, 1927, 43, 563.l5 H. Briick, 1928, 51, 707; also V. Zdanow, A. Erschow, and G. Galachow,l6 Acta Physicochim. U.R.S.S., 1338, 9, 803; M. F. Mamotenko, ibid.,l7 Physical Rev., 1930, 36, 57.ibid., 1935, 84, 241.1939, 11, 225; A. A. Schuchowitzky, ibid., 1935, 2, 81.F. London, 2. physikal. Chern., 1931, B, 11, 222 ; H. S. W. Massey andR. A. Buckingham, Proc. Roy. Irish Acad. Sci., 1938, 45, A, 31 ; J. K. Knipp,Physical Rev., 1938, 53, 734; K. Cohen and H. C. Urey, J . Chem. Physics,1939, 7 , 157ORR : INTERMOLECULAR ENEMIES. 27on the relative orientation of the molecules, the average energyover all orientations being zero.For molecules with low momentsof inertia, such as those considered here, this effect would never bephysically appreciable but it is possibly significant for the analysisof kinetic data involving highly asymmetrical molecules.The second-order perturbation calculation yields results equivalentto a variation calculation, the energy so calculated correspondingalways to an attractive potential. At distances apart of two ormore molecular diameters it is permissible to neglect electronexchange between the particles and t o simplify the calculation byexpanding the CouIomb electrostatic energy in a Taylor's seriesin the internuclear distance R. The successive terms in the ex-pansion, varying as R-3, R-4, etc., give rise to the dipole-dipole,dipole-quadrupole, etc., energies, which vary as R-6, Adetailed review of the calculation of these van der Waals forces upto 1939 has been made by H.Margenau,19 and only a brief summaryof the results will be given here. The first successful calculationsapplicable to a number of atoms and molecules were made by F.who, using a perturbation method, obtained an expressionfor the dipole-dipole energy, in terms of the constants of the opticaldispersion formula, for the interacting particles. A more general,but less accurate expression, also due to F. London,18 involves theionisation potentials and polarisabilities of the reacting particles.H. Margenau 21 has extended this method to calculate approximatelythe corresponding dipole-quadrupole energies.These calculationsand those of R. A. Buckingham22 agree in showing that the R-8energy terms contribute about 20% to the total van der Waalspotential at the energy minimum, and that the R-10 term can ingeneral be neglected. Variation calculations employing pro-gressively more adequate wave functions have been considered byJ. C. Slater and J. G. K i r k ~ o o d , ~ ~ J. G. K i r k ~ o o d , ~ ~ J. P. Vinti,25H. Hellmann,26 and R. A. Buckingham.22 The results of thesecalculations have been discussed by J. K. Knipp 27 in relation tohis own more general treatment. Knipp uses the central fieldapproximation with Hartree eigen-functions to calculate the dipole-dipole energies of argon atoms.A significant feature of the calcul-ations is that 98% of this energy is contributed by the outer electronshells.Special attention has been devoted to calculating the van derl9 Rev. Mod. Physics, 1939, 11, 1.21 J . Chem. Pbsics, 1938, 9, 896.23 PhysicccZ Rev., 1931, 37, 682.'L5 Physical Rev., 1932, 41, 813.2 6 Acts Physicochim. U.R.S.S., 1935, 2, 273.2 7 Phy&ical Rev., 1939, 55, 1244.Z . Physik, 1930, 63, 245.22 Proc. Roy. Soc., 1937, 160, A, 113.24 Physikal. Z . , 1931, 33, 6728 GENERAL AND PHYSICAL CHEMISTRY,Waals energy of two helium atoms. The most accurate calculationsof the who, using accurateHylleraas lY, obtains the value 1.368 x l0-12erg A.G. The successiveR-8 and coefficients have been calculated approximately byH.Margenau 29 and by G. H. Page.30 While the theoretical positionremained as described above, it was not permissible without furtheranalysis t o add together the calculated “ exchange ” repulsivepotential and the van der Waals potentials, calculated with neglectof exchange, to obtain the potential curve in the neighbourhood ofthe minimum, as was frequently assumed. H. MargenauY3l how-ever, calculates the second-order exchange energy at the minimumand finds that, although this forms a very important contributionin the case of two hydrogen atomsF2 yet it yields only a smallattractive potential in the case of two helium atoms. It is con-sidered probable that the relative importance of this term will bemuch less for heavier atoms.The best available self-consistentquantum-mechanical calculation of the interaction energy of twohelium atoms, valid over a wide range of internuclear distances, isgiven by the following expression, which combines J. C. Slater’scalculation l2 of the repulsive with H. Margenau’s calculations 31of the attractive potentials :$ = [770 exp.(- 4.60R) - 560 exp.(- 5.33R) - 1.39/R6 - 3.0/R8](b) Thermodynamic Data.-The analysis of thermodynamicdata in terms of force fields implies the existence of an adequatestatistical description of the system involved. Although usefulsupplementary information can be deduced from the properties ofsolid and liquid phases, the greatest detail is obtained a t presentfrom gas-phase data. Owing to the work of J. E.M a ~ e r , ~ ~ it isnow possible to expand the classical configuration integral whichoccurs in the expression for the Helmholtz free energy of a gas ofN molecules confined in a volume v, as a power series in the densityNlv, the successive terms representing the interaction of moleculesin groups of two, of three, and so on. The combinatory analysishas been discussed and clarified by various authors.34 The first28 Proc. Carnb. Phil. SOC., 1930, 26, 542; 1931, 27, 66; T. D. H. Baber andH. R. Hass6, ibid., 1937, 33, 253.29 Physical Rev., 1931, 38, 747.go Ibid., 1938, 53, 426.g2 H. Hellmann and K. W. Majewski, Trans. Paraday SOC., 1937, 33, 43.g3 J. Chem. Physics, 1937, 5, 67 ; J. E. Mayer and P. G. Ackermann, ibid.,p. 74; J. E. Mayer and S.F. Harrison, ibid., 1938, 6, 87; S. F. Streeter andJ. E. Mayer, ibid., 1939, 7, 1025; J. E. Mayer, J . Physical Chem., 1939, 43, 71.34 M. Born, Physica, 1937, 4, 1034; M. Born and K. Fuchs, Proc. Roy.SOC., 1938,166, A, 391 ; B. Kahn and G. E. Uhlenbeck, Physicu, 1938,5,399.coefficient are those of H. R.x 10-l2 erg . . (1)s1 Ibid., 1939, 56, 1000ORR : INTERMOLECULAR ENERGIES. 29three terms of the equation of state derived therefrom are asfollows : 35+(R) being the intermolecular potential and g,(R) an integral dueto the simultaneous interaction of three molecules. Since thepresent accuracy of the experimentally measured virial coefficients[B and C in the expression, Pv = RT(l + B/v + C/v2) . . .] is toolow to permit a direct determination of +(R) by inversion of theabove integrals,36 these must be integrated directly or numericallyfor particular analytic +(R) whose unknown parameters are to befixed by comparing the calculated and experimental behaviourof B and C as functions of the temperature T.The first detailedanalyses of this kind were carried out by J. E. Lennard-J~nes,~~who employed the bi-reciprocal function, + = - p/Rm + v / D . Withm = 6 it was found that equally good fits were obtained with valuesof n varying from 8 to 14. The experimental sublimation energiesand interatomic distances of the crystals were used to fix theseconstants more precisely.379 38 J. Corner 39 has shown, in the caseof neon and argon, that both crystal and virial data are well repre-sented by unique potential functions of this kind.The effect ofexponential repulsive potentials, which are suggested by quantumtheory, has been examined by R. A. Buckingham,40 J. A. Wasast-jernafl and others.42 A calculation of the third virial coefficientsfor a number of gases, using a bi-reciprocal potential curve withn = 12, has been carried through by J. de Boer and A. Michels 35, 43and, although laborious, forms a useful check on the constants ofthe potential curve calculated from the second virial coefficients.J. G. Kirkwood, J. Chern. Physics, 1935, 3, 300; J. de Boer andA. Michels, Physica, 1939, 6, 97.36 S. C. Collins and F. G. Keyes, J . Physical Chem., 1939, 43,5; R. E. A. C.PaIey and N. Wiener, Amer. Math. SOC. Coll. Pub., 1934, 19, Chap. 3.37 Proc.Physical SOC., 1931, 43, 461; Physica, 1937, 4, 941. A furtherreview of this work is contained in R. H. Fowler and E. A. Guggenheim’s“ Statistical Thermodynamics,’’ Camb. Univ. Press, 1939, Chap. 7.38 M. E. Hobbs, J . Chem. Physics, 1939, 7, 318.3D Trans. Faraday Xoc., 1939, 35, 711.40 Proc. Roy. Soc., 1938, 168, A, 264.Soc. Sci. Fenn. Phys. Math., 1932, 6, NOS. 18-22; 1935, No. 20; Phil.42 J. G. Kirkwood and F. G. Keyes, Physical Rev., 1931, 37, 832;48 Physica, 1938, 5, 946.Trans., 1938, 237, A, 105.K. Herzfeld, ibid., 1937, 52, 37430 GENERAL AND PHYSICAL CHEMISTRY.Joule-Thomson coefficients,36, 44 which can now be measured toan accuracy of 1%, provide a second convenient source of thermo-dynamic data for force field analysis. J.0. Hirschfelder, R. B.Ewell, and J. R. Roebucka5 calculate potential curves takingaccount of quantum corrections for helium, neon, and argon byusing such measurements. Further similar inductive analyseshave been made by W. Wen-Po *6 and J. Corner.*’ A third sourcefrom which sufficiently accurate virial coefficients can be calculatedis provided by the measurement of sound velocities in gases, afield which is being developed by A. van Itterbeek and his co-w o r k e r ~ . ~ ~ This method is especially useful in that it considerablyextends the present temperature range of data available.For gases such as hydrogen, deuterium, and especially helium,a t low temperatures where the de Broglie wave-length A = hldmkTis of the order of magnitude of a molecular diameter, the above-described classical treatment is no longer adeq~ate.~g The influenceof the mutual field +(R) of a pair of particles of mass m, energyE, and angular momentum (h/2x)2/Z(Z + 1) on their relative motionis obtained by solving the radial wave equation 5Owhere K~ = 4x2mE/h2, the proper asymptotic solution of which is$ = R-lsin (KR - 32x + Sl).I n terms of the phases S,, thus defined,and the discrete energy levels E,, and weights gn of the function+(I$), the second virial coefficient is 5 1-d2(R$)/dR2 + [ K ~ - (4n2m/h2>+(B) - Z(Z+ l)/R2](R#) = 0 . (3)8zgneE*lkTj n . . (4)When the atoms contain an even number of elementary particles4 4 J. R. Roebuck and H. Osterberg, J . Chem. Physics, 1940, 8,627 (containsa complete list of references to their earlier work).45 Ibid., 1938, 6, 205; J.0. Hirschfelder and W. E. Roseveare, J . PhysicalChem., 1939, 43, 15.4 6 Phil. Mag., 1939, [vii], 26, 225.4 7 Trans. Paraday Soc., 1940, 36, 781.A. van Itterbeek and P. Mariens, Physica, 1938, 5, 153; 1940, 7, 125;A. van Itterbeek a d 0. van Paemel, ibid., 1938,5,593, 845; A. van Itterbeekand L. Thys, ibid., pp. 640, 889; A. van Itterbeek, P. de Bruyn, and P.Mariens, ibid., 1939, 6, 511.E. Wigner, Physical Rev., 1932, 40, 749; J. C. Slater, ibid., 1931, 38, 237.6o H. Faxen and J. Holtsmark, 2. Physilc, 1927, 45, 307; N. F. Mott andH. S. W. Massey, “The Theory of Atomic Collisions,” 1933, Chap. 2;K. Schiifer, 8. physikal. Chem., 1937, B, 36, 85; B, 38, 187.61 L.Gropper, Physical Rev., 1936, 50, 963; 1937, 51, 1108; E. Beth andG. E. Uhlenbeck, Physiccc, 1937, 4, 916; B. Kahn, Diss., UtrechtORR : INTERMOLECULAR ENERGIES. 31(BoseEinstein statistics), E = + 1, the summations are takentwice over even values of I , whereas, when they contain an oddnumber (Fermi-Dirac statistics), E = - 1, and the summationsaxe taken twice over odd values of 2. At high temperatures theseexpressions converge t o the classical one. Series exp+nsions havebeen developed for use in this region.52 At low temperatures,however, the and E, must be calculated numerically for particularforms of +(R). Such calculations have been made in the case ofhelium by H. S. W. Massey and R. A. B~ckingham,~~ who use a Slater-Kirkwood 23 potential; by L.G r ~ p p e r , ~ ~ who adds to this potentialthe and R-l0 terms calculated by H. Margenau 29 and concludes,from a comparison of the experimental 55 and the calculated valuesof B(T) in the liquid helium temperature range, that the truepotential curve must lie between these two; and finally, by J. deBoer and A. M i ~ h e l s , ~ ~ who obtain close agreement with experimentusing a bi-reciprocal potential curve determined by inductiveanalysis from high-temperature helium virial data.43 They furtherfind that this curve gives somewhat closer agreement than is obtainedby using the most recent purely theoretical curve already mentioned,vix., equation (l).31 The present degree of concordance betweenthe best theoretical and inductively determined potential curvesfor helium is shown in the figure.Although it is possible to obtain potential curves accurate overa wide range of values of R by the inductive analysis of thermo-dynamic data on gases covering an extensive temperature range(and thus a, wide range of collision diameters), yet similar analysesof the available data on condensed phases, which are restricted tomuch more limited density ranges, give precise information appro-priate only to the region of the potential minimum.An accurate statistical theory of the liquid phase has not yetbeen obtained.The " statistical cage " model has been improvedby the explicit introduction of actual intermolecular potentials byJ. E. Lennard-Jones and A. F. De~onshire,~' who successfullycalculate various physical properties and constants of simple52 J.G. Kirkwood, Physical Rev., 1933, 44, 31; G. E. Uhlenbeck and53 Proc. Roy. SOC., 1938,168, A, 378; 169, A , 205.54 Physical Rev., 1939, 55, 1095.5 5 W. H. Hewom and H. H. Kraak, Physica, 1935, 2, 37; W. H. Keesomand W. K. Walstra, ibid., 1939, 6, 1146.5 6 Ibid., p. 409.5 7 Proc. Roy. SOC., 1937, 163, A, 53; 165, A , 1 ; 1939, 169, A , 317; 170,A, 464; A. F. Devonshire, ibid., 1940, 174, A, 102; 3. E. Lennard-Jones,Proc. Physical Soc., 194-0, 52,729; J. E. Lennard-Jones and J. Corner, Trans.Faraday SOC., 1940, 36, 1156; W. J. Archibald, Physical Rev., 1939, 58, 926.E. Beth, Physica, 1936, 3, 72932 GENERAL AND PHYSICAL CHEMISTRY.liquids in terms of the parameters of the potential curves.Inversely,such data can be interpreted on the basis of this model to giveinformation regarding the force fields of the molecules concerned.Recent developments, both theoretical and e~perirnental,~~ nowenable one to calculate the atomic distribution function, g(R), forquasi-monatomic liquids from the experimental X-ray scatteringI. qqR) = yg - - '::) x 10-la. (Ref. 45.)11. +(R) = (* -=) x 10-12. (Ref. 43.)R12 R6curves, so that it is possible by combining various purely experi-mental data t o determine the force fields by induction. J. H.Hildebrand 59 uses the relation2rcN21m c#@)g(B)B2dR = Ev(where v and E are the molal volume and the energy of evaporation),in conjunction with the experimental g(R) curves at a series of5 8 B.E. Warren and N. 8. Gingrich, Physical Rev., 1934, 46, 368; B. E.Warren, J. Appl. Physics, 1937, 8, 645.SD J . Chem. Physics, 1933, 1, 817; 1939, 7, 1 ; J. H. Hildebrand, H. R. R.Wakeham, and R. N. Boyd, ibid., 1939, 7, 1094ORR : INTERMOLECULAR ENERGIES. 33temperatures,m to derive the constants of a bi-reciprocal potentialin the case of liquid mercury. Further analyses of this kind onargon 61 and other similar simple liquid systems would provideinformation on the important question of the structure of liquidsas well as on intermolecular potentials.62A rigorous statiskical theory of the thermodynamic behaviour ofcrystals is being developed and applied to calculate various physicalproperties, notably the melting points and behaviour under strainin terms of a general intermolecular force field, by M.Born and hisc o - ~ o r k e r s . ~ ~ The full development of this work, particularly inits application to the solid rare-gas atoms, should lead to accuratedeterminations, by inductive analysis, of the position and, inparticular, the curvature of the potential curves in the vicinityof the minimum. Calculations of this kind, but based on theassumption of a Debye spectrum of frequencies, have already beenmade by various workers,64 who obtain fair agreement withexperiment by using force fields otherwise determined.(c) Kinetic Data.-In a recent treatise S. Chapman and T. G.Cowling 65 have described in detail the classical derivation of themolecular distribution law for non-uniform gases in its fully de-veloped form, which is mainly due to D.Enskog,66 S. Chapman,67and D. Burnett.68 In terms of this law, such free-path phenomenaas the viscosity, thermal conductivity, and molecular and thermaldiffusion of simple gases can be treated on the basis of certainsimple molecular models. A general treatment of all thesephenomena can at present best be given in terms of a law of mutualmolecular repulsion, 'uix., + = v/Rn. On this model, the temperaturevariation of the viscosity, 3, of a gas should be q/q0 = (T/T,JS, whereX = 8 - 2/n. The actual values of n obtained, which vary from6o R. N. Boyd and H. R. R. Wakeham, J. Chem. Physics, 1939, 7,958.61 A. Eisenstein and N. S. Gingrich, Physical Rev., 1940, 58, 307;K. Lark-Horovitz and E.P. Miller, Nature, 1940, 146, 459.62 C. N. Wall, Physical Rev., 1938, 54, 1062; C. A. Coulson and G. S.Rushbrooke, ibid., 1939, 56, 1216; J. Corner, Proc. Physical SOC., 1940, 52,764.63 J . Chem. Physics, 1939, 7, 591; Proc. Camb. Phil. SOC., 1940, 36, 160;M. Born and R. D. Misra, ibid., pp. 173, 466; M. Born and R. Furth, ibid.,p. 454; Nature, 1940, 145, 741.64 K. Herzfeld and M. Goeppert-Mayer, Physical Rev., 1934, 46, 995;G. Kane, J. Chem. Physics, 1939, 7, 603; W. Wen-Po, Phil. Mag., 1936,[vii,] 22, 49, 281; 1937, 23, 33; 24, 466; 1938, 25, 111.6 5 " The Mathematical Theory of Non-uniform Gases," Camb. Univ.Press, 1939.66 Diss., Upsala, 1917; Svensk Vet. Akad. Arkiv. Mat. Ast. B'ys., 1921, 16,l.6 7 Phil. Trans., 1916, 216, A, 279; 1917, 217, A, 115.68 Proc.Lond. Math. SOC., 1935, 39, 385; 40, 382; Proc. Camb. Phil. SOC.,REP.-VOL. XXXVII. B1937, 33, 359, 36334 GBNERAL AND PHYSICAL CHEMISTRY.13.5 for helium to 6.4 for argon, indicate the much greater inter-penetration of the electron shells that OCCUFS at comparable tem-peratures in collisions involving the lighter atoms. A slight trendof n with temperature is due to the approximate nature of thepotential law assumed. A further indication that a more elaboratemodel is required is the fact that the average values of nlz for pairsof gases, obtained from the analyses of diffusion coefficients, areuniformly lower than the means of the individual values of n, andn, determined from viscosity measurements.In the case ofviscosity, the experimental data have also been analysed in terms ofvarious other models, the most general of which are those of J. E.Lennard-J~nes,~g who combines a repulsive potential with a R-2attractive potential, and H. R. Has& and W. R. who usea R-4 attractive term. The Lennard-Jones model gives a two-constant formula, which represents the experimental data accuratelyover a wider range of temperature than the simpler model, butdiverges a t low temperatures, as would be expected since, in thisregion, in addition to quantum effects in the cases of hydrogen andhelium, the contribution of the attractive potential is imp0rtant.7~A summary of the experimental and theoretical work in this fieldup to 1939 is given by S.Weber.72 A. van Itterbeek and W. H.Keesom 73 have initiated a further new series of experimentalmeasurements 74 on gas viscosities.Extensive measurements on the thermal diffusion of gases,which prove to be convenient and accessible sources of force-fielddata, have been carried out by T. L, Ibbs and his c o - ~ o r k e r s . ~ ~Moreover, wide prospects now exist for the further developmentof this field in the use of isotopic mixtures, since the mathematicalproblems involved in the interpretation of the data are therebymuch simplified. 76 Theoretical expressions for the thermal diffusionProc. Roy. SOC., 1924, 106, A, 441 ; 1925, 107, A , 157 ; J. E. Lennard-Jones and W. R. Cook, ibid., 1926,112, A, 214.70 Proc. Roy. SOC., 1929, 125, A, 196; also W.Wen-Po, Phil. Mag., 1938,[vii], 25, 865.71 A. van Itterbeek and 0. van Paemel, Physica, 1938, 5, 1009.72 Ibid., 1939, 6, 562.73 Ibid., 1933, 1, 128; 1935, 2, 97; 1938, 5, 257.74 W. H. Keesom and G. E. Macwood, ibid., p. 749; A. van Itterbeek andA. Clam, ibid., p. 938; W. H. Keesom and P. H. Keesom, ibid., 1940, 7, 29;A. van Itterbeek and 0. van Paemel, ibid., p. 273; also A. B. van Cleave and0. Maass, Canadian J. Res., 1935, 12,57 ; 13, 384.7 5 T. L. Ibbs, Physica, 1937, 4, 1133 (bibliography up to 1937); B. E.Atkins, R. E. Bastick, and T. L. Ibbs, Proc. Roy. SOC., 1939, 172, A, 142;A. A. Hirst and G. E. Harrison, ibid., 169, A, 573; R. E. Bastick, H. R.Heath, and T. L. Ibbs, ibid., 173, A, 543.76 S. Chapman, (a) Phil. Mag., 1917, 34, 146; 1919, 38, 182; ( b ) Nature,1940, 146, 607ORR : INTEEMOLEUULBR ENERUIES.36coefficients of isotopic mixtures have been made, to the same degreeof accuracy as the present viscosity calculations, for differentmoleculcGr models, by R. C. Jones,77 and the relative accuracyof these models is estimated by comparing the memured thermaldiffusion coefficients, for zb given temperature interval, withtheoretical values calculated from the experimental viscositiesin the same interval. Such intercomparisons for isotopio mixturesof methane 78 and of neon 7B indicate that this method is rathersensitive t o the proper form of molecular field and may prove avaluable additional method of determining the actual con-stant~.~6b* 77s 80 A further important application of these calcul-ations of thermal diffusion coefficients lies in their connexion withK.Clusius and G. Dickel's method81 of separating isotopes.Theoretical calculations 82 of the efficiency of such a system forseparating the isotopes of carbon, methane being used, have recentlybeen confirmed by experiment. 83The generalisation of the classical treatment of D. Enskog andS. Chapman to include quantum statistics has been made by E. A.Uehling,@ who shows that the equations of transport are unaltered,but that the experimental coefficients require correction. Theseare exactly analogous to those which occur in the quantum-statisticalexpression for B(T) (equation 4) and can be evaluated when thephases due to the relative motion of the molecules are calculated.The introduction of the quantum corrections to the calculation ofthe viscosity coefficients of helium, even on the rigid elastic spheremodel, makes an immediate improvement in the agreement obtainedon the classical theory.Mb9 85 Intercomparisons of the experimentaldata and rigorous calculations such as those carried out by H.S. W.Massey and R. A. B~ckingham,~3 using a, Slater-Kirkwood 23potential, and E. A. Uehling and E. J. Hellund,B*c using an in-7 7 Physical Rev., 1940, 58, 111.78 Idem, ibid., 1940, 57, 338.Natumiss., 1938, 26, 546; 1939, 27, 148, 487; 2. physikab. Chem., 1939,B, 44,397, 451.82 W. H. Furry, R. C. Jones, and L. Onsager, Physical Rev., 1939, 55,1083; R. C. Jones and W. H. Furry, ibid., 1940,57,547.T.I. Taylor and G. Glockler, J. Chem. Physics, 1940, 8, 843 (containsa complete summary of earlier references on this method). Compare also F.T. Wall and C. E. Holley, ibid., p. 949; W. Krasny-Ergen, Physical new.,1940, 58, 1078; J. Bardeen, ibid., 1939, 57, 35; 1940, 58, 94.(a) E. A. Uehling and G. E. Uhlenbeck, Physical Rev., 1933, 43, 562;(b) E. A. Uehling, ibid., 1934, 46, 917 ; (c) E. A. Uehling and E. J. Hellund,ibicl., 1938,54,479; (4 E. J. Hellund and E. A. Uehling, ibid., 1939, 56, 818;( e ) E. J. Hellund, ibid., 1940, 58, 278.85 H. S. W. Massey and C. B. 0. Mohr, Nuture, 1932, 130, 276; Proc. Roy.SOC., 1933,141, A, 434; 1934,144, A, 188.A. 0. Nier, ibid., 1939, 58, 1009.H. Brown, ibid., 58, 66136 GENERAL AND PHYSICAL CHEMISTRY.ductively determined bi-reciprocal curve, in the case of helium,provide useful criteria for determining the relative accuracy of thepotential curves assumed.A special consequence of the Bose-Einstein statistics, which apply to helium atoms, is that the viscosityshould become appreciably pressure-dependent a t 1" or 2" K., butmeasurements have not yet revealed this effect.73MoZecuZur Buys.-Possibly the most direct experimental approacht o the measurement of intermolecular potentials is through thestudy of molecular rays. Technical difficulties, which have so farhindered developments in this field, are now being overcome.H. S. W. Massey and C. B. 0. Mohr 85 first pointed out that, sincein general the intermolecular potential will decrease more rapidlythan R-3 at large distances, the quantum theory predicts, in contrastt o the classical theory, finite cross-sections for collision which canbe calculated in terms of the phases 8, (equation 3). Thus it ispossible to deduce the form of +(R) from scattering experimentswhen these are carried out with sufficiently high resolution.More-over, since the collision diameter of, e.g., argon molecules movingwith ordinary thermal velocities is approximately twice the inter-nuclear distance, such data give a direct measurement of theasymptotic B-6 van der Waals coefficient. Values of such constantsfor rare-gas atoms have been calculated by H. S. W. Massey andR. A. Buckingham86s22 from the scattering data of S. Rosin andI. I.Rabi,87 which are in fair agreement with values calculated inother ways. On the other hand, other investigators 88 havedeveloped the technique of measuring the scattering of atomsmoving with energies of several hundred electron-volts. Suchenergies may correspond to collision diameters of less than half theinternuclear distance, and so provide information on the mutualrepulsive potentials. W. J. C. 0.3. ELECTRON DIFFRACTION BY GASES AND VAPOURS, ANDELECTRIC DIPOLE MOMENTS.A Report on the theory and chemical applications of the diffractionof electrons by vapours appeared in 1936.1a Report on the Structure and Stereochemistry of Simple OrganicMolecules elaborated some of the matters discussed previously, andraised fresh ones. Since then there have been developments in thetheory of the method and in its practical application which are8 6 Nature, 1936, 138, 77.88 I.Amdur and H. Pearlman, J . Chem. Physics, 1940, 8, 7, 998; R. L.I n the following year8 7 Physical Rev., 1935, 48, 373.Mooney, Physical Rev., 1940, 58, 871.S . Glasstone, Ann. Reports, 1936, 33, 65.L. 0. Brockway and T. W. J. Taylor, ibid., 1937, 34, 196SUTTON : ELECTRON DIFFRACTION BY GASES AND VAPOURS. 37important because they either mark substantial advances in itsactual power or indicate potentialities. There have also been manyfurther applications of it to chemical problems.Although there has been no general Report on the measurementand the chemical applications of electric dipole moments of moleculessince 193lY3 there have been recent Reports on special topics.In1935* there was one on the use of electric moments to determinevalency angles; and in 1936 one on the effect of the solvent in themeasurement of moments. During the last decade there has beena vast amount of work on the theory of electric polarisability, itsmeasurement, and its use as a tool in structural investigations or asa concept in many physicochemical, organochemical, or even bio-chemical problems. It is hardly an exaggeration to say that mostsubstances which will dissolve in non-polar solvents or can bedistilled safely in a water-pump vacuum have been examined; andcontinued efforts have been made to measure less tractable materials.It is clearly impossible to give a full account of so much work in thecompass of this Report ; so it is proposed to give a summary of mostof the main developments in the theory of the method, and toaugment this by more detailed discussions of some special topics.Concerning the applications to chemical problems of these twoexperimental methods, it is important to bear in mind that neitherone is so powerful and unambiguous that it alone should be reliedupon when a question of molecular structure is under examination ; 6whenever possible all the methods which can be applied should bebrought to bear on the problem.Although this can hardly be donethoroughly in this Report, it will be done whenever it is important tocompare or contrast the conclusions from bond-length data andelectric dipole-moment data.For this reason the discussion of suchconclusions will be given in a separate section, tied neither to onemethod nor to the other.No special attempt has been made to give complete references tothe very large number of relevant papers : but those given, togetherwith the references which they themselves contain, should constitutean adequate guide to the literature.(i) Electron Diffraction by Gases and Vapours.The Development of the Methods of Analysis.Although an account of the methods then in use for interpretingelectron-diffraction photographs was given in an earlier Report,' aN. V. Sidgwick and E. 5. Bowen, Ann. Reports, 1931, 28, 365.S. Glasstone, ibid., 1935, 32, 126.L. E. Sutton, J., 1940, 544.Idem, &bid., 1936, 33, 117.7 L.0. Brockway, Rev. Mod. Physics, 1936, 8, 23138 GHNERAL AND PHYSICAL CHEMISTRY.brief restatement is included here as an introduction to later develop-ments. Several reviews of the subject, including descriptions ofapparatus, appeared re~ently~7-1~ but some of them are not readilyavailable, and most of them are in certain respects out of date.There is need of a fresh review, more detailed than this Reportcan be.When moving electrons undergo collision with other chargedparticles, they behave as if they have wave-character, A = h/mv( h = Planck’s constant, rn = electronic mass, v = velocity). Conse-quently, when large numbers of them collide with a body composedof particles which are separated by distances of the same order as A,there is interference between the electron waves scattered by theseveTa1 particles, and a diffraction pattern results which can beobserved by photographic or other means.This is very similar tothe diffraction which occurs when X-rays are scattered by sucha body : and the photographs of X-ray or electron scattering by,e.g., gold foil are strikingly similar in qualitative character.It has been shown * that when the diffracting agent is anassemblage of gas molecules, the following expression gives therelative intensity of scattering as a function of the angle 0 betweenthe initial beam and any direction in which we are interested (0 isknown as the diffraction angle and may have any value between0 and X ) :Iw = kXiZj#t4j (sin x z j ) / x i j .. . (1)k is a scale constant, and the functions t,h express the scattering powersfor electrons of the several atoms, while xii = (474, sin @)/A = Zijjs,A being the electron wave-length, and Zij the distance betweenthe ith and the j t h atom. If we are concerned with scatter-ing by diatomic molecules, this expression therefore tells us thatthere will be scattering by the ith and ,jth atoms themselves, kt,hi2and as Zij --+ 0 isunity], with interatomic scattering expressed by k(&,bj ++,+j)(sin xij)/xij superimposed on it. The intensity of the atomscattering varies with the angle only if t,ht and $ j do so, whereas theoccurrence in the interatomic term of the function (sinxij)/xij, oralternatively [sin (aij sin -p)J/aij sin @, means that as 8 increasersrespectively [since the limit of (sin* H.J. Emeldm and S. Miall, Chem. and Ind., 1936, 952.9 P. Debye, Angew. Chem., 1937, 50, 3.lo G. B. Kistiakowsky, J . Physical Chem., 1937, 41, 175.l1 Th. Schoon, Angew. Chem., 1939, 52, 245, 260.l2 J. A. A. Ketelaar, Nederland. Tijd8chr. Natuurkunde, 1938, 5 , 233.l3 J. Y . Beach, Pubns. Amer. Assoc. Adv. Sci., No. 7, ‘‘ Recent Advancesin Surface Chemistry and Chemical Physics,” 1939, 88. * See Refs. in Ref. (1)SUTTON : ELECTRON DIFFRACTION BY GASES AND VAPOURS. 39from 0 to x the interatomic scattering must oscillate like a sinecurve but with steadily diminishing amplitude; and so it musteither augment or diminish the atomic scattering and give periodiovariations in the total intensity.The greater Zij is, the more fre-quently will the function oscillate over a given range of e, and con-versely. In a more complex molecule, each pair of atoms gives riseto a periodic intensity term, so the total diffraction pattern is thesum of all these and of all the atomic terms. In addition to thescattering so far considered, termed " coherent " because theelectrons undergo no change of wave-length by collision, there is" incoherent " scattering which falls rapidly as 0 increases, shows noperiodic character, and adds to the background.The essential task, therefore, in the use of electron scattering fordetermining the interatomic distances in amolecule is to try toanalyse8 complex curve into its periodic components and then, knowing h,to calculate the distances corresponding to each one.A completelist of the distances could then be compiled. There are two waysof attempting this. The earlier one is the method of direct com-parison, wherein several models are assumed for the molecule underexamination, the theoretical curves of intensity against s [i.e.,(4x sin &e)/h] is calculated for them, and comparison of these withthe experimental intensity curve is made. The model giving thebest-fitting theoretical curve is taken as the solution.The later method, called the radial distribution method,14 is basedupon the idea that, strictly, a molecule should not be considered ascomposed of scattering points only but of scattering materialspreading throughout space, though varying greatly in density.Consequently, it theoretical intensity curve should be an integral,not merely a summation; and the curve giving the radial distri-bution of material density is related by a process of Fourier integralinversion to the curve which expresses the intensity of scattering asa funation of scattering angle.Just as the former can be used tocalculate the latter if a molecular model be assumed, so the formershould be calculable from the latter, experimental curve. Theapproximation which is made when a model is assumed, vix., thatscattering is only from points, is very good; but the correspondingassumption which was at first made in calculating the radial distri-bution curve from the scattering curve, vix., that the interatomicscattering is discontinuous, and gives only sharp maxima or minima,is not so good; and it constitutes a considerable drawback to theapplication of the method.Nevertheless, as will be shown later, themethod has been very useful.Before any relation of the distribution of matter with the way inl4 L. Pauling and L. 0. Brookway, J . Amer. Chena. SOC., 1936, 67, 268440 GENERAL AND PHYSICAL CHEMISTRY.which it scatters electrons can be attempted, we must know some-thing about its ability to scatter, ie., about the scattering functions9 for the atoms concerned. For fast electrons accelerated by apotential of 30 kv. or more, it may be shown that for any atom iwhere Zi is the atomic number of i, and fi is its ability to scatterX-rays (8 and h have their earlier meanings).Now X-rays arescattered because the electric component of the wave inducesvibrations in the electronic clouds round the atoms, but the waveletsscattered from different parts of the diffuse cloud round any atommay interfere destructively and this happens increasingly as 8increases; f i itself varies with 8, increasing as 0 decreases, and when0 = 0 it equals 2,. The meaning of the electron-scattering factorsis therefore that the screening of the nuclei by the electronic cloudsround them is less the greater 8 is, i.e., that the scattering a t lowvalues of 0 is predominantly by the electrons whereas a t large valuesit is predominantly by the n~clei.1~ Although 2, - j 6 increases as8 increases, the denominator [(sin +0)/hl2 increases so much fasterthat +% falls very rapidly.So rapid is the fall in the atomic back-ground and in the interatomic factors that the maxima and minimain the periodic functions on which the factors operate are reduced tomere inflexions on a rapidly falling curve, except a t small valuesof (sin +e)/h. Moreover, unless the relative importance of theinteratomic terms to the atomic ones is rather large, these inflexionsare shallow ; and in a microphotometer record of a diffraction photo-graph they are further obscured by the zigzag caused by emulsiongrain. Although this last difficulty can be mitigated by rotatingthe plate about the central spot when taking the microphotometricrecord,16 it is difficult to correlate more than five inflexions with thetheoretical curve, and furthermore, the calculation of the latter isextremely tedious unless the molecule is very simple.Although afew substances have been examined in this elaborate manner,l5p 1 7the rapid application of the method to large numbers of compounds,many of them quite complex, would have been impossible unlesssimpler approximate methods had been developed.The earliest and most drastic of these has proved, with but littlemodification, to be the most satisfactory. Although the blackeningof a, diffraction photograph diminishes steadily from the central spotoutwards, by a singularly impressive optical illusion the human eyel5 M. H. Pirenne, J. Chem. Physics, 1939, 7, 144.16 C. Degard, J. PiBrard, and W. van der Grinten, Nature, 1935, 136, 1421 7 L.Pauling and L. 0. Brockway, J. Chem. Physics. 1934, 2, 867SUTTON : ELECTRON DIFFRACTION BY GASES AND VAPOURS. 41sees dark and light rings superimposed on a slowly falling background,and these do not diminish in intensity nearly so rapidly as would beexpected from the full theoretical curve. Roughly speaking, theeye appears to subtract the background and divide by it. Now ifthis be done literally in the proper expression for the intensity,17we obtainIf further we assume that f is the same function of 8 for all atomsconcerned, say ZF(0), we obtain by cancellation of [l - F(8)I2,Ic" being a scale constant.Empirical tests have, in fact, shown that this expression reproduceswith surprising accuracy what the eye observes.Where this curveshows maxima, minima, double maxima, or shelves, the eye seesthem. Consequently, by observing visually the values of (sin @)/Aat which maxima or minima occur, and then comparing them withthose read off from the graph of expression (2) as a function of(sin 4 0 ) / ~ or (4x sin @)/A, one of the two steps necessary in correlat-ing the experimental and the theoretical curve is achieved. Theother step is the comparison of the observed and the calculatedintensities of the several maxima and minima. This cannot be donenearly so satisfactorily by eye especially if, as with the older cameras,there is, in addition to the coherent atomic scattering and theincoherent scattering, an " experimental " background caused byaccidental X-rays, or by scattering occurring outside one very smallvolume element near the vapour jet owing to inadequate pumpingand condensation.If a maximum appears denser than one imme-diately inside, it can be safely concluded that this feature shouldappear on the theoretical curve, but if the outer maximum appearsthe fainter, then this does not necessarily mean that the theoreticalpaxima must bear this relation, for the eye is conscious of some fallin background. Estimates of relative intensity are equally necessarywhether the method of analysis is that of direct comparison or radialdistribution, and in view of the limitations of the human eye as aninstrument for this purpose it is fortunate that, as a rule, intensity isless important than position.It is obviously desirable both in principle and in practice todevelop a method for observing diffraction which is objective andis simple to apply.Clearly, what is wanted is a means of obtainingan experimental intensity curve which can be correlated with theconvenient, simple theoretical intensity curve given by expressio42 GENERBL AND PHYSICAL CHEMISTRY.(2) ; a, means, in fact, of actually subtracting the background and ofdividing by it, so as to show the intensity of the fluctuations relativeto the background. Attempts to approximate to this were madeby various methods; by making a, microphotometer record with agraded positive, designed to compensate for the background, placedover the electron-diffraction negative,l8 or by using a rotatingsector in place of the positive.19 The use of a differential photo-meter has also been expIored.20 All these methods have onecommon disadvantage, however, in that they employ the usualdiffraction photograph as a basis.The range of intensity of scatter-ing within the range of 8 which is important is, however, very great,and commonly exceeds the limits of linear response of the photo-graphic emulsion. Attempts have therefore been made to photo-graph the electron scattering itself through a compensatingsector.21n 22 This necessitates having rotating mechanism in a high-vacuum chamber, which raises technical problems, but the results sofar published appear very promising, for several real maxima andminima, have been obtained on a microphotometer record even forhydrides such as ammonia and acetylene.If this method can besatisfactorily developed it offers a prospect of more accurate anddetailed analysis than has yet been possible without very greattrouble, and also of power to determine the structure of lightermolecules than could formerly be dealt with. It would also justifythe use of a rather more accurate expression for the scatteringintensity than (2). It will be remembered that in the latter theassumption was made that the X-ray scattering factor f is the samefunction Z F ( 8 ) of the scattering angle for all atoms in the molecule.Although this is a good approximation if the atoms are all in thesame period, it is far from satisfactory if they come from widelyseparated periods, and the difference can affect the relative intensi-cOies,15 or even to some extent the positions, of theoretical maximaand minima for any but diatomic molecules.Correction for thishas been made in some of the most recent determination^,^^ butprobably it is hardly necessary until the shape of the experimentalcurve can be measured more satisfactorily. If this can be done,.then, since the actual electron-scattering factors for the lighterelements decrease proportionately less rapidly with increasing 8 than18 V. E. Cosslett, Trans. Farachy SOC., 1934, 30, 981.19 P. Trendelenburg, Naturwiss., 1933, 21, 173; F. Trendelenburg andH. Franz, Ver6ff. Siemens-Konz., 1934, 13, 48 ; F. Trendelenburg, Physikal.Z., 1939, 40, 727.2O F. W.Seam, J . Opt. Xoc. Amer., 1936, 25, 162.21 H. Mark, private communication.P. Debye, Physikal. Z., 1939, 40, 66, 404, 507.2* D. P. Stevenson and V. Schomaker, J. dmeT. Chem. SOC., 1940, 62, 1913SUTTON : ELEUTRON DIFF'RACTION BY GASES AND VAPOURS. 43do those for the heavier elements, it should be possible to determinethe positions of the former in a molecule more easily than would beexpected if the atomic numbers are taken to be the scatteringfactors as in expression (2). Conversely, X-ray diffraction shouldemphasise the contributions of the heavier atoms, and consequentlythe two methods should be complementary. In principle, it iseven possible to determine the distribution of electrons withina molecule by studying the electron diffraction at small angles,when electron screening is very marked (p.40).15 Realisationof these theoretical possibilities depends, however, upon technicaldevelopments.A further complication in analysis, which is proving to be of somepractical importance, is the effect of thermal vibration on the diffrac-tion pattern. It has been shown ' 9 24 that if two atoms, i and j, arevibrating with a mean square amplitude (SZij2)Ay their scattering isexpressed bywhere Aij = Sn2(SZi,2),,[(sin~8)/h]2, and $i and xij have theirprevious meaning. The effect of the exponential term is to decreasethe amplitude of the intensity fluctuations, and this effect is moremarked the greater ( S Z i j 2 ) ~ v is, as might be expected, and thegreater 0 is.Consequently, it becomes essential to allow for this temperatureeffect when trying to match a theoretical curve to observed intensi-ties if 0 is large [for most molecules it is necessary if s = (4x sin + 0 ) / ~> 151 or if certain distances may be considerably affected byvibration even at low temperatures, i.e., if vibrational force constantsare low.It has been observed that the pattern for the stannous orplumbous iodides shows practically no effect of the 1-1 diffractionterm, and this is attributed to a large temperature effect arisingfrom the relative ease with which a valency angle can bend.25Terms which are affected by " free rotation " are likely to be greatlydimini~hed.~~ For a complex molecule, wherein it is possible thatthe force constants may vary greatly, the periodic terms for differentpairs of atoms will therefore vanish in order as 8 increases, the" soft '' distances going first and the " hard " ones last.This mayhelp in the analysis of the pattern, for it may make possible someseparation of parameters ; the " hard " distances being determinedfirst, from the outer rings, and then taken as known in order thatthe " soft " distances may be derived from the inner rings. Thegeneral considerations underlying present and potential methods ofI = Ee- A{j#z$3 (sin xi j) /xi j24 R. W. James, Physikal. Z., 1932, 33, 737.25 Work by Dr. M. W. Lister at Oxford, shortly to be published44 GENERAL AND PHYSICAL CHEMISTRY.analysis having been presented, it remains to say something abouttheir detailed application, and to assess their reliability.The radial distribution method is, in principle, a much neaterway of making use of the same data, vix., the positions and therelative intensities of maxima and minima, than the method ofdirect comparison; but it is every whit as empirical in the formused,26 and moreover, since it uses the same experimental data itcan give no more information though it may give it more easily.In the original form of the method the visual intensity curve wasused, and there was no attempt to take into account its full shape aswas required by the integral form of the radial distribution expression.ooD(r) = E‘”i, s61(s)[(sin sr)/sr]ds,Instead it was assumed to consist only of instantaneous maxima andminima, ie., a summation of a few terms was used instead of anintegral in order to make the method practicable.This procedure,as judged by the other more certain though more pedestrian method,was not very satisfactory, for it gave poor resolution of distances andunreliable values. A simple explanation of this was that it gavetoo much importance to the early, intense maxima which were knownto be difficult to place accurately. More detailed theoretical con-siderations indicated the desirability of multiplying the visualintensities by an exponential weight factor sk2e- “k’, sk being theposition of the Eth maximum, and a being chosen so that the ratioof the weighted intensities of the strongest and the weakest maximumis as 10 : 1.27 * This increases the relative importance of the outermaxima, and is found to give improved resolution.More recentlyan attempt has been made to take into account more fully the shapeof the curve by considering maxima and minima, not as lines, butas segments of cosine curves.28 This treatment leads to a moreelaborate but quite convenient approximate expression which is26 For discussion of the validity of the treatment, see P. Debye and M. H.Pirenne, Ann. Physik, 1938, 33, 617; P. Debye, Physikal. Z., 1939, 40, 573;and (27).27 C. Degard, Bull. SOC. chim. Belg., 1938, 47, 770.28 J. Walter and J. Y . Beach, J . Chem. Physics, 1940, 8, 601.* (Added in proof, 18/2/41.) This procedure is the one followed at Oxfordin accordance with the tradition brought over by Dr.L. 0. Brockway in1937. It appears not to be the same as that which is indicated in the paperby J. Walter and J. Y . Beach 2 8 and which is said to be the common practiceat Pasadena and at Princeton. The report of the A.C.S. meeting at Balti-more in 1939, to which American authors sometimes refer, gives no exactaccount of the procedure. It could be wished that a full description mightbe published in a readily accessible journalSUTTON : ELECTRON DIFFRACTION BY GASES AND VAPOURS. 45found to give still better resolution and placing of the peaks ofthe r2D(r)-r curve when applied to some theoretical intensity curvesand compared with the older procedures. The radial distributionmethod in one form or another is very useful for a preliminaryanalysis, for it makes possible the elimination of certain modelsand gives a useful guide to those which should be investigatedby trial and error.Direct comparison of theoretical intensitycurves with the visual appearance of the photographs must,however, still be regarded as the more tried and trusted method;and it is still necessary for settling the final model and forestablishing the limits of accuracy in determination of the severalparameters,The use of a set of (sinax)/ax strips29 in conjunction with anelectric calculating machine, or of sets of punched cards in a machineof the International Business Machine or Hollerith type, brings thelabour of computation within reasonable limits.The subjective nature of the most-used methods of interpretationof electron-diffraction photographs has caused the results to bereceived with some scepticism by exponents of other methods ofmeasuring interatomic distances ; but experience has shown that itcan be as accurate as any.are thenumber of important, independent parameters in the molecule,* therelative importance of the interatomic and the atomic scatteringterms,t the physical characteristics of the material-for veryvolatile substances, being less easy to condense, give less clearphotographs, whereas non-volatile ones may show large temper-ature effects when vaporised-and the accuracy with which voltage(for determining A) and distances in the camera and on the platescan be measured.The tests of the method are of two kinds, wix., direct, by comparingresults obtained for the same molecule or the same bond by electron-diffraction and by other methods; and indirect, by comparingvalues obtained for interatomic distances with those predicted forthem from atomic covalent radii which were derived through othermethods.Some comparisons of the first type are given in the follow-ing table which, though not exhaustive, should suffice to show thatthe agreement is on the average within 1 yo and that the discrepanciesare not noticeably systematic.28 Originally published by Prof. P. C. Cross of Brown University, Provi-dence, Rhode Island, U.S.A. * The importance of the parameter ZiJ is determined by the magnitude ofn#& compared with the corresponding products for other interatomic distances,where n is the number of occurrences of Zu.t The factor determining the clarity of the rings in a diatomic molecule is2(#i#j)/(#6s +- t,hj2), which has a maximum value when #$ = tjj.The factors which decid46Bond.Cl-CIBr-Br1-1I-GI s-sP-PAs-AsC-CGe-Ge .c-cCaL4.cCaL-Car.car.-car.CClSi-ClGENERAL AND PHYSICAL CHEMISTRY.Electron diffraction.Substance. value, A. c1* 2.009 l7Br2 2.2891:7251 2.642*3Oa;'2.1 SdV)2.21 a32.441.52-1.552-41 :l1.341.341.461.54 a'1.391-75 l72.0101 - ;her method. - Value.1.9842%11 :' 2.662.3153;'2.102-082.25 342.501.542.44 :: 1.331.331-462,9411-531.39 4a1.75 l o1.98 l67Method.*SP.SP.SP.SP.X-Rayx-Rayx-Rayx-RayX-Rayx-Raysp.SP.SP.X-RayX-Rayx-Rayx-RayDiff., %.+ 1.24-0.35-0.75 - 0.65 --- 1.8 - 2.5- 1.25 + 0.75 + 0.75 - 0.1 + 0.650 .o0.0 + 1-5-* Sp. = spectroscopic, (c) = crystalline, (1) = liquid, (v) = vapour.Under the second heading come many measurements ; especiallythose of bonds between sulphur atoms in H,S, and (CH3)2S2,abetween possible pairs in carbon, nitrogen, and oxygen, and betweencarbon and most of the elements with which it forms covalentbonds.35 The agreement, usually to within 1%, between so manysuch values and the corresponding series of the atomic covalent radiiwhich were mostly derived from spectroscopic or X-ray measurementsmay be considered to show both that the radii are accurate and thatthe electron diffraction method is reliable within the limit given.There are, it is true, some larger discrepancies to which attentionhas been directed, especially for the carbon-halogen bonds inmethyl fluoride and chloride for which the electron-diffractionvalues are 1.42 and 1.77 A.as against 1.385 and 1.66 A. by spectrumanalysis; 45 i.e., they are 2.5% and 6.6% greater than the latter.This disagreement has not yet been resolved,2 but it is perhapssignificant that later spectroscopic work may be interpreted as givingL. R. Maxwell and V. M. Mosley, Physical Rew., 1936, 49, 199.31 B. E. Warren and J. T. Burwell, J . Chem. Physics, 1935, 3, 6 .32 N. S. Gingrich, ibid., 1940, 8, 29.33 L. R. Maxwell, V. M. Mosley, and S. B. Hendricks, ibid., 1935, 3, 698.34 C.D. Thomas and N. S. Gingrich, ibid., 1938, 6, 659.35 L. Pauling, " The Nature of the Covalent Bond," Cornell, 1939, Chap. V.36 L. Pauling and L. 0. Brockway, J. Amer. Chern. SOC., 1937, 59, 1223.37 L. Pauling, A. W. Laubengayer, and J. L. Hoard, ibid., 1938, 60, 1605.3t3 W. L. Penney, Proc. Roy. Soc., 1937, A, 158, 306.39 E. H. Eyster, J . Chem, Physics, 1938, 6, 580.40 L. Pauling, H. D. Springall, and K. J. Palmer, J . Amer. Chern. Soc.,41 G. Herzberg, F. Patat, and H. Verleger, J . Physical Chem., 1937, 41,42 L. 0. Brockway and 5. M. Robertson, J., 1939, 1342.43 L. 0. Brockway and J. Y. Beach, J . Amer. Chein. Soc., 1938, 60, 1836.44 D. P. Stevenson and J. Y. Beach, ibid., p. 2872.45 G. B. B. M. Sutherland, Trane. Paraday Xoc., 1938, 34, 385.1939, 61, 927.123; R.M. Badger and S. H. Bauer, J . Chem. Physics, 1937, 5, 599SU!PJ!ON : ELECTRIC DIPOLE MOMENTS. 47a value of 1.71 A. for GC1 in methyl chloride,*6 which reduces thediscrepancy to 3.6%.The investigations of bond lengths in elements could not beaffected by possible error due to ignoring the different rates ofvariation of the X-ray scattering factor, f, with 8 for differentelements, but the second class of investigations might be, for itconsists of measurements on compounds, and these were nearly allmade without any correction. The fact that they show as goodagreement as the results in the first class indicates that the effect isnot serious.This method of investigating the distribution of matter in amolecule is clearly a very potent one already ; and it will steadilybecome mom effective.We may&hope, not only that it will givedetailed information about small molecules, but that useful approxi-mate methods of analysing the scattering by large, complex mole-cules will be developed, so that a limited but useful amount ofinformation can be derived from them. A step in this directionhas been taken in the examination of pp‘-di-iododiphenyl ether forthe purpose of measuring the 1-0-1 angle; 33 when instead oftaking, as is usually done, the length of an arbitrary bond, or anAngstrom unit, as the unit of length in which all the coefficients ofs in the (ZiZ’sin Zijs)/Z,js terms am expressed, the distance betweeniodine atoms was chosen as the unit in any particular model.Ob-vioualy, if the scattering were due only to the two iodine atoms, themaxima and minima in the graph of theoretical intensity against swould then come a t the same s values, whatever the value of the1-0-1 angle, for the curve is only a function of this angle if thelatter comes into the expression for EII in whatever units are chosen :and if ZII itself is the unit it does not. Hence, if the intensity curvesfor a series of models be so plotted, maxima due primarily to 1-1scattering will remain stationery but others will not. In this waythe former can be detected, and the iodine-iodine distance can becalculat ed.It is possible, also, that the rotating-sector device could be used t ocompensate for the scattering of all the terms save one, such as aniodine-iodine term, so that this alone would be photographed andused to determine one distance of particular interest.(ii) Electric Dipole Moments.Developments in MethocEs of Determination and in the Theory ofPoktrisation.In the last ten years there has been no startling development inexperimental methods of measuring electric dipole moments ; and‘13 G.B. B. M. Sutherland, J . Chern. Physics, 1939, 7, 106648 GENERAL AND PHYSICAL CHEMISTRY.as yet there has been no quantitative determination of the quadripolemoment of a molecule, although qualitative indications of suchmoments have been observed.47 By far the most important stillare the early methods based on the study of the variation withtemperature af the electric polarisability of a molecule, or on itsvariation with the frequency of the applied field as in the commonmethod of measuring it a t medium radio frequencies, through thedielectric constant, and a t visible frequencies, through the refractiveindex.48 There have been improvements in details of apparatusand experimental technique, especially in the design of constant-frequency oscillator^.^^ Greater emphasis has been placed onmeasurements in the vapour phase owing t o the difficulties nowknown to be inherent in measurements made in solution (see p.51).The “ beam ” method, in which the dispersion of a beam of moleculesby an inhomogeneous field is studied, has been developed 50 but isstill far from being readily or generally applicable.The early quantitative theory thereof was developed for diatomicmolecules, and made possible determinations of moments of someion-pairs which could not be attempted by any other m e ~ n s .~ 1Recently, the theory for more complex molecules with the momentparallel to a possible axis of rotation has been developed; 52 and ithas been shown that the dispersion of these ought to depend uponthe first power of the field strength, as well as upon the square asfor molecules with moments perpendicular to possible axes ofrotation. Experimental application of the theory to ammonia 53showed the existence of this linear effect but gave a value of 0.5 D.for the moment, which is only one-third of that (1.46 D.) given byother methods.There was one development which may become of considerable47 F.H. Miiller, Wiss. Verofl. Xiernens-Werken, 1938, 17, No. 1 , p. 20.48 (a) P. Debye, “ Polar Molecules,’’ New York, 1929; ( b ) C. P. Smyth,“ Dielectric Constant and Molecular Structure,” New York, 1931 ; (c) P.Debye, Trans. Faraday SOC., 1934, 30, 679; (d) H. A. Stuart, “Molekul-struktur,” Berlin, 1934; ( e ) 0. Fuchs and K. L. Wolf, “ Dislektrische Polaris-ation,” “ Hand- und Jahrbuch der chemischen Physik,” Band 6, I, Leipzig,1935; (f) R. J. W. LeFBvre, “ Dipole Moments,” London, 1938.49 ( a ) J. D. Stranathan, Rev. Sci. Instr., 1934, 5, 334; 1938, 6, 396;( b ) L. G. Groves and S. Sugden, J., 1934, 1094; (c) I. E. Coop and L. E.Sutton, J . , 1938, 1269; ( d ) L. G. Groves, J . , 1939, 1144; also 48(f) and aforthcoming paper by G.I. M. Bloom and L. E. Sutton.50 ( a ) R. Fraser, “Molecular Rays,” Cambridge, 1931; ( b ) I. Estermannand R. G. 5. Fraser, J. Chem. Physics, 1933, 1, 390.51 (a) H. Scheffers, Physikal. Z., 1934, 35, 425; ( b ) W. H. Rodebush,L, A. Murray, and M. E. Bixler, J . Chem. Physics, 1936, 4, 372; (c) R. G. J.Fraser and J. V. Hughes, ibid., p. 730.The reason for this discrepancy is not yet clear.52 H. Scheffers, Physikal. Z., 1940, 41, 89.b8 Idem, ibid., p. 98SUTTON : ELECTRIC DIPOLE MOMENTS. 49importance. It has long been known that in principle it is possibleto derive values of the electric dipole moment of a diatomic moleculefrom the absorption coefficients for its pure rotation lines in the farinfra-red, or of a single bond in a symmetrical molecule, such as theC-H bond in methane, if the absorption coefficient of the linescorresponding to its optically " active " bending frequencies can bedetermined.The requisite measurements are extremely difficult,but a rather easier alternative is to study the variations of therefractive index in the region of the spectral lines.= The mostrecent investigations of the latter kind have given a value of 1.18 D.for the moment of hydrogen chloride,55 which agrees quite well withthe most recent value of 1.085 D. from the polarisation of thev a ~ o u r , ~ ~ and of 0.307 D. for the moment of the C-H bond.57 Anestimate of the latter value has also been made, on the basis of aless exact theoretical treatment, from measurements of absorptioncoefficients for rotation-vibration overtones in chloroform, and isin good agreement with the above value, being 0.3-0.4~.~* Atheoretical calculation by the molecular orbital method has given0.53 D., but this method is likely to yield a high value.59 Theimportance of such a determination of one single bond-moment in acomplex molecule is that it makes possible the evaluation of themoments of the other bonds : otherwise, oiily differences relative to,say, the C-H bond are known.The theory of the polarisability (commonly called polarisation)methods has been considerably developed in the period underreview.The first consists offinding how a polar molecule will behave in an applied uniformelectric field of known strength; the second consists of finding howthe strength of the field acting on a molecule in a mass of dielectricmaterial is related to that of the applied external field.The firstquestion has been answered with a great deal of certainty by theapplication of quantum rnechanic~.~~"~ 54 It is satisfactory to findthat, except at low temperatures, the result obtained initially by asimple classical treatment is valid. The predicted divergence fromthe classical curve is just detectable in tlie case of hydrogen anddeuterium chlorides.56 The second question is more complex andis less susceptible of precise treatment. The original treatment byClausius and Mosotti involved the assumption that the field in54 J. H. Van Vleck, " The Theory of Electric and Magnetic Susceptibilities,"Oxford, 1932.5 5 R.Rollefson and A. H. Rollefson, Physical Rev., 1935, 48, 779.5G R. P. Bell and I. E. Coop, Trans. Faraday SOC., 1938, 34, 1209.5 7 R. Rollefson and R. Havens, Physical Rev., 1940, 57, 710.5 8 B. Trimm and R. Mecke, 2. Physik, 1935, 98, 363.59 C. A. Coulson, Proc. Clamb. Phil. SOC., 1940, 36, 509.There a,re two parts to the problem50 GENERAL AND PHYSICAL CHEMISTRY.question could be made up of three parts : (1) the primary field,(2) the field from the charges on the outside surface of the dielectricmass and on the inside of a spherical cavity centred about the testmolecule, (3) that from the substance inside the cavity. I n thesimple treatments, the third force was assumed to be zero. Further-more, the dipole molecule was assumed to have no action upon thesurrounding medium, and therefore this effect could have no reactionupon the dipole even if it were polarisable. In the ideal case of adilute gas these assumptions are perfectly adequate; but as thedensity of the material increases they become increasingly invalid,and the departures from the simple theory are commonly ascribedto dipole-dipole interaction or association.If the dielectric materialis a solution, there is also dipole-solvent interaction.I n order to take account of these effects a number of treatmentshave been proposed which have had some measure of success. Nodetailed account of them will be given here. Early developmentswere fully reported in 1936,5 and recent developments, the mostimportant of which are concerned with applying and developingL.Onsager's treatment,60 have not come very much nearer to solvingthe practical problem, although they have helped greatly to elucidatethe character of intermolecular forces.61 The general nature of theproblem is now clear and it is obviously difficult to solve exactly.The forces between molecules at short distances depend upon severalfactors including some which are difficult to express simply, such as" shape," and the spatial distribution of positive and negativocharges ; or which are difficult to determine with sufficient accuracy,such as " nearest distance of approach " or partial " saturationeffects " when molecular orientation occurs in local fields.Thesedifficulties can be overcome to some extent by introducing con.veiiient experimental parameters such as the Kerr constant, but notusually with complete quantitative success. Progress has becomesemi-empirical. It was once hoped to find a solution of the problemwhich would make it possible to measure the polarisation of a polar69 J . Amer. Chem. SOC., 1936, 58, 1486.61 See (a) W. H. Rodebush and C. R. Eddy, J. Chem. Physics, 1940, 8,424; (b) J. Norton Wilson, Chem. Reviews, 1939, 25, 377; (c) A. Piekara,Proc. Roy. Soc., 1939, A, 172, 360; ( d ) J. G. Kirkwood, J . Chem. Physics,1939, 7, 911; (e) C. P. Smyth, J. Physical Chem., 1939, 43, 131; (f) M. E.Hobbs, J . Chem. Physics, 1939, 7, 849; (9) C. J. F. Bottcher, Physica, 1939,6, 59; (h) A.H. White, J. Chem. Physics, 1939, 7, 758; (i) R. H. Cole, ibid.,1938, 0, 385; (j) K. L. Wolf, H. Frahm, and H. Harms, 2. physikal. Chern.,1937, By 38, 237; (k) F. H. Miilk, Physikal. Z., 1937, 38, 498; (E) P. Debye,Physikal. Z., 1935, 36, 100, 193; (m) F. R. Goss, J . , 1940, 1752; (n) E. G.Cowley and J. R. Partington, J., 1938, 1598; ( 0 ) G. Thomson, S., 1938, 460;K. Hignsi, Sci. Papers Inst. Phys. Chem. Res. Tokyo, 1937, 31, 311; andfurther references thereinSUTTON ; ELECTRIC DIPOLE MOMENTS. 61solute in any chosen solvent at any arbitrary concentration (or evento use a pure polar liquid) and then, by introducing experimentalparameters of solvent and solute which could readily be measured,t o calculate accurately the moment of a free molecule such as wouldbe obtained from the polarisation of a dilute gas.This goal has notyet been achieved and it is now clear that in general no simplemethod can be perfect, nor perfect method simple; but if thesolvent be non-polar, the solution infinitely dilute, and if, further,the solute be similar in shape to one for which there are sufficientlyextensive data, a reasonably satisfactory answer can usually beobtained .The lack of a satisfactory theoretical basis for treating solutions inpolar solvents is felt particularly in the investigation of large andhighly polar molecules, especially those of biological interest suchpeptides and proteins which can only be examined in aqueoussolution. An empirical relation between solute polarisation anddielectric increment has been developed from studies o f solutions ofthe amino-acids, for which the moment can be estimated, if certainassumptions about the structure be granted, or derived from otherexperimental data.62 When this relation is applied to proteinsolutions, it gives values of several hundred Debye units for themoments of the proteins.These results are discussed later (p. 56).Certain unusual solute-solvent interactions have been discovered.The moments of the hydrogen halides are greater in solution thanin the vapour phase, possibly because the importance of the ionicstrucfure is raised by the increased dielectric constant of themedium.63 Donor molecules and certain polar hydrogen corn-pounds, such as chloroform, form what appear t o be weak hydrogenbonds.aAnother matter in which there has recently been considerableclarification of ideas is that o f the effect of intramolecular vibrationsupon the electric polarisability.There appear to be two possiblecases to consider, according to whether the non-vibrating moleculeis polar or not. In the former case the effect of vibration is likelyto be of minor importance, but in the latter it may be major. Thesubject is sufficiently important and timely to warrant a separate,detailed discussion in a later section; and here only a brief recitalof conclusions and results will be given. It would be expected thatin hydrogen-containing polar molecules such as water or ammonia,62 J. Wyman, Chem. Reviews, 1936, 19, 213; S. Arrhenius, Physikal.Z.,1939, 40, 534.F. Fairbrother, Trans. Paraday SOC., 1937, 33, 1507.84 ( a ) S . Gksstone, ibid., p. 200; (b) D. L1. Hamrnick, A. Norria, and L. E.Sutton, J., 1938, 175552 GENERAL AND PHYSICAL CHEMISTRY.in which the valencies may be executing anharmonic vibrationsrelative to each other, the substitution of deuterium for protiumwould, through decreasing the amplitude of vibration, reduce themean intervalency angle and hence increase the moment. Such adifference has been observed between light and heavy ammonia,65but in water the smaller predicted difference has not been detected.A related isotope effect is to be expected in vibrating diatomicmolecules, for in hydrogen chloride the interatomic distance shouldbe greater than in deuterium chloride, and the moment should bechanged by Ar .dp/dr.Now, dispersion measurements in theinfra-red give values of (dp/dr)2, but do not give dp/dr; so althoughthe magnitude of the difference can be predicted the sign cannotbe obtained from these data alone. If the bond is essentiallycovalent, then when it is stretched indefinitely the molecule dissoci-ates into two neutral atoms, which form a non-polar system, andconversely, if the internuclear separation is diminished to zero,another non-polar system is obtained. Between these two limits thesystem must be polar, so there must be a distance for which themoment is a maximum; but it is impossible to predict whether thiswill be greater or less than the equilibrium distance, and thereforewhether dp/dr is positive or negative.The numerical agreement ofthe theoretical and experimental differences for hydrogen chlorideand deuterium chloride is as good as can be expected, and further-more the algebraic difference leads to the interesting conclusion thatdp/dr is negative.66 On the modified ion-pair model of the mole-cules with a partial electronic charge on each atom it should ob-viously be positive; and consequently the failure of this predictionshows that so simple a model is far from adequate.*Consideration of the effect of vibration in non-polar moleculessuggests that there are two possible effects : one is that the fieldbends the symmetrical molecule and induces a " distortion " moment ;the other that the molecule bends under thermal impact and orientsas a polar molecule while still bent.The latter is the physical6 5 J. M. A. De Bruyne and C. P. Smyth, J . Amer. Chena. SOC., 1935, 57,G 8 R. P. Bell and I. E. Coop, Trans. Parad&y SOC., 1938, 34, 1209.* (Added in proof, 18/2/41.)1203.It has been pointed out by Prof. M. Polanyi,in a discussion, that either the relative positions given for the minima of thepotential energy curves for H-CI and H+Cl- (see Pauling, ref. 35, p. 43, orJ . Amer. Chem. SOC., 1932, 54, 988) or the fact that the sum of the ionicradii (0 + 1.81 = 1.81 A.) is greater than the normal interatomic distance(1.28 A.) should mean that the ionic character increases when the ordinaryHCl bond is slightly extended. The observed negative value of dp/dr is notin agreement even with this more sophisticated molecular model, and istherefore of great interestSUTTON : ELECTRIC DIPOLE MOMENTS.53picture which has usually been employed as a basis for treatments ofthe effect of “free rotation” in such molecules as ethylene di-chloride. More careful consideration shows that the two processesare statistically indistinguishable, and it may therefore be best thatwhat were called “ atom polarisation ” and the “ polar character offlexible molecules ” should both be called “ vibration ” polarisation.Provided that the vibration be harmonic and of small amplitude, theresulting polarisation is likely to be independent of temperature.There is little doubt now that most of the anomalous polarisationsof molecules which should on the usual structural principles benon-polar, are really large vibration polarisations, and are not dueto a solvent effect as was once suggested (Section iii).It had long been realised that the difficulty of measuring the atompolarisation in polar molecules was a bar to the accurate determin-ation of electric moments.If the polarisation of the vapour of thesubstance can be studied over a suficient range of temperature,usually lOO”, this dieculty can be overcome, but otherwise it isnecessary to make an empirical allowance for PA as 5-15% ofPE,67s 68 or to study the infra-red dispersion. A number of directmeasurements of vibration polarisation have recently been carriedout thus, by following the changes of refractive index of liquids, or ofsolutions, in passing from the visible frequencies through the infra-red region until the effect of orientation absorption becomes mani-fest.69 They have provided several instances where the values forthe atom or vibration polarisation PA are somewhat different whenmeasured in this way and by determination of the sum of the atomand the electron polarisation from the curve of P plotted against1/T. It is suggested that the error is probably in the dielectricmeasurements.One factor which has frequently complicated the interpretationof dipole-moment data is the relatively large moments which a polarbond or group may induce in other parts of a molecule. Such waspresumed to be the reason for the increase in moment with chainlength in homologous series, the incorrectly large values calculatedfor the angles between polar valencies, and part of the differ-ences between the moments of aromatic and aliphatic com-po~nds.~~b* 48e* 701 71 In general, it is necessary to allow for theseinduction effects in quantitative stereochemical applications, and8 7 See ref.48 ( e ) , p. 299.6 8 L. G. Groves and S. Sugden, J., 1937, 1779.6Q C. H. Cartwright and J. Errera, Proc. Roy. Soc., 1936, A, 154, 138; also‘O H. M. Smallwood and K. F. Herzfeld, J. Amer. Chem. Soc., 1930, 52,71 L. E. Sutton, Proc. Roy. Soc., 1931, A, 133, 668.refs. 54, 55, 57.191954 GENERAL AND PHYSICAL CHEMISTRY.in attempts to calculate individual bond moments, or to detect" abnormalities " in moments of a whole molecule.The difficulty of accurately so doing has been a great nuisance.Several attempts have been made, and considerable success isclaimed for the most recent of them. The problem consists essen-tially of finding the field acting over a given element of volume,finding the polarisability of the matter in that element, multiplyingthem together, and then integrating the resulting vectors throughoutspace. The primary dipole is com-monly regarded either as a partial ion-pair (which we have alreadyseen is inadequate for another purpose, p.52), or as a mathematicaldipole placed somewhere along the valency, but it is improbable thateither approximation holds well at close quarters, i.e., within 1 or2 A. The field acting at a point in the molecule will not be that ofthe primary dipole alone, but of it and all the moments which it hasinduced elsewhere in the molecule.Finally, the field induced bythe primary dipole may react on the bond in which this occurs, whichis always polarisable in some degree, and so gives in effect a modifiedprimary dipole. The mean polarisability of bonded atoms in weakuniform fields is known; but the fields near polar bonds are notweak, and moreover, because they are not uniform it is reallynecessary to know the variation of the polarisability over the domainof the atom. The primary dipole has usually been treated as amathematical one in order to be able to use the simpler field expres-sion, and some reasonably plausible rule for placing it along thevalency has been adopted.The field a t a point was a t first taken asthat due to the primary dipole in a vacuum; later, some allowancewas made for the effect of the intervening medium by dividing thefield, in a vacuum, by the dielectric ~onstant,'~ ie., assuming thatthe primary dipole and the point under consideration are embeddedin a mass of uniform material. The effect of the field on the polaris-able substance is then treated in one of several ways. The atommay be regarded as a polarisable point located a t the nucleus, inwhich case the external field at its centre is multiplied by the atomicpolarisability; or it may be assumed that, since the polarisation ismainly due to displacement of the outer electrons, the polarisablematerial is located a t points in the valencies, and the field a t thesepoints along the valencies, determined by some plausible rule, is thenmultiplied by polarisabilities calculated for the bonds.73 Recentlya more elaborate treatment has been devised ; it is designed to allowNone of these tasks is easy.72 L. G. Groves and S. Sugden, J., 1937, 1992.73 See inter aEia papers by (a) (Mrs.) C. G. LeFbvre and R. J. W. LeFBvre,J . , 1936, 1130; G. C. Hampson and A. Weissberger, J., 1936, 393; C. P.Smyth and G. L. Lewis, J . Amer. Chem. SOC., 1940, 62, 721SUTTON : ELECTRIC DIPOLE MOMENTS. 55for the variation of the external field over the domain of the atomwhich is being p~larised.'~ The projection of the supposedlyspherical atom is divided into squares, the field in each of which iscomputed, and then the moment induced in an anchor ring generatedby rotation of the circular projection about the axis of the primarydipole is derived ; finally, it is assumed that the ratio of the momentinduced in the spherical atom to that induced in the anchor ring isthe same as that of the volumes.Granting this assumption, madein order to convert a triple integral into a double one which canbe evaluated graphically, it follows that the angle between theinduced and inducing dipoles is the same in the anchor ring and theatom.It is clear that no method has so far provided an exact treatment,and indeed there is Little hope t h a t this could be done with the presentdata and without more developed computing techniques than cancommonly be employed.It is therefore important to assess thevalue of these comparatively easily applied methods. When thematerial in which the dipole is being induced is 5-6 A. away fromthe primary dipole, as at the other end of a benzene ring, they allgive much the same result, so all of them can be trusted; butwhen the separation is only 1-2 A. they all give different results,and may therefore all be untrustworthy. An alternative statementis that if the induced moment is only one-tenth of the primarymoment it can be accurately calculated, but if its ratio is 0.5 to 1-0then probably it cannot be. The choice between the methodsmust be largely empirical, i.e., it must be decided by the internaland external consistency of the results; but it is not easy toapply tests.It is encouraging that the anchor ring method ofGroves and Sugden gives tetrahedral values for the valency anglesin methylene dichloride and chloroform,74 in agreement withelectron-diffraction results, when applied to the data for methylchloride and these two compounds, although the other methodsdo not : but it leads 72 to a value for the C-0 bond moment of2.28 which is less than that of the C-0 bond, 2-30, contrary to allexpectation.The phenomenon of dipole absorption by liquids occurs in thefrequency range of the applied field between the limits where a polarmolecule can still orient freely and follow the changing phase of thefield perfectly, and where the friction between it and the surroundingmolecules is so great that it has completely stopped orienting.Suchabsorption means that electromagnetic wave energy is being degradedt o heat. It can give information about the size of the dipolarmolecule, the viscosity of the liquid, and the magnitude of the74 L. G. Groves, J., 1938, 119756 GENERAL AND PHYSICAL CHEMISTRY,orienting dipole.75* 7 6 ~ 77 An interesting example of its f i s t applic-ation comes from studies on some hydroxylic compounds, in whichit is found that absorption takes place a t frequencies so high thatthe molecule as a whole cannot rotate and therefore that it arisesfrom the rotation of the hydroxyl group independently of the mainbulk of the molecule. The information about the effective micro-scopic viscosity (which may differ from the bulk viscosity) shows thatit is relatively high for hydroxylic compounds, as is to be anticipatedfrom the known tendency of such compounds to associate throughhydrogen-bond formation.75* 76 It has occasionally proved useful asan alternative method of measuring the moment ; as, e.g., in the caseof certain proteins, for which results in good agreement with thosecalculated from the dielectric increment were obtained." Althoughboth methods incorporate the same empirical relation, they are stillsufficiently different for this agreement to be impressive. The verylarge values obtained, of several hundred Debye units, are striking,but they do not necessarily give much information about the struc-ture of the protein molecule, because it is not known over whatportion of the molecule a single-electron transference may be madeto give the betaine structure responsible for the large moment, orindeed if only one electron is transferred.If the molecule were linear, and one electron were transferred fromend to end, a moment of 500 D., such as is found (the range is 300-1200 D.), would correspond to a length of about 100 A., i.e., about100 bond lengths projected on the axis of a zigzag chain.Contri-butions from side groups being neglected, the molecular weight ofsuch a molecule would be 1400-2000, which is only 2-3% of theactual value. This simple model of the dipolar molecule is thereforeincorrect. If the molecule were a square sheet, with every groupbonded to four others, the molecule weight would indicate a side ofcu.70 A. ; if the pattern were a more open one, based on hexagonalrings, the side would be about 100 A. A single-electron transferencealong a side or diagonal would then give a moment of the observedorder of magnitude. If the molecule formed a hollow cube, the cube7 5 See ref. 48 (a), 48 (c), and 61 (cz-Z).7 6 ( a ) P. Debye, Chem. Reviews, 1936, 19, 171; ( b ) E. Fischer, Physikal.Z., 1939, 40, 645; ( c ) E. Fischer and G. KLges, ibid., p. 721; ( d ) A. Bud6,ibid., p. 603; ( e ) K. Schmall, Ann. Physik, 1939, 35, 671; (f) H. Fricke andL. E. Jacobson, J . Physical Chem., 1939, 43, 7 8 8 ; ( 9 ) E. Fischer and 3'. C.Frank, Physikal. Z . , 1939, 40, 345; (h) H. Muller, Ergebn.exakt. Naturw.,1938, 17, 164; (i) P. Girard and P. Abadie, Physikal. Z., 1938, 39, 691;(j) E. Plotze, Ann. Physik, 1938, 33, 226; ( k ) E. Keutner and G. M. Pota-penko, Physikal. Z . , 1937, 38, 635; ( 1 ) J. Malsch, ibid., 1936, 37, 849; (m)G. Martin, ibid., p. 665.7 7 (a) J. L. Oncley, J . Amer. Chem. SOC., 1938, 60, 1115; ( b ) J. D. Ferry andJ. L. Oncley, ibid., p. 1123SUTTON : ATOM POLARISATION. 57side would be about 3 0 ~ . or 42 A. for closed or open structuresrespectively. One-electron transference along a side, a face diagonal,or even cube diagonal, giving moments of 150-350 D., would sufficeto explain the smaller moments observed, but not the medium orlarge ones. The results for circular or spherical models would be ofsimilar orders of magnitude.If, therefore, the moments have realsignificance they show that there is not an electron transferencealong the whole length of a linear molecule, nor only a single-electrontransference in a cubic model : but they do not by themselves allowof any choice between the other possibilities.Dipole orientation can occur in solids, both crystalline and non-crystalline, but does not necessarily lead to dipole absorption. Thepolar molecules in certain solids, such as some forms of hydrogenbromide and hydrogen sulphide, have surprising freedom to orientin an external field, for dielectric constants as bigh as 20 or 40 areobserved. This phenomenon is obviously of great value in studyingthe physics of crystals. Although there is a fair degree of correlationwith the thermodynamic properties, and with molecular parameters,there is as yet no satisfactory quantitative theory of the effect.78-81Orientation in non-crystalline solids such as organic glasses or waxescan give rise to absorption.The phenomenon is of great practicalimportance, as one cause of dielectric losses in insulators at radiofrequencies.82* 83* 84 A quantitative theory of it, too, is as yet lacking.(iii) Atom Polarisation, the Orientation Polarisation of FlexibleMolecules, and Molecular Vibrations.If a molecule is placed in a uniform electric field it becomes" polarised " ; i.e., relative to its former state it develops an electricmoment which reduces its potential energy in the field.The polarisation is commonly considered to arise by three inde-pendent processes, known as electron, atom, and orientat ion polaris-ation.The first occurs because the electrons are pulled one way andnuclei another, so the original distribution of electrons round nucleiis elastically distorted. The second occurs if the molecule contains7 8 See (a) C. P. Smyth, Chem. Reviews, 1936, 19, 329; (b) A. Eucken, 2.Elektrochem., 1939, 45, 126.79 (a) W. 0. Baker and C. P. Smyth, J . Amer. Chem. SOC., 1939, 61, 1695;( b ) idem, ibid., p. 2063; ( c ) idem, J . Chem. Physics, 1939, 7, 574; ( d ) idem,J . Amer. Chem. SOC., 1939, 01, 2798.(a) A. H. White and W. S. Bishop, ibid., 1940, 62, 8 ; ( b ) A. H. White,B. S. Biggs, and S. 0. Morgan, ibid., p. 16.A. Muller, Proc.Roy. SOC., 1937, A, 158, 403; 1938, A, 166, 316.82 E. B. Moullin, J . Inst. Elect. Eng., 1940, 86, 113.83 D. R. Pelmore and E. L. Simons, Proc. Boy. SOC., 1940, A, 175, 253, 468.84 F. C. Frank and W. Jackson, Trans. Faraday SOC., 1940, 36, 44058 GENERAL AND PHYSICAL CHEMISTRY.dipolar bonds; for then the atoms bearing an effective negativecharge will be pulled in the opposite way from those bearing aneffective positive charge. Hence, a dipolar bond may change itslength, or the angle between two such bonds may be changed.Both of these processes are functions of the field strength and ofelastic constants of the molecule alone.The third type of polarisation occurs if the molecule as a whole hasa permanent electric dipole moment, which is there even whenthere is no external electric field.It occurs by the twisting of thepolar molecule in the field under the action of a couple. If all themolecules were to twist freely, the smallest external field wouldsuffice to align them all and bring about a high degree of order,Thermal collisions tend to preserve disorder, however, and hence itis said that thermal effects oppose the field. Prom these consider-ations it may readily be seen that, whereas electron and atom polaris-abilities depend only upon the nature of the molecule, orientationpolarisability should decrease as the temperature rises. Further-more, owing to the very different damping influences which wouldaffect electrons, atoms, and molecules if an alternating field wereapplied, it would not be expected that the three kinds of polarisationshould all contribute at all frequencies.Orientation should cease inthe ultra-short radio-wave region in liquids or in the far infra-red ingases, atom polarisation should cease before the visible frequenciesare reached, but electron polarisation could go on until the hard X-rayfrequencies are reached.The expected dependence of polarisation * upon temperature andupon frequency of the applied field is realised ; for if the polarisationof a molecule is not differeqt at infra-red and at medium radiofrequencies, then neither is it dependent upon temperature, andconversely ; and, furthermore, while the polarisations of dihomo-atomic molecules such as those of hydrogen, nitrogen, and oxygenare not appreciably different at the visible and the infra-red fre-quencies, those of polar molecules such as hydrogen chloride, or ofnon-polar molecules with opposed polar bonds (e.g., carbon dioxide)are different.The foregoing picture of polarisation as occurringby three independent processes thus appears adequate andacceptable.”” 54When the polarisation of a compound such as ethylene dichlorideis considered, fresh problems seem to arise. If the molecule werefixed in the tram-position, by dipole-dipole or by “exchange”forces, it would be non-polar, and would undergo only electron andatom polarisation, as does tram-dichloroethylene ; but if the two* The words “ polarisability ” and “ polarisation ” are commonly usedinterchangeablySUTTON : ATOM POLARISATION. 59halves could rotate freely relative to each other, then there would bemolecules with every conceivable configuration, and therefore withdipole moments varyhg continuously from zero to a maximumvalue pmax..It can easily be shown that if all configurations areequally probable, the mean polarisation 4 x w / 9 k T y correspondsto a mean moment I f rotation were not perfectlyfree, but if fixation in the tram-position were nevertheless incom-plete, E. would have a value intermediate between 0 and ( 1/2/2)prnaz.Moreover, since the configurations of higher potential energy arealso those of higher moment, as the relative importance of thesewould increase with rising temperature, so too would the moment.The theoretical work which has been done on this problem falls intoseveral periods.At first, the aim was to derive an expression, by classical methods,for the mean moment or polarisation as a function of temperatureif it were assumed that the main forces between the rotating groupsare between the dipoles therein.1 Later work comisted in elaborat-ing the classical treatment to take into account the momentum co-ordinates which showed that if there is a change in the moments ofinertia with configuration some configurations will be more favouredthan others.Crudely speaking, it is found that the spinning ethylenedichloride molecule, to take a concrete example, tends to spin asnearly as possible about the line joining its main masses, the chlorineatoms, and to reduce its moment of inertia about this axis; sothe trans-configuration is favoured.The factor representing thiseffect, by which the exponential term representing that of potentialenergy is multiplied, is called the statistical weight factor. Quantummechanics was also applied to the problem, and besides providingan equivalent weight -factor representing the variation in the densityof quantum states with configuration, it showed that at low temper-atures the effect of zero-point energy is important. In all theselater treatments the potential function taken had to be one withsimple mathematical properties, rather than one derived from aphysical model of the molecule. It was hoped that the applicationof these theoretical results to experimental data would make possiblethe calculation of parameters expressing the intramolecular forcesbetween the rotating groups.It was realised that these are prob-ably not dipole-dipole forces alone, and that it is most desirable toinvestigate them experimentaIly.2, 3,Seo refs. (ii) 48 (d, e ) ; and article by K. L. Wolf and 0. Fuchs in Freuden-berg’s ‘‘ Stereochemie,” Vienna, 1932, p. 774.J. E. Lennard-Jones and H. H. M. Pike, Trans. Pur&y Soc., 1934, 30,830.W. Altar, J. Chem. Physics, 1935, 3,460.J. Y. Beach and D. P. Stevenson, ibid., 1938, 6, 635.of (1 / 4 2 ) h a x . 60 GENERAL AND PHYSICAL CHEMISTRY.I n this concentration upon one part of the problem, the nature ofthe polarisation process as a whole in flexible molecules was over-looked.There seems to have been no clear realisation that atompolarisation and the orientation polarisation of molecules bent bythermal collision are one and the same thing. On a crude dynamicview it appeared indeed that they are distinct, for it would seem that,if the period of bending were large compared with the time of relax-ation of the molecule, there should be both orientation and atompolarisation; whereas if it were small there should still be atompolarisation (in addition to electron polarisation) .5 The absurdityof this view became obvious, however, on considering what thepolarisation of a freely rotating molecule would be; for, since it isgreater the smaller the force constant resisting distortion? in such amolecule it should obviously be infinite, which it certainly is not.The expression for atom polarisation can be derived in classicalmechanics, which is adequate for the purpose,(ii) 54 by finding theaverage displacement Z caused by a field P, and multiplying this by4xNei/9P, where ei is the effective charge moved.Thus.+ mwhere V(x) is the potential energy for a displacement x.The expression for the orientation polarisation of a molecule suchas ethylene dichloride is obtained by substituting a mean value ofp2 in the relation between Po and p.The latter is obtained by oneaveraging process expressed thus :-where 8 is the inclination of the instantaneous moment of the mole-cule to the axis of the applied field F , and + is the remaining angularco-ordinate for the three-dimensional problem.48 Another averagingprocess gives2 @e - V(V) lkTd#-2 - P max P -4' .. . . (3) I.- V(v)'kTd+where # is the angular displacement from the trans-configuration, andV(+) expresses the potential energy in terms of it.(a) A. E. Finn, G. C. Hampson, and L. E. Sutton, J., 1938, 1254; (b)D. L1. Hammick, G. C. Hampson, and G. I. Jenkins, ibid., p. 1263; (c) I. E.Coop and L. E. Sutton, ibid., p. 1269SUTTON : ATOM POLARISATION. 61Now the general expression for the complete, mean polarisationof a flexible molecule is :where p(#) expresses p as a function of#.This includes the effect of internal potential energy, field, per-manent moment [now defined as the constant term in p($)] andtemperature. It therefore gives the whole of the polarisation exceptthe electron polarisation, which would require further internal co-o r h a t e s and potential functions and has been omitted for simplicity.It can immediately be seen that (4) is the general form of (l), andalso that it is composed of the two stages represented by (2) and (3).Consequently the polarisations visualised from the two differentdynamical pictures are statistically indistinguishable.I n thespecial case when p = pl# and V(#) = V0#12/2, P = 4xNpI2/9V0.For oscillations of small amplitude, when both of these conditionsare usually satisfied, the polarisation is independent of temperature,but if depends less simply upon #, or if there is any anharmonicity,the polarisation is temperature-sensitive. This clarification of theprinciples is a necessary preliminary to the consideration of detailedresults, which may now be undertaken.It has long been realised that undetermined atom polarisation maygive rise to appreciable, and even serious error in the determinationof electric dipole moments by the common method of taking theorientation polarisation as the difference of the total polarisationsat radio and visible frequencies.Extrapolation of the latter, i e . ,the molecular refractivity, to infinite wave-length from observationsin the visible range obviously cannot remove the error, because thesedata are unaffected by absorptions in the infra-red region, whereatom polarisation is first manifested. Unless, therefore, the polaris-ation can be studied over a range of temperature, or the refractivitycan be followed through the infra-red, as has been possible so faronly for relatively few of the many substances investigated (p.53),atom polarisation must be either ignored or assessed empirically.From studies of the polarisations of compounds with known PA andPE, various authors concluded that PA is about 5-15% of PE.(ll) 679 68Hence, it came about that atom polarisation was frequently regardedas a nuisance, but hardly ever as a major phenomenon. It is nowknown, however, that this is not true, for such polarisation can belarge and of possible importance as a means of determining certainmolecular parameters. It can be large, not only in molecules of the6 N. R. Davidson and L.E. Sutton, J., 1939, 34762 GENERAL AND PHYSICAL CHEMISTRY,ethylene dichloride type, in which small moments are held opposedhy relatively weak forces, but also in those wherein strongly dipohrbonds or groups are held opposed to one another by ordinaryvalency forces.From the early days of dipole measurements it was known thatsome molecules which ought not to be polar appeared in fact to beso ; i.e., they showed differences between their radio- and visible-frequency polarisations which, since they were 2 0 4 0 % of PE,appeared not to be atom polarisation. Examples of this are poly-nitroberuenes, p-benzoquinone, metallic acetylacetonates, and otherco-ordination compounds. The measurements had been made insolution, and it was therefore suspected that the anomalies might bedue to a solvent effect.Two possible ways in which this could comeabout were suggested.', In 1938,5(~) however, it was shown thatthese anomalous polarisations were just as great in the vapour phitse,and furthermore, that they were independent of temperature.It was therefore concluded that they must after all be atompolarisations. *This conclusion was further supported by a quantitative examin-ation of the data. Molecules which contained the same polar groups,able to oscillate in the same way, should have equal atom polaris-ations. The anomalous polarisations for cyanogen, trans-dicyano-ethylene (fumaronitrile), and p-dicyanobenzene are in fact 8.3, 9.8,and 11.9 c.c., respectively;5(c)s 99 l o and those for several symmetricalp - benzoquinones , t etrame t hylc ycZo but anedione, and carbon sub-oxide are all 8-10 C.C.Next, when the appropriate relation be-tween PT-PE, p (the vibrating moment), and Vo (the force constantof bending) was used to calculate any one of these quantities fromexperimental values of the other two, at value in agreement with anexperimental value of it, or of reasonable magnitude, was obtainedin almost all instances. Finally, by using the force constant derivedon the assumption that the polarisation in p-benzoquinone arisesfrom vibration of the C=O bonds perpendicular to the plane of thedouble bond, and taking the appropriate reduced mass, a period of3.0 x 10-13 see. was calculated for this mode of vibration. Theexperimental value is 2.8 x 10-13 secS5(c)The general conclusions are that, since the mechanism of atom or7 (a) H.0. Jenkins, J . , 1936, 862; ( b ) S. H. Bauer, J . Chem. Physics, 1936,8 F. C. Frank and L. E. Sutton, Trans. Faraday Soc., 1937, 33, 1307.B H. E. Watson and R. L. Ramaswamy, Proc. Roy. SOC., 1936, A, 156, 144.10 See forthcoming paper by G. I. M. Bloom and L. E. Sutton.* At that time the unreality of the distinction between this and thepolarisation of molecules bent by thermal impact was not realised; SO therelative merits of the two theories were gravely debated.4, 458SUTTON : ATOM POLARISATION. 63vibration polarisation is unrelated to that of electron polarisatioii,no simple numerical relation can be expected. Although theempirical rule that PA is 5-15% of PB holds reasonably well forcompounds containing polar bonds with less than 2.5 D., it fails ifthe bond moments itre greater.In general, too, it is not an additivefunction of atoms, although rough additivity may be observed in aseries of related compounds such as the di-, tri-, and tetra-acetylace-tonates of certain metals. Compounds with essentially the sameoscillating systems have the same vibration polarisation. Fromobserved vibration polarisations it is frequently possible to calculatethe force constant of vibration, and by combining this with an ob-served frequency of vibration it may be possible to elucidate themode of vibration. To allow for PA in a, determination of dipolemoment, the best procedure is to calculate it roughly by using therelation PA = 4xNpl2/9VO (pl = vibrating moment, Vo = forceconstant) for each one-dimensional oscillation in the molecule ; butif the component moments are less than 2.5 D., the 5% rule can beapplied with sufficient accuracy.If the term atom polarisation is to be retained it is best applied tothe temperature-invariant polarisation arising from vibrations ofsmall amplitude, as has been done in the f'oregoing section.Theterm " vibration polarisation " may then be used in a more generalsense to include both this and the temperature-variant polarisationarising from oscillations of large amplitude. Some of the problemsconnected with the analysis of experimental data in order to derivea potential function for these oscillations have already been indi-cated.When the important mode of vibration is a partial rotationabout a single bond the potential function is likely to be complicated.The ideal procedure would be to derive it directly from the data andthen to analyse it into its constituent terms, such as dipole-dipoleinteraction, dispersion forces, and exchange forces. This has beendone only in so far as that if the observed moment is (1 /fi)pmS, andis temperature-invariant, the function is presumed to be one withvalues always much less than kT. In general, it is impracticable :so usually, a function is assumed, possibly as a result of consideringthe probable interatomic forces; it is then tested, and if it does notfit it may be adjusted until an improvement is effected.2s3*4s10Such a trial-and-error method could give a unique solution (cf.theprocedure of direct comparison for analysing electron-diffractionphotographs) if enough curves were tried, and if the vibrations weresufliciently large for the whole of the curve to be explored. Theavailable temperature range is, however, usually so small that onlya part can be followed, and therefore there is little point in tryingmany curves. This limitation has been realised with increasin64 GENERAL AND PHYSICAL CHEMISTRY.clearness by workers in the subject, but it may be further emphasised.The figure shows typical potential curves which have been tested 4s loagainst the data for ethylene dichloride-so far the most carefullyinvestigated substance of this type.The clear area represents theregion within which kT>V(t,h). Although the importance of themolecules with the larger values of $ is relatively greater, becausethey make larger contributions to the polarisation, the exponentialfalling off in the number of molecules in these configurations isdecisive, and consequently little is known about the shape of thecurve above the fringe of the shaded “ veil of ignorance ”. Never-theless, in spite of this limitation, which applies also t o electron-diffraction studies, useful conclusions have been reached fromdeterminations of the lower portions of the curves. The combinedresults of the two experimental methods show that the repulsionsbetween the groups in the ethylene dihalides are not predominantlydue to C-Hal dipole-dipole forces but to exchange forces betweenthe several possible pairs of atoms, commonly called “steric ”forces, They are greater in ethylene chlorobroniide than in ethylenechloride, and still greater in ethylene dibr~rnide,~, 11 which is theorder of increase of halogen diameter but not of C-Hal moment,Vibration polarisation thus far seems a satisfactory explanationl1 (a) J.Y. Beach and K. J. Palmer, J . Chem. Physics, 1938, 6, 639; ( b )J. Y. Beach and A. Turkevich, J . Amer. Chem. SOC., 1939, 61, 303SUTTON : ATOM POLARISATION. 65of the facts. The Raman spectra have, however, been interpretedto mean that there is an equilibrium between molecules of thesesubstances fixed in cis- and trans-configurations.12 Because thishypothesis seemed inherently less probable, the polarisation datahave not hitherto been used to test it, but an interesting result isfound when this is done.It is clear13 that if V,, is the difference of energy between thetruns- and the cis-form, and s is the ratio of the weight factors,then Nmns/Ncz8 = seVdRT whence log, N,/N,-Iog,s = V,,/RT, so ifloglo NJN, is plotted against 1/T a straight line should be obtained.This can be done because N,/Nc can be calculated fromPobs.if assump-tions be made about the magnitude of the rotating moment; andit results in surprisingly good straight lines, both for solution andfor vapour data, which intersect the ordinate axis to give values fors favouring the trans-form, as is expected.10 Hence it appears thatexisting polarisation data support this hypothesis; but it can beshown that they are really somewhat ambiguous.On the trunsoscillation theory the equivalent plot would be of loglo (1 - p)/( 1 + p) *against Vm/RT, and the form of this curve can be calculated forpossible potential functions. It is found that, although the cosine,linear, or parabolic functions do not give rectilinear curves, thecurvature over ranges which correspond to the V; calculated andthe AT employed is usually small, and might be barely detectable.1°It is therefore clear that, until polarisation measurements over awider range of temperature are available, no choice between thetheories can be made. Electron-diffraction studies are similarlyindecisive, because the proportion of the cis-isomer would be small.A parallel suggestion, also based on Raman spectra studies, hasbeen made that acetylene tetrachloride consists of a cis-form andtwo others formed by relative rotation of the groups through 120"either way, and that the difference of potential energy is 1.1 kg.-cals.13 This can be definitely rejected, for the molecule would thenbe much more polar than in fact it is; and this conclusion is sup-ported by electron-diffraction re~u1ts.l~ The latter, also, are in dis-agreement with the further suggestion that cyclohexane is flat.15Polarisation measurements have not so far given any informationabout the forces between the non-bonded hydrogen atoms in ethane-like molecules.l2 A.Langseth, H.J. Bernstein, and B. Bak, J. Chern. Physics, 1940,8,415.l3 A. Langseth and H. J. Bernstein, ibid., p. 410.l4 V. Schomaker and D. P. Stevenson, ibid., p. 637.l5 A. Langseth and B. Bak, ibid., p. 403.* p (see 4) is an average value of cos +, where C$ is the angle between themoments of the two halves of the molecule projected on to a plane perpendicularto the axis of rotation.REP.-VOL. XXXVII. 66 GENERAL AND PHYSICAL CHEMISTRY.Other points of interest in relation to interatomic forces andmolecular structure which have recently emerged are the following.If the radius about which bound chlorine atoms rotate is suilicientlylarge, as in 00'-dichlorodiphenyl, it appears that London dispersionforces of attraction may cause the cis-position to be favoured so faras steric forces allow.16Measurements on benzil have been interpreted to mean that themolecule has the two benzoyl groups fixed skew a t approximately90"; because the moment is equal to that for free rotation, vix.,( 1 / 1 / 2 ) b L , and is temperature-invariant, yet in the analogous stil-bene dichlorides and hydrobenzoins rotation is far from being freeas is shown by the difference between the mean moments for theracemic mixture of enantiomorphic forms and for the meso-form.17The electric moment of hydrogen peroxide in dioxan solution agreeswith the skew right-angled configuration predicted for it from theinteraction between the electron clouds round the oxygen atoms,(i)but it is not by itself decisive.An X-ray examination of the liquidstructure is, however, more definitely favourable.l* Both theelectron-diffraction and the electric-moment evidence relating to thefreedom of rotation about the S-S bond in hydrogen disulphide,dimethyl disulphide,lg~ 2o and sulphur monochloride 20, 21 are in-conclusive.The moment of hydrazine agrees with that predictedfor a fixed skew configuration,")4 as do those of some alkyl and arylderivatives.22 The infra-red absorption of hydrazine 23 also leads tothe conclusion that it has the configuration required by Penney andSutherland's theory.")Inhibition or reduction of rotation about a single bond, owing toresonance with a double-bonded structure, apparently occurs incertain esters, but not in all, as is shown by the magnitude and thetemperature characteristics of the polari~ation.~~.25 The factsthat the diacetyl potential function is a parabolic one with the largepotential barrier (V,,) of ca. 23 kg.-cals., indicate that there is similarstiffening in this molecule.10 Electron-diffraction evidence shows16 G. C. Hampson and A. Weissberger, J . Amer. Chem. Soc., 1936, 58,2111.1 7 C. C. Caldwell and R. J. W. Le Fbvre, J., 1939, 1614.18 J. T. Randall, PTOC. Roy. SOC., 1937, A, 159, 83.19 D. P. Stevenson and J. Y . Beach, J . Amer. Chem. Xoc., 1938, 60,*O C. P. Smyth, G. L. Lewis, A. J. Grossmrtnn, and F. B. Jennings, 111,21 K. J. Palmer, ibid., 1938, 60, 2360.22 H. Ulich, H. Peisker, and L. F. Audrieth, Ber., 1935, 68, B, 1677.Zs N. F. Fresenius and J. Kazweil, 2.physikal. Chem., 1939, B, 44, 1.24 R. J. B. Massden and L. E. Sutton, J., 1936, 1383.26 S. Mizushima, and M. Kubo, Bull. Chem. SOC. Japan, 1938, 13, 174.2872.ibid., 1940, 62, 1219SUTTON : STUDY OF MOLECULAR STRUCTURE. 67the fixation even more definitely, both in this compound and inglyoxal.26 It is not possible to show its existence in ccy-b~tadiene.~~The mutual repulsion of rotating groups falls rapidly if the separ-ation is increased above one bond length ; although certain complic-ations arise from doing this in practice, since it necessitates eitherintroducing further axes of rotation or placing a benzene ring betweenthe groups. The rate of fall depends partly upon the polarity of thegroups ; and, as might be expected, it is less if they are highly polar(e.g., CH,*CN) than if they are but weakly so (e.g., CH2Br, OCH,,It is premature to say that the experimental data have beencompletely interpreted; they frequently are too meagre as yet togive very definite answers to all the questions which it is hoped theymay settle.The full nature of the forces between rotating groups,and the way in which they are affected by the medium when asolvent is present, have yet to be fully elucidated.NH2).28. 29,30(iv) The Application of Electric Dipole Moment Measurements andof the Diflraction of Electrons by Vapours to the Study of Molecular8tructure.The potentialities of the foregoing experimental methods havefrequently been explained in review article~,~l, 32 and to someextent they have been taken for granted in the previous section,but for convenience they may be summarised afresh.The information forthcoming from an electron-diffraction studyis comprised in the perfect radial distribution curve.It is a listof all the possible interatomic distances in which each item is taggedaccording to the scattering powers of the two atoms to which itrefers. Since the scattering factors are known, it is frequentlypossible, in simple molecules, to discover to what pair of atoms eachdistance refers, and thus to elucidate the complete spatial distribu-tion of the nuclei. In favourable cases it is therefore possible togive very exact and complete answers to questions of stereochemistryinvolving interatomic distances though not of order in space, forenantiomorphs obviously give the same diffraction pattern.Forlarge molecules the method usually is not so useful, for it gives so26 J. E. Lu Valle and V. Schomaker, J. Amer. Chem. Soc., 1939, 61, 3520.27 V. Schornaker and L. Pauling, ibid., p. 1769.28 Ref. (ii) 48 (e), p. 402 ; also (ii) 48 (f), Chap. V.J. A. A. Ketelaar and K. J. Palmer, J. Amer. Chent. SOC., 1937, 59,2470.P. Trunel, Ann. Chim., 1939, 12, 93.31 See refs. (Section i) 1, 2, 3, 6, 7, 9, 14, 35; 4, 5, 48.32 (a) C. P. Smyth, J. Org. Chem., 1936, 1, 17; (b) idem, Pubns. Amer.Assoc. Adv. Sci., No. 7, “ Recent Advances in Surface Chemistry and ChemicalPhysics,” 1939, 9868 GENERAL AND PHYSICAL CHEMISTRY.many data that they cannot be disentangled. When the moleculeis highly symmetrical this dficulty may, however, be overcome.It is also possible, as a general rule, to discover whether two atomsare bonded covalently or not : if they are 3 A.or more apart theyare not bonded effe~tively.~~ The usefulness of this criterion isgreatly augmented by the empirical rule that the covalent radius ofan atom is constant to a first approximation. Exceptions to thisrule have been interpreted to mean that the bond is not what it wasthought to be, i.e., that it is not of the supposed order (single,double, or triple), or that the orbitals used in it are not those used inthe majority of the covalent bonds which the two atoms in questionusually form, e.g., that it may be ionic, or that, let us say, d orbitalsmay be used in addition to s and p orbitals.Electron-diffractionstudies have therefore been very extensively and intensively em-ployed to investigate bond character and so to detect the phenomenonof wave-mechanical " resonance " in molecules.The electric dipole moment measures the asymmetry of distribu-tion of the negative charge relative to the positive in a molecule.It is a vector quantity, and for many purposes may be consideredas the vector sum of moments characterising the individual bonds.They are not particularly characteristic, for they can often react oneach other very markedly by induction : 35 nevertheless, it is possibleto visualise and in some degree to calculate the primary, undisturbedmoment of a bond. The vector character of the total momentobviously makes it a function of the molecular symmetry.Stereo-chemical problems can therefore be elucidated even in quite largemolecules, provided that the problem in question is related in a simpleway to the symmetry of the molecule; examples of this will occurlater .Since the moment of a polar molecule, group, or bond is a measureof the electrical asymmetry therein, it has often been used to studyproblems related to the distribution of valency electrons. At-tempts have been made to correlate the moments of bonds with thedifferences between electronegativity co-ordinates characteristic ofthe atoms : 36 they have met with considerable qualitative success,but it is not yet possible t o predict the moment of a bond withnearly as much accuracy as its length.The magnitude of the33 See ref. (i) 35 (p. 154).34 N. V. Sidgwick, Chem. Reviews, 1936, 19, 183.35 L. E. Sutton and L. 0. Brockway, J . Amer. Chem. SOC., 1935, 57, 473.36 ( a ) J. G. Malone, J . Clzem. Physics, 1933, 1. 197; ( b ) M. G. Malone andA. L. Ferguson, ibid., 1934, 2, 99; ( c ) C. P. Smyth, J . Physical Chem., 1937,41, 209; ( d ) idem, J . Arner. Chem. SOC., 1938, 60, 153; ( e ) F. T. Wall, ibid.,1939, 61, 1051; 1940, 62, 800; (f) H. C. Brown, ibid., 1939, 61, 1483; (9)K. L. Wolf and H. Harms, 2. physikal. Chena., 1939, By 44, 359SUTTON : STUDY OF MOLECULAR STRUCTURE. 69moment of a bond, group, or molecule may also be used to elucidatemajor problems concerning their structure, as in the cases of theisocyanides 34 and the proteins (see p.56).By comparing the observed moment of a system with the valueanticipated for it from the roughly known values of bond momentsit is possible to detect abnormal electron distributions. The alge-braic differences of moment between aliphatic and aromatic,ethylenic, or acetylenic compounds have been correlated with theelectron displacements consequent upon wave-mechanical reson-ance.37 The logical basis of the use of abnormalities of either bond-length or electric moment to establish the reality of this phenomenonis the same. If, in a large number of compounds abnormalities arefound which are in the sense predicted from the resonance hypothesisthen, although other effects may cause abnormalities which occa-sionally confuse or obscure the issue, we may reasonably concludethat the hypothesis is correct, and that it is an essential part of thetruth though not necessarily the whole truth. Agreement in a fewisolated cases means nothing, but agreement in a large numberis significant : and agreement between independent experimentalmethods is still more significant.") 6 It is on such a statistical basisof substantial agreements that the experimental case for the exist-ence of resonance rests. One further general remark about the char-acteristics of dipole moments and of bond lengths may be made.The moment of a bond is generally more sensitive to disturbancesthan is its length.The percentage change caused in the former byresonance (unless it introduces no fresh electronic asymmetry) isgenerally greater than in the latter, and the presence of nearby polarbonds also has much more effect.This means that such disturbancesare easier to detect, but also easier to confuse : and so it lays furtheremphasis on the desirability of not relying on one method alone.8tereochemistry.-There have been several interesting recentadvances not yet reported upon.It has been possible from electron-diffraction studies to assignspatial configurations to many simple organic and inorganic mole-cules of the type AB,, AB,, AB,, AB,, and AB6.38 These configur-ation are supported by the electric moments when both data areavailable. A recent example is that tellurium tetrachloride hasbeen shown to have an apparent moment of 2 .5 4 ~ . in benzene'' ( a ) Refs. (ii) 48 (f), 72 ; ( b ) L. E. Sutton, Trans. Paraday SOC., 1934,30,789;( c ) L. E. Sutton and R. J. B. Marsden, J., 1936, 599; ( d ) D. I. Coomber andJ. R. Partington, J . , 1938, 1444; ( e ) H. L. Goebel and H. H. Wenzke, J .Amer. Chem. SOC., 1937, 59, 2301; (f) idem, ibid., 1938, 80. 697; ( 9 ) J. A. C.Hugill, I. E. Coop, and L. E. Sutton, Trans. Paraday SOC., 1938, 34, 1518.38 See N. V. Sidgwick and H. M. Powell, Proc. Roy. SOC., 1940, A, 176,15370 GENERAL AND PHYSICAL OHEMISTRY.solution,39 which, since it is too large to be explained as atom polaris-ation on probable values of v, or Vo (p. 62), is probably real and showsthat the molecule is unsymmetrical. The electron-diffractionpattern indicates a distorted pyramidal structure, such as mightbe derived from a trigonal-bipyramidal arrangement of five valencieswith one of them-probably equatorial-occupied by an unsharedpair of electrons.40The configurations are in general agreement with quantum-mechanical theories of directed valencyY41 but they may also besuccessfully correlated on simple principles which are more readilyapplied by the chemist.38 These are (1) that a fully-shared valencygroup has the most symmetrical of the possible arrangements;(2) if the stereochemical arrangement corresponding to a fully-sharedvalency group is known, the configuration of a molecule with a groupof the same size, but with some of the electrons unshared, can bederived by supposing that each pair of unshared electrons fills theplace of a covalency. These rules are very satisfactory except forcompounds of the transitional elements, for which it is less easy todefine the term " valency group ".Electron-diffraction work on compounds of the type SOX,,S0,X2,21 and PSX,,43 where X is a halogen, indicate that thevalencies are not directed along the axes of a regular tetrahedron,but that the XSX or XPX angles are less, while the OSO angle isgreater, than 109" 28'.This change is the opposite of that to beexpected from steric repulsion forces, and therefore it is presumedto mean that the bonds are not all single. This conclusion issupported by the evidence of bond lengths and bond moments (p. 73).It may be mentioned that in the silicon analogues of methylenechloride and chloroform the Cl-Si-C1 angle was found to be 110" 31" 44 and therefore not to show the effects of repulsion between chlorineatoms which are observed in the carbon compoundS, where the angleis 121'5 2°.35 This is to be expected, since the central atom islarger and the C1-C1 distance is consequently greater in the formercompounds.One more of the classical stereochemical problems in organicchemistry has finally been settled by dipole-moment measurements.It has been shown that the labile and stable forms of several sub-C.P. Smyth, A. J. Grossmann, and s. R. Ginsburg, J . Amer. Chem. SOC.,1940, 62, 192.40 D. P. Stevenson and V. Schomaker, ibid., p. 1267.41 ( a ) See ref. (i) 35, Chap. 111; (b) G.E. Kimball, J . Chem. Physics, 1940,42 L. 0. Brockway and J. Y . Beach, J . Amer. Chem. Soc., 1938, 60, 1836.43 J. Y. Beach and D. P. Stevenson, J . Chem. Physics, 1938, 6, 75.44 I;. 0. Brockway and I. E. Coop, Trans. Faraday SOC., 1938, 34, 1429.8, 188SUTTON: STUDY OF MOLEOULBR STRUCWRE. 71stituted phenyldiazocyanides are stereoisomers of each other andthat the configurations allotted to them from their chemical pro-perties are correct, the former being cis and the latter trans.45 Theexperimental method employed was that of substituting a linearpolar group in the para-position and determining the effect on themoment. It was fortunate that the cis-form of azobenzene had beendiscovered a short time previously,46 since its moment was necessaryfor the proper interpretation of the other data.Its high polarity(p = 3.0 D.) agrees with the X-ray evidence 47 that it is indeed the&-compound.The proof, from the polarisations of the vapours, that p-dinitro-benzene and beryllium acetylacetonate are non-polar 6 ( ~ ) shows that,as required by the wave-mechanical concept of resonance, thenitro-group has an axis of symmetry along the bond joining it toother groups, and that the acetylacetonate ring is planar and hasan axis of symmetry passing through the metal atom and the middlecarbon atom. The symmetry of the nitro-group is demonstratedfurther by dipole and electron-diffraction studies on tetra-nitr~methane.~sIt is to be expected that compounds of the type (I) will be flat if’the valency angle of X is greater than 120°, and bent along theline X-X if it is less than 120°, and if X is CH,,0, S, or Se 49 it is clear that a flat structure means anon-polar molecule, while a bent one will be polar\/\ \/ if the Ph-X bonds are polar.Dipole evidenceshows that except when X is 0, the molecule is(1.1 bent. The exception has been explained as aresult of resonance. From this it might be expected that if X is>NMe, a similar result would be f0und.~0 The evidence is incon-clusive, however, for there is a small moment of 0-4 D.49 which mayindicate that the ring system is bent, or that this is flat but that thenitrogen bonds are non-planar and there is cis-trans-isomerism, withthe cis-form polar.The development of methods whereby the valency angle of anatom or group X in the compound XPh, can be determined, andallowance made for certain interactions which cause error, has beenreported upon previ~usly.~ One further possible cause of error, the46 R.J. W. Le FBvre and H. Vine, J., 1938, 431.46 G. S. Hartley, J., 1938, 633; G. S. Hartley and R. J. W. Le FBvre, J.,47 J. M. Robertson, J., 1939, 232.*@ (IWiss) I. G. M. Campbell, (Mrs.) C. G. Le Fbvre, R. J. W. Le FBvre,6o L. E. Sutton and G. C. Hampson, Tram. Farachy SOC., 1935, 31, 945.’” n;) X1939, 531.A. J. Stosick, J. Amer. Chem. SOC., 1939, 61, 1127.and E. E. Turner, J., 1938, 404; see also ref. (4) (p. 37)72 GENERAL AND PHYSICAL CHEMISTRY.effect of solvent, has now been eliminated by a series of polarisationmeasurements on the vapours of diphenyl ether and suitablederivatives, which give a value of 124" &- 5°.51 This is a reasonableagreement with that of 128" & 4" from solution measurements 5Oand also with one of 118" 3" obtained from an electron-diffractionstudy of pp'-di-iododiphenyl ether.52 The angle is therefore cer-tainly near 120" in these ethers, and not 110" or less as in aliphaticethers, or other oxygen compounds.This agrees with one of thedeductions from the " resonance '' hypothe~is.~~ There is as yetlittle evidence to show whether or not another deduction-that thebenzene rings may be coplanar-is correct, but some dipole evidencehas been interpreted to mean that they are not, and that the planeof one is perpendicular to that of the other.53The structures of several large molecules which prove to have highsymmetry have recently been established by electron-diffractionstudies, notably those of hexamethylenetetramine, phosphorustri- and pent-oxide, arsenious oxide,54 phosphorus s~lphoxide,~~and tetranitr~methane.~S The first five of these have structuresbased on a tetrahedral cage with nitrogen, phosphorus, or arsenicatoms at the four corners and oxygen atoms or methylene groupsbridging the six sides. The compounds in the " quinquevalent "states have four more oxygen or sulphur atoms attached outside tothe phosphorus atoms.The high degree of symmetrymade possiblevery good determinations of bond lengths in these compounds,some of which will be mentioned in the following sub-section.Tetranitromethane has the nitro-groups arranged in regular tetra-hedral fashion; they are symmetrical about the C-N bonds, andvibrate with an amplitude of ca.20" about skew mean positions.Electric moments have proved valuable for confirming the squarearrangement of four valencies attributed to the nickel, palladium,platinum, and gold atoms in certain of their complex compounds.This has usually been because the moments enable the cis- andtrans-forms to be differentiated and identified, the former being themore polar of the two. The first example was that of nickel in theglyoximes. Only one form of the p-chlorophenyl-n-butyl derivativecould be isolated : it had a moment nearly the same as that of nickelglyoximes which did not contain polar substituents in the hydro-carbon radicals, and it was therefore presumed to be the trans-51 I.E. Coop and L. E. Sutton, J., 1938, 1869.52 See ref. (i) 33, p. 705.53 K. Higasi and S. Uyeo, Bull. Chem. Xoc. Japan, 1939, 14, 87.54 (a) L. R. Maxwell, S. B. Hendricks, and L. S. Deming, J . Chem. Physics,1937, 5, 626; (6) G. C. Hampson and A. J. Stosick, J. Amer. Chern. SOC.,1938, 60, 1814.65 A. J. Stosick, ibid., 1939, 61, 1130SUTTON : STUDY OF MOLECULAR STRUCTURE. 73form.56 Complexes of the type [NiX,(PR,),] have been isolated inonly one form which is apparently trans if X = C1, Br, or I and R =Et, Pra, or Bus, but cis if X = Both cis- and trans-platino-complexes of formula [PtX,Y,] have been isolated, wherein X maybe one of several electronegative atoms or groups and Y is analkyl or aryl cyanide, isocyanide, sulphide, phosphide, arsenide, orantimonide : the isomers have large differences of moment whichmake their identification easy.58 Only the trans-forms of the similarpallado-complexes have been i~olated.5~ The observed momentsof [Bua3P)PdC1J2 and of its arsenic analogue indicate a state of tauto-merism between the three possible stereoisomeric structures.6OX-Ray crystallographic determinations of the structures of[Et,AuBr], and of [Pra,AuCN], have shown that these molecules aresymmetricaL6l In common with the trans-forms of some of thecomplexes previously mentioned, they have apparent dipole momentsof 1.0-1.5 D., which can now be attributed to atom polarisation(see p.62) arising from the large, opposed dipoles which theycontain.Structure.-Considerations of space make it impossible to discussmore than very few of the problems of molecular structure on whichlight has recently been thrown by electric-moment or electron-diffraction studies ; and they compel a very arbitrary choice of sub-ject. The two selected are (1) the nature of the bonds formedbetween oxygen and other elements, when the higher valencystate of the latter is developed thereby; (2) the cause of the dis-crepancies between observed and predicted values of the bondsformed by halogens.The success withwhich the formuh of many compounds could be systematised if thevalency group of electrons were assumed to be an octet 62 led to thegeneral acceptance of electronic formuh inwhich octetswere ascribedto atoms whenever possible.This tendency was strengthened by theinterpretations of parachor measurements and of some stereo-chemical observation^.^^ Thus it came about that, in compounds(1) Bonds between oxygen and other elements.66 H. J. Cave11 and S. Sugden, J., 1935, 621.67 K. A. Semen, 2. anorg. Chem., 1936, 229, 265.6B (a) Idem, ibid., 1935, 225, 97; ( b ) 1936, 229, 225; 1937, 231, 365.58 F. G. Mann and D. Purdie, J., 1935, 1549.81 A. Burawoy, C. S. Gibson, G. C. Hampson, and H. M. Powell, J., 1937,62 G. N. Lewis, “Valence and the Structure of Atoms and Molecules,”83 (a) N. V. Sidgwick, “ The Electronic Theory of Valency,” Oxford, 1927;Idem, ibid., 1936, 873.1690.New York, 1923.( b ) S.Sugden, “ Parachor and Valency,” London, 193074 GENERAL AND PHYSIOAL CHEMISTRY.where the valency of an atom has been developed to one of itshigher states by combination with oxygen, 7 the bond attaching the additional oxygen atomwas presumed to be a co-ordinate link, andJ. J. not a double bond as in earlier valency0 0 theory. Phosphoryl compounds were writtenas X,P+ 0, thionyl and sulphuryl ones X,S+ 0and X,SN0, and chlorine heptoxide as (I), corresponding struc-tures being ascribed to related compounds, including the oxy-acidsand their ions. As newer physical methods were applied, the validityof these formulations became less certain.The first clear challenge came when, on the basis of their (‘ adjacentcharge rule,” L.Pauling and L. 0. Brockway 64 suggested that theoctet formulations of several oxy-acids must be wrong, and thatactually there is resonance between several possible structureswith double bonds holding one or more of the oxygen atoms. Theysupported their argument by pointing out that the bond lengths inthe common oxy-acid ions, determined by X-ray examination ofcrystals, are much less than those to be expected for single bonds.Either, therefore, co-ordinate links must be shorter than normalsingle bonds, or the bonds in question cannot be co-ordinate links.Electron-diffraction studies soon provided further examples ofanomalous ‘I co-ordinate ” links, and also, together with X-raycrystal-structure data, demonstrated that the true co-ordinate linkhas the same length as a normal bond,e5 in agreement with a theoreti-cal prediction.66 This was shown particularly clearly in trimethyl-amine 0xide,~~(6) where, since the donor atom is of the first shortperiod, its covalency is limited to four; so the fourth atom, theoxygen, can only be held by a co-ordinate link.The N-tO bondlength is 1.36 A. ; the sum of the single-bond covalent radii is also1 . 3 6 ~ . In contrast to this, the bonds with oxygen formed byelements in later periods, which are not restricted to quadri-covalency, are sometimes as short as or even shorter than triplebonds, as may be seen from the following table, which supplementsthe earlier data €or oxy-acid ions.Electric dipole-moment data emphasise the difference still further.The moment for a co-ordinate link 1.36 A.long, if an electron werecompletely transferred, would be 6.43 D. In actual fact the N+Olink is not quite so polar : recent measurements on amine oxides64 J . Amer. Chem. SOC., 1937, 59, 13.66 See (a) K. J. Palmer and N. Elliott, ibid., 1938, 60, 1852; (b) M. W.6 6 N. Elliott, J. Amer. Chem. Soc., 1937, 59, 1380.O+Cl-0- 1+0(1.1if0Lister and L. E. Sutton, Tram. Faraday SOC., 1939, 35, 495SUTTON : STUDY OF MOLEUULAR STRUCTURE. 75Compound.NOMe,SO,Me,so2 so3SO2C1, :2:i POF,Cl,POFCl,, POCl,PSCl,PSF,Distance, A.7 ~ -I--I ___---_ILb1c.re h \lobs., A. single. double. triple.1.36 f 0.03 65(b) 1.36 1.18 -1.44 f 0.03 6 S ( b ) 1.70 1-52 1.471.43 f 0.01 6 7 ,, 9 9 ¶ ¶1.43 f 0.02 21 ,, ? 9 9 91-43 f 0.02 21 ,, 9 9 9 91.43 f 0.02 6 8 9 ,1.55 f 0.03 42 l-”?G 1.57 1 *241.39 f 0.02 54@) ,, 9 , 991-94 f 0.03 43 2.14 1.95 1-811.85 f 0.02 6 8 ,, 9 , 3sshow that it is approximately 4-38 D.69 But the SO, PO, and PSbonds which should be more polar (6.86-9*31), even shortened asthey are, are actually much less so.The most recently calculatedvalues are 2.0-2.5, 3.5, and 2.5 D., respectively.70* 71There is but one step lacking to make the proof complete; this isto show that when sulphur or phosphorus does form genuine co-ordinate links a highly polar bond is generated. This has now beendone by measuring the moments of complexes between trimethyl-amine, trimethylphosphine , diethyl ether , die thy1 sulphide, andboron trichloride or trifluoride.’l It is impossible for the boroncompound to combine save by accepting a co-ordinate link from thenitrogen, phosphorus, oxygen, or sulphide atoms.In all cases thereis a large difference between the moment of the complex and thesum of those of the uncombined constituents. The abrupt drop inthe Merence which occurs from nitrogen to phosphorus compoundswhen an oxygen atom is attached is not observed in these complexes.Hence there can be little doubt that the bonds from phosphorus orsulphur to oxygen are not of the same type as those to boron, i.e.,they are not co-ordinate links.The extreme shortness has been explained as the result of reson-ance with triple-bonded structure, e.g., S s ; but a large contri-bution, such as is essential on this theory in most of the oxy-com-pounds examined, would reduce the moment of the SO and PObond to zero, or even reverse it.No such drastic effect is actually8 7 V. Schomaker and D. P. Stevenson, J. Arner. Chern. SOC., 1940, 62, 1270.6 8 D. P. Stevenson and H. Russell, ibid., p. 3264.68 E. P. Linton, ibid., p. 1945 ; these measurements agree reasonably wellwith unpublished observations made by J. S. Hunter and N. J. Leonard atOxford.70 I. E. Coop and L. E. Sutton, Trans. Faradccy Xoc., 1939, 34,505.Unpublished observations by G. M. Phillips at Oxford76 GENERAL AND PHYSICAL CHEMISTRY.observed : the bond is still decidedly polar, and the negative pole istowards the oxygen. It has therefore been suggested that the SO,PO, and PS bonds are really double bonds, but that the ordinaryradii do not apply.7o These are stable bonds, with large heats offormation; and it is possible that they therefore involve bondingorbitals of a different kind from those which would correspond to thecovalent radii used.The change of bond is thus attributed to achange in bonding orbitals, but to one which does not affect themultiplicity of the bond. Such changes have already been pos-tulated occasionally, e.g., to explain the difference between the ter-and quadri-covalent radii for boron; 72 and it now appears possiblethat they may frequently be responsible for changes of bond-lengthwhich have hitherto been attributed to change in multiplicity.Halogen compounds are frequently covalentand conveniently volatile.Many have therefore had their structuredetermined from the electron-diffraction patterns. It has beenfound that, although the lengths of the halogen bonds are usuallyequal to the sum of the appropriate radii in saturated organic com-pounds, they are frequently different from it in unsaturated ones, orin inorganic ones.The organic compounds have been extensively discussed in aprevious Report .73 The general conclusions were that the chloro-,bromo-, and iodo-derivatives of saturated hydrocarbons show noabnormalities in bond lengths, but that the fluorine compounds haveshort C-F links if there are two or more fluorine atoms on one carbonatom. Chloro-, bromo-, and iodo-derivatives of unsaturated hydro-carbons with the halogen attached to an unsaturated carbon atomalso have short carbon-halogen bonds.Later work has shown thatthe acetylenic derivatives are more abnormal than the ethylenicones which are affected to much the same degree as the benzenoidones.37(e, 9) The only Unsaturated fluorine derivative so far examinedis flu~robenzene,~~ in which the C-F bond (IcF = 1.34 -j= 0.04 A.) isreported to be shortened to about the same extent as are those incarbon tetrafl~oride.7~ .These facts have been interpreted to meanthat there is resonance between the ordinary structure and others inwhich two unshared electrons of the halogen atom form a doublebond with the carbon atom.Essentially the same explanation had previously been offered forthe reductions in the electric moments of unsaturated halogen(2) Halogen bonds.72 H.A. LBvy and L. 0. Brockway, J . Airier. Chem. SOC., 1937, 50, 2085;73 See refs. (i) 1 and (i) 2 .74 H. Osaha, Bull. Chem. SOC. ,7apan, 1940, 15, 31.75 See ref. (i) 2.see also ref. (i) 35, p. 160SUTTON : STUDY OF MOLECULAR STRUCTURE. 77compounds compared with those of saturated ones; 37(b) and there isfair quantitative agreement in the conclusions from the two sets ofdata about the relative importance of double-bonded structures inchloro-, bromo-, and iodo-comp~unds.~~(f) Objections have beenraised against this e~planation,?~ however, on the grounds that suchdonation, by going counter to the normal tendency of the halogen toattract electrons and so destabilising the double-bonded structure,should reduce its importance as a component structure in theresonance hybrid.This effect should be particularly marked forfluorine and least for iodine; and it is a most striking fact that,judging from the chemistry of its compounds, fluorine differs fromthe other halogens in showing no inclination whatsoever to formmore than one single bond. Yet although there is some tendencyfor the changes in length or moment to increase in the order Cl<Br< Iit is not very marked, and they certainly do not vanish in fluorinecompounds. If, therefore, fluorine forms a double bond it must bethat this has a greater heat of formation than two single bonds, justas that of N=N is greater than 2(N-N).80 It has been suggestedalternatively that there is no double-bonding, but that there is someelectrostatic effect which is as yet imperfectly understood.77 Thefacts are that when halogens are attached to unsaturated carbonatoms the C-Hal link is shortened, and that, as judged from bothelectric moment data and chemical behaviour, electrons appear tomove from the direction of the halogen atom into the body of theunsaturated system.Although no complete and entirely acceptableexplanation seems yet to have been given, it seems on the wholeprobable that resonance with double-bonded structures is the maincause in the chlorine, bromine, and iodine compounds, but that someother cause may be predominant; in the fluorine compounds.The inorganic compounds have not been systematically discussedin a Report, but the salient points may be briefly summarised asfollows.In the halogen derivatives of saturated carbon compounds,with the exceptions previously noted, the halide bond lengths arenormal; but in the compounds of most other elements they areabnormal. The N-Cl,") 23 0-C1, O-F,359 (i) 14 F-F 78 or S-Br,7g bondsare too long ; the bonds between halogens and boron or any elements,save the halogens themselves, in the second short period or Iaterperiods are in some degree shortened.80* *1 The fact that there is76 See ref. (i) 1. '' G. Baddeley, G. M. Bennett, S. Glasstone, and B. Jones, J., 1935, 1827.78 L. 0. Brockway, J. Amer. Chem. SOC., 1938, 60, 1348.7e D. P. Stevenson and R. A. Corey, ibid., 1940, 62,2477.See ref.(i) 35, Chap. VII.See A. H. Gregg, G. C. Hampson, G. I. Jenkins, P. L. I?. Jones, andL. E. Sutton, Trans. Farachy SOC., 1937, 33, 85278 GENERAL AND PHYSICAL CHEMISTRY.contraction only if the central atom belongs to a later period thanthe first, e.g., silicon, or has an incomplete octet when exercising itsnormal covalency, e.g., boron, led to the theory that the contractionsare all due to resonance with double-bonded structures in which theoctet is completed or, in the heavier elements, exceeded. It wouldthen be expected (cf. p. 77) that the contractions would increase inthe order F<Cl<Br<I, although this might be opposed by thepossibly greater tendency of the lighter halogens, like other lightelements, to form double bonds. In actual fact, the halides of theelements which have incomplete octets in their normal valency stateshow much the same percentage contraction in chlorides, bromides,and iodides, although the fluoride is considerably shorter ; but thoseof the other elements show a marked decrease in contraction, i.e., inabnormality, from fluoride to iodide.This suggests that there maybe two causes of contraction,81* 82 the double-bonding, postulatedabove, which operates in the first class of halide, and another, sofar not understood, which certainly operates in the second class andpossibly in both. The decrease of contractions in the orderF>Cl>Br>I and the existence of expansions in the N-CI, 0 4 1 ,S-Br, 0-F, and F-F bonds suggest that there is a qualitative relationbetween the strength and the length of a bond, and since the‘‘ electronegativity co-ordinates ” in Pauling’s electronegativitymap 83 are derived from bond energies, this may be expressed bysaying that large differences in electronegativity tend to make thebond short, whereas small ones tend to make it long.This con-dition alone is not sufficient, and consequently it has been regardedrather as permissive.8* If double-bonding can occur when thebonds are highly polar, then there is contraction. No explanation ofthe expansions is provided on this theory, however. This lack andthe objections to any double-bonding mechanism make it desirableto look for possible alternatives; and the following is worthy ofconsideration.The shortening or lengthening of halide bonds maybe connected with variations in the bonding orbitals used by thehalogen and the central atom, these being of a kind which does notaffect the multiplicity. If a large heat of bond formation canpromote hybridisation of atomic orbitals to give new bonding orbitals,it may give rise to new and smaller covalent radii, whereas a lowheat of formation, by discouraging such hybridisation, may increasethe radii. When fluorine combines with itself or oxygen, oxygenwith fluorine or chlorine, sulphur with bromine, or nitrogen withchlorine, it is possible that the bonding orbitals are largely derived82 A table of percentage contractions, including new results, will appear inS3 See ref.(i) 35, Chap. 11.a forthcoming paper by M. W. Lister and L. E. SuttonRITCHIE : KINETICS : PHOTOCHEMISTRY. 79from atomic p orbitals ; but when these atoms combine with otherelements to form stronger bonds they use sp hybrid orbitals. Sincethe latter bonds are the commoner, they give rise to the smaller" standard " covalent radii.* Furthermore, while carbon can onlyuse sp3 hybrid bonding orbitals, silicon could possibly use hybridsinvolving d orbitals as well in a strong bond such as the Si-F bond;and these strong bonding orbitals might correspond to considerablyreduced radii. Thus the conditions which permit double-bondingalso permit this change in hybridisation.One further general rule, that contractions diminish with increas-ing weight of the central atom when it is attached to the same halo-gen, may be explained on the grounds that the effect of changes inthe valency shell has steadily less effect upon the electronic arrange-ments of the atom as a whole with increase of atomic number of thelatter.It will be noticed that this suggestion unifies not only the con-tractions and the extensions of halide bonds, but also the con-tractions of the oxide bonds in the higher oxides and oxy-acids, whichwere discussed previously.This is an attractive feature, but untilthe idea has been developed beyond its present crude, incomplete,and qualitative form its real value can hardly be judged.L. E. S.4. KINETICS.Photochernistr y.Primary Processes.-In the previous Report on Photochemistry,lit was pointed out that a photo-activated molecule may react inseveral ways at comparable rates, the procedure favoured beingdetermined by a variety of factors.In the last few years, a numberof extensive investigations, more particularly of organic molecules,has facilitated a, more complete recognition of these processes inrelation to the absorption spectra; it has become clear, however,that the characteristics of the absorption spectrum alone are nocertain assurance of the type of process occurring.Ann. Reports, 1937, 34, 64. * (Added in proof, 25/2/41.) V. Schomaker and D. P. Stevenson ( J . Arner.Chern. Soc., 1941, 63, 37) suggest increases of the standard covalent radii foroxygen, nitrogen, and fluorine and give an empirical relation between bondlength and difference of electronegativity values.83 The hypothesis thatelectronegativity difference alone is suflicient to cause contraction ww aban-doned some time ago as inadequate.Moreover, the uncertainty of thesi@cance of electronegativity values [ref. (i) 61 makes it necessary to con-sider what other factors which may strengthen the bond will also shorten it,as is done here80 GENERAL AND PHYSICAL CHEMISTRY.I n general, absorption of a light quantum by a molecule mayresult in the production of one of several excited states. Whenenergy is absorbed at a particular bond by a process of transitionto a repulsive energy level or to a sufficiently high point on anattractive level, rupture of the bond results within one vibrationperiod.define a simple ruptureprocess as one in which, on the average, decomposition occurs a tthe locus of absorption in less than one vibration period after thecreation of the excited state; on the other hand, if, during thelifetime of the excited state, some shift of energy or position withinthe molecule occurs from the initial excited state so that rupturefinally takes place, either at some other bond or at the initiallyabsorbing bond, we deal with a predissociation process, appearingin the spectrum in the form of the well-known diffuse absorptionbands. Such predissociation may then be divided into two classes,spontaneous and induced, and a distinction made by consideringthe effect of altered pressure, but in general it is diEcult to determineexactly the actual process involved.Unambiguous interpretationsof such principles are usually confined to the relatively simplecase of the diatomic molecule. In sulphur, S,, for e~ample,~ theabsorption bands in the region 2435-2615 A. are blurred even atlow pressure, and a spontaneous predissociation is hence concerned.The absorption spectrum a t 2615-2799 A. is sharp at low pressures,but becomes blurred at higher pressures ; the previously forbiddentransition is now made possible by the fields introduced by neigh-bouring molecules. In iodine, relatively few iodine atoms areproduced in the discrete region of the spectrum, but the yield isgreatly increased when a magnetic field is applied or the pressureis increased by the addition of otherwise inert gas molecules.Inother words, the discrete spectrum is here associated with an inducedpredissociation.I n polyatomic molecules, where the term vibration period itselfcannot be so exactly defined, the properties of hypersurfacescorresponding to different energy levels may be considered, theintersection of surfaces being comparable to the intersection oflines in the potential-energy diagram for a diatomic system. De-composition may in simple cases occur as before as a result ofabsorption from the normal state to a repulsive or weakly attractivesurface, in which case a continuum would be observed in the spectrum.Similarly, other concepts derived from the study of diatomic systemsmay be applied directly to the polyatomic ones.In complexmolecules, however, the spectral characteristics generally are not2 J . Chem. Physics, 1938, 6, 416.3 Cf. G. Herzberg and L. G. Mundie, ibirE., 1940, 8, 263.M. Burton and G. K. RollefsoRITCHIE : KINETICS : PHOTOCHEMISTRY. 81so clear. Diffuseness, for example, is not always sufficient in decidingbetween a spontaneous and an induced process. R. G. W. N ~ r r i s h , ~in a, discussion of the relationship of fluorescence to photolysisin gaseous systems, has pointed out that the equation TAV - 1,connecting the breadth of an absorption line (Av) and the life-time(7) of the excited molecule, indicates that a t 3 0 0 0 ~ . a reductionof the life-time from sec., the normal life-time of the fluorescingmolecule, to 10-13 sec., corresponds to an increase in the naturalbreadth of the line by a factor of lo5; the previously sharp linesbecome diffuse, and the onset of diffuseness in this case correspondsclosely to the photolytic threshold.In a diatomic molecule, theremay be thus a sharp demarcation between fluorescence and pre-dissociation as shown by the spectrum. On the other hand, in apolyatomic molecule, the life-period may be much longer by reasonof the complexity of the vibration; if 7 be now 10-lO sec. instead of10-13 sec., the natural breadth of the line at 3000 A. is only lo-,instead of 10 A. Diffuseness is not now apparent, yet the probabilityof decomposition is 100 times ( L e . , 10-8/10-10) that of fluorescence.Under these conditions the efficiency of photolysis may then benearly unity, fluorescence being almost entirely absent, and yetsharp rotational structure is to be observed in the spectrum.Dis-appearance of fluorescence is then on this basis a better criterion ofpredissociation than diffuseness.Care must therefore be taken in extending the spectral interpre-tations applicable in the case of simple molecules to those of greatercomplexity, and in particular, relatively greater attention must bepaid to the nature and amounts of the products of photolysis for thegiven conditions of wave-length and temperature. In this respectthe simpler aldehydes and ketones have again been the subjectsof much investigation. Absorption here occurs a t the CO link,but rupture never follows at this point ; predissociation processesare therefore involved.The main problem is then the determinationof the relative importance of the three primary reactions :CORIR, -+ RIR, + CO ; CORIR, + R1+ COR,;CORIR,-+ R, + R, + COWith formaldehyde, a fine-line absorption has been reported forthe region 3570-2750 A., followed by a diffuse region 2750-2500 A.,and a continuous background from approximately 2670 A. to lowerwave-lengths. The products of photolysis are almost entirelyhydrogen and carbon monoxide; even in the Schumann region' Trans. Farday SOC., 1939, 35,21; for a general discussion of fluorescenceand photochemical kinetics in polyatomic molecules, see W. A. Noyes, junr.,and F. C. Henriques, junr., J. Chem. Physics, 1939, 7, 76782 GENERAL AND PHYSICAL CHEMISTRY.there is no evidence of CH2 radicals or oxygen atoms.6 Fluorescencehas been observed a t 3530,3400, and 3270 A., indicating a life-periodfor the excited molecule of to 10-7 sec. Addition of air to100 mm.pressure does not diminish the fluorescence appreciably;this predissociation process must then be spontaneous in this spectralregion. The sharp-line spectrum might then indicate that, whilethe molecules may remain in a stable hypersurface for some timebefore decomposing in that surface, a transition from that surfaceto another is forbidden. The decomposition reaction would thenbe best represented byH*CHO*+H,+CO . . . - (a)By experiments involving the use of iodine vapour to remove anyhydrogen atoms formed, E. Gorin finds, however, that at 3130 A.as well as at 2537 A,, the permanent-gas product is almost exclusivelycarbon monoxide, the HI/CO ratio being nearly 2 at these wave-lengths ; the main reaction is on this basisH*CHO+H+CHO .. . - (b)At 3650 A., a considerable quantity of hydrogen is formed and thedata here indicate that the probability of reaction (a) to that ofreaction (b) a t the temperature concerned (100") is 0.3 to 0.7.Such results serve to emphasise the difliculty of determining theexact process of decomposition occurring in such complex moleculesfrom examination of the spectra alone ; diffuseness does not neces-sarily indicate a spontaneous process, while discrete absorptionmay be associated with an induced or spontaneous process.Gorin deduces 7 from his results that the heat of dissociationfor the first hydrogen atom in formaldehyde is considerably smallerthan that formerly recognised.From the predissociation limit,which, however, can only give an upper limit, approximating tothe true value the more complex the molecule (see ref. 2), R. Mecke 8deduced the value of 102 kg.-cals. ; the wave-length 3650 A. (above)corresponds to an energy of 78 kg.-cals. The stability of the formylradical has been previously noted : according to the above, 26kg.-cals. are required to dissociate it. It is to be pointed out that,according to E. Bergmann and R. Samue1,lOin certain cases two bondsR. G. W. Norrish and W. A. Noyes, junr., Proc. Roy. SOC., 1937, A, 163,221.6 G. Herzberg and K. Frau, 2.Physik, 1932,76,720; cf. G. H. Dieke andG. B. Kistiakowsky, PTOC. Nat. Acad. Sci., 1932, 18, 367.J . Chem. Physics, 1939, 7 , 256.* 8. Elektrochem., 1930, 36, 589.F. Patat, 2. physikal. Chem., 1936, B, 32, 274; M. Burton, J . Amer.Chem. SOC., 1938, 60, 212.lo Nature, 1938, 141, 832RITCHIE : KINETICS : PHOTOCHEMISTRY. 83of an absorbing molecule may be broken simultaneously withoutinvolving two single-bond energies, by the direct transition from theground state to a repulsive term, correlated to a saturated moleculeof lower valency. It is not always clear from present experimentalevidence how the energy of one particular bond appears in relationto the energy required in the photodecomposition.At 2537 A., the primary processHCHO-+H+H+CO .. .becomes energetically possible, but in presence of iodine the yieldof hydrogen iodide would be double that expected from reaction (b).There is, however, no significant change in the values at 2537 ascompared with those at 3130 A., and it is thus to be concluded thatreaction (c) does not play any important part in the formaldehydephotolysis.In acetaldehyde, the discrete absorption spectrum extends from3484 to 3050 A., where the diffuse spectrum begins. A continuumunderlies the whole. On the formaldehyde analogy, we may havetwo predissociation processes in action simultaneously, one re-ferring to the production of free radicals and the other givingdirectly methane and carbon monoxide. One cannot assume,however, that the production of free radicals will in all cases befavoured by decrease in wave-length. Experiments to determinewhich process is favoured by given conditions have been carriedout l1 on the basis that free radicals catalyse the chain thermaldecomposition of acetaldehyde ; at high temperatures the chainlength increases rapidly and accounts for practically all the decom-position.At high temperatures, therefore, determination of theratio of quantum yields for two wave-lengths gives the ratio ofconcentrations of free radicals a t these wave-lengths. It hasthus been concluded that, at room temperature, more radicals areproduced a t 2652 A. than a t 3132 A. ; but as the temperature israised, the yield of radicals a t 3132 A. relatively increases, and a t100" the yields are equal.Fluorescence fades correspondinglyas the temperature is raised, this being in agreement with the viewthat at 3132 A. the dissociation is aided by thermally excitedvibrational energy. The processes concerned may be analysed bythe principles already cited. On the basis of only one excited statefrom which the molecule may dissociate, there may be a slow de-composition in the same hypersurface into ultimate molecules, thisoccurring when the increase in energy is not sufficient to break theC-C bond to give free radicals. If a Sufficient energy is absorbed,transition from the initially excited state to a weakly attractivel1 G. K. Rollefson and D. C. Grahame, J . Chem. Physics, 1939, 7, 77584 GENERAL AND PHYSICAL CHEMISTRY.state will lead to such dissociation.G. K. Rollefson and M. Burtonhave shown, however,12 that under these conditions the probabilityof free-radical production is highest when the amount of energycontained by the molecule in the excited state is a minimum inexcess of that required to cause free-radical decomposition ; thus,rise in temperature decreases the free radical yield a t 2652 A.and increases it at 3132 A. Further, at still shorter wave-lengths,e.g., 2536 A., free-radical production ought to be less than at 2652 A.E. G ~ r i n , ~ ' l3 in fact, finds from the photodecomposition in presenceof iodine a t 100" that the two primary reactions occur simultaneously,but at 2 5 3 6 ~ . dissociation into molecules is favoured, the ratiobeing 2.9 : 1.At 3130 A. the free-radical mechanism is 2.6 timesas probable as the decomposition into ultimate molecules. Furtherconfirmation of such trends has been obtained by tellurium mirrorremoval methods. l4 Fluorescence in acetaldehyde has beenascribed to the combination of acctyl radicals to give diacetyl; l5that only a very faint fluorescence has been observed at 2652 A . ~ ~is not thus in disagreement with this view. It is noteworthy thatthe green fluorescence of acetone illuminated by light of 3130 A.,identical with that found in diacetyl itself excited by wave-lengths3650, 4047, and 4358 A., appears only after continued radiationand is therefore to be attributed to the diacetyl formed; l7 sincediacetyl does not itself fluoresce appreciably when irradiated with3130 A.unless acetone is present, the simplest picture is that excit-ation arises by some collision of the second kind.F. E, Blacet and D. Volman l8 find that the volatile productsof decomposition of acetaldehyde over the range 3340-2380 A.are solely methane, hydrogen, and carbon monoxide, the ratioH,/CO increasing as the wave-length is decreased. Discussionis given, for room-temperature photolysis, of a series of reactionsbased on the free-radical mechanism similar to that proposed forhigh-temperature photolysis l9 and to that advanced for the thermaldecomposition.20 This chain mechanism allows explanation of thealtered ratio HJCO with change in pressure, light intensity, andl2 J . Chem. Physics, 1938, 6 , 674.l3 Acto Physicochim.U.R.S.S., 1938, 9, 681.14 P. A. Leighton and D. Volman, J. Chem. Physics, 1939, 7, 781.15 W. A. Noyes, junr., and F. C. Henriques, junr., J . Chern. Physics, 1939,l6 D. C. Grahame and G. K. Rollefson, ibid., p. 98.l7 G. M. Almy and S. Anderson, ibid., p. 805.l9 J. A. Leermakers, ibid., 1934, 56, 1537; D. C. Grahame and G. K.2o F. 0. Rice and K. F. Herzfeld, J . Amer. Chem. SOC., 1934, 56, 284.7, 767; see G. M. Almy, H. Q. Fuller, and G. D. Kinzer, ibid., 1940, 8, 37.J . Amer. Chem. SOC., 1938, 60, 1243.Rollefson, J. Chem. Physics, 1940, 8, 98RITCHIE : KINETICS : PHOTOCHEMISTRY. 85temperature. It is possible that the effect of altered wave-lengthis partly connected with the excess energy remaining in the CHOradical after the primary dissociation.Further information re-garding polymerisation processes involving CHO and CH, is de-sirable. No polymerisation was found in presence of i ~ d i n e , ~indicating free radicals as the cause of polymerisation ; the smallerpolymerisation observed normally a t low wave-lengths is in agree-ment with the effects of altered wave-lengths illustrated above.The decomposition of acetaldehyde, both photochemical and thermal,has been discussed in detail in a number of papers ; 21 it is concludedthat the C-C bond strength is of the order 75 kg.-cals. The photo-lysis of acetyl bromide has been investigated and discussed22 onthe basis of three primary steps, two yielding free radicals and thethird involving direct decomposition of the activated molecule intothe final products.V. R.Ells and W. A. Noyes, j ~ n r . , , ~ conclude that methyl ethylketone a t 25" dissociates mainly to give ethyl radicals in the nearultra-violet (3000 A.) but at shorter wave-lengths (1880-2000 A.)methyl and ethyl radicals are produced in nearly equal amountsfrom the primary process. The quantum yield of carbon monoxideformation from diethyl ketone is approximately unity in both wave-length regions. In higher aldehydes and ketones a further typeof decomposition has been previously noted, e.g.,CH,*CH,*CH,*CHO + CH,*CHO + C,H, . . (d)From its probability to that of other reactions this may be inter-preted as an induced predissociation process.2 E. Gorin 7 hasconcluded that in methyl ethyl ketone vapour a t 70" the primaryquantum efficiency in the decomposition into free radicals is of theorder unity, and thus suggests that the type (d) reaction may inthis case occur by secondary processes involving interaction betweenthese radicals.Against this, it is to be pointed out that C. H.Bamford and R. G. W. Norrish 24 find the production of moleculesaccording to type (d) mechanism to occur in paraffioid solutionas readily a t room temperature as a t loo", whereas decompositionyielding free radicals is completely suppressed in ketone photolysisa t room temperature, this being ascribed to recombination ofradicals by the principle of primary recombination following restraintby a shell of solvent molecules. Only at temperatures of 70-100"21 M.Burton, J. Chem. Physics, 1939,7, 1072 ; T. W. Davis and M. Burton,ibid., p. 1076; M. Burton, H. A. Taylor, and T. W. Davis, ibid., p. 1080;H. A. Taylor and M. Burton, ibid., p. 414; F. 0. Rice and W. R. Johnstone,J. Amer. Chem. SOC., 1934, 56, 214.22 D. H. Etzler and G. K. Rollefson, ibid., 1939, 61, 800.28 Ibid., p. 2492. 24 J., 1033, 1531, 164486 GENERAL AND PHYSICAL CHEMISTRY.do the free radicals show hydrogenation at the expense of thesolvent, which thus shows equivalent unsaturation. It is ditlicultto assess possible effects of solvent on the relative rates of thetwo possible decompositions from the point of view of inducedpredissociation, and it may be argued that direct comparisonbetween solution and gas phase is here not justifiable.It is, forexample, possible that the free-radical mechanism, involving adifferent hypersurface from that of the ultimate molecule formation,is suppressed a t low temperatures in a process which produces nofree radicals, in which case application of the primary recombinationprinciple would be unnecessary. Comparison of reactions in thegas phase with those in solution has been given by R. G. Dickinson ; 25there are, however, few other recent investigations from this pointof view.Reference has been made in previous Reports to various methods(e.g., use of metallic mirrors or of nitric oxide) of confirming andidentifying free radicals in general. Primary and secondaryaliphatic amines yield hydrogen atoms and alkylamino- or dialkyl-amino-radicals respectively in the primary act, the hydrogen atomsbeing detected by their reaction with propylene.26 Irradiationof hydrazine in the presence of propylene confirms similarly theproduction of atomic hydrogen in the initial decomposition forwave-lengths of ca.2200 A., and nitric oxide may be used to detectN2H, radicals in this photolysis as well as the NH, radicals producedin the photolysis of ammonia.,' The addition of nitric oxide assiststhe photodecomposition of methyl iodide 28 by removal of methylradicals involved in the reverse reaction.A variety of photochemical methods have thus been employed inproducing free radicals in order to study their subsequent reactions,e.g., photolysis of methyl ethyl ketone,29 of such compounds asdiethyl-mercury and -~inc,~O dimethylmercury,3l of alkyl iodides,s2of alkyl iodides in the presence of mercury to remove iodine atoms,=and by the mercury-sensitised production of hydrogen atoms in thepresence of olefins.34 Methyl radicals are thus found to be more25 J .Physical Chem., 1938, 42, 739.26 C. H. Bamford, J., 1939, 17; cf. dimethyl- and diethyl-nitrosoanlines,27 Idem, Trans. Paraday SOC., 1939, 35, 668.28 T. Iredale, ibid., p. 458.3o Idem, ibid., p. 396.31 J. P. Cunningham and H. S. Taylor, ibid., 1938, 6, 359.32 W. West and L. Schlessinger, J. Amer. Chem. SOC., 1938, 60, 961.33 J. C. Jungers and L. M. Yeddanapalli, Trans. Faraday SOC., 1940,36,483.34 W. J. Moore, junr., and H. 8. Taylor, J . Chem. Physics, 1940, 8, 504.idem, ibid., p.12.W. J. Moore, junr., and H. S. Taylor, J . Chem. Physics, 19.10, 8,466RITCHIE : KINETICS : PHOTOCHEMISTRY. 87effective than ethyl radicals in polymerising ethylene,29 in accordancewith other results; disappearance of the radicals in this polymeris-ation is accounted for by saturation by hydrogen-atom capture andby mutual recombination. Detailed discussion of reactions ofmethyl radicals has been given by H. A. Taylor and M. Burton.35In the case of the metallic alkyls, further products may arise byinteraction of radicals with the alkyl molecules them~elves.~~ Furthercheck on the actual concentrations of such radicals is desirable.Primary Quantum E$iciencies.-The overall photochemical yieldsin the above-mentioned organic molecule decompositions are as arule less than unity, and there are two ways in which such lowvalues may be explained.In the &st, the primary quantumefficiency may be unity but the secondary reactions, e.g., recombin-ation of the radicals initially produced, may reduce the overall yield.Alternatively, the efficiency of the primary reaction, chemicallyspeaking, may not be unity, by reason of fluorescence, collisionaldeactivation, or internal degradation of energy. It might beexpected that in more complex molecules, with their greater possi-bilities for internal degradation, the low quantum efficiencies mightbe ascribed t o this cause.36 Some recent investigations, however,allow of measurement of the primary quantum efficiencies apartfrom the overall yield; the methods of attack on the problem andthe results obtained are of considerable interest.employed iodinevapour to remove as alkyl iodides or hydrogen iodide the radicalsprimarily formed, the method depending on the fact that thesaturated hydrocarbons that may be formed by direct splitting ofthe aldehyde and ketone do not react with iodine in the temperaturerange investigated. Analysis of the reaction products then dis-tinguishes quantitatively primary reactions giving free radicalsfrom those giving saturated hydrocarbons. The principle of themethod may be exemplified by acetone.Here the overall quantumyield for given conditions of temperature, wave-length, light intensity,and pressure is known. A series of experiments with and withoutiodine vapour is carried out.The ratio of carbon monoxide moleculesin absence of iodine to methyl iodide molecules, multiplied by theknown overall quantum yield in the absence of iodine, gives thequantum efficiency of the primary stage. With acetone, Gorinfinds the quantum efficiency of the formation of methyl iodide tobe approximately unity, in the unfiltered light of the mercury arc.The primary process must then beCH3*CO*CH3 + CH, + CH3*C036 J. Chem. Physics, 1939, 7, 675.s6 See R. G. W. Norrish, Tram. Farday SOC., 1939, 35,22.In aldehyde and ketone photolysis, E. Gori88 GENERAL AND PHYSICAL CHEMISTRY.with an efficiency of unity. In the presence of iodine, only asmall amount of carbon monoxide is formed below SO", this con-firming the stability of the acetyl radical; the low overall yield inabsence of iodine is then to be ascribed to the recombination ofmethyl and acetyl radicals.I n acetaldehyde, as discussed previously, a t 3130 A.and loo",dissociation into free radicals is 2.6 times as probable as that intomolecules. The sum of the quantum yields for both reactionswas found to be unity. The normal low overall quantum yieldmust then be due to recombination of the radicals CH, and CHO.In formaldehyde, the primary process a t 3130 A. is dissociationinto free radicals, and the quantum efficiency of hydrogen iodideformation is of the order unity, this again indicating unit efficiencyfor the primary stage.E. Gorin finds the decomposition scheme for methyl ethyl ketoneto be somewhat similar to that of acetone ; in the presence of iodine,the amount of carbon monoxide is relatively very small whencompared with the amount of alkyl iodide. The principal primaryreactions are thus :C,H,*CO*CH, + C2H5 + CH,*COand C,H,*CO*CH, -+ CH, + C,H,*COthere being other evidence 37 to show that these have approximatelyequal probabilities; propane is formed in small amount only (5%)by the primary processC,H,*CO*CH, + C3Hg + COComparison of photolysis rates for CH,*CO*C,H5 and CH,*CO*CH,in the presence of iodine, corrected for different absorption co-efficients, indicates also that the primary dissociation here has aquantum efficiency of about unity.It is thus to be observed that illumination of these organic mole-cules in the presence of iodine vapour leads generally to dissociationwith a primary quantum efficiency of order unity.In someapparently simpler molecules, however, such unit efficiency is notobserved. The photolysis of ammonia may be here considered.The kinetics of this decomposition have already been the subjectof considerable experiment and disc~ssion.~~ The main ditEicultyis the explanation of the low overall quantum yield, the predis-sociation spectrum and the absence of fluorescence being takenas indicating the dissociation of the molecule in 10-13 sec. after37 R. G. W. Norrish and (Miss) M. E. S. Appleyard, J., 1934, 874; see alsoref. (23).s8 See, e.g., H. S. Taylor, J. Physical Chem., 1938, 42, 783RITCHIE : KINETICS : PHOTOCHEMISTRY.89absorption of the quantum. Since the decomposition is repressedby the addition of atomic it was first assumed thatammonia is partly re-formed by combination of the hydrogen atomsand NH, radicals initially produced by absorption, while hydrogenis formed by hydrogen-atom recombination on the walls of thevessel or in the gas phase. Actual measurement of the hydrogen-atom concentration by the para-hydrogen conversion method 4Ogave, however, a value much lower than that expected on theabove basis. Since hydrazine is formed in small amounts in thephotolysis and is known to react with atomic hydrogen, it wassuggested 41 that the low atom concentration in the ammoniaphotolysis was due to reaction with this product. E.A. Birse andH. W. Melville 42 have recently determined the efficiency of the re-action of hydrogen atoms with hydrazine in the temperature range20-200", the atoms being produced by mercury sensitisation andtheir concentration measured by the para-hydrogen conversion.The energy of activation is about 7 kg.-cals. and the steric factor isabout from these results and the conditions of the ammoniaphotolysis, it is concluded that hydrazine is not responsible for thelow stationary atom concentration. In further in~estigation,~~it has been shown that photo-decomposing ammonia has noappreciable effect in removing hydrogen atoms a t an abnormallyfast rate, even when conditions are such that inhibition of am-monia decomposition by hydrogen atoms is to be observed, andthe conclusion is drawn that the primary dissociation is not soefficient as has hitherto been supposed.The determination of the primary reaction efficiency has thenbeen approached in the following way. In a mercury-sensitisedconversion of para-hydrogen, the rate of conversion is given by :The first term refers to conversion brought about by the chainatomic exchange reaction, while the other two terms represent rateof conversion due to dissociation of the para-hydrogen molecules,which is equal to the sum of the rates of removal of hydrogenatoms, by diffusion to the walls and by the gas phase triple collisionprocess respectively.In a certain range of fairly low pressures,the second term is predominant ; hence39 H. W. Melville, Proc.Roy. SOC., 1932, A., 138, 374.40 L. Farkas and P. Harteck, 2. physikal. Chem., 1934, B, 25,357.41 W. M u d and A. van Tiggelen, Bull. SOC. chirn. Belg., 1937, 46, 104.42 Proc. Roy. SOC., 1940, A, 175, 164.43 Ibid., p. 18690 QENERAL AND PHYSICAL CHEMISTRY.At the same time the hydrogen-atom concentration is determinedby that of the excited mercury atoms; if all the incident lightIin. is absorbedwhere k3[H2][Hg*] is the rate of removal of Hg* by collision withhydrogen, and k4-l is the mean life of the excited mercury atom.Since the rate of production of hydrogen atoms 2k3[H,][Hg*] mustbe balanced by R[H]/[H,]Hg" = lin./(Jh[H21 + k4)If k4 is small compared with k3[H2], the absolute rate of para-hydrogen conversion will be numerically equal to the rate of pro-duction of hydrogen atoms.Investigation showed that in anarrow pressure range (approx. 0.4-2 mm. for the given set-up)the rate of conversion was, indeed, independent of the hydrogenpressure. The primary quantum efficiency of ammonia, decom-position by zinc-spark radiation was then obtained by the followingprinciple. By adjustment of mercury-vapour lamp and sparkintensities, the rate of the mercury-sensitised para-hydrogen con-version a t 50 mm. pressure can be made equal to that obtainedfor 30 mm. of ammonia and 50 mm. of para-hydrogen by the spark.The rate of hydrogen-atom production is then the same for bothsystems, and is numerically equal to the rate (8) of mercury-sensitised para-hydrogen conversion found for the same intensityof the mercury lamp for a para-hydrogen pressure of 1 mm.Thenumber of spark quanta absorbed (labs.) by the ammonia can bedetermined from the rate of the spark decomposition of ammoniaand the already known overall quantum yield, viz., 0.3 ; theefficiency of the primary process is then 8/I&&. The values obtainedby E. A. Birse and H. W. Melville 43 range from 0-68 to 0.48, with amean value of 0.58.H. J. Welge and A. 0. Beckman 44 found that a t 1990 A., for smallamounts of ammonia decomposition, the product was exclusivelyhydrogen, and that under certain conditions the overall quantumyields were much higher than 0.3, ranging from 0-74 to 0.95. Thisis in apparent contrast to the primary value 0.58 (above). Thelatter value was obtained for 2100 A., and it is known that the yieldincreases with decreasing wave-length, by an effect on the primaryprocess.It may be noted in this connection that E. 0. Wiig45found the overall yield of trideuteroammonia decomposition to begreater in the discrete region of absorption a t 2138 A. than in the44 J . Anzer. Chem. SOC., 1936, 58, 5461. 4 5 Tbid., 1937, 59, 827RITCHIE KINlTIaS : PROTOUHEMISTRY. 91diffuse region a t 2100 A. As in the case of formaldehyde, previouslydiscussed, discrete absorption does not necessarily imply induceddissociation ; spontaneous processes occurring after differenttimes may be concerned, and the low primary efficiencies may bedue to simultaneous degradation of energy. In trideuteroammoniathe possibilities of decomposition are (1) ND, --+ ND, + D,(2) ND, --+ ND + D, ; M.Burton 46 points out that, in NH ascontrasted with NH,, all the electrons are paired and that NH maythus resemble CH, in behaving more as a reactive molecule than as afree radical. There is then a competition between decompositioninto free radicals and one into ultimate molecules in the primaryact,12 in which case the free-radical mechanism will be favouredby decrease in wave-length. If the primary quantum efficiencyof the free-radical mechanism is less than that of the decompositioninto molecules (reaction 2 requires less energy than reaction l),the variation of overall yield with wave-length may be explained.The efficiency of secondary reactions will, of course, partly determinethe result.Application of the para-hydrogen technique to tri-deuteroammonia a t 2100 A. gave 43 a primary quantum efficiency(ND,-+ ND, + D) of 0-28, on the assumption that no error isintroduced by taking the exchange conversion initiated by deuteriumatoms as kinetically identical with that by hydrogen atoms. Theoverall yield was 0.21, in agreement with previous mea~urement.~~In the direct photo-decomposition of ph~sphine,~' the overallquantum yield was found, by the exchange technique, to be approxi-mately the same (0.50) as that of the secondary processes, and hencethe primary quantum efficiency is unity. On the other hand, inthe mercury-sensitised decomposition, the over-all yield is 0.27.As would be expected, the yield of the secondary processes is asbefore 0.50, and the primary efficiency is therefore 0.54; everyexcited mercury atom which is deactivated does not lead to de-composition of the deactivating phosphine molecule.Similarcalculations lead to mercury- sensitised primary efficiency values of :NH, = 0.12, ND, = 0.02, N2H4 > 0.33. It is obvious that in alarge number of mercury-sensitised reactions the efficiency of theprimary decomposition may be much smaller than has been pre-viously supposed. The overall quantum yield for the mercury-sensitised decomposition of methane has been calculated 48 as 0.2 a t196" and 0.008 at 98". It is therefore probable that a t room temper-ature methane merely quenches the resonance radiation without4 6 J .Chem. Physics, 1938, 6, 680.47 H. W. Melville and J. L. Bolland, Proc. Roy. SOC., 1937, A, 160, 384.48 K. Morikawa, W. S. Benedict, and H. S. Taylor, J . Chern. Physics,1937, 5, 21292 GENERAL AND PHYSICAL CHEMISTRY.decomposition. The yield of the mercury-sensitised decompositionof ethane at 35" is 0.12, and that of butane 0.55 49; in the former,rise of temperature from 100' to 475" approximately doubles thethis being attributed to the greater efficiency of the reactionH + C2H, + C,H5 + H,. It is concluded that the mechanismat high temperature is not sensibly different from that at 100".No mercury- sensitised decomposition of carbon monoxide 51 is founda t 2537 A., but a t 1849 A. the quantum yield is of the order unity.Photosensitisations involving metallic vapours other than mercuryhave previously been limited in number, either for theoreticalreasons or by technical difficulties.Details have, however, beengiven 52 of a cadmium lamp suitable for photochemical purposes.In mercury-hydrogen mixtures, two hydrogen atoms are producedfrom one excited mercury atom; although HgH does exist in suchsystems, inability to secure HgH resonance radiation has beeninterpreted 53 as evidence that the weak band radiation of HgHis due to molecules formed in excited states and dissociating im-mediately. In suitable illuminated cadmium-hydrogen mixtures,CdH resonance radiation is observed; an excited cadmium atomin collision with a hydrogen molecule produces CdH in the normalstate.The initial action in the case of ethane (3261 A.) is thenand the overall quantum yield 54 in the initial stages of the reactionis of the order 0.4 at 278". Similar values were obtained for propaneand butane. The cadmium-sensitised reactions of ethylene inabsence and in presence of hydrogen have also been studied.55A general table of other quantum yields has been given byF. Daniels.56 It may be emphasised that recent work on the chloro-acetic acid hydrolysis, often used as a standard in photochemicalwork, has shown that the quantum efficiency is less than the unitvalue previously accepted. It. N. Smith, P. A. Leighton, and W. G.Leighton 57 report a value for 2537 A. of 0-31 at 25", increasing asthe temperature rises to 0.69 a t 69", and L. B.Thomas 58 finds in4 9 E. W. R. Steacie and N. W. F. Phillips, J . Chem. Physics, 1938, 6, 179.50 E. W. R. Steacie and R. L. Cunningham, ibid., 1940, 8, 800.5 1 J. E. Cline and G. S. Forbes, J . Amer. Chem. SOC., 1939, 61, 716.52 E. W. R. Steacie and R. Potvin, J . Chem. Physics, 1939,7,782; Canadian53 L. 0. Ollsen, J . Chem. Physics, 1938, 6, 307.54 E. W. R. Steacie and R. Potvin, ibid., 1939, 7, 783.55 Idem, Canadian J . Res., 1940,18, B, 47.58 J. Physical Chem., 1938, 42, 711.5 7 J. Amer. Chem. SOC., 1939, 61, 2299.5 8 Ibid., 1940, 62, 1879; see 1,. Kuchler and H. Pick, 2. physikal. Chent.,J . Res., 1938, B, 16, 337.1939, By 45, 116RITCHIE : KINETICS : PHOTOCHEMISTRY. 93confirmation 0.34 a t 25". Although this reaction may, as before,be used as a standard of actinometry from the point of view ofindependence of quantum efficiency on moderate variations ofconcentration and light intensity, 59 care must obviously be takenin its application to the determination of other efficiencies.Theuranyl oxalate decomposition is reported similarly to have anefficiency value of 0.60.Reaction Kinetics.-At one time it was customary to dividephotochemical reactions into two groups, the direct dissociationtype and the activated molecule type, where, in the latter, de-composition followed the collision of the activated molecule with asecond unactivated molecule of similar kind. The number of re-actions to which such a mechanism can be now applied is very small.The quantum efficiency of the trideuteroammonia decomposition,examined by E.0. Wiig on this basis,45 may be explained in otherways, as previously discussed.46 Also, the quantum efficiencyof the nitrosyl chloride decomposition a t wave-lengths greater than5300 A. is independent of its pressure down to 7 mm. pressure and isunaffected by added nitrogen or carbon dioxide 61; accordingly,the dissociation mechanism is favoured for this region as well as forthe continuum for wave-lengths less than 5300 A.I n the halogens, the results of absorption of a quantum of radi-ation are generally well known for gaseous systems, and interestin reactions involving halogen absorption has usually centredin the kinetics of the secondary chemical processes following dis-sociation. The form of the expression for the overall quantumefficiency is frequently determined by the recombination processesof the halogen atoms.It has been largely photochemical evidencewhich has established the fact that the recombination of twobromine or two iodine atoms requires in effect a triple collisioninvolving a third partner, but it is doubtful if in illuminated chlorinethe analogous process will serve to explain the experimental results.I n the last Report mention was made of the possible removalof chlorine atoms by collision with chlorine molecules to give C1,complexes, these decomposing reversibly as well as by mutualrecombination. The necessity for the introduction of such com-plexes has not been universally admitted; M. Bodenstein, W.Brenschede, and H. J.Schumacher remark62 that there are nogrounds for such assumption on the basis of the other numerousSee, however, ref. (50), p. 802.W. G. Leighton, R. N. Smith, and P. A. Leighton, J . Amer. Chem. SOC.,1938, 60, 8566.6l G. L. Nsttanson, Acta Physicochim. U.R.S.S., 1939, 11, 521.6a 2. physikal. Chem., 1938, B, 40, 12594 UENERAL AND PHYSIUAL CHEMISTRY.investigations involving illuminated chlorine, and that the workof A. J. Allmand and his co-workers 63 is to be explained on otherlines. The postulation of the CI, complex here depends largelyon the lower quantum yield observed in the hydrogen-chlorinecombination a t higher chlorine concentrations, this being attributedto a retardation of the rate by chlorine by the reaction Cl + C1, +Cl,.W. A. Alexander and H. J. Schumacher 64 point out that if thelight absorption is great and the reaction velocity large, convectionmay be an important factor,65 in which case wall action may beresponsible for the reduced rate by surface removal of chain carriers.On the other hand, study of the Budde effect in chlorine in thepresence of other added gases 66 shows that, although convectiondoes occur as judged by measurement of relative thermal conductiv-ities of the gas mixtures in the same vessel under the same pressureand temperature conditions, yet such convection effects wereapproximately the same for all the added gases, and hence thedifferences observed in the Budde effect, after allowance for diffusionand thermal conductivity, are such as to support the formationof C1, by triple collisions of different efficiencies (C1 + C1, +M --+ C1, + M).It is intended shortly to publish further work 67 on the hydrogen-chlorine reaction from this point of view; results are interpretedas supporting the Cl, hypothesis, and show that under certainconditions retardation to different degrees results in the gas phasefrom the presence of other added M molecules, as well as by hydrogenchloride acting in this capacity. The apparent discrepanciesbetween the results and formulze obtained by Merent workersin this and other reactions may be due to the non-recognition of thepossible effects of such third-paxtner molecules.It has beengenerally assumed that the collision efficiency of such triple collisionsis unity; experiment has in some cases shown this to be true,68but in certain other instances, some small degree of activation maybe necessary.The photochemical formation of carbonyl chloride from carbonmonoxide and chlorine has been exhaustively reviewed in relationto the thermal reaction by M.Bodenstein, W. Brenschede, andH. J. Schumacher 69 on the basis of the reversible equilibriumG3 G. V. V. Squire and A. J. Allmand, J., 1937, 1869; H. C. Craggs, G. V. V.Squire, and A. J. Allmand, ibid., p. 1878.64 2. physikal. Chem., 1939, B, 44, 322.6 5 See W. Franke and 11. J. Schumecher, ibid., 1939, B, 42,320, 321.c 6 M. Ritchie and R. L. Smith, J . , 1940,394. 6 7 M. Ritchie and D. Taylor.6 8 Recombination of bromine atoms ; K.L. Muller and H. J. Schumacher,6B Ibid., 1938, B, 40, 121.Z. physikal. Chem., 1939, B, 42, 327. RITCHIE : KINETICS : PHOTOCFLEMISTRY. 95formation of COCl 70 from CO, C1, and a third molecule. Thechlorine-sensitised oxidation of carbon monoxide has been furtherinvestigated a t low total pressures (10-100 mm.) 71; a8 at higherpressures, the rate is proportional to the square root of the monoxidepressure. Discussion of observed ratios of carbon dioxide tocarbonyl chloride formation in such systems leads to the postulationof the intermediate CO,Cl formed from oxygen and COCl,72 the modeof decomposition of such a complex depending on the temperature.The reaction between chlorine and CC1,Br 73 involves C1 and CCI,as chain carriers ; the chlorine-sensitised oxidation of CC1,Br ismuch faster than the corresponding bromine reaction, both beingof complex character 74 and yielding carbonyl chloride, chlorine,and bromine.The reaction between chlorine and chloral 75 producescarbon tetrachloride, carbon monoxide, and hydrogen chloride, alsoby a chain mechanism involving CC1,CO as a chain carrier, stableunder the given conditions (70-90") ; chains are broken by itsbimolecular decomposition. The corresponding oxidation 76 yieldscarbonyl chloride, carbon monoxide, and hydrogen chloride, the ratebeing determined by the amount of absorbed light only; peroxideformation of the type CCl,,CO,O, is here assumed.The photobromination of dichloroethylene,77 although complicatedby the reverse reaction, resembles in many respects the photo-synthesis of hydrogen bromide.At low pressures, bromine atomsare removed a t the vessel surfaces, the rate thus rising as diffusionto the wall is prevented by the addition of carbon dioxide or helium ;a t high pressures a retardation is found owing to the recombinationof bromine atoms by triple collisions, with a, corresponding changein the labs. factor of the rate expression. The radical involved in thechain mechanism is C,H2C1,Br ; a quantum yield of lo3 is recordedfor certain conditions, that of the reverse reaction being of theorder 10. In the bromination of acetylene,78 which is somewhatsimilar in kinetics, a quantum yield of over lo4 has been recorded;increased pressure raises the rate to a maximum by prevention ofdiffusion of bromine atoms to the wall, but a t high pressures, byreason of the high rate of reaction, convection effects are such as to70 See also W.Brenschede, 8. physikal. Chem., 1938, B, 41, 237.71 W. Franke and J. H. Schumacher, ibid., 1938, B, 40, 115.72 W. Brenschede, ibid., 1938, B, 41,254.73 H. J. Schumacher, ibid., 1939, B, 42, 324; cf. H. G. Vesper and G. K.Rollefson, J . Amer. Chern. Xoc., 1934, 56, 1455.74 W. Franke and H. J. Schumacher, 8. physikal. Chern., 1939, B, 42,297.7 6 W. A. Alexander and H. J. Schumacher, ibid., 1939, B, 44, 57.7 6 Idem, ibid., p. 313.7 7 K. L. Muller and H. J. Schumacher, ibid., 1939, B, 42,327.76 Idem, ibid., 1938, B, 30, 36296 GENERAL AND PHYSICAL CHEMISTRY.keep the rate approximately proportional to Iaabs., with surfaceremoval of bromine atoms still predominant.There is no necessityfor consideration of a Br, complex. 79 The photochemical formationof trichlorobromomethane from chloroform and bromine is againsomewhat similar to that of hydrogen bromide and to the reactionbetween chlorine and chloroform. 8OThe gas-phase kinetics of the photolysis and iodine-sensitiseddecomposition of C2H,12 81 probably resemble those of the corre-sponding reactions in carbon tetrachloride solution 82 ; both involvethe radical C2HpI. Thermal reaction is considerable in both cases.Analysis of kinetics of photohalogenation in solution is in generalcomplicated by the fact that an appreciable concentration of halogenatoms will not be reached even when the radiation is such as toproduce complete dissociation in the gas phase, by reason of thecage-eff ect of solvent molecules and the resulting primary recombin-ation.The nature of the reacting species may then be uncertain.Quantum yields will tend to be low. A value of approximately0.01 is recorded 83 for the action of illuminated bromine on bromo-benzene in carbon tetrachloride ; although there is no appreciablethermal reaction at ordinary temperatures, the postulated mechanisminvolves two secondary thermal reactions following the primaryaddition reaction between activated bromine molecules and thebromobenzene. No substitution reactions were here observed.The photochemistry of di-iodoacetylene and tetraiodoethane inhexane has been studied 84; in aqueous solution, the photoreactionbetween bromine and hydrogen peroxide 85 and the photo-oxidationof the nitrite ion by bromine 86 have also been investigated.A comprehensive review of photochemistry in the Schumannultra-violet has been given by W.G r ~ t h . ~ 'The analysis of the kinetics of such systems is, of course, greatlyfacilitated by the ease with which the light intensity may be altered,the concentration of chain carriers or other intermediates beingthus varied by a method which largely avoids the introductionof other complicating factors. Some recent extensions of suchtechnique may be briefly considered. H. W. Melville and his co-79 Cf. J. E. Booher and G. K. Rollefson, J . Amer. Chem.SOC., 1934, 56,80 V. Braunwarth and H. J. Schumacher, Kolloid-Z., 1939, 89, 184.8 1 W. H. Janneck and E. 0. Wiig, J . Amer. Chem. SOC., 1940, 62, 1877.82 Cf. H. J. Schumacher and E. 0. Wiig, 8. physikal Chem., 1930, B, 11,45.83 D. L. Hammick, J. M. Hutson, and G. I. Jenkins, J., 1938, 1959.84 J. W. Tamblyn and G. S. Forbes, J . Amer. C'hem. SOC., 1940, 62, 99.8 5 A. E. Callow, R. 0. Grifith, and A. McKeown, Trans. Paraday SOC.,86 Idem, ibid., p. 559.2288.1939, 35, 412.87 Z. Elektrochem., 1939, 45, 262RlTCHIE : KINETICS : PHOTOCHEMISTRY. 97workers have made considerable use of the “arc plus spark”technique, where the combined effects of light of two wave-lengthsare considered in relation to the effects of each observed separately.For instance, in the photolysis of ammonia,43 where the removalof hydrogen atoms by some reacting species is a possibility, examin-ation has been carried out on the principle of decomposing ammoniain the presence of hydrogen atoms, and by observing, by the para-hydrogen method, whether the hydrogen-atom concentration isreduced thereby.Hydrogen atoms were thus produced fromhydrogen by mercury vapour and the mercury arc giving the re-sonance line 2537 A., while in the same vessel ammonia was directlydecomposed by 2100 A. radiation from a zinc spark. In this par-ticular case the ratio of the combined rate of para-hydrogen con-version (spark plus arc) to the sum of the rates produced by sparkand arc separately was unity, indicating that photodecomposingammonia had no appreciable effect on removing hydrogen atomsat an abnormally fast rate.Similar technique has been employed in attempts to determinethe mean life-time of the NH, radical for given conditions.Theuse of the rotating sector for such general purposes (e.g., the life-time of the fluorescing diacetyl molecule ss) is well known. In theammonia decomposition, the basis of the method was to generatehydrogen atoms and NH, radicals a t known time intervals. Atsufficiently small intervals, NH, radicals, if generated first, will beremoved by subsequently formed hydrogen atoms, by the reactionH + NH, --+ NH, ; a t larger intervals the NH, radicals may beotherwise removed and no inhibition of the ammonia decompositionwill be found.At some intermediate interval inhibition might justbe observed, this then giving an indication of the life-time of theradical. The light beams from the mercury lamp and the sparkwere at right angles to each other and entered the reaction vesselthrough a slit in a concentric revolving cylinder. With this arrange-ment the ratio of rate of spark alone to the difference betweencombined rate and arc alone [X/(XA - A ) ] was observed to be greaterthan unity, thus indicating inhibition, not only a t small timeintervals but also a t the large interval of 70 sew.An interesting extension of such methods of determining the life-time of short-lived molecules has been described for the case of themethyl acrylate photopolymerisation.s9 If two sharply definedbeams of light of equal intensity and separated by a considerabledistance are allowed to fall on the vapour, the total rate of polymeris-ation will be the sum of the rates produced by each separately.88 G.M. Almy and S. Anderson, J . Chem. Physics, 1940, 8, 805.8Q T. T. Jones and H. W. Melville, Proc. Roy. SOC., 1940, A, 175, 392.REP .-VOL . XXXVII . 98 GENERAL AND PHYSICAL CHEMISTRY.If, however, the active species carrying on the chain process isremoved by mutual recombination (rate of removal of P cc P2),superposition of the beams will give a total rate 4% times that ofeach separately. As the distance between the beams is increased,there will be a gradual increase in rate to d% times the superimposedvalue, and hence the distance at which the rate has the mean value+(1 + d%) will be a measure of the distance of diffusion of theactive species, this in turn being connected with the mean life-time.The theory of the relationship between diffusion and mean life-timehas been worked out in detail for the case of hydrogen atoms, wherethe various factors of diffusion, recombination, and mean life-timeare known, and comparison then made between calculated and knownvalues.Extension of such calculation to the polymerisation ofmethyl acrylate indicates a mean life-time of the active polymeras 1.8 secs. for given conditions; the absolute value of the reactionvelocity coefficient between active polymer and monomer and thesteric factor may consequently be estimated.Reactions in Solids.-In the last few years considerable advancehas been made in the theoretical and experimental study of absorp-tion in solids,g0 but there are few direct measurements of the quantumefficiency of photochemical reactions under such conditions, mainlybecause of the diflticulties of measuring the light energy absorbedand of ensuring uniform exposure of the solid.C. F. Goodeve andJ. A. Kitchener 91 describe a method in connection with the de-composition of the dye Chlorazol Sky Blue adsorbed on titaniumdioxide, whereby the quantum efficiency has been found for lightof 3650 A. and the change of efficiency studied for the approximaterange 7200-3200~. The quantum efficiency in the extreme redis negligible but rises to a steady value of approximately 2 x 10-5for the region in which the dye absorbs strongly. A sharp rise isfound at the shorter wave-lengths corresponding to the thresholdof absorption by the dioxide.Measurement at 3 6 5 0 ~ . indicatedin the earlier stages of illumination some inhibitory action, succeededby a rise in quantum efficiency dependent on the concentrationof " vulnerable " dye molecules; in the later stages, the efficiencydecreased more rapidly than the total concentration of unbleacheddye, this being taken to indicate that some dye molecules, by reasonof their arrangement on the surface, do not undergo photosensitis-ation by the dioxide.92 The photochemical oxidation of ammonia,to nitrous acid in aqueous solution is also sensitised by titaniumdioxide.g3O1 Trans.Faraday Soc., 1938, 34, 570. 90 Cf. Ann. Reports, 1938, 35, 85.91 Cf. C. F. Goodeve and J. A. Kitchener, ibid., p. 902.98 G. G. Rao, 2. phyaihl. Chem., 1939, 184, 377LAWRENCE : COLLOIDS. 99Barium and strontium azides are decomposed by ultra-violetlight a t room temperature, the thermal decomposition being acceler-ated by suitable pre-ill~rnination.~~ The threshold for the absorp-tion of ultra-violet light by azide ions in solution and in the solidstate is 2600-2700 A., and that for the photochemical reaction isin the same region; prolonged illumination with light of wave-length less than 2360 A. produced nuclei of metallic barium. Para-hydrogen is formed by mercury-arc illumination of solid hydrogeniodide at low temperature, but normal hydrogen is formed by thephotolysis of formaldehyde and methyl alcohol under the same~onditions.9~ This is taken to indicate that the two hydrogenatoms forming a molecule come from the same formaldehyde ormethyl alcohol molecule.New PubZicatio~zs.-Mention may be made of " Photochemistryand the Mechanism of Chemical Reactions " (Prentice-Hall Inc.,New York) by G.K. Rollefson, reviewed by P. A. L e i g h t ~ n . ~ ~General photochemical technique is described in " ExperimentalMethods in Gas Reactions " (Macmillan & Co., Ltd.) by A. Farkasand H. W. Melville. M. K.5. COLLOIDS.Within the last ten years the whole face of colloid chemistry haschanged. Previously, direct application of the methods of chemistryand physical chemistry were rare, such as the work of Loeb, andwere little more than groping in the dark. To-day the molecularweights of molecules examined and synthesised have increased t o atleast a million-a thousandfold increase.This is due primarily tothe recognition that polymer molecules are built up from small simplerepeating units and that naturally occurring substances such as theproteins are also definite giant molecules whose molecular weightcan be determined. It is true that this advance has been largelydue t o the introduction of new methods of attack such as the ultra-centrifuge and X-rays, but the classical colloid property of highviscosity has also played an important part. At the same time,advances in surface-film research have opened up an entirely newfield, but it is a field which does not properly belong t o colloids,although the association is close.The surface properties of long-chain amphipathic molecules are properties of molecularly dispersedsubstances, whereas the true colloidal ones are due t o the micelles.04 W. E. Garner and J. Maggs, Proc. Roy. SOC., 1939, A, 172,299.95 A. Farkas and L. Farkas, J . Amer. Chem. SOC., 1939, 61, 3393.96 Ibid., p. 2567100 GENERAL AND PHYSICAL CHEMISTRY.It is somewhat ironical that the micellar colloids, such as the soaps,which were formerly expected to throw light on more complexlyophilic colloids, are now relegated to a comparatively small sub-group. At the same time serious attempts are being made t oprovide an exact theoretical treatment of hydrophobic sols on thebasis of a diffuse double layer and surrounding ionic atmosphere,the Debye-Huckel equation being used.It is easy to see how farwe have advanced from the old facile phrase that “ Colloid chemistryis the chemistry of surfaces,” in which was meant the reduction ofinterfacial energy by adsorption.To see the present position, it is necessary to consider the sizesof colloid particles of the various types. Formerly, the size wasso large that there was no continuity from normally sized moleculesto colloid particles. Now, the range is continuous. Formerly,Brownian motion was often used as a size-limiting criterion. Now,we have continuity from small Brownian motion in the largestparticles to ordinary heat motion in the smaller ones, so that wecannot make the arbitrary division between crystalloid and colloidon this ground any more.Emulsions contain particles with diameterfrom 0.5 to 5p ; hydrophobic sols range from a few pp to more thanIOOpp. About 5p is the upper limit for observable Brownianmotion. Micellar colloids, such as the soaps, have “molecularweights ” of the order of 15,000. On the other hand, polymers rangefrom less than 100 to more than a million. Hydrophobic sols arewell above the million mark. The polymer molecules illustrateparticularly that mere size alone is not enough; their shape, orlength/breadth ratio, is of the greatest importance. Here we getcontinuity between the long polymers and the classical cases ofnon-spherical, hydrophobic sols such as vanadium pentoxide.Our present position is that we enquire how molecules aggregateto form colloid particles; i.e., what shape is the particle and whyis it stable? The first part still requires the idea of reduction offree surface energy in hydrophobic and micellar colloids.Inpolymers, reaction kinetics provide the answer. Stability of theaggregation colloids depends upon their charge and the surroundingionic atmosphere, whereas in polymers solubility is true solubility,though we must regard the layer of solvent around the particles asmore closely bound than in ordinary crystalloidal solutions. Wethus return to a conception of solvation which, however, is quitedifferent from and less vague than the older theories of solvation.Hydrophobic 801s.The problem of the stability of hydrophobic sols, of their electricalproperties, and of their coagulation has received much attentioLAWRENCE : COLLOIDS.101lately.follows :H. R. Kruyt 1 has summarised the position excellently as1. Why do particles of a mere suspension tend to adhere to2 . What must be done to prevent this? (question of peptis-3. What promotes coalescence ? (question of flocculation).each other 1 (question of attraction).ation).He then points out that Question 1 was in the past answered bythe concept of reduction of free surface energy, but that the van derWaals forces responsible fall off as a high power of the distance fromthe surface. This shortness of range is difficult t o reconcile with muchexperimental evidence.As regards Question 2 , the older conceptof the electric double layer paid insufficient attention to the originof the double layer and to its nature. The problem of coagulationis complicated by the opposite explanations: by the adsorptiontheory, whereby all changes of stability are explained as changes ofthe number of charging ions and adsorption of discharging ions;or by Gouy’s restoration of the Helmholtz concept of the diffusedouble layer which explains the changes by disturbance of the“gegenion” atmosphere. E. K. Rideal,2 puts the matter in anotherway. The colloid particle may be regarded as a very large ion,whose charge arises from surface ionisation. Hence, all suchparticles must be ionogenic. On the other hand, we may assumethat there is adsorbed at the interface an electrolyte as ion-pairs ofwhich one ion is held more firmly than the other which wanders awayinto the solution.It is obvious that, in a sol free from coagulatingelectrolyte, stability is determined solely by the behaviour of theparticles towards each other.The logical origin of the present attempts to visualise the state ofaffairs and to provide a theoretical treatment is to be found in therecognition of the difference between E and < potentials. The Epotential is the potential from the interface to the bulk of the disper-sion medium, whereas the electrokinetic potential is that betweenbulk of solvent and exterior of the layer of liquid adhering to theinterface. The < potential is therefore that across part of a diffusedouble layer (Gouy, von Smoluchowski, Stern).H.C. Hamaker has drawn curves in two dimensions of the profileof the potential/distance-from-particle relation, and has outlinedthe types. S. Levine and G. P. D ~ b e , ~ I. Langm~ir,~ B. Derjaguin,gNed. Chem. Ver., Symposium on Hydrophobic Colloids, 1937, 7.Trans. Paraday SOC., 1940, 36, 1.Trans. Faraday SOC., 1939, 35, 1125; 1940, 36, 215; S. Levine, Proc.Roy. SOC., 1939, 170, A , 145, 165; J. Chem. Physics, 1939, 170, 165.Ibid., 1938, 6, 873. Trans. Faraday SOC., 1940,36, 203.Ned. Chem. Ver., 1938, 16102 GENERAL AND PHYSICAL CHEMISTRY.and others have followed this method to calculate the conditions ofstability and coagulation of hydrophobic sols.The potential curveis the sum of the van der Waals and the electrical forces. Thegeneral result is that, at some distance from the particle surface,there is a minimum. The energy function exhibits a maximum closeto the particle and a minimum further out. Inside the maximumthe particles are attracted ; outside, repulsed. There is naturally atendency to correlate this minimum with the appearance of thixo-tropy, and some workers have assumed that orientation of aniso-dimensional particles is a requisite for gelation due to these forces.However, we see later (p. 119) that this picture is impossible onexperimental grounds, but the nature of the forces of attraction isnot considered in the problems of irregular packing of the particleswhich is shown to be the first requisite.The connection betweenthe minimum and the properties of pastes is obvious. The originalpapers of Levine and Dube and of Derjaguin and the discussion 7should be consulted. E. J. W. Verwey also suggests that, in casesof emulsions, in which the charge cannot be raised to a value largeenough for stability, an emulsifier shifts the double layer potentialdrop towards the outer phase. Solid emulsifiers are also thought toact in the same manner.Micellar Colloids.The best known micellar colloids are those formed by the soapsand similar " paraffin-chain " salts. Tbis name has been introducedto cover a large number of synthetic substances used during recentyears as detergents and wetting agents. G. S. Hartley's epithet" amphipathic '' admirably describes these substances which, asJ.W. McBain first recognised, are molecules consisting of polar andnon-polar parts in a molecule large enough for each to exert its ownproperties. For instance, the paraffin-chain is insoluble in waterwhile the end groups -CO,Na, -HS03, -NH,, etc., are soluble.Aggregation of the paraffin chains reduces the interfacial energy a tthe paraffin-water interface, and micelles are built up with a water-soluble exterior. Their size seems to be well defined, probablybecause it is determined by the packing of the exterior polar groups.Some of these are ionised. A valuable account of these substancesand their synthesis has been published by H. K. DeangThe constitution of solutions of paraffin-chain salts mentioned in7 Trans.Faraday Soc., 1940,36,711.3 Report of Symposium on Wetting and Detergency (Leather TradesChemists Society) (A. Harvey, 1937). See also F. E. Bartell, Ind. Eng.Chem., 1939, 36, 31 ; and, as emulsifiers, W. Clayton, " Technical Aspects ofEmulsions," p. 9 (Leather Trades Chemists Society) (A. Harvey, 1935).Ned. Chem. Ver., 1937, 58LAWRENCE : COLLOIDS. 103the Annual Reports for 1936 10 is now generally accepted, anda considerable amount of further evidence has been found for theabrupt change of properties at the critical concentration a t whichmicellar aggregation sets in. C. R. Bury’s demonstration that thelaw of mass action requires a very rapid increase in colloidal associ-ation at this point has been confirmed by G.S. Hartley and others.11If C is the concentration of soap, x the fraction aggregated, and Kthe equilibrium constant, thenThe form of this equation explains the steep rise of solubility ofsoaps which occurs over a few degrees.*N. I(. Adam and H. L. Shute l2 measured the surface tension ofdilute soap solutions and its variation over long periods by meansof the sessile drop method. In very dilute solutions, Q falls slowlyfor a week or more, finally very nearly reaching the value for con-centrated solutions. For a chain of sixteen carbon atoms thecritical concentration, c, was ~/1000, and for the twelve-carbonchain, ~ / 1 0 0 . Salts reduce c but without altering the final valueof Q. J. Powney and C. C. Addison l3 also measured surface tensionagainst air and interfacial tensions against xylene for some soaps.Table I gives their values for c; 0.1% of sodium hydroxide wasadded to suppress hydrolysis, since this affects the value of c.K = 1z[C(1 - x ) ] ~ / C X .TABLE I.c (molar).soap.From o against air. From u against xylene.Sodium laurate .................. 0.02 1 0,020,, myristate ............... 0.0029 0.0029,, palmitate ............... 0.00065 0.00066,, stearate ............... 0.000075J. Powney and L. J. Wood l4 found the electrophoretic mobility toreach a maximum a t the critical concentration---M/lOO for sodiumlaurate. Variations were found with chain length and also markedtemperature changes. E. E. Wark and I. W. Wark,l5 working withtrimethylcetylammonium bromide, sodium cetyl sulphate, andpotassium laurate, found that flotation ceases to be effective at c.R.C. Palmer and E. K. Rideal l6 measured adhesion numbers a t45” and found minima a t :Dodecyl sodium sulphate ............ ~ / 1 0 0Cetyl sodium sulphate ............... 3~/1000Cetylpyridinium bromide ............ 3~/1000l3 Ibid., p. 372.-10 P. 110. l1 ‘‘ Paraffin-chain Salts.” (Hermann, 1936.)l a Tram. Farctday SOC., 1938, 34, 758.14 Ibid., 1939, 35, 420. * First observed by Krafft, and since named the ‘‘ Krafft point.”l5 Nature, 1939, 143, 856. 16 J., 1939, 573104 GENERAL AND PHYSICAL CHEMISTRY.J. Stauff 1' determined the hydroxyl-ion activities of sodium soapspotentiometrically and found the activity to increase with increasingconcentration up to the critical value : log c falls linearly withincrease of chain length.K. Hess, W. Philippoff, and H. Kiessig l8concluded from X-ray examination of solutions that the soap ispresent as molecules, amorphous micelles, and crystalline micelles.G. S. Hartley and D. F. Runnicles l9 deduced from diffusion experi-ments that the micelle of cetylpyridinium salts has a diameter of52 A.The more concentrated solutions of the soaps and the puresoaps themselves continue to provide problems. The soaps areperhaps uniquely complicated among the paraffin-chain salts ;nevertheless, their behaviour is of theoretical and practical import-ance. J. W. McBain's phase-rule studies of soap-water systemshave been continued for sodium palmitate-sodium chloride-waterat 90°,20 sodium laurate-sodium chloride-water,21 and sodiumpalmitate-sodium laurate-sodium chloride-water atA.S. C. Lawrence23 has prepared a number of soaps in purecondition. On heating, sharp melting points are not obtained, butwhen the isotropic liquid is cooled most of the substances passthrough a plastic state which on further cooling changes into micro-crystalline solid. Cases examined by X-rays gave only a sidespacing in the plastic region, but showed the low-temperaturemodification to be fully crystalline. Of the stearates, those of lead,zinc, and thorium did not exhibit the plastic phase, and zinc oleatealso failed to show this modification. It is shown that many of themelting points recorded for the metal soaps are in~orrect.~4 R.D.Vold and M. J. Vold 25 have shown that sodium palmitate exists insix phases between 70" and 320". Their existence was establishedby a dilatometric method, supplemented by microscopic investig-ation in polarised light. Four of these phases are believed to beliquid-crystalline, but their nature so far is obscure. The phasesdistinguished with certainty are shown in the table, together withtransition temperatures and mean coefficients of expansion.Curb fibres to sub-waxy soap ............ 117' 0-0085Sub-waxy soap to waxy soap ............ 135 0.0 148Waxy soap t o sub-neat soap ............ 208 0.027Sub-neat soap to neat soap ............... 253 0.0035Neat soap to isotropic liquid ............ 292 0.002 1Phase change.Temp. Vol. increase, c.c./g.17 8. physikal. Chem., 1938, 183,55.Is Proc. Roy. SOC., 1938, A, 168, 420.2o J . Amer. Chem. SOC., 1938, 60, 2066.22 Ibid., 1939, 61, 30.24 J. Braun, " Die Metallseifen," Leipzig 1932.25 J . Amer. Chem. SOC., 1939, 61, 808.Kolloid-Z., 1939, 88, 40.21 Ibid., p. 1870.23 Trans. Paraday SOC., 1938, 34, 660LAWRENCE : COLLOIDS. 105These results may be compared with those of sodium laurate-water systems,26 and also with the mesophase postulated in mono-layer~,~7 although it is impossible to see how a monolayer can beliquid-crystalline in either the classical smectic or the nematic state.R. D. Vold 28 has described the liquid-crystalline phases in sodiumoleate-waher systems, and has attempted to correlate these withother liquid-crystalline systems.Illustrations are given, but furtherexamination, perhaps by X-rays, is needed before their constitutioncan be determined. The following table summarises the results forsodium laurate-water.Composition range.Phase. Soap, wt. 74. Soap, mols. %.Isotropic solution ..................... 0-35.8 0.00-4.4Middle soap ........................... 38-52 4.80-7.95Neat soap .............................. 5 5 . 6 7 5 9.1-1 9.5Waxy soap ........................... 77-88 21-2-36.3Curd fibre .............................. 91.3-100 45-2-100The intermediate phases in the anhydrous soap may be comparedwith J. D. Bernal's results 29 for thallous stearate, which shows asmectic phase well defined by its layer flow.K. Hermann 30 madean X-ray examination of the smectic and the solid phases, butBernal's closer optical examination of the transition disclosed muchgreater complexity. At a few degrees below the melting point, thesubstance becomes uniaxial; on further cooling, the axes cross andit becomes increasingly biaxial. At about 100" there is an abruptchange and another form with smaller birefringences but with thesame orientation appears. This is the form stable a t room temper-ature. Unpublished photomicrographs taken in polarised light bythe Reporter disclose four textures between liquid and solid. TheReporter also observed that the colour of the first phase appearingbelow the melting point was invariably yellow (with crossed Nicols),and that the intermediate phases of sodium stearate are also yellow-an observation recorded by Vold and Vold in the case of the waxyphase of sodium palmitate. This curious uniformity in sampleswhose thickness varied from a small fraction of a mm.up to 0-5 mm.may be due to orientation at the glass surfaces-a phenomenonalready observed to occur most persistently in other nematicliquid- crys t alline subst a n ~ e s . ~ l2 6 J . Amer. Chem. Soc., 1939, 61, 37.2 7 D. G. Dervichian, J . Chem. Physics, 1939, 7, 931.28 J . Physical Chem., 1939, 43, 1213.29 Trans. Paraday SOC., 1933, 29, 1074.31 M. Q. Friedel, Ann. Physique, 1922, 18, 271, 474.Ibid., p. 973106 GENERAL AND PHYSICAL CHEMISTBY.Internal Solubility in Soap Micellas and Peptisation.Water-insoluble Substances.-S.U. Pickering 32 stirred oil intoconcentrated soap solutions and obtained systems which he describedas emulsions. The oil content was 99% in one case. His method ofanalysis contains an error which reduces this value to 95.2%, but,in actual fact, his “ clear lumps of gel ” were not emulsions at all.Under the microscope they can be seen to be pastes of very thinplates of solid soap dispersed in the oil. Pickering also found,however, that, on diluting his oil-soap-water pastes, he obtainedtwo layers-one of emulsion and one of clear soap solution whichstill contained oil. A direct analysis of these sols has been madeby A. S. C. Lawrence,33 who found that in one case 100 C.C. of clearsol contained 2-74 g. of potassium oleate and 2 C.C.of “Nujol.”This, however, was not the maximum amount of oil which couldbe dissolved. Lawrence used a simpler and more direct method ofdetermining solubilities of oil in which 90% alcohol was used assolvent.* In the absence of soap the “ Nujol ” is insoluble, so theamount dissolved by the soap can be determined directly. Thesaturation values for 100 g. of soap dissolved in 2000 C.C. of 90%alcohol were :soap. “ Nujol ” dissolved, C.C.Sodium stearate ................................. 110Potassium stearate .............................. 170Caesium stearate ................................. 200Sodium laurate .................................... 80G. S. Hartley34 measured the solubility of trans-azobenzene inaqueous solutions of cetylpyridinium salts by a colorimetric method,and found that it was proportional to the concentration of para&-salt ions over a wide range of concentration of the soap.m esolubility fell off abruptly below the critical concentration at whichmicelle formation starts. J. W. McBain and T. M. Woo 35 measuredthe solubility of a water-insoluble t dye in laurylsulphonic acidsolutions. The dye was dissolved either from the solid state or fromtoluene solution. Maximum solubility (per unit weight of soap) wasfound in the most dilute solution. Solubilities in other soap solutionsare also given as empirical “ dye numbers,’’ which were derived fromthe distribution of dye between soap and toluene solutions. Incontrast to the above result, solubility was zero in the homologouspotassium soaps of the saturated carboxylic series up to the octoate.That is, solubility appears only when micelles are present.McBain32 J., 1907, 91,2001 ; 1917, 111, 95.33 Trans. Paraday SOC., 1937, 33, 816.34 J., 1938, 1968.* The soap is still colloidally dissolved in this mixture.7 Seep. 108.85 J. Physical Chern., 1938, 42, 1099LAWRENCE : COLLOIDS. 107and Woo criticise the experiments of other workers on the groundthat formation of unstable von Weimarn sols is not guarded against.The experiments of Hartley and Lawrence show that equilibrium wasreached without supersaturation, so the criticism is invalid.Wuter-solubk Substances.-As early as 1886 36 it was shown thatsoap lowers the consolute temperature of phenol-water and cresol-water systems.A. S. C. Lawrence3’ pointed out that glycerolimproves soap solutions for blowing bubbles by peptising them, andthat the improvement cannot be due to an increase of viscosity, aspreviously assumed, since this is reduced. E. Angelescu and D. M.Popescu 38 examined the effects of a number of soaps upon theo-cresol-water system and found that the solubility of the cresol wasincreased. They found the viscosity of the ternary system to passthrough a maximum, after which it fell to a small value on furtheraddition of cresol. The maximum was attributed to an intermediatedegree of dispersion of the soap. They suggested that the finaldispersion of the soap was molecular. E. Lester Smith39 alsoobserved increased solubility in water of cresols and ethyl acetatein the presence of soap, and concluded that the phenomenon couldnot be due to adsorption and that the increase of solubility wasmutual. Lawrence40 found that sodium soap solutions werepeptised by alcohols, amines, and phenols.Peptisation occurs firstin the homologous series at the C, soap. Sodium stearate solutionsare saturated by amyl alcohol at 6 mols. of alcohol per mol. of soap ;for hexyl and heptyl alcohols the figure was about 4, and for oleyland cetyl about 1. For aniline, the value was 7, but from setting-point curves it was concluded that peptisation ended at 4 mols. andthat the extra three were dissolved internally. This suggestion wassupported by the ease with which the first four were taken up ascompared with the sluggish solution of the last three.Fig. 1 4 1shows diagrammatically the possible cases : (a) is that of peptisationby a water-soluble substance, (b) that for a water-insoluble one, and( c ) that for the intermediate case, which is the commonest. Case (a)applies to the lower alcohols, glycerol, and cane sugar. Measure-ments cannot be made with these liquids which are miscible withwater already. Case ( b ) is that of paraffins, azobenzene, etc.; andcase (c) is that of aniline, intermediate alcohols, and other substancesnot completely miscible with water but containing hydroxy-, amino-,and other polar groups. There is a distribution of the added sub-3a W. AlexBefY, Wied. Ann., 1886, 28, 305.37 J .Physical Chem., 1930, 34, 263; Proc. Roy. SOC., 1935, A , 148, 59.38 Bul. SOC. Chim. Rodnia, 1930, 12, 58.39 J . Physical Chem., 1932, 36, 1401.40 Trans. Faraday SOC., 1937, 33, 325.41 A. S. C. Lawrence, “ Wetting and Detergency,” p. 203108 GENERAL AND PHYSICAL CHEMISTRY.stance between soap and water which is largely in favour of thewater in case ( a ) ; e.g., some 50% of glycerol is needed to peptise5% soap. At some stage the consolute temperature of the waterand added liquid will fall below that a t which the experiments aremade. Nicotine has two consolute temperatures and the phasediagram is a ring. Addition of soap reduces the size of the ring untilit vanishes at,about 8% of sodium stearate a t room temperature.Peptisation is attributed to formation of loose complexes betweenthe -C02Na group and the polar group of the added peptiser.Thedifference between the results of McBain and Woo and those of otherworkers on insoluble substances is explained by the fact that McBainand Woo’s water-insoluble dyes (described in their paper only bytheir trade names) were amines which naturally formed salts withFIG. 1.(a) Peptisation alone. (b) Internal solution ( c ) Internal solution andin the soap micelle. peptisation.their soap-laurylsulphonic acid-which dissolved in the water.For instance, yellow AB is l-benzeneazo-2-naphthylamine, and theBB dye is the tolueneazo-compound.A. S. C. Lawrence42 has examined solutions of soaps in theinert solvent, “ Nujol.” These are interesting in that they are thereverse case of aqueous solutions.The soluble part is now thehydrocarbon chain, while the polar groups are insoluble. Solutiondoes not occur until thermal separation of the polar groups is reached,the temperature for which approaches that of the melting point ofthe pure soap to isotropic liquid. About this temperature a fluidsolution is obtained; below it, the soap solution sets to a gel. At alower temperature, corresponding with the transition in the soapfrom plastic to crystalline, the gel is transformed into a paste ofmicro-crystals. The gel region therefore corresponds with the plasticstage in the soap. Its temperature limits are lowered by additionof polar s~bstances,~~ of which the fatty acids are the most efficient.42 Trans.Faraday SOC., 1938, 34, 660. Ibid,, 1939, 35, 702LAWRENCE COLLOIDS. 109Calcium and barium soaps are peptised by small amounts of water,which forms fatty acid by hydrolysis. Phenols are also active.Emulsions.Discussion of emulsions is frequently complicated by inclusion ofvery dilute suspensions of oil in water which are stabilised by acharge and not by an emulsifying agent. It is most desirable thatthe name should be restricted to those systems in which a thirdsubstance acts as emulsifier. The dilute suspensions are typicalhydrophobic sols. As such they are of special interest since theparticles are spheres with a uniform adsorbing surface. Excludingthese suspensions, we have the two classes of emulsions-oil inwater, and water in oil.In general, particle size is fairly uniformand large, the diameter being from 0-5 to 5p. G. S. Hartley44has attempted a theoretical discussion of particle size and stability,starting from the model of oil dissolved inside the micelle. Thistreatment suffers, however, from the difficulty that there seems tobe no continuity from this model to emulsion droplets. Saturationof the micelle by oil represents an equilibrium. Some abrupt changemust occur then, and very much larger droplets become the stableones in the emulsion. Pickering’s in which there wasa large excess of soap available for the micelle swelling through theintermediate stage, show the two processes of internal solution andemulsification proceeding independently and without continuity .The chief property of emulsions considered hitherto is that ofphase inversion. With soaps or similar amphipathic water-solublesubstances as emulsifiers, oil-in-water systems are obtained.Addi-tion of salts of calcium, barium, thorium, etc., causes inversion.The wedge 45 theory attempted an explanation on the unsupportedand highly improbable idea that the bivalent soap molecules had awedge form which determined the nature of the emulsion. Thesimple expedient of using a univalent kation (e.g., silver or thallous)which forms a water-insoluble soap without giving the wedge formdisproves this theory, since phase inversion does occur. J. H.Schulman and E. G. Cockbain 46 have investigated this question,using complex formation in the emulsifier layer to alter its nature.Substances which had already been studied in monolayers werechosen as emulsifiers ; e.g., cholesterol dissolved in paraffin gives noemulsion when shaken with water, and sodium cetyl sulphate in thewater gives a very poor one, but when both substances are present44 “ Wetting and Detergency,” 1937, 153.45 P.Finkle, H. D. Draper, and J. H. Hildehrand, J . Amer. Chem. SOC.,4 G Trans. Paraday SOC., 1940, 36, 651.1923, 45, 2780110 GENERAL AND PHYSIOAL CHEMISTRY.excellent emulsions are formed and their stigness is proportional tothe rigidity of the complex formed in monolayers. The stability ofthe emulsion prepared with various substances which reacted withmonolayers of cholesterol was in the same order as the reactivity.It appears that, for stable emulsScation of oil in water, the oildroplets must be charged (by the emulsifying agent a t the interfacewhich also moves the double-layer potential drop further out intothe water).The emulsifying layer must be close-packed andconsequently the interfacial tension is small. The amount of water-soluble substance present must be greater than that required toform the monolayer, but when a complex-forming substance is addedin the oil no such excess is required. The stability of the emuIsionis attributed to van der Waals attraction between the non-polarparts of the emulsifying molecules, and is also affected by interactionbetween the poIar groups of complex-forming substances. Furtherexperiments showed that the emulsion type was oil-in-water whenthe droplets were charged (as with simple amphipathic emulsifierand with some complexes) but water-in-oil when they were un-charged.Phase inversion can be brought about by addition ofcalcium salts to a sodium soap-stabilised emulsion, and a similarresult is found when sodium hydroxide is added until the pH is 14.Similarly, heptadecylamine-stabilised emulsions are inverted a t verylow pH values. It is concluded that water-in-oil emulsions requirethat the emulsifier layer should be uncharged and that it shouldpossess a considerable rigidity.The picture of emulsions, although not complete theoretically, isclear. The two types, oil-in-water and water-in-oil, are funda-mentally different : the former owe their stability to the charge onthe oil droplets, and this charge is due to the ionised layer of emulsi-fier. Stability will be poor unless this layer is close-packed-eitherby choosing a suitable substance or by coupling one, which givesonly a gaseous film, with a substance which forms a close-packedcomplex.Such emulsifiers will also lower the interfacial tension atthe oil-water interface, but it seems that the actual value of theresidual interfacial tension is not a stability-determining factor, andonly affects the work required to form the emulsion. J. L. van derMinne47 remarks that pairs of liquids such as ether-water andphenol-water do not form emulsions in the neighbourhood of theirconsolute temperature although the interfacial tension is lower herethan in most stable emulsions. It should be noted, however, thatthis failure to form emulsions is perhaps due to mutual solubility.When water is the continuous phase, the charged oil droplets repelone another since the medium is ionised.With oil as continuous47 Symposium of Hydrophobic Colloids, Ned. Chem. Ver., 1938, 138LAWRENCE : COLLOIDS. 111phase this factor does not operate, since there is no ionic atmospherearound the droplets. It seems that the stability of water-in-oilemulsions is due to rigidity of the emulsifying layer. Against this,the residual interfacial tension is insuflicient to pull the droplets intospheres as in oil-in-water systems. Such emulsions, therefore,contain irregular sack-shaped particles of water whose size dependsupon the mechanical treatment of the emulsion as opposed to thetrue equilibrium which may be reached in oil-in-water emulsions.R.C. Pink48 has shown that magnesium and calcium oleates,formed by metathesis in inversion of sodium oleate-stabilisedemulsions, are hydrated soaps insoluble in both phases. They actas emulsifiers, therefore, as finally divided solids. Van der Minne hasexamined emulsions stabilised by ferric oxide flocculated from a sol.A. King and L. N. Mukerjee 49 have measured the rate of hydro-lysis of amyl acetate emulsions, using a number of emulsifyingagents. The rate was always higher in the emulsions, but the rateper unit area of interface was greatest in the absence of an emulsifierand varied with different emulsifiers.The Viscosity of Colloidal Systems.Graham first recognised high viscosity as a characteristic ofhydrophilic colloids, and named his volume efflux viscometer a“ Colloidoscope.” Later, gelation was recognised also as character-istic of such colloids.Still later it was found that many such solshad an anomalous or non-Newtonian viscosity; i e . , the viscositycalculated by the usual equations, which assume Newtonian flow,varied with the rate of flow. At low rates of shear, the value wasvery large, decreasing then with increasing rate of shear. Thisanomaly, combined with W. F. Darke, J. W. McBain, and C. S.Salmon’s observation 5O that gelation of soap sols was unaccom-panied by any change in osmotic pressure or electrical property, ledto the idea that anomalous sols were gels too dilute to supportthemselves under the deforming force of gravity.” It was alsofound that many sols whose viscosity was anomalous becamebirefiingent when in flow : these were sols in which the particleswere rod-shaped.If the rods are themselves birefringent, orient-ation by flow makes the sol appear birefringent. By a curiousmuddling of ideas, the orientated arrangement showing flow bi-refringence was associated with the “ structure ” causing theviscous anomaly. Meanwhile A. Einstein 51 had published his well-** J., 1938, 1252; 1939, 619. J. SOC. Chern. Ind., 1938, 57, 431.6O Proc. Roy. SOC., 1921, A, 98, 395.61 Ann. Physik, 1906, [iv], 19, 289. * This is perpetuated in the German name equivalent to “structureviscosity.’ 112 GENERAL AND PHYSICAL CHEMISTRY.known equation, which is quoted in the text books in the sterileform rsol = r ( l + 2-5V), where rsol is the viscosity of the sol,7 that of the pure dispersion medium and V the total volume of thedisperse phase.I n other words, the increase of viscosity of the soldue to the colloid particles is proportional to their total volume only.Much work has been carried out recently upon the flow character-istics of colloids, but the larger part has been concerned withmeasurements of the flow of pastes and similar complex mixturesin which the ‘‘ body ” is a commercially useful attribute. Thescience of “ rheology ” is mainly concerned with such systems andhas been summarised in a recent Atthe same time, however, theoretical and ex-perimental investigations have been made ofthe problem, less complex systems beingused, and the results constitute an importantadvance in our knowledge and perhaps astarting point for the examination of themore complex systems.If we consider thesimplest case of spherical particles too largeto show Brownian motion and dispersed toa dilution sufficient for absence of any inter-action between the particles, then in flow thevelocity gradient rotates the particles. Eachis then surrounded by a region of disturbed,,of,/e forwtonianf / o ~FIG. 2.Idealised picture of case of flow, in which the liquid is travelling in adirection different from that of the stream- laminar$ow of paste tntake.line flow of the undisturbed bulk of the dispersion medium.Fromthese considerations Einstein derived his equation. When, however,concentration is increased, complications set in. Fig. 2 shows anidealised picture of the cross-section of a tube through which asuspension is flowing in laminar or telescopic flow. It is clear that,as shown, each particle of volume xd3/6 occupies a volume of d3, sothat the concentration is ca. 50%.In laminar flow each annulus of particles must move with avelocity which increases from zero at the wall t o a maximum at theaxis of the tube. Shear can take place only across the liquidseparating the annuli. In Fig. 2 , flow would be vanishingly smalland the apparent viscosity very large. Now, according to theEinstein equation there must be no interaction, either directlybetween the particles or between the regions of disturbed flowaround them.Viscosity will therefore have started to increaserapidly above the Einstein value before the stage shown in Fig. 2.When the particles overlap, flow conditions will become mores2 G. W. Scott Blair, “ Industrial Rheology,” 1938LAWRENCE : COLLOIDS. 113complex still, and the measured viscosity greater. If there areforces of attraction, whether long-range electrical or van der Waals,the paste will have a rigidity and a yield point. It must be recog-nised that there will be difficulty at high concentrations in distin-guishing between real rigidity and very high transference of momen-tum which w i l l reduce the velocity gradient towards zero and raisethe observed viscosity towards infinity.It cannot be urged toostrongly that the viscosity of the dispersion medium is unchangedfrom its normal value. It is the conditions of flow which aredisturbed. The rotational movement of the particles and theFIG. 3.Orientation of rods in velocity gradient.resulting disturbance of flow are sometimes spoken of as an extradissipation of energy, but it is clearer to regard this as a “side-tracking ” of some of the energy which the ordinary equations forcalculating viscosity assume to be used only for Newtonian shearingof a homogeneous system.A spherical particle can have no special orientation with respectto the streamlines of flow. An infinitely thin rod, however, is quitedifferent. When it lies across the velocity gradient, its interferencewith laminar flow will be greatest, and the velocity gradient willexert a couple upon it which will rotate it until its long axis liesalong the streamline (Fig. 3).If it is infinitely thin, it will, in thisposition, have no effect upon the flow or the viscosity. Particles s114 GENERAL AND PHYSICAL CHXMISTRY.orientated will, however, show their maximum birefringence. Thiswas established first by J. Robinson,53 who measured viscousanomaly and streaming birefringence in the same apparatus--acoaxial cylinder viscometer with a transparent bottom for theoptical observations. Fig. 4, taken from his results for a sol of thetobacco mosaic virus, shows that maximum viscous anomaly isassociated with minimum birefringence and vice versa.The earlierworkers regarded hydrodynamic orientation as irreversible so longas the shearing was continued. RotatoryBrownian motion is a disorientating force, and the flow conditionsThis, however, is not so.Radiizns prr sec.FIG. 4.042% Tobacco mosaic virus.Variation of relative viscosity (0) and double refraction of 3020, ~ at 14.4",--- at 19-6", with rate of shear.depend upon the balance of the two forces. Brownian motion willcause a rotational diffusion of rods. If there is any orientation,i e . , more in one position than in any other, then there will be morediffusion away from the position of orientation because there aremore particles there. Now, particles diffusing clockwise will tend tobe restored to their original position (Fig.3) by hydrodynamicorientation, but those moving counter-clockwise will have theirmovement accelerated and will perform a complete rotation. Theposition at which angular velocity is at a minimum will contain moreparticles at any moment and will therefore be the observed directionof orientation. This will move in a clockwise sense with increase ofrate of shear. The intensity of the birefringence also increases with53 Proc. Roy. SOC., 1939, A, 170, 519; also A. S. C. Lawrence, ibid., 1937.A, 163, 323LAWRENCE : COILOIDS. 115rate of shear. We can see that reduction of size of rods or rise oftemperature increases disorientation at any fixed rate of shear andvice versa.It is noteworthy that rise of temperature increases thedisorder and therefore increases the viscous anomaly. The viscosityof the dispersion medium retains its normal negative temperaturecoefficient which may be larger than the increase of viscous anomaly,but, if the comparison is made between viscosities relative to thoseof the solvent a t the temperatures chosen, then the increase ofanomaly is shown. Robinson has verified this surprising conclusion.Several theoretical treatments of the high viscosity of suspensionsof rod-shaped particles have been attempted. W. Kuhn 54 andM. L. Huggins s5 have shown that the viscosity will depend uponthe particle shape, but they assume random orientation and so failto explain viscous anomaly. Most of the experimental workershave assumed complete orientation. This is only true when theparticles are very large and elongated, i.e., when Brownian dis-orientation is negligible compared with hydrodynamic orientation.For example, X-ray examination of such suspensions in flow hasshown that there is almost complete orientation.56 E.Guth 57 andP. Boeder 58 have given theoretical treatment on the lines discussedabove but failed to calculate the additional dissipation of energy dueto the precessional motion of the rods. Robinson surmounted thisdifficulty by using the optical data which give both the numberorientated and their direction * with respect to the streamlines. Asalready mentioned, the orientation is statistical-the preferredposition is that through which the particles pass most slowly.Theactual motion of a rod-shaped particle is precessional owing to itsrotation in the velocity gradient plus its linear flow along the stream-lines. From the experimental data for the orientation, the rate ofdissipation of energy has been calculated, and the apparent viscosityof the sol so obtained is in good agreement with the experimentalvalues.This work has not been extended to the very important systemsin which the particles are not rigid, e.g., polymers. This problemis complicated by the non-linear form of chemically linear polymermolecules. As Kuhn has shown, a velocity gradient imposes atension on an anisodimensional particle. An aperiodically coiledpolymer molecule will therefore be pulled out as well as rotated.64 2.physikal. Ghem., 1932, A, 161, 1, 427.66 J . Appl. Physics, 1939, 10, 700.5 8 I(. Hess and J. Gundermann, Ber., 1937, 78, 1800; K. Hess, H. Kiessig,and W. Philippoff, Naturwiss., 1938, 26, 184." Kolloid-Z., 1936, 74, 147; 75, 15. " Z. Physik, 1932, 75, 259.* More exactly, that of the optic axis, but in the cases chosen it coincidedwith the long geometrical axes of the rods116 GENERAL AND PHYSICAL CHEMISTRY.When, however, its long axis lies along the streamline, it will tend tocoil up again. During these complex movements it will carry withit a, certain amount of the dispersion medium immobilised inside thecoil. Shear may cause strain birefringence, and the straightenedout particles may then show orientation birefringence.Theobservations of A. von Muralt and J. T. Edsall 59 on myosin solsshowed the two effects in the same sol, the strain birefringenceoccurring in the less highly sheared part, which behaved like a weakgel, while the more highly sheared part was fluid and showed theorientation effect.The general and very important conclusion is that viscous anomalyis the result of interference with the normal laminar flow of thesolvent by the precessional motion of H e particles. The viscosity isanomalous (i.e., it varies with the rate of shear) because the inter-ference depends upon the balance of hydrodynamic orientationand Brownian disorientation. Other conditions being unchanged,variations of rate of shear alter the hydrodynamic orientation.Further complexitiesare introduced when there is mutual interferencebetween the particles, either via the dispersion medium or by directadhesion. The consequences of interference and the resultingphenomena have not been recognised sufficiently by workers on thissubject.The need for working at low concentrations has beenignored, especially by the rheologists.C. F. Goodeve 6O has proposed a theory of anomalous viscosity andthixotropy based upon particle adhesion, the flow properties of thesystem being determined by the rate of breaking and remaking thesel i n k s . This theory is obviously inadequate to explain the type ofviscous anomaly described above, where there is no mutual inter-action at all, but it would seem to be of particular value as applied topastes.The experimental evidence from the better-known t k o -tropic systems shows that the rate of re-formation of links is quiteslow compared with the rate of breaking. He concludes that suchsystems have a “scaffolding” structure. This name seems un-fortunate, since it suggests somewort of cubic lattice-like structurerather than the picture of a random mesh which has been acceptedfor many years.Gehtion and Thixotropy.The confusion already mentioned between anomalous viscosityand gelation makes it necessary to discuss the latter before summaris-ing the position. All gelation is due to a reduction of the solubilityof the solute : by cooling sols of hydrophilic colloids and by additionof electrolytes to hydrophobic ones in amounts less than those59 J .Biol. Chem., 1930, 89, 289, 315; Trans. Faaraday Xoc., 1930, 26, 837.6o Ibid., 1939, 35, 342LAWRENCE : COLLOIDS. 117required for coagulation. In a few cases, e.g., soaps, the gel stagecan also be reached by warming the less soluble curd fibre phase inwater. This is important in so far as it shows that the gel stage isnot merely a metastable system. Thixotropic gels, which havereceived the most attention recently (partly because they are mostamenable to experimental investigation), are gels which are brokendown easily to fluids by shaking and reset on standing for a shorttime. Attempts have been made to explain thixotropy by thepeculiar electrical conditions (cf. p. 102). H. FreundlichY6l H. C.Hamaker,3 I.Langmuir,62 and S. Levine have tried to explainthixotropy by means of the potential curves for charged particles.This method, however, is unsatisfactory for two main reasons.First, it does not explain the resetting after a time lag; and,secondly, it fails to explain that thixotropy is usually a phenomenondue to non-spherical particles. ’ For example, alumina sols show nothixotropy when prepared by precipitation and subsequent peptis-ation. When, however, prepared by Crum Brown’s method ofboiling the basic acetate with water, the sol is thixotropic. In theformer case the particles are spherical; in the latter, they are thincrystalline plates. It may be that this criticism is due to the quitefortuitous fact that the maximum concentrations of sphericalparticles attainable in true hydrophobic sols are too small for a con-tinuous structure to be built up, whereas filamentary or plate-shapedparticles can build such a structure.On the other hand, the criticismmay be real on the grounds that the partial coagulation required forthixotropic gelation results from discharging parts of the surfaces ofthe rods or plates leaving the remainder charged-a position notrecognised in the potential-curve theories. If this were the case, thelocally discharged patches would adhere to similar patches in otherparticles and a continuous structure could be built up provided thatthe particles were themselves fairly large. A. I. Rabinerson 63distinguishes between sols in which the particles have their t: poten-tial generally lowered, and those where it is lowered at points only.In the first case, aggregation is compact, and in the latter extended.We may, however, consider first the gelation process without anyreference to the nature of the forces of adhesion, merely making thereservation that an elastic gel does require adhesion at the points ofcontact in the mesh structure of particles whatever their shape maybe.Certain facts must be recognised. Gels are optically anisotropic.Thixotropic gels are being sheared when broken down to fluid byshaking. This clearly implies orientation of the anisodimensionalparticles. Examination of the gelation in polarised light reveals61 “ Thixotropy,” 1935, p. 13.63 Acta Physicochim. U.R.S.S., 1938, 8, 733.62 J .Chem. Physics, 1938, 6 , 873118 GENERAL AND PHYSICAL CHEMISTRY.this clearly. On re-setting, the flow birefringence in the sol, andtherefore the arrangement of particles in groups each with commonorientation, is destroyed. This observation is in complete agree-ment with the anomalous viscosity of sols of anisodimensionalparticles described above, and a further connection is found in thetemperature effect. Rate of gelation increases with temperatureprovided that there be no solubility change. I n the soaps, ofcourse, there is a marked solubility change, and K. H. Meyer andA. van der Wijk 64 have suggested that gelation of gelatin is alsodue to a sudden change of solubility. Where there is no such changeof solubility, as in the thixotropic systems, the temperature effecthas been confirmed.N. Sata and N. Naruse 65 have observed theeffect in alumina sols. M. Prasad and D. M. Desai 66 have alsoobserved it in complex inorganic systems. Bentonite also shows theeffect.67 I. Langmuir 68 has measured the rate of disappearance ofstreaming birefringence in bentonite sols as a function of temper-ature, and has found that the time decreases as temperature rises,He pointed out that the loss of birefringence is due to rotationalBrownian motion. Using the rotational diffusion equation and thetemperature coefficient of the rate of loss of birefringence, hecalculated the energy of activation, but, at the lowest concentrationused by him, two-thirds of this is due to the change of viscosity ofthe water.The Einstein equation for the Brownian rotation (ofspheres) is O2 = KTt/4xr3q, where 8 is the average angle movedthrough in the time t, r is the radius of the particles, and q theviscosity of the dispersion medium at the absolute temperature T.The energy barrier appeared a t a concentration of 1.4%. Langmuirdeveloped a, theory which required the particles, which are thinplates, t o be orientated in a pseudo-cubic lattice and held in positionby long-range electrical forces of adhesion. Although this distanceis large, vix., 5000a., it is not large compared with the size of theplates, which have a diameter of 4500 A. If we calculate the volumefilled by these discs rotating, we get an effective volume filled whichbrings them into contact a t about 3%.From the criterion of touch-ing a t points, the effective volume will be d3 instead of nd3/6. Thisbrings the critical concentration above which the particles are notfree to rotate to ca. 1.5%. The effective volume is obviously lessthan d3 but will be somewhat increased again by addition of the fieldof disturbed flow around the particles. This correction, however,84 Helv. Chim. Acta, 1937, 20, 1331. 66 K0110id-Z., 1939, 86, 10.2.6 6 J . Indian Chem. SOC., 1939, 16, 117.67 E. A. Hauser and C. E. Reed, J . Physical Chem., 1936, 40, 1169; 1937,8.9 J . Chem. Physics, 1938, 6, 873.41, 911; E. A. Hauser and D. S. le Beau, Kolloid-Z., 1939, 86, 105LAWRENCE : COLLOIDS. 119will not be large. Langmuir observed experimentally 1.4% as thecritical concentration.It seems probable, therefore, that theenergy barrier is simply due to interference with Brownian dis-orientation and that, in sufEciently dilute sols, the whole of thetemperature coefficient would be covered by the change of viscosityof the water. Hence it seems that Langmuir’s results are explicableon the random-mesh theory of gel structure. This explanation alsoexplains the anisotropy of the gel, whereas Langmuir’s pictured gelshould have a birefringence which is the sum of the individualbirefringences of the orientated particles. It has been suggestedthat such gels might be micro-crystalline 69 and therefore effectivelyisotropic, but such a picture fails to explain adhesion qf the micro-crystalline units.Incidentally, working on bentonite suspensions,Hauser 67 has experimentally confirmed the disorientation theory ofgelation.The phenomenon of “ rheopexy ” 70 is also explicable on the samelines. Xheopexy is the increased rate of resetting of a thixotropicgel as a result of gentle motion-rolling a test-tube of the liquidgently backwards and forwards between the hands. It is clear thatsuch a movement is one which causes mild turbulence and is a dis-orientating one and not a steady shearing. It is therefore anaddition to the disorientating action of rotatory Brownian motion.When the particles are flexible, the picture becomes more complex.The breakdown of a felted mass of intertwined fibrils, whetheradhering at points of contact or not, is obviously very much morediflicult than the corresponding case of rigid rods. A reduction ofapparent viscosity by continued shearing has been observed ingelatin solutions.71 In these cases, it is particularly important thatthe limitations of the Ostwald or other volume efflux viscometersshould be realised. If the system takes an appreciable time to reachsteady flow, it will have passed through the instrument long beforethis steady state is reached. Table I1 gives a survey of possiblecases.True emulsions are not included in this table, since they show apeculiar anomaly due to distortion of the droplets and work is doneagainst the interfacial tension.The size distribution is also altered.This, incidentally, is the principle of the homogeniser, which appliesa very high rate of shear to the emulsion either by forcing it througha small orifice or by shearing between two plates close together androtating in opposite directions at high speed.Rigidity of particles6Q Trans. Paraday Soc., 1940, 36, 730.70 H. Freundlich and F. Juliusburger, ibid., 1935, 31, 920.71 B. Kandelaki, G. Kikodze, and N. Dolidze, J. Phys. Chern. Buss., 1937,10, 524120Particles.SphericalNearlyisodi-mensionalAnisodi-mensional.FlexiblefibrilsRigid rodsRigid platesGENERAL AND PHYSICAL CHEMISTRY.Examples.Oil hydrosnlsAu, C, clay solsTABLE 11.Behaviour under shear.Gelatin, agar,soaps, polymersBenzopurpurinV,O,, tobaccovirus, Hg sulpho-salicylic acidFe 0 ,A1,0,, ben-;o:i teDilute sols. Concentrated sols.Einstein’s case Plastic pastesEinstein constant 72 7, Y9abnormally large ;flow Newtonianor slightly ano-malousPossible orient- Gelation or highation with de- viscositygradationorientationHydrodynamic Thixotropic gelsis defined as rigidity sufficient to withstand deformation by Brownianmotion.Rigid rods then rotate, whereas flexible ones are distorted.Benzopurpurin forms thin crystalline needles which can be seenunder the ultra-microscope to be partially flexible. Before thistable can be regarded as an accurate scheme, it is necessary to relatethe size factor with viscous behaviour, and also to define concen-tration in an entirely new manner.Size Factor.-As we have seen, the behaviour of anisodimensionalparticles is determined by the balance between hydrodynamicorientating forces and rotational Brownian disorientation.TheReynolds number gives the upper limit to which the rate of flow canbe increased without turbulence setting in. It follows, therefore,that particles below a certain size and length/breadth ratio cannotbe orientated by but the viscosity of their sols will be largerthan those of spherical particles of similar size. The flow will beNewtonian, but the Einstein factor will be unduly large. Hugginshas discussed the viscosity in the homologous series of fatty acids.When particles are so large that Brownian disorientation is neg-ligible, a new disorientating factor appears. The finite cross-section of the rods is itself a source of instability, since one side tendsto move faster than the other.There is, therefore, still a torquewhich, in one sense, causes precession. In this connection, it is wellknown that rods falling through a liquid at rest deviate from astraight-line path. J. W. McBain and (Mrs.) M. E. L. McBain 7* havecarried out an ingenious experiment in which rates of sedimentation72 E. W. J. Mardles, Trans. Paraday SOC., 1940, 36, 1007.73 A. Peterlin, 2. Physilc, 1938, 111, 232; KoZEoicE-Z., 1939, 38, 230.74 J. Amer. Chew. Xoc., 1937, 59, 342LAWRENCE : COLLOIDS. 121of fibrils of quartz, and of the spheres formed by fusing the fibrils,were measured in water. Great elongation in the fibrils causedonly a 13-fold increase of viscous resistance.The authors concludedthat the high viscosity of suspensions cannot be due to length alone.This explanation is obviously incorrect. The particles are fallingthrough a fluid across which there is no velocity gradient. There isstill a shape effect, but not approaching that found in viscositymeasurements where the particles increase the viscosity by theirprecessional motion which, as we have seen, requires the velocitygradient. McBain and McBain's case is similar to the conditions inthe ultracentrifuge.H. G. F. Winkler 75 has found thixotropy in suspensions in waterand organic liquids of non-spherical mineral particles of microscopicsize. Spherical particles, however, showed no thixotropy. E. A.Hauser and D. S. Le Beau 76 have measured the viscosity of suspen-sions of fractionated montmorillonite and bentonite and found thatat very low concentrations the value decreases with increasingparticle size. They attribute the high viscosity of the finest sols tohydration, but it is possible that the increased Brownian dis-orientation with decreasing size is sufficient to explain these results.S. A. Glickman 77 detected anomaly a t low rates of shear in solsof cellulose nitrate. The critical stress for disappearance of anomalyfell with increase of particle size. Roughly, we may say that(with a given concentration) the viscosity will rise to a maximumwith increase of particle size and then decrease again. The changesof viscosity of any system will then obviously depend upon theoriginal position on the curve of the starting material.Degradation of particles by shear is important. Two differenttypes of degradation must be recognised-breaking long ones intoshorter lengths, and breaking down bundles of particles which maybe in a nematic or para-crystalline arrangement, i.e., with long axesparallel but with no other regularity. Again, the starting pointdetermines whether breakdown increases or decreases the viscosity.R. Signer 78 has observed degradation of polymer molecules a t veryhigh rates of shear. Soap solutions probably show degradation of thesecond kind, but it should be noted that formation of gel is a quiteseparate case. E. Heymann 79 has shown that the sol-gel change insodium oleate-water systems is unaccompanied by any volumechange. This is not surprising, however, if we regard gelation asthe linking of particles already present in the sol (and possiblyincreasing in number) and not as any change in the packing of the7 5 Kolloid-Beih., 1938, 48, 341. 7' Kolloid-Z., 1939, 86, 105.77 J . Phys. Chem. RUSS., 1938, 11, 825.7 8 Helv. Chirn. Acta, 1936, 19, 1324. 79 Trans. Paraday Soc., 1938,34,689122 GENERAL AND PHYSIUAL GHEWSTRY.molecules in the micelles. M. Ulmann,*O however, has found that thespecific volumes of this soap fall into three well-defined regions withtransitions at 0.03% and at 0.9%.Eflective Volume and Concentration.-The picture of the preces-sional motion of an anisodimensional particle in a velocity gradientwill be in three dimensions a sort of solid double cone. Now, it isclear that some concentration will be reached at which the totalvolume effectively filled by these solid cones will be equal to the totalvolume of the solution. Further increase of concentration will causeinterference, higher transmission of momentum, and much higherviscosity. The hydrodynamic analysis of the results will no longerbe accurate. The longer the particles, the lower will be the criticalconcentration. It is clear that the effective volume will be less thanxE3/6, where l is the length of the rod, but greater than x(&l)2e,which is H. Staudinger’s assumed value, 8 being the thickness. Itis also clear that this critical concentration will vary with the rateof shear. With complete orientation, for example, the effectivevolume has fallen t o the actual volume. Staudinger assumed,however, that his “ limiting ” concentration was constant. Heused Ostwald viscometers in which rate of shear and thereforecritical concentration varied continuously across the tube. Simi-larly, a system of anisodimensional particles a t rest will retainstreaming birefringence permanently unless the particles havesufficient space to disorientate themselves under Brownian motion.The relation between the latter critical concentration (which is nota completely abrupt change) and the critical concentration in flowwill not be definite, since the latter is not a constant a t all. Never-theless, it is of the greatest importance that its sigmficance should berealised-namely, that all measurements of viscosity of solutions ofanisodimensional particles should be made below it. Slightlyabove it, the overlapping of the particle volumes is shown by a verylarge deflection on setting the Couette viscometer in motion. Thedeflection then falls and reaches a steady value. If, with a sol ofthis concentration, the outer cylinder is rotated by hand through afew degrees, the inner (hanging) one follows as if linked rigidly. Onstanding, however, the hanging cylinder gradually returns to restduring some hours. From the rate of return, viscosities can bemeasured at rates of shear as low as one revolution of the outercylinder in six months. The return shows that the sol is liquid, butwith a velocity gradient so small that its viscosity approaches infinity.The rate of return may be compared with Maxwell’s relaxation time.Only in the case of the larger polymer molecules is anomalousviscosity observed. Below a molecular weight of 100,000, flow isso 2. phvsikal. Chem., 1938, 182, 18LAWRENCE : COLLOIDS. 123Newtonian and sometimes remains so to even higher values.81Flow birefringence has been observed at much lower values,82 e.y.,in a polystyrene of molecular weight 5000, but the orientation ofsuch particles was only slightly different from random orientation.Not until the molecular weight rises above 50,000 does seriousorientation begin. At 630,000 the angle of isocline has only reachedca,. 75" (random orientation gives 45", and complete orientation 90").Signer explains his results by a partially coiled form. Increasing therate of shear tends to straighten these, and the observed flow bire-fringence increases more rapidly than required by the simple orienta-tion theory as developed for rigid rods.82, 83 Signer's results for poly-styrenes and cellulose nitrate solutions support the hydrodynamicaltreatment of anomalous viscosity, and he rejects Kuhn's theory.For Newtonian solutions, the Staudinger equation is generallyaccepted as giving results of the correct order of magnitude, but hisattempt to justify it on hydrodynamical grounds is open to seriouscriticism. G. Gee ** has shown that molecular weights of fraction-ated crepe rubber determined by the Staudinger viscosity methodagree well with those obtained by osmotic pressure measurementsfrom values of 60,000 to 350,000.I n the Second Report of the Academy of Sciences of Amsterdamon Viscosity and Plasticity, J. M. Burgers 85 summarises the workon the viscosity of suspensions of small particles of elongated form.He concludes that Staudinger's formula is an approximation to oneof the theoretical formulze, but it is clear that no general formula hasyet been found. It is generally agreed also that the results forpolymer molecules, especially those of higher molecular weight, canbe explained only by some assumptions of deformation. The com-plexity of the position in this respect has already been indicated.It may be noted, however, that combination of temperature effectsupon viscous and optical streaming properties may provide anexperimental approach.86 It is apparently satisfactory to applyhydrodynamical treatment. This is not surprising in the case of thelarge rigid rods in thixotropic systems, and it is easy to understandin the case of polymers on account of the very close association ofthe solvent with the surface of the polymer It may beDie hochmolekularen organischen Verbindungen," 1932,p. 188.Staudinger,82 R. Signer and C. Sadron, Helv. Chim. Acta, 1936, 19, 1324.83 R. Signer and P. von Tavel, ibid., 1938, 21, 535.84 Trans. Faraday Xoc., 1940, 36, 1163.8 6 P. 113 et sep.8 6 A. S. C. Lawrence, Zoc. cit., ref. (53).87 Cf. J . N. Brensted and K. Volyvartz, Trans. Paraday AsIoc., 2940, 36,619.Both this summary and the main report are very impor-tant and should be consulted124 GENERAL AND PHYSICAL CHEMISTRY.noted that the most important assumption underlying the hydro-dynamic treatment is that there shall be no slip a t the solvent-particle interface.I n the above-mentioned report C. J. van Nieuwenburg remarksthat, although much experimental work has been carried out inmeasuring the plastic properties of various materials, the greaternumber of the instruments used do not allow an unambiguousrelation between rate of shear and shearing stress; further, thatmeasurement of “ plasticity ” cannot be expected to yield usefulresults unless plasticity is itself defined and complications due toother properties allowed for or avoided, e.g., elasticity, thixotropy,wall slip, etc.Finally, it cannot be emphasised too strongly that the investig-ation of flow properties should be carried out in a suitable viscometer.Where viscous anomaly exists, it has been pointed out that the Ost-wald type is unsatisfactory since the sol is moving with differentviscosities in different parts of the tube. The same criticism appliesto observations of streaming birefringence in flow through a tube.Some workers, however, have used the Couette instrument for theoptical work and the Ostwald for viscosity and then compared theresults. Even in the absence of anomaly, the Ostwald type is notentirely satisfactory, since the theoretical treatment requires theabsence of inertia effects (connected with the accelerations ordecelerations of particles or elements of volume of the liquid). It hasbeen shown that the Couette viscometer gives confirmation of theEinstein equation, whereas the Ostwsld gave values slightly toolow.88 For transparent solutions, A. S. C. Lawrence’s coaxialcylinder instrument 89 is the most satisfactory. For plastic pastes,C. F. Goodeve has described an ingenious instrument in which therate of shear is varied by alteration of the spacing of two truncatedcones which take the place of the orthodox cylinders. J. Pryce-Jones 91 has described an important “ thixotrometer ” with whichhe has examined the characteristics of paints.A. S. C. L.A. S. C. LAWRENCE.H. W. MELVILLE.W. J. C. ORR.M. RITCHIE.L. E. SUTTON.8 8 F. Eirich and 0. Goldschmid, KoZZoid-Z., 19.37, 81, 7.89 See J. Robinson, Zoc. cit., ref. (53).90 J. Sci. Instr., 1939, 16, 19.91 J. Oil Colour Chern. ASSOC., 1936, 295

ISSN:0365-6217    

DOI:10.1039/AR9403700023

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

INORGANIC CHEMISTRY.1. INTRODUCTION AND GENERAL.DURING the year under review there has been a steady flow ofpublications covering very many topics in inorganic chemistry.In spite of the comparatively large volume of published work,however, it has proved a dficult task to select any special contri-butions as constituting outstanding advances. The year haswitnessed rather a general advance, made up of small additions toour knowledge here and there, of the continuation of systematicexperimental studies which have been in progress for some time,and of the fuller utilisation of special technique.In this Report brief summaries of a number of diverse experi-mental papers covering practically the whole of the Periodic Tableare included in the first section. No systematic account is given,however, of recent publications on complex salts, nor has the verywide subject of heterogeneous equilibria been considered. In eachof these fields the inaccessibility of certain publications would havegreatly impeded the preparation of a satisfactory survey.Sub-sequent sections of the Report deal with the boron hydrides, withthe fluorination of inorganic compounds, and with the thermaldiffusion method of isotope separation. The review of these topicswas thought to be timely, for in each instance a considerable volumeof important experimental material had accumulated, which hadnot, as far as the Reporters were aware, been sumrnarised in recentpublications.A method of preparing pure deuterium peroxide has been described 1in which deuterium oxide vapour is blown through a mixture ofdeuterosulphuric acid and potassium persulphate at 70-90".Theoperation is conducted in an all-glass apparatus and the mixture ofD20 and D202 formed is fractionally condensed and further enrichedin D,02 by fractional distillation. The dilute peroxide solutionresulting is again sent through the acid mixture, and the cycle ofoperations repeated until lOOyo D202 is obtained. The compoundHDO, is synthesised from equimolecular quantities of H,02 andD202. The method is stated to be suitable also for the preparationof H20, of high purity.A further contribution to the controversial subject of lead sub-oxide has been made by L. L. Bircurnshaw and I. Harris.2 AccurateF. Feher, Ber., 1939, 72, 1789.J., 1939, 1637126 INORGANIC CHEMISTRY.analyses were made of the gases evolved when lead oxalate isdecomposed in a vacuum at 300". These indicate that the reactionis represented by the equation 3PbC20, --+ 2Pb0 + Pb +4C0, + 2C0, and that lead sub-oxide is not formed. This con-clusion is supported by X-ray photographs of the residue and bymeasurements of its electrical conductivity, from which it appearsthat the grey-black powder is an intimate mixture of lead and thered tetragonal form of lead monoxide.New thiobasic mercury salts have been described. The com-pound Hg(HgS)SO, is prepared from mercuric sulphide by mixingit with equal parts of perhydrol and concentrated sulphuric acid,warming until the reaction product is white, and pouring intowater.3 It is shown that Hg(HgS)SO, gives a characteristic X-raydiagram, and is not a mixture of sulphide and sulphate.Thesecond compound has the formula Hg,S,Cr,O,, and is prepared byheating freshly precipitated mercuric sulphide on a water-bath withan aqueous solution of chromium trioxide. Here again, the X-raydiagram is that of a new substance, the structure assigned to whichThe dielectric constant of aluminium bromide has been mea~ured,~and it has been shown that the dipole moment is zero for the liquidand for its solutions in bromine, carbon disulphide, or benzene.These results indicate the existence of symmetrical Al,Br6 moleculesunder the above conditions. Measurements by I. Poppick and A.Eehrman6 of the parachor of this compound in benzene solutionshow that the molecular complexity is the same in solution as inthe molten state.Products containing approximately 75% ofanhydrous aluminium perchlorate have been obtained by the actionof anhydrous perchloric acid on pure aluminium chloride, followedby evaporation of unchanged acid and sublimation of unchangedchloride.' It was shown that aluminium could not be depositedfrom solutions of the perchlorate in organic solvents.Cyanates of silicon, phosphorus, and boron are described byG. S. Forbes and H. H. Anderson.s Silicon tetrachloride reactedon refluxing with a suspension of silver (iso)cyanate in benzene,98% of the product consisting of Si(NCO), (m. p. 26.0", b. p. 185.6")and the remainder of Si(OCN), (m.p. 34.5", b. p. 247.2"). Theidentification of these compounds rested on analogies with ethyl3 G. L. Chabovski and E. Potavian, BuZ. Chim. SOC. Romtine, 1938, 39, 27.4 GI-. L. Chabovaki and T. Trifrirescu, ibid., p. 65.5 V. A. Plotnikov, I. A. Scheka, and Z. A. Jankelevitsch, J . Ben. Chem.6 J . Amer. Chem. SOC., 1939, 61, 3237.7 E. G. Hackenberg and H. Ulich, 2. anorg. Chem., 1939, 243, 99.8 J. Arner. Chern. SOC., 1940, 62, 761.is [Hg (HgS) 2lCr2O 7'Rwsia, 1939, 9, 868EMELI~US AND WELCH : INTRODUCTION AND GENERAL. 127isocyanate and cyanate. Phosphorus trichloride yielded P(NCO),(m. p. - 2.0", b. p. 169.3") as the only volatile product, and borontrichloride yielded a white solid of low solubility which decomposedon moderate heating and, from its analysis and properties, wastaken to be B( OCN),.These authors describe unsuccessful attemptsto prepare boron chlorobromides, fluorobromides, oxychloride, andoxybromide.Interesting new addition products of boron trifluoride withsulphates and phosphates are described by P. Baumgarten and H.He~mig.~ Potassium sulphate absorbs boron trifluoride at about300" and forms the compound K2S04,BF,; further addition of thetrifluoride to this compound occurs with some difliculty, andabsorption ceases before the composition K,S04,2BF3 is reached.Cssium sulphate readily forms Cs2S0,,2BF, a t 250-300", butNa,S0,,BF3 is obtained with difficulty at 310-330". Lithiumsulphate absorbs very little boron trifluoride, and the alkaline-earthsulphates are inwerent, although thallous sulphate yieldsT12S0,,BFp These sulphate-borofluorides are regarded as trueadditive compounds.All evolve boron trifluoride when heated,and axe immediately decomposed by water into the sulphate andboron trifluoride, which is further hydrolysed to boric and fluoboricacids. As in the similar series of additive compounds of sulphurtrioxide, combination between a sulphate and boron trifluorideprobably occurs by completion of the boron electron octet with alone pair of electrons from an oxygen atom in the sulphate ion.Analogous compounds, K,P,0,,4BF3 and K3P04,3BF,, are formedfrom boron trifluoride and potassium pyro- and ortho-phosphate,respectively, at about 400" ; the sodium salts yield simiIar products.Under similar conditions sulphur trioxide gives no additive com-pounds, but decomposes the phosphates.Indications have been obtained of the formation of a boron oxy-fluoride, (BOF),, in the reaction between boron trifluoride andtrioxide at 100-340°.10 The supposed oxfluoride is stable in thegaseous state at these temperatures, but on cooling, it decomposesrapidly with deposition of an unidentified colourless solid.Theoxfluoride is also stated to be a primary product of the reaction ofboron trifluoride with certain salts of oxy-acids; with potassiumorthoborate, nitrate, and carbonate the changes are representedby the following equations : 3KB02 + 6BF3 = SKBF, + 2(BOF),;6KNO, + 9BF, = 6KBF4 + (BOF), + 3N2O5; 3K2C0, + 9BF3 =GKBF, + (BOF), + 3C0,.Similar reactions occur with alkaline-earth salts, except that metallic fluorides are the final solid products.10 P. Baumgarten and W. Bruns, aid., p. 1763.Ber., 1939, 72, 1743128 INORGANIC CHEMISTRY.Magnesium oxide reacts thus : 3Mg0 + 3BF, = 3MgF2 + (BOP),.A similar reaction occurs with calcium oxide, but less readily.During the past year the chemistry of the rare-earth elementshas again attracted considerable attention. J. Ant-Wuorinen 11 hasdescribed a method of separating these elements by fractionalhydrolysis of aqueous solutions of their azides, which is effectedelectrolytically in a diaphragm cell with platinum electrodes. G.Mannelli l2 has found that all the rare-earth metals can be precipit-ated quantitatively as their 8-hydroxyquinoline compounds, andthat the differences of p , necessary for precipitation of cerium,thorium, and yttrium enable these elements to be separated from acrude rare-earth mixture.The electrolytic separation of cerium asceric phosphateI3 has been improved by I. A. Atanasiu and M.Babor l4 and rendered suitable for direct application to a sulphuricacid extract of monazite ; in a supplementary paper 15 these authorshave described the preparation and properties of hydrated andanhydrous ceric phosphate.The preparation and study of pure compounds of individualmembers of the rare-earth group have been continued; W. Feit l6has obtained pure holmium compounds by fractional crystallisationof rare-earth bromates and basic nitrates, and H.Bommer l7 hasexamined the crystal structure of metallic holmium, which has ahexagonal close-packed lattice (a = 3-557 A., c = 5.620 A. ; d =8.764; at. vol. = 18.65 c.c.). Magnetic measurements by Bommershow that holmium obeys the Curie-Weiss law a t 90-515" K.The calculated magnetic moment is 10.6 Bohr rnagnetons. Thepurification of gadolinium from traces of europium, which isreduced to the bivalent state with strontium amalgam, is alsodescribed.ls New compounds of the rare earths include meso-periodates of yttrium, cerium, and erbium of the type M10,,4H20,and a hydrated diorthoperiodate of yttrium, Y412013,11H20.1gContinued interest in the recently-developed chemistry of galliumis shown by a paper dealing with the chlorides. A. W. Lauben-gayer and F.B. Schirmer20 have prepared GaC1, and GaCl, in astate of purity and measured the vapour pressures and vapourdensities over wide temperature ranges. The trichloride, preparedby passing chlorine, diluted with nitrogen, over heated gallium,11 Suomen Kern., 1940, 13, B, 1.12 &ti X Congr. intern. Chim., 1938, 11, 718.l3 J. W. Neckers and H. C. Kromers, J . Amer. Chem. SOC., 1928, 50, 955.1 4 Bull. Acad. Sci. Rournaine, 1939, 20, 27. l5 Ibid., p. 32.16 2. anory. Chem., 1940, 243, 276. l7 Ibid., 1939, 242, 277.l8 L. Rolla, Atti X Congr. intern. C'him., 1938, 11, 766.lo R. K. Bahl and S. Singh, J. I n d i a n Chem. SOC., 1939, 16, 375.2O J . Amer. Chem. SOC., 1940, 62, 1678EMEL$US AND WELCH : INTRODUCTION AND GENERAL. 129melts at 77.0" and boils at 200.0"/760 mm.; below 200" its vapourconsists chiefly of Ga,C& molecules, but the degree of dissociationinto GaCl, increases from 3-44,% a t 205.5" to 88.26% at 498.3".The dichloride was prepared by heating the trichloride with excessof gallium in a vacuum at 175", and distilling off the product at225".Traces of the trichloride formed by decomposition weresubsequently removed by distillation at 160". The lower chlorideforms apparently dimorphous, colourless crystals which melt a t170.5", and slowly decompose into the trichloride and gallium above200". Although the diamagnetism of the solid21 suggests that itis composed of Ga,Cl, molecules, the vapour a t 400470" containsan appreciable proportion of GaCl,, in which gallium exhibitsanomalous bivalency.No evidence was obtained of t'he existenceof the monochloride, GaC1.H. Schulten 22reports the production of rhenium halogeno -pent acarbonyls,Re(CO),X, where X = C1, Br, or I. These are obtained whenK2ReX, or other rhenium halogen compounds are heated in carbonmonoxide a t 230"/200 atm. They are stable in air and insoluble inwater, but will dissolve in organic solvents and may be sublimedin an atmosphere of carbon monoxide. The same author hasstudied the preparation of [Co(CO) J2 from anhydrous cobalt halidesand carbon monoxide.23 The reaction occurs most easily withthe iodide, less so with the bromide and chloride, and not a t all withthe fluoride. Cobalt iodide and carbon monoxide react a t 200 atm.even at room temperature, and the resulting addition compoundCoI,,CO is appreciably volatile. Further reaction, leading to theformation of the carbonyl, occurs through the vapour a t the copperor silver walls of the autoclave.Admixture of finely-dividedmetals with the cobalt halide greatly increases the yield of carbonyl,the effect increasing in the order Au < Pt < Ag < Cu < Cd or Zn.A. A. Blanchard and P. Gilmont 24 have studied the preparation ofcobalt carbonyl, nitrosylcarbonyl, and carbonyl hydride by thecyanide method. Carbon monoxide was found to be absorbedalmost quantitatively by an alkaline suspension of a cobalt salt,forming KCo(CO),, though this reaction took place only when asmall amount of cyanide wits present.Nitric oxide displaced thevolatile cobalt nitrosylcarbonyl, Co(NO)(CO),, from the resultingsolution even when it was distinctly alkaline, whereas acids dis-placed the volatile cobalt carbonyl hydride HCo(CO),. TheMetallic carbonyls have been further studied.21 W. Klemm and W. Tilk, 2. anorg. Chewb., 1932, 207, 175.22 Ibid., 1939, 243, 164.23 Ibid., p. 145.24 J . Amer. Chem. SOC., 1940, 62, 1192.REP .-VOL . XXXVII . 130 INORGANIC CHEMISTRY.tetracarbonyl [Co(CO),], was obtained by allowing the hydride todecompose spontaneously a t room temperature.Reductionof tungsten hexachloride and molybdenum pentachloridea t - 10" to 0" by means of carbon monoxide and iron or zinc dust in amixture of ether and benzene yields the hexacarbonyls, W(CO), andMo(CO),.~~ Increased initial pressures of the monoxide improvethe yields of the carbonyls, which are otherwise of the order of 10-14%; W(CO), may also be prepared from WOCl, in ether by theabove method.Hitherto the, existence of complexes of germanium other thanthe hexafluorogermanates has been doubtful, but evidence has nowbeen put forward26 which shows chlorogermanic acid and thehexachlorogermanates to be capable of existence.When germaniumtetrachloride was dissolved in concentrated hydrochloric acid,migration experiments showed the movement of germanium to theanode, presumably in the form of GeC1," ions. Czsium hexachloro-germanate, Cs,GeCl,, has also been prepared by the addition ofgermanium tetrachloride to a solution of caesium chloride in 1 vol.of ethyl alcohol and 2 vols.of 12~-hydrochloric acid; it was in-soluble in 12N-hydrochloric acid, but dissolved readily in water andwas rapidly hydrolysed. It was stable in dry air a t room tem-perature, but was decomposed above 160" without melting. Thecrystals had a face-centred cubic lattice, with a Ge-C1 distance of2.35 A.The existence of a tribromide of nitrogen has not been reporteduntil recently, although the compounds NH2Br 27 and NHBr, 28have been prepared. M. Schmeisser 29 has now obtained an ammineof nitrogen tribromide, NBr3,6NH,, by mixing brominz vapour withexcess of ammonia at 1-2 mm. total pressure a t 20". Ammoniumbromide is deposited and the new compound condenses as a purple-red solid in a trap cooled to - 95".It decomposes explosivelyabove - 70", according to the equation NBr,,GNH, = N, +3NH,Br + 2NH,. This reaction establishes the constitution, forno other substances likely to be formed under the conditions given(e.g., an ammine of bromine) would yield these decompositionproducts in the observed ratio.25 K. A. Kotscheschkov, A. N. Nosmcjanov, M. M. Nadj, I. M. Rossinskaja,and L. M. Borisova, Compt. rend. Acad. Xci. U.R.X.X., 1940, 26, 54; K. N.Anisimov and A. N. Nesmejanov, ibid., p. 58.2 6 A. W. Laubengayer, 0. B. Billings, and A. E. Newkirk, J . Amer. Chem.Xoc., 1940, 62, 646.2 7 W. Moldenhauer and M. Burgor, Ber., 1929, 63, 1615.28 G. H. Coleman, C. B. Yagor, and H. Soroos, J . Amor. Chem. SOC., 1934,56, 965.Naturwiss., 1940, a$, 63EMEL~US AND WELCH : INTRODUCTION AND GENERAL. 131Metallic phosphides have been the subject of three recent papers.J. L.Andrieux and M. Ch6ne30 have prepared MOP and a newsubphosphide, Mo3P, by the electrolysis of a fused mixture ofsodium metaphosphate, sodium chloride, and molybdenum tri-oxide at 800" between carbon electrodes. Both phosphides resistattack by hydrochloric acid, but are decomposed by sulphuric ornitric acid, fused alkalis, or oxidising agents. M. Heimbrecht andW. Biltz31 have not obtained any evidence for the existence ofFeP,, but have prepared an unstable greyish-black phosphideby treatment of a tin-iron alloy (5% of iron) with phosphorusunder pressure. This gives an X-ray diagram distinct from thatof FeP,.The system iridium-phosphorus has also been studied ; 32the only compounds formed are IrP, and 1r2P.R. Holtje and H. Schlege133 have shown that the action ofphosphine on cuprous, argentous, and aurous halides affords thecompounds CuC1,2PH3, CuBr,2BH3, CuI,PH,, AgI,PH3, andAuI,PH3 in addition to the known compounds CuC1,PH3, CuBr,PH3,CuI,2PH3, and 2AgI,PH3.Continuing their studies on the constitution of complex metallicsalts, F. G. Mann and his co-workers 34 have prepared and examineda wide range of compounds of tert.-alkyl-phosphines and -arsineswith cadmium and mercuric halides. The simplest of the cadmiumcomplexes, [(R3P),CdX2] (class l), almost certainly has a tetrahedralstructure. Two further types of cadmium compound,[(R3P) 2(CdX2) 21and [(R3P)3(CdX2)2] (classes 2 and 3), have been obtained; class 2has the bridged structure typified by (I), and the new type ofstructure, indicated without differentiation of the planes of thebonds in (11), has been tentatively assigned to class 3.Br Br >Cd<pEt3] [Br Br(1.1 (11.)The mercuric halides form five classes of compounds, three of which(classes A, B, and E) are entirely analogous to the cadmium com-plexes.Class C, [(R,P),(HgX,),], includes salts with two differenttypes of structure; some members of the series, e.g.,and [(Bua,As),(HgBr2),], are shown by X-ray analysis to be molecularcoinpounds of class B complexes with the mercuric halide. The30 Compt. rend., 1939, 209, 672.32 K. H. Soffge, M. Heimbrecht, and W. Biltz, ibid., 1939, 243, 297.33 Ibid., p.346.34 R. C. Evans, E'. G. Mann, H. S . Peiser, and D. Purdie, J . , 1940, 1209.[(Pra3As> 2(HgC12) 3131 2. anorg. Chem., 1939, 242, 233132 INORGANIC CHEMISTRY.molecules of the compound [(Et,As),(HgI,),], however, do notpossess a centre of symmetry; their structure is possibly (111).Similar alternative structures (a molecular compound of a class Bcomplex with 2 mols. of mercuric halide, or a structure of type I11with an additional ring) appear possible for class D, [(R,P),(HgX,),],but a preliminary X-ray examination indicates that the latteralternative is unlikely, at least for the compounds [(Et,P),(HgBr,),]and [(Et,As),(HgCl,),]. I n all of the above classes both phosphineand arsine derivatives have been obtained.Bridged complexes containing both cadmium and mercury havealso been prepared.35 The compound [ ( Pra3P) ,CdHgI,], forexample, may be obtained (a) from [(Pra3P),Cd12] and mercuriciodide or ( b ) by boiling an alcoholic solution containing equivalentamounts of [(Pra,P),(CdI,),] and [(Pra,P),(HgI,),]. The secondmethod of preparation indicates that the parent cadmium andmercury compounds of classes 2 and B, to which the product isprobably analogous in structure, are partly dissociated in hotalcoholic solution to give the free radicals (IV) and (V).When thePra PI(IV.) >Cd/'Pra P(V.)I'solution cools, these radicals unite in pairs to give the cadmium-mercury complex, which is less soluble than either of the parentcompounds. Bridged cadmium-mercury complexes containingmixed halogens, e.g., [(Pra3P),CdHgBrzI2], or mixed arsine andphosphine groups, e.g., [(Pra,P)(Pra,As)CdHgI,], are also described.An interesting type of " mixed " bridged complex,has also been prepared ; its molecule is dissymmetrical and containsmetal atoms with both planar (Pd) and tetrahedral (Hg) valencydistribution (VI).This type is similar to the copper complexes[(Pr",As),PdHgBr,l,obtained by D. P. Meller, G. J. Burrows, and B. S. Morris.36 Noanalogous palladium-cadmium complexes were formed, and35 F. G. Mann and D. Purdie, J . , 1940, 1230.3G Nature, 1938, 141, 414EMEL~US AND WELCH : INTRODUCTION AND GENERAL. 133attempts to prepare mixed cuprous- or argentous-mercuric com-plexes, containing two tetrahedral units, were also unsuccessful.A chemical proof of the linear configuration of the complexes[Et,P+AuX] and [Et,As+AuX] (X = C1, I), which is sup-ported by X-ray analysis,37 has been attempted.38 These com-pounds, together with the corresponding bromide, combine readilywith a further molecule of halogen to give 4-covalent auric com-plexes of the type [Et,P+AuX,], which should have a planarstructure.If the two halogen atoms enter the vacant trans-positions of the linear 2-covalent complex, two isomerides of acompound such as [Et3P-+AuBrI,] should be formed by (a)addition of iodine t o [Et,P+AuBr], and (b) addition of iodinemonobromide to [Et,P+AuI]. In the several cases tried the twomethods of preparation gave identical products. It was concludedthat the co-ordinated groups in the auric compound are mobile andtherefore adopt positions giving maximum stability, as in the caseof certain 4-covalent palladous complexes.The chemistry of the oxides of niobium has been clarified byrecent work.According to G. Grube, 0. Kubaschewski, and K.Z~iauer,,~ the oxides NbO,, NbO, and Nb,O may be obtained byreduction of the pentoxide with a mixture of hydrogen and watervapour under appropriate conditions; NbO, is also obtained byheating the pentoxide in argon a t 1150-1300". The individualityof all these oxides is proved by Debye-Schemer photographs, butthe supposed oxide Nb,O, is shown to be a mixture of NbO andNbO,. Although the claim to have reduced Nb,O, to Nb20 withhydrogen has been contested by G.Brauer,40 this reaction has sincebeen confirmed.41 Brauer finds that the pentoxide exists in a tleast two forms, one of which is obtained when niobic acid, precipit-ated from aqueous solution, is dehydrated a t 600-800". Whenthis form is heated above 800" a modification of the oxide giving adifferent X-ray pattern results. The m. p. of niobium pentoxidein an atmosphere of oxygen is 1460". NbO, was prepared byBrauer by heating Nb205 with niobium or hydrogen, and thisoxide, when heated with the metal a t 1750" in argon, gave NbO.Intermediate phases of the type NbO, (x = 2.0-2.5) were preparedby heating mixtures of Nb,05 and NbO, a t 1400". P. Sue4, hasstudied the equilibrium Nb,05 + C += Nb,O, + CO at 715-840°, and has found the heat of formation of Nb,O, t o be 393 kg.-cals.37 F.G. Mann, A. F. Wells, and D. Purdie, J . , 1937, 1828.38 J., 1940, 1235.40 Naturwiss., 1940, 28, 30.41 0. Kubaschewski, 2. Elektrochem., 1940, 46, 284.4 2 J . Chim. physique, 1939, 36, 280.39 2. Elektrochem., 1939, 45, 885134 INORGANIC CHEMISTRY.The hydrolysis of sulphur mono- and di-chlorides, known pre-viously to be a reaction of considerable complexity, has recentlybeen studied in some detail. According to G. H0lst,4~ treatment ofeither chloride with potassium hydroxide in 95% ethyl-alcoholicsolution results in immediate neutralisation of 2 mols. of the alkali,of which a further 2 mols. are neutralised in a subsequent slowreaction. The initial rapid reaction is ascribed to the formation ofS,(OH), and S(OH),, from the mono- and di-chloride, respectively,and the slower change, which is almost inhibited in 99.5y0 alcohol,to the further hydrolysis of these compounds.B. S. Ra0,44 con-fining his attention to the monochloride, has studied the hydrolysisproducts in greater detail. The reaction occurs rapidly and com-pletely at the interface between a solution of sulphur monochloridein carbon tetrachloride and aqueous sodium hydroxide, to whichcadmium chloride is added to react with any sulphide formed. Theproducts are free sulphur, sulphide, sulphite, and thiosulphate, withsulphate and trithionate if dilute alkali is used. All these sub-stances can be formed by secondary reactions of S20 or S,(OH),which, in accordance with Holst's conclusion, is regarded as theprimary hydrolysis product.The compound S,(OH), may corre-spond in constitution with the hypothetical thiosulphurous acid.The existence of disulphur monoxide, S,O, has been confirmed byR ~ o , ~ ~ who prepared it by passing sulphur monoxide into dry carbontetrachloride at - 12" to - 18". The reactions of the resultingdeep yellow liquid were thought to indicate the presence of S,O,formed by the reactions 2SO = S,O,; SO + S20, = S,O + SO,.The S,O decomposes on keeping according to the equation 2S,O =3s + SO,. Cryoscopic measurement's show that the carbon tetra-chloride solutions actually contain S,O in the form of a complexwith dissolved sulphur, to which the formula S,,xS,O (x = 3 or 4) isassigned.The hydrolysis products of these solutions correspondwith those of similar solutions of S,Cl,.The preparation and properties of sulphur monoiodide andthionyl iodide have been examined by M. R. A. R ~ o . ~ ~ Both areobtained by treating a carbon tetrachloride solution of the corre-sponding chloride with dry potassium iodide ; the resulting solutionsin carbon tetrachloride are unstable a t room temperature and theirdecomposition is promoted by light. The hydrolysis products ofthe monoiodide are analogous to those obtained with sulphurmonoiodide, but thionyl iodide and alkali give sulphur, sulphite,thiosulphate, iodide, and hypoiodite, the latter resulting from43 Bull. SOC. chirn., 1940, [v], 7, 276.4 4 Proc.Indian Amd. Sci., 1939, 10, A, 423.4 5 Ibid., p. 491. 4 6 Ibid., 1940, 11, A, 162, 185EMEL~US AND WELCH : INTRODUCTION AND GENERAL. 135iodine formed in the primary reaction. The amount of thio-sulphate formed is greater than that produced in the hydrolysis ofthionyl bromide. The decomposition of sulphur monoiodide andthionyl iodide and their absorption spectra have been studied insomeThe existence of a selenium iodide, SeJ, (n undetermined), isindicated by the absorption spectrum of mixtures of selenium andiodine dissolved in carbon disulphide, and by a study of the equi-librium between the two elements in carbon tetrachloride solution.48The pure hydroselenides of sodium, potassium, rubidium, andcaesium have been prepared by W. Teichert and W.Klemm49 bythe action of hydrogen selenide on the ethoxides in anhydrousalcohol, with exclusion of moisture and oxygen. The crystalstructure of these compounds has been studied, and they have beenshown to resemble the hydrosulphides. The sodium, potassium,and rubidium compounds have a rhombohedrally distorted sodiumchloride structure at room temperature, and the sodium chloridestructure at higher temperatures, whereas the czesium compoundhas the caesiurn chloride structure. The radii of the SH' and SeH'ions are 1.98 and 2.11 A., respectively.Recent studies on affinity by W. Biltz and his collaborators haveincluded the system uranium-~ulphur,~0 in which the compoundsUS,, U,S,, US,, and U,S, are formed, the last two being new;US3 is prepared by the action of S on US, at 600--800", excesssulphur being removed either by volatilisation in a vacuum a t 300"or by extraction with carbon disulphide.The formula of the sub-sulphide U4S, has been confirmed by X-ray analy~is.~l A new oxideof uranium, U,05, has been obtained by R. Lyden 52 by prolongedtreatment of U,08 with potassium hydrogen carbonate solution at100". The reaction is represented by the equation U,08 +4KHC0, = K,[UO,(CO,),] + U,O, + CO, + 2H,O, and appears tosupport the formula UO~,U,OF, for U308. This formula is alsoindicated by measurements of the magnetic sus~eptibility.~3A simple method has been described for conversion of platinum,iridium, rhodium, and palladium into their double czsium chloride^.^The metal is mixed with caesium chloride and treated with ammon-ium chloride vapour and oxygen a t a red heat; attack is due to4 7 M.R. A. Rao, Proc. Indian Acad. Sci., 1940, 11, A , 175, 201.4 8 J. D. McCullough, J . Amer. Chem. SOC., 1939, 61, 3401.4 9 2. anorg. Chem., 1939, 243, 86.50 E. F. Strotzer, 0. Schneider, and W. Biltz, ibid., 1940, 243, 307.51 M. Zumbusch, ibid., p. 322.52 Finska Kern. Medd., 1939, 48, 124.53 H. Haraldsen and R. Bakken, Naturwiss., 1940, 28, 127.54 M. Deldpine, Bull. SOC. chim., 1939, [v], 6, 1471136 INORGANIC CHEMISTRY.chlorine formed from hydrogen chloride and oxygen. Electrolyticreduction of solutions of ruthenium trichloride in hydrochloric acidgives solutions containing Ru" or Ru', or metallic Ru, accordingto the conditions; 55 Ru" is moderately stable in 2~-hydrochloricacid, but Ru' readily gives Ru" and Ru.J. Meyer and H. Kienitz 56have described the preparation of RhF,,6H20, M,RhF, (M =K, Rb, Cs), RhBr,,2H20, and RhI,, together with numerous newdiazido- and triazido-complex salts, and have also studied theconstitution of solutions of rhodium trichloride in various solvents.Renewed interest in polyiodides and related compounds is shownby the appearance during the year of papers dealing with ammon-ium, sodium, potassium, and rubidium polyiodides. In each casethe compounds formed were identified by a systematic study ofthe ternary system alkali iodide-iodine-solvent, the customary wet-residue method being used. In a particularly thorough isothermaland polythermal study of the ammonium c0mpounds,~7 with water assolvent, the existence of the solid phases NH,I,, NH41,,3H,0, andpossibly NH,I,,H,O has been established; NH,I, also appears as asolid phase in the binary system ammonium i~dide-iodine.~~ Theknown polyiodides of rubidium and caesium occur only in theanhydrous state or as organic solvates, whereas those of lithium,sodium, and potassium appear to exist mainly in hydrated forms :ammonium, giving both types, provides a transition between thesetwo classes.The sodium iodide-iodine-water system 59 is unusualin that the polyiodide phases which exist at O", viz., NaI,,2H20 and(probably) Na12,3H,0, apparently contain an even number ofiodine atoms combined with each sodium atom; the preparation ofa few other even polyhalides (e.g., CsBr4, CSI,) has been claimed, butthe sodium di-iodide is unique.It is suggested that the even poly-halides may be molecular compounds, the components of whichcontain odd numbers of halogen atoms, e.g., NaI,NaI,,BH,O orNaI,,Na15,4H,0. A study of the potassium polyiodides,60 con-ducted on similar lines to that of the ammonium compounds, showsthat the solid phases KI,,H,O, KI,,H,O, and two polymorphicforms of K13,2H,0 occur in aqueous systems; the first two com-pounds had been obtained previously by N. S. Grace.61 With5 5 G. Grube and H. Nann, 2. Elektrochem., 1939, 45, 874.56 2. anorg. Chem., 1939, 242, 281.57 T. R. Briggs, K. H. Ballard, (Miss) F. R. Alrich, and J. P. Wikswo,M,5 8 T.R. Briggs, and K. H. Ballard, J. Physical Chcrn., 1940, 44, 322.59 G. H. Cheesman, D. R. Duncan, and I. W. H. Harris, J., 1940, 837.Go T. R. Briggs, K. D. G. Clack, K. H. Ballard, and W. A. Sassaman,6l J . , 1931, 604.p. 325.J . Physica,Z Chem., 1940, 44, 350EMELBUS AND WELCH : INTRODUCTION AND GENERAL. 137toluene as solvent the only rubidium polyiodide stable in the solidstate at 6" or 25" is RbI,, but with benzene the solvated compoundsRb17,4C6H6 and Rb18,4C6H6 can be obtained a t either of thesetemperatures. 62 A polyhalide K16CNs,4C6H6, which is analogousto Rb17,4C6H6 and contains the thiocyanate radical as a pseudo-halogen, has also been obtained in the system potassium thio-cyanate-iodine-benzene at 6". Attempts to prepare compounds ofiodine with sodium bromide, potassium bromide, ammoniumbromide, thallous bromide, sodium thiocyanate, or ammoniumthiocyanate, with benzene or toluene as solvent at 6", were un-successful, but ctmium bromide gave the known compound C S B ~ I , .~ ~Attempts have also been made 64 t o prepare the higher iodides ofcopper, iron, silver, and thallium by treating metallic copper oriron, silver iodide, or thallous iodide with iodine and benzene ortoluene a t 6"; in the first three cases FeI,, CuI, and AgI were theonly iodides which resulted, but thallium gave the compoundsT1618 and TII,.There has been sustained interest in reactions in non-aqueoussolutions. It has been found, for example,65 that, in presence ofcertain metallic oxide catalysts, potassamide reacts slowly withpotassium nitrate and yields the nitrite : 3KNH, + 3m0,3KOH + N, + NH, + 3KN0,.The best catalysts are ferricoxide, cobaltic oxide, and nickelous oxide, although cupric oxide andMn,O, are moderately active. I n all instances the catalyst isslowly attacked by the potassamide and a potassium ammono-metallate, e.g., CuNK,,xNH, is formed. Reaction between potas-sium nitrite and potassamide in liquid ammonia is very slow,although a t higher temperatures and in absence of a catalyst thefollowing reaction occurs : KNO, + KNH, + 2KOH + N,.The action of liquid ammonia on the sulphur trioxide additioncompounds with pyridine, dimethylaniline, dioxan, hydrogenchloride, and sodium chloride results in ammonolysis, the chiefproduct being ammonium sulphamate, NH,*SO,NH,.66 Ammon-ium imidodisulphonate is also formed, its amount increasing as thestability of the sulphur trioxide addition compound diminishes.In this connexion, the same authors have also studied the productionof sulphamic acid in the reaction between hydroxylamine sulphateand sulphur dioxide.W. Klatt 68 has found that, although hydrazoic acid and hydrogenG2 H. W. Foote and M. Fleischer, J . Physical Chem., 1940, 44, 633.63 Idem, ibid., p. 640. 64 Idem, ibid., p. 647.6 5 F. W. Bergstrom, J . Amer. Chem. Soc., 1940, 62, 2381.6 6 H. H. Sisler and L. F. Audrieth, ibid., 1939, 61, 3392.6 7 Ibid., p. 3389. 2. physikal. Chem., 1939, 185, 306138 INORGANIC CHEMISTRY.cyanide are almost insoluble in anhydrous hydrofluoric acid andmost azides and cyanides are decomposed, silver and mercuricazides and mercuric cyanide dissolve and can be recovered unchanged.From the boiling-point elevations produced, it is deduced thationisation takes place into F’ and cations such as (AgN,H)’ andEHg(CN),Hl’*Further support for the view that pyridine and compounds ofthe same type function as bases in selenium oxychloride solution isafforded by the isolation of addition compounds of pyridine, quinoline,and isoquinoline with selenium oxy~hloride.~~ With pyridine thereare two such compounds, C,H,N,GeOCl, and (C,H,N),SeOCl,,ionisation of the first yielding (C,H,NSeOCl)‘ and Cl’.H.J. E.A. J. E. W.2. THE BOROX HYDRIDES.The classical investigations of A.Stock and his collaborators onthe boron hydrides added enormously to our knowledge of thisimportant group of compounds. Indeed, it appeared a t one timethat little remained to be completed. During the last ten years,however, there has been a marked revival of interest in these com-pounds, both from the point of view of preparative chemistry andin so far as problems of valency and molecular structure are con-cerned. Stock’s contributions to this subject have been adequatelysummarised elsewhere,l but it would seem useful to review in somedetail the more recent developments.The method used by Stock in preparing boron hydrides was thedecomposition of magnesium boride, obtained by heating together amixture of magnesium and boric oxide, with either 10% hydrochloricacid or, better, as E.Wiberg and K. Schuster subsequently found,8~-phosphoric acid. The use of phosphoric acid increases the yieldunder the best conditions from about 4 to about 11%. The volatileproducts obtained are the hydride B4H10, mixed with small amountsof B,H,, B6Hlo, Bl0HI4, carbon dioxide, hydrogen sulphide,silicon hydrides, and phosphine, and the separate hydrides areobtained from this complex mixture by a laborious process offractional distillation and condensation in an all-glass vacuumapparatus. Diborane, B,H,, which may be regarded as the chiefhydride, is not formed directly in the reaction, but is obtained69 J. Jackson and G. B. L. Smith, J. Amer. Ghem. SOC., 1940, 62, 544.1 See A. Stock, “Hydrides of Boron and Silicon,” Cornell University Press,1933.Ber., 1934, 67, 1805EMEL~US : THE BORON HYDRIDES.139together with B,H,, B1,-,H14, and possibly also a further hydride ofthe formula B6HI2, by heating tetraborane, B4HI0, at 180".For reference purposes, the formulae, physical proper ties andstability of these hydrides are tabulated below :B,H,. B,H,,. B,H,. B,H,,.- B. p. .................. -92*5" 18" -V. p. at 0", mm. ... - 388 66 63.0Dscomp. by H,O ... Very Slow Slow -M. p. .................. -165.6' -120" -46.6" -123.3"rapidStability ............ High Low High VerylowB6H10' B10H14'7.2 - - ca. 213"-65.1' 99.6"Slow VeryslowLow VeryhighIn 1931, A. B. Burg and H. I. Schlesinger developed an entirelynew and considerably more efficient method for preparing thesimplest of these hydrides, diborane.A mixture of boron trichloridevapour and hydrogen was passed a t a high streaming rate and a t apressure of 5-10 mm. through an electrical discharge formedbetween water-cooled copper electrodes. The products escapingfrom the discharge zone consisted of much unchanged boron halide,mixed with hydrogen, hydrogen chloride, monochlorodiborane,B2H,C1, and a small amount of diborane. In addition, a resinousyellow-brown solid was deposited near the discharge. This evolvedhydrogen when treated with alkali and contained ill-defined, solidboron hydrides of the type described by Stock, which are formed inall experiments involving pyrolysis a t moderate temperatures. Theisolation of the comparatively non-volatile mixture of boron tri-chloride and monochlorodiborane is easy, and when kept a t 0" thesecond of these compounds disproportionates according to theequation : 6B2H,CI = 5B,H6 + 2BC1,.The resulting diborane isagain easy to isolate owing to its high volatility.Burg and Schlesinger noted that boron tribromide undergoes asimilar reaction with hydrogen in the discharge, but the advantagesof using the bromide in place of the chloride were first emphasisedby A. Stock and W. Siitterlin,4 who found that it was difficult to freediborane from hydrogen chloride and recommended the use ofmetallic potassium for this purpose. Boron tribromide, on theother hand, is more easily manipulated than the trichloride, owingto lower volatility, and hydrogen bromide and diborane are alsoseparated with ease.Either of these halides gives a much betteryield of hydride than is obtained from magnesium boride; indeed,it may be as high as 80% of the halide rea~ting,~ and there is noJ. Arner. Chem. Xoc., 1931, 53, 4321.Ber., 1934, 67, 407 ; see also A. Stock, H. Martini, and W. Sutterlin, Ber.,A. Stock and W. Sutterlin, ibid., p. 410.1934, 67, 396140 INORGANIC CHEMISTRY.doubt that this new preparative method has been an importantfactor governing the success of recent studies in this field.Burg and Schlesinger's method yields only diborane, and it isimportant to determine how this may be converted into the otherhydrides. Progress in this direction was reported in 1933,6 when itwas shown that other hydrides were formed by heating diborane tomoderate temperatures.A flow method was developed, diboranebeing passed a t 1 6 3 0 c.c./min. through a glass U-tube held a t100-120". The two hydrides B4H10 and B5H11 are produced andmay be isolated. The thermal decomposition of B5H11 was investig-ated next, and a t 100" it was found to yield H,, B,H,, B4Hlo, B,Hs,B10H14, and a non-volatile yellow solid. The yields of B5Hs andBloHl, increased with the period of heating, but addition of a largeexcess of hydrogen to B5H11 retarded the formation of the lessvolatile products and made it possible to obtain reasonable yields ofB,H,, and B,H6. The importance of these observations lies in thefact that, owing to the complex equilibrium between the varioushydrides of boron, any member of the series may be synthesised,while a t the same time the improved method for preparing diboraneis utilised.Burg and Schlesinger's subsequent studies in this field havecentred round the problem of the structure of the hydrides andhave dealt particularly with the investigation of factors which affectthe stability of the boron-boron bond.An early investigation ofthis sort resulted in the preparation of dimethoxyborine, BH( OMe),,by the interaction of diborane and methyl alcoh01.~ These sub-stances react readily a t room temperature with evolution of hydro-gen. The product, which boils a t 25.9" and is not associated in thegas phase, is rapidly hydrolysed, BH(OMe), + 3H,O = B(OH), +2MeOH + H,, and also decomposes reversibly according to theequation 6BH(OMe), + B,H, + 4B(OMe),.No volatile mono-methoxyborine was formed in the reaction with alcohol, but a whitenon-volatile solid was obtained which was thought to be a polymerof this substance. Other alkoxyborines were isolated in a sub-sequent investigation of the reaction of diborane with organiccompounds containing a carbonyl group,8 and are referred to later.Methyl-substituted derivatives of diborane are formed in thereaction between diborane and trimethylboron a t roam temperat~re.~By varying the proportion of the alkylboron, four derivatives,B2H5Me, B2H4Me2, B,H,Me,, and B,H,Me,, are obtainable, but6 A. B. Burg and H. I. Schlesinger, J. Amer. Chem. SOC., 1933, 55, 4009.'7 Idem, ibid., p.4020.H. C. Brown, H. I. Schlesinger, and A. B. Burg, ibid., 1939, 61, 673.H. I. Schlesinger and A. 0. Walker, ibid., 1935, 57, 621EMEL~US : THE BORON HYDRIDES. 141even with a large excess of the alkyl there is no indication of theformation of either the penta- or the hexa-methyl derivative.The structure of the methylboranes was determined by vapour-density measurements, hydrolysis, and oxidation. On treatmentwith water a t room temperature they break down at the boron-boron bond, yielding 2 mols. of boric or methylboric acids, e.g.BHMe,*BH,Me --+ B(OH)Me, + B(OH),MeThe unsymmetrical dimethyl compound tends always to be formedin greater quantity than its homologues. For instance, with equalvolumes of the two reagents a mixture was obtained consisting of8.17(0 of mono-, 81.6% of unsymmetrical di-, S.Oyo of tri-, and 2.3%of tetra-methyldiborane. I n the trimethyl derivative two of themethyl groups are attached to one boron atom and one to the other,while the tetramethyl derivative yields 2 mols.of dimethylboric acidon hydrolysis and is therefore symmetrical.Monomethyldiborane is the least stable of these substances andhas a marked tendency to revert to diborane and trimethylboron.This accords with the observation, already noted, that monochloro-diborane disproportionates to form diborane and boron trichloride.These observations as a whole also support Stock’s conclusion thatthe boron-boron bond can exist only so long as each boron atom isattached to one hydrogen atom.The symmetrical dimethyl derivative of diborane, which is notformed in the above reaction, is prepared by treating the mono-methyl derivative with dimethyl ether a t - 8O0.l0 Fission of theboron-boron bond occurs, and the borine radical, BH,, forms a co-ordination compound, Me,O+BH,, with the ether.The liberatedBH,Me radicals, which are unable t o form a similar compound,combine to produce (BH,Me) 2, the symmetrical dimethyldiborane.Other alkyl derivatives of diborane are entirely analogous to themethyl derivatives. The four ethylboranes and mono- and di-n-propyldiborane are obtained by the reactions of triethylboron andtri-n-propylboron, respectively, with the parent hydride.11 Therewas evidence of the formation of further-substituted propyl deriv-atives, but the lack of volatility, a serious handicap in work of thistype which utilises in the main the special technique developed byStock, prevented their isolation.Direct evidence for the transitory existence of the borine radical,BH,, is afforded by the formation of the compound borine carbonyl,BH,CO, by the interaction of diborane and carbon monoxide.12lo H.I. Schlesinger, N. W. Flodin, and A. B. Burg, J. Amer. Chem. Soc.,l1 H. I. Schlesinger, L. Horvitz, and A. B. Burg, ibid., 1936, 58, 407.l a A. B. Burg and H. I. Schlesinger, ibid.. 1937. 59, 780.1939, 61, 1078142 INORGANIC CHEMISTRY.When these two substances are heated together an equilibrium isset up (B2H6 + 2CO =+ 2BH,CO), the position of which is suchthat a sample of borine carbonyl a t 200 mm.pressure would be 95%decomposed a t 100". The time for equilibrium to be established isof the order of 15 mins. a t loo", but at room temperature reactionis very slow. As a result it is possible by rapid cooling of a mixturein equilibrium at 100" to isolate borine carbonyl, which has b. p.- 64", and has been proved by vapour-density determinations tohave the monomeric formula. The compound is formulated witha co-ordinate bond between the carbon and the boron atom.Borine carbonyl is decomposed by heating with water a t 100"[BH,CO + 3H20 = B(OH), + 3H2 + CO]. With gaseous am-monia it forms an addition compound BH,(CO)(NH,),, a fairlystable white solid which reacts with a solution of sodium in liquidammonia, evolving hydrogen.Prom the quantity of hydrogen, itis concluded that the compound contains two ammonium ions permolecule of borine carbonyl, but its exact nature is not yet known.It was as a result of the study of this reaction with ammonia,which had been expected to displace carbon monoxide from theborine carbonyl, that the interaction of borine carbonyl and tri-methylamine was subsequently studied. This amine was found todisplace carbon monoxide rapidly and completely at room tem-perature according to the equation :BH,CO + NMe, = Me,N,BH, + COThe reaction takes place only a t temperatures which are sufficientlyhigh to cause dissociation of the carbonyl itself. Thus, althoughthere is no observable displacement of carbon monoxide at - SO",reaction is rapid at 20".As would be anticipated from this result,diborane also reacts readily with trimethylamine t o give the sameproduct, Me,N,BH,, which is a solid of m. p. 94-94.5" and (extra-polated) b. p. 171".The reaction of diborane with ammonia t o form the diammoniate,B,H6,2NH,13 and the decomposition of this when heated in asealed tube, with formation of the volatile cyclic compound B,N,H6,are already well known. The analogous reactions of the methyl-boranes were investigated by H. I. Schlesinger, L. Horvitz, andA. B. B ~ r g . 1 ~ The ammonia derivatives were prepared by con-densing the methylborane under examination with an excess ofammonia, holding the mixture at - 115" for 15 mins., and thenraising the temperature gradually to - 78-5", a t which point theunreacteil ammonia was pumped off.Loss of ammonia ceased inl3 A. Stock and E. KUSS, Ber., 1923, 56, 807.l4 J. Arner. Chem. XOC., 1936, 58, 409EMEL~US : THE BORON HYDRIDES. 143each instance when the molecular ratio of combined ammonia tomethylated borane was approximately 2 : 1, showing that com-pounds analogous to the diammoniate of diborane were formed.They were white powders, which were more stable than the methyl-diboranes themselves, although less stable than the ammoniate ofthe unmethylated hydride. The diammoniate of monomethyl-diborane may be heated to 50" without change, but above thistemperature evolution of hydrogen begins : the correspondingcompound of dimethylborane loses hydrogen a t lo", and the othersbegin to decompose at - 35".The ammonia derivatives decompose rapidly a t 180-200", andby these reactions aminodimethylborine, BMe2*NH2, and the cycliccompounds B3N3H,, MeB3N3H,, Me,B3N3H,, and Me,B3N3H3 havebeen prepared.An alternative method is to heat the methylatedhydride directly with ammonia a t 180-200". Aminodimethyl-borine is one of the chief products of the pyrolytic action of ammoniaon di-, tri-, or tetra-methyldiborane. The other compounds arederivatives of B3N,H,, in which one or more hydrogen atoms of theBH groups are replaced by methyls. The m. p.'s and b. p.'s ofthe four ring compounds are tabulated below.B3N3H, ......... -58" 53" B,N,H,Me, ... -48" 107'M.p. B. p. M. p. B. p.B3N3H5Me ......-50 87 B,N,H,Me, ... 31.5 129The chief evidence on which these compounds are formulated is aquantitative study of the hydrogen evolution on hydrolysis, it beingassumed that free hydrogen results only from hydrolysis of theB-H bonds.The synthesis of N-methyl derivatives of triborinetrianiine, asdistinct from the B-methyl derivatives mentioned above, calls fora different method, but has been achieved by heating togethermixtures of diborane, ammonia and meth~larnine.1~ Such mixtureswere heated in sealed tubes in an oil-bath a t 200" for 15-30minutes. The product contained hydrogen, triborinetriamine, andits N-methyl derivatives in proportions depending upon the ratioof methylamine to ammonia, and was analysed by fractional con-densation of the components of the mixture in the vacuum apparatus.Three methyl derivatives were isolated and fully characterised, theirformulte and b.p.'s being tabulated below together with those ofthree NB-methyl compounds obtained by the reaction of trimethyl-boron with niono-N-methyltriborinetriamine. These substitutionproducts are formed simultaneously; the fact that only six deriv-l 5 H. I. Schlesinger, D. M. Ritter, and A. B. Burg, J. Amer. Chem. SOC.,1938, 60, 1296144 INORGANIC CHEMISTRY.atives exist is to be regarded as confirmation of the cyclic structureof B3N3H6.B . p. B. p.N-MeB,N,H, ............... Go NB-Me,B3N,H4 ............ 12a0N-Me,B,N,H, ............... 108 NBB'-Me,B3N3H, ......... 139N-Me,B,N,H, ...............134 NBB'B"-Me,B,N,H2 ...... 158The structures of the cyclic compound B3N,H6 and of its sub-stitution products have been conclusjvely settled by electron-diffraction experiments,l6 but they have become far more compre-hensible as the result of a further publication by H. I. Schlesingerand A. B. Burg l7 on the structure of the diammoniate of diborane.E. Wiberg l8 considered this compound to be a true salt and assignedto it the formula (NH4),++(H2B=BH,)--. On the other hand,1 g.-mol. of a true diammonium salt when dissolved in liquidammonia should react with 2 g.-atoms of sodium and liberate 2 g.-equivs. of hydrogen, whereas Burg and Schlesinger found that only1 equiv. of hydrogen was set free. Oiily when the reaction tem-perature was allowed t o rise above - 77" was further hydrogenevolved as a result of secondary reactions.The ammoniate wasaccordingly formulated as NH,+[BH,. NH, . BH,]- , a structurewhich contains the B-N-B skeleton and provides a clue t o theformation of the cyclic molecule B,N3H6 by thermal decomposition.The use of excess of diborane in preparing the diammoniate givesrise to a new volatile compound B2H7N (b. p. 76.2" ; m. p. - 66.5").A detailed study of this substance l9 showed that it was most readilyprepared by exposing the diammoniate to diborane at 85-loo",whereby, with a flow method, 19-33% yields of B2H7N based onthe diborane used were recorded. This compound may be kept forsome days a t room temperature but decomposes fairly rapidly a t45". With trimethylamine a t - 80" it forms a stable additionproduct Me3N,B2H,N, and heating of this with excess of trimethyl-amine leads to the removal of a BH, group with the production of1 mol.of borine trimethylamine per rnol. of B,H,N originally used.Ammonia, like trimethylamine, adds on to B2H,N and forms asolid monoammoniate, B2H7N,NH,, which, when rapidly heatedto 200", gives a 45% yield of B3N,H6. This is a somewhat greateryield than is obtained by heating the diammoniate of diborane.When B,H,N,NH, is dissolved in liquid ammonia and treated withsodium, 1 g.-equiv. of hydrogen is liberated per mol. of B,H,N used.l6 A. Stock and R. Wierl, 2. anorg. Chem., 1931, 203, 228; S. H. Bauer,J. Amer. Chem. SOC., 1938, 60, 524.l7 Ibid., p. 290.Ber., 1936, 69, 2816.H.I. Schlesinger, D. M. Ritter, and A. B. Burg, J . Amer. Chem. SOC.,1938, 60, 2297EMELBUS : THE BORON HYDRIDES. 145The structures (I) and (11) are thought to represent B,H,N and itsammoniate, the NH, molecule in (11) being bound by a co-ordinatelink to the boron atom.H H H . . . . . .(1.) B:N:B:H . . . . . .H H HH H H H . . . . . . . .H : N : B : N : B : H (11.) . . . . . . . .H H H HThe reaction between phosphine and diborane affords an interest-ing parallel with the ammonia reaction, for 2 vols. of phosphinereact a t 0" to - 20" with 1 vol. of diborane, forming the solid com-pound B,H6,2PH,.20 This is less stable than the diammoniate;it dissociates above - 30°, the dissociation pressure at 0" being200 mm. The compound separates from the gaseous phase as long,white, needle-shaped crystals which inflame spontaneously in theair, and are decomposed rapidly by water to form hydrogen, phos-phine, and boric acid.When it is heated rapidly to 200°, hydrogenis evolved and a non-volatile residue is obtained which was notidentified. It is clearly established, however, that no volatilesubstance corresponding t o B,N,H, is produced.The compound B2H,,2PH, dissolves slowly in liquid ammoniawithout apparent reaction. It should be noted that a true phos-phonium salt would dissolve with complete evolution of phosphine(e.g., PH4Br 3 PH, + NH,Br). The solution of B2H6,2PH, inliquid ammonia evolves phosphine slowly at - 60", but the amountdoes not exceed about 55% of the total combined.When the solventammonia is pumped away, a solid residue remains which approxim-ates to B2H6,PH,,NH,, and it is suggested that this should beformulated as NH,(BH,. PH, . BH,). Gaseous ammonia also dis-places 21-57% of the phosphine from the addition compound, theamount depending on the conditions, while hydrogen chloride reactsto form chlorinated derivatives. Under controlled conditions thereaction isB,H6,2PH3 + 2HCl = B,H4C12,2PH3 + 2H2The product is a clear non-volatile liquid, which will react furtherwith hydrogen chloride, forming a white crystalline solid,B2H2C1,,2PH,. By treatment with hydrogen chloride underpressure this is converted into BCI,,PH,, a compound which wasoriginally prepared by J. A. Besson21 by the interaction of borontrichloride and phosphine below 20".Methyl derivatives of diborane react with trimethylamine in the20 E.L. Gamble and P. Gilmont, J . Arner. Chem. SOC., 1940, 62, 717.21 Cornpt. rend., 1890, 110, 516146 INORGANIC CHEMISTRY.same way as diborane itself.22 The fully methylated compoundBMe,,NMe, is prepared by treating trimethylboron with a slightexcess of trimethylamine, and is a stable solid which melts at 120"and can be sublimed in a vacuum. The vapour is 70% dissociateda t 80". In a similar way trimethylamine reacts with tetramethyl-diborane and with symmetrical dimethylborane, forming the com-pounds BHMe,,NMe, and BH,Me,NMe,. In these derivatives thestability (as measured by the dissociation of the vapour) decreasesas the number of methyl groups substituted in the borine radicalincreases.There is also a step-wise gradation in thg rates ofreaction of these compounds with gaseous hydrogen chloride :only with the trimethyl derivative, BMe,,NMe3, is the B-N linkbroken by this reagent. The equations representing the reactionin the other instances are :BH,,NMe, + HCIBH,Me,NMe, + HC1= BHMeCl,NMe, + H,BHMe,,NMe, + HC1 = BMe,Cl,NMe, + H,Methylboric acid, which is of interest as being a hydrolysis productof certain methylated boron compounds, has been prepared recentlyby A. B. using the reaction between methylmagnesiumiodide and methyl borate. This method has been applied in thepreparation of other alkylboric acids.24 The crude acid was de-hydrated by passage over dehydrated calcium sulphate, and theresulting anhydride was found to melt a t - 38" and to have an(extrapolated) b.p. of 79". Pure methylboric acid was formed bythe addition of water to the anhydride, and was found to beextensively dissociated in the vapour phase, the b. p. being about100".The anhydride was found to have the trimeric formula (BMeO),,and in this respect resembled phenylboric anhydride 25 and higheralkylboric anhydrides.,G Ammonia and trimethylamine formedaddition products without destruction of the trimeric character.The ammoniates had the formulze (BMeO),,NH, and (BMe0),,2NH3,the first being stable and the second unstable. The only trimethyl-aminate was (BMeO),,NMe,, and this was a fairly stable solid,capable of sublimation in a vacuum.Burg suggests a cyclicstructure for the anhydride, similar to the structure of B,N,H,, butmade up of alternate BMe groups and oxygen atoms. I n the** H. I. Schlesinger, N. W. Flodin, and A. B. Burg, J. Amer. Chem. SOC.,1939, 61, 1078.23 Ibid., 1940, 62, 2228.34 H. R. Snyder, J. A. Kuck, and J. R. Johnson, ibid., 1938, 60, 105.2 5 C. R. Kinney and D. F. Ponte, ibid., 1936, 58, 197.2 6 H. R. Synder, J. A. Kuck, and J. R. Johnson, Zoc. cit., ref. (24).= BH,Cl,NMe, + HEMEL~US : THE BORON HYDRIDES. 147mono-addition compounds there are co-ordinate links from nitrogento one boron atom, and t o explain the unstable diammoniate, theconception of hydrogen bonding is invoked.I n the course of attempts t o use boron trifluoride to produceaddition compounds with methylboric anhydride, it was found thatthis reagent causes a breakdown of the trimer according t o the(idealised) equation (BMeO), + 2BF,+ BMeF, + B,O,.The newcompound, methylboron fluoride, BMeF, (b. p. - 62.3"; m. p.- 130.5"), is obtained in good yield in this reaction, and by ananalogous cleavage of dimethylboric anhydride a second product,dimethylboron fluoride, BMe,F (b. p. - 42-2"; m. p. - 147.4") isobtained. The discovery of these two substances has completed theseries of methylated boron fluorides BF,, BMeF,, BMe,F, and BMe,.A new and potentially very important link between organicchemistry and the chemistry of the boron hydrides has been estab-lished in a recent study by H. C. Brown, H.I. Schlesinger, and A. B.Burg2' of the reactions of diborane with organic compounds con-taining a carbonyl group. Certain simple molecules only were used,the first being acetaldehyde, which gave good yields of the newcompound diethoxyborine : 4CH,*CHO + B2H6 = ZBH(OEt),.This, like its homologue dimethoxyborine, decomposes on standinginto diborane and the trialkyl borate. It is readily hydrolysed t o1 mol. of boric acid, 1 mol. of hydrogen, and 2 mols. of ethyl alcoholper mol. of the compound. With acetone the main reaction productis diisopropoxyborine : 4COMe2 + B2H6 = 2BH(OPrP),. The re-action with trimethylacetaldehyde was examined in order to deter-mine whether enolisation plays any part in these processes, but theresult was similar t o that with acetaldehyde or acetone, as thefollowing equation shows :4CMe,*CHO + B2H6 = ZBK(O*CH,*CMe,),With methyl formate the only products which could be isolated weredimethoxyborine and methyl borate, and it was supposed that thelatter was formed together with diborane by disproportionation ofdimethoxyborine.The reaction with ethyl acetate was similar,but no reaction wds observable between diborane and acetylchloride, carbonyl chloride, or chloral. In the course of thesestudies it was pointed out that there was a certain correlationbetween the ability of the carbonyl group to react with diboranearid the stability of the boron trifluoride addition compound whichis formed by those substances containing a carbonyl group.A novel type of boron hydride derivative has been described87 J.Amer. Chena. SOC., 1939, 61, 673148 INORGANIC CHEMISTRY.recently,28 vix., A1B3H12, formed by treatment of trimethylaluminiumwith excess of diborane a t temperatures up to 80". This substancemelts a t - 64.5", has a vapour pressure a t 0" of 119 mm., and anextrapolated b. p. of 44". It undergoes slow polymerisation in theliquid state a t room temperature. The reactions have so far beenreported on only in a preliminary manner, but with methyl etherthe compound AlB,H,,,Me,O results, while with ammonia a seriesof products containing up to 4 mols. of ammonia is obtained. Withtrimethylamine there is again a mixture of products, one of thesebeing borine trimethylamine. The authors also mention the interest-ing fact that diborane reacts with alkyls of metals other thanaluminium, giving in some cases alkyldiboranes and also otheras yet unidentified products.2PaUp to this point the experimental work reviewed has been ofinterest primarily from the point of view of preparative chemistry,but in the last ten years there has also been a considerable outputof research of a more physical character bearing on the question ofthe structure of the boron hydrides.It was in 1931 that A. Stockand R. Wierl 29 first established by electron-diffraction measure-ments that the compound B3N,H, has a hexagonal structure.This pioneer investigation has been followed up by S. H. Bauer, andto-day we know the shapes and dimensions of a fair number of thecompounds to which reference has already been made.30Electron-diffraction photographs obtained from diborane are verysimilar to those from ethane and are incompatible with an ethylene-like model.31 The B-B and B-H bond distances are 1.86 &- 0.04 A.and 1.27 If: 0-03 A., respectively, the valency angle being tetra-hedral within 3".For this model G. N. Lewis 32 has suggested thatthere might be six electron-pair bonds resonating among sevenpositions, each bond having six-sevenths single-bond and one-seventh no-bond character. This structure would be stabilised bythe resonance energy of the molecule among its seven forms. Theabove bond distances are larger than the respective single-bond28 H. I. Schlesinger, R. T. Sanderson, and A. B. Burg, J .Amer. Chem. SOC.,1939, 61, 536.Recent publications (H. I. Schlesinger, R. T. Sanderson, and A. B. Burg,J . Amer. Chem. SOC., 1940, 62, 3421 ; H. I. Schlesinger and H. C. Brown, ibid.,p. 3429; A. B. Burg and H. I. Schlesinger, ibid., p. 3425; J. Y. Beach andS. H. Bauer, ibid., p. 3440) describe borohydrides of aluminium, lithium,and beryllium.2. anorg. Chem., 1931, 203, 228.30 See also L. Pauling, "The Nature of the Chemical Bond," Cornell31 S. H. Bauer, J . Amer. Chem. SOC., 1937, 59, 1096; T. F. Andehon and32 Ibid., 1933, 1, 17.University Press, 1939.A. B. Burg, J. Chem. Physics, 1938, 6 , 586EMELBUS : THE BORON HYDRIDES. 149values of 1.76 and 1.18 A,, a result which is in keeping with Lewis'ssuggestion. L. Pauling 3, points out, however, that one-electronbond structures also contribute to the resonance effect.The electron-diffraction investigation of the structures of tetra-borane, B4H10, and pentaborane, B5H11, showed that the first ofthese substances has a butane-like structure with B-B and B-Hbond distances of 1.84 -j= 0.04 and 1.28 & 0.03 A., re~pectively.~~The hydride B5Hl1 resembles pentane or isopentane in its structureand has B-B and B-H bond distances of 1.81 & 0-03 and 1.26 -+0.03 A., respectively.In each of these hydrides Bauer's resultsindicate that there is free rotation about the boron-boron bond.There are four electrons less than the number needed to provide oneelectron pair for each bond, and the interatomic distances are inkeeping with the hypothesis of resonance among structures of theLewis type and structures with four one-electron bonds.The hydride B,H, was found by S.H. Bauer and L. Pauling 35 t ohave the cyclic structure (I) with the following interatomic distances :B-B, 1.76 & 0.02; B-H, 1.17 & 0 . 0 4 ~ .H\B/ \HH\ B/H H\B* HH/" H . B-B<~>-B/ \B-B/H \B-B . HH . H\BH/B<B/ \H H/H/ ' H H ' \H H ' \H(1.) (11.1Here again there is believed to be resonance, and the same islikely t o be true of the hydrides B,H,, and BiOH14, which have not,however, been examined experimentally. The structure (11) hasbeen suggested for B,,H,,, with the modification that the electrondeficit is not localised in certain singlet linkages, but is distributedthroughout the molecule.Reference has been made already to the structure of triborine-triamine, but this substance was re-examined by S.H. Bauer.36The benzene-like form was confirmed, and the B-N distances foundto be 1-44 The compound B,H,N, to which referencehas already been made, was likewise investigated and was assignedthe structure (111), with a B-N bond distanceH/ \E/ \H bond value of 1 . 5 9 ~ . Bituer's observationsdo not differentiate conclusively between theforms BH,-NH-BH, and BH,*NH,*BH,, butthey establish definitely the existence of the34 S. H. Bauer, J. A,mer. Chem. SOC., 1938, 60, 805.0.02 A.3 B B /H . H of 1-56 & 0.03 A., which is close to the single-I H(111.)B-N-B unit, the angle of which is tetrahedral within 4".36 Ibid., 1938, 60, 524.33 op.cit.35 Ibid., 1936, 58, 2403150 INORGANIC CHEMISTRY.Borine carbonyl and borine trimethylamine have also beenexamined by the electron-diffraction method.37 The former isfound to have the B-C-0 group arranged linearly, with the hydrogenAtoms completing the tetrahedron round the boron (B-H, 1-20 &0.03 ; B-C, 1-57 & 0.03 ; C-0, 1.13 -J= 0.03 A.) ; and the latter hasa similar structure (B-N, 1.62 & 0.15 ; N-C, 1.53 & 0.06 A.), which,together with that of borine carbonyl, is discussed by Bsuer interms of the theory of resonance.H. J. E.3. FLUORINATION OF INORGANIC COMPOUNDS.This section of the Report contains a summary of recent public-ations on the fluorination of inorganic halogen compounds.Although a number of fluorinating agents have been used in thepast, attention has hitherto been directed mainly to the preparationof fully fluorinated substances.A feature of the more recentresearch in this field has been the successful production of com-pounds containing fluorine with other halogen atoms.The fluorinating agent most commonly employed is antimonytrifluoride. It was first introduced by F. Swarts towards the endof last century for the treatment of organic halogen compounds,and has since been widely used for this purpose. I t s generalapplicability in the field under discussion was first emphasised byH. S. Booth and C. F. Swinehart,l who have since studied its reactionwith a considerable number of inorganic chloro- and bromo-com-pounds.Antimony trifluoride may be used either alone or in con-junction with antimony pentachloride or bromine as a catalyst.The general procedure in performing a fluorination is very simple andconsists of heating the halogen compound to be fluorinated with theantimony compound and the catalyst. Ease of fluorination variesfrom one compound to another, and it may be necessary to carryout the reaction under pressure in order to attain the neededtemperature. Booth has studied in detail the conditions forreplacing part only of the halogen initially present by fluorine,and also the methods of separation of the complex mixtures offluorinated products which often result, and his publications shouldbe consulted in this connexion for experimental details.The fluorination of trichlorosilane was the first reaction studiedby H.S. Booth and W. D. Stillwell.2 The products isolated were :SiHF3 (b. p. - 97.5"), SiHF,Cl (b. p. approx. - 50"), and SiHFC1,(b. p. - 18.4"). The first had been prepared by 0. Ruff and37 S. H. Bauer, J . Amer. Chem. SOC., 1937, 59, 1804.l Ibid., 1932, 54, 4750. Ibid., 1834, 56, 1531EMEL~US : FLUORINATION OF INORGANIC COMPOUNDS. 151C. Albert by the action of stannic fluoride or titanic fluoride ontrichlorosilane. Silicon tetrachloride, when treated in a similar waywith antimony trifluoride, gave the compounds SiF, (b. p. - 9 5 ~ 7 ) ~SiF,Cl (b. p. -70.0°), SiF,Cl, (b. p. - 32.2'), and SiFCl, (b. p.l 2 ~ 2 " ) . ~ Germanium tetrachloride gave GeF,, GeF,Cl (b.p.-20-3"), GeF,CI, (b. p. - 2.8"), and GeFC1, (b. p. 37.5°).5 In theirgeneral chemical behaviour these substances are intermediatebetween the tetrachlorides and the tetrafluorides. The chlorofluoro-derivatives of monogermane show a marked tendency t o dispro-portionate into germanium tetrafluoride and tetrachloride.Partly chlorinated silicon hydrides may also be fluorinated bymeans of antimony trifluoride, and from SiH,Cl and SiH,Cl, thefluorides SiH,B (b. p. - 98") and SiH,F, (b. p. - 77.5") have beenobtained.6 W. C. Schumb and E. L. Gamble 7 were successful influorinating hexachlorodisilane completely by reaction with zincfluoride a t 50-60°, although hexachloroethane does not exhibit acomparable reaction. The product, Si,F, (v. p. = 1 atm.at- 19.1"), undergoes a mildly explosive reaction with chlorine whenthe containing vessel is heated locally,s resulting in disruption ofthe Si-Si bond and the formation of Sip,, SiF,Cl, SiF,Cl,, andprobably also traces of SiFC1,.have prepared SiF3Br (b. p. - 41.7'), SilF',Br, (b. p. 13.7"), andSiFBr, (b. p. 83.8"). The first two of these compounds are formedtogether with SiF4 in the reaction between bromine and hexafluoro-disilane. Antimony trifluoride reacts with silicon tetrabromidewithout a catalyst, and also yields the three mixed halides. It isnoteworthy that the fluorinating reaction of silver fluoride in thisinstance is recorded as being too violent and that of titanium tetra-fluoride too slow.Fluorochlorobromides of silicon have also been described bySchumb and Anderson,lo the compounds isolated being SiFClBr,(b.p. 59.5") and SiFC1,Br (b. p. 35.4'). Three methods of prepar-ation were used, vix., the partial fluorination of the silicon chloro-bromides SiClBr, and SiCl,Br,, the chlorination of silicon fluoro-tribromide (4SiFBr, + 3c1, = 2SiFC1Br2 + 2SiFC1,Br + 3Br,), andthe action of antimony trichloride on silicon fluorotribromide,resulting in partial replacement of bromine by chlorine.The fluorination of a variety of phosphorus halides has also beenBer., 1905, 38, 53, 2222.H. S. Booth and C. F. Swinehart, J . Amer. Chem. Soc., 1935, 57, 1333.H. S. Booth and W. C. Morris, ibid., 1936, 58, 90.A. G. Maddock, H. J. Emelbus, and C. Reid, Nature, 1939,144, 328.J .Amer. Chem. Xoc., 1932, 54, 583.Ibid., 1936, 58, 994.W. C. Schumb and H. H. AndersonFJ W. C. Schumb and E. L. Gamble, ibid., p. 3943.In IbirJ., 1937, 59, 651152 INORGANIC CHEMISTRY.exhaustively studied by Booth and his collaborators, the m. p.'s andb. p.'s of certain new compounds obtained being tabulated below.Parent Deriv- Parent Deriv -halide. ative. M. p. B. p. halide. ative. M. p. B. p.PCI, 1 -151.5" -101.1" POBr,4 POF,Br - 84.8' 30.5"g$Cl -164.8 - 47.3 POFBr, -117.2 110.1PBr, PF,Br -135.8 - 16.1 PSFsC1 - 155.2 6.3PFCI, -144.0 13.85 PSCl, PSF, -148.8 - 52.3PFBr, -115.0 74.8 PSFCI, - 96.0 64.7POCl, POF, - 39.4 - 39.8POF,Cl - 96.4 3-1POFCI, - 80.1 52.9H. S. Booth and A. R. Bozarth, J . Amer. Chem. SOC., 1939, 61, 2927.H. S.Booth and S. G. Frary, ibid., p. 2934.H. S. Booth and F. B. Dutton, ibid., p. 2937.H. S. Booth and C. G. Seegmiller, ibid., p. 3120.H. S. Booth and M. C. Cassidy, ibid., 1940, 62, 2369.I n the majority of these reactions antimony trifluoride in con-junction with a catalyst has been employed. It is notable, however,that Booth has found calcium fluoride to be a satisfactory fluorin-ating agent. This material has also been employed for the fluorin-ation of aliphatic halogenated hydrocarbons ; 11 it is prepared inpellet form and packed into a heated tube, through which the vapourto be fluorinated is passed. There is no reason to suppose that thesewould be the only satisfactory fluorinating agents. Thus, for ex-ample, zinc fluoride will fluorinate phosphorus tri- or penta-chloridecompletely and, were it desired to limit the extent of reaction, thereis little doubt that the intermediates could be isolated.An interesting reaction of the tervalent phosphorus fluorochloridesis the addition of halogens: PF,, PF2C1, and PFC1, all reactwith chlorine or bromine, forming compounds, which, however, tendto disproportionate.The compounds PF2Br and PFBr, also showa considerable tendency t o disproportionate.The partial fluorination of sulphur halides by the methods underdiscussion has not been described, but interesting results have beenobtained with sulphuryl chloride and thionyl chloride. Treatmentof sulphuryl chloride a t 6 atm. with antimony trifluoride andpentachloride gives the fluorochloride S0,FCI (m.p. - 124.7", b. p.7.1") as the sole product.12 Sulphuryl fluoride itself, although notformed in this reaction, is formed by the interaction of sulphurdioxide and fluorine l3 or by the action of heat on bariumfluorosulphonat e .14 The fluorochloride is intermediate in chemicalreactivity between SO,Cl, and SO,P,. Thionyl chloride with thel1 E. I. du Pont de Nemours & Co., F.P. 730874, 1932.l2 H. S. Booth and C. V. Herrmann, J . Amer. Chem. SOC., 1936, 58, 63.l3 H. Moissan and P. Lebeau, Compt. rend., 1901, 132, 374.l4 M. Trautz and K. Ehrmann, J . p r . Chem., 1935, 142, 79WELCH : THE SEPARATION OF ISOTOPES. 153same reagent yields thionyl fluoride and the fluorochloride SOFCI(m. p. - 139-5", b. p. l2-2").l5Fluorination of polymerised phosphonitrile chlorides has beenaccomplished with the aid of lead fluoride.The polymer (PNCI,),reacts with lead fluoride at 130-340", and by fractional distillationof the product the compound P4N,C1,F, (b. p. 105.8") is obtained.16The compound P4N,C1,F4 (b. p. 130.5") has also been isolated fromthe products of this reaction, and it has been observed that itsdegree of dissociation when heated is less than that of P,N,C1,F6and much greater than t h a t of P4N4C18, showing that progressivereplacement of chlorine by fluorine diminishes the stability of thepolymer. When heated under pressure at lOO", P,N,CI,F, istransformed into a rubber-like mass, which decomposes when heatedto 250400", giving P,N3C1,F, (b. p. 115-117") and P3N3C1,F,(b.p. 140-142").17H. J. E.4. THE SEPARATION OF ISOTOPES BY THERMAL DIFFUSION.Thermal Diffasion in Gases.I n the Annual Report for 1938 0. J. Walker referred briefly tothe application of the principle of thermal diffusion to the separationof isotopic gas mixtures. Very promising separations of theisotopes of some of the lighter elements have since been obtainedby this method, and there is no apparent reason why it should notbe applied successfully to heavier elements for which efficientchemical or physical methods of isotope separation are not yetavailable. Some forty papers dealing with this application ofthermal diffusion have appeared during the last two years, and thesubject merits a separate section of this Report.It was shown by D.Enskog that the application of a temperaturegradient to a gaseous mixture of molecules of different mass shouldproduce a small gradient in the relative concentrations of theconstituents; this effect cannot be accounted for by simple kinetictheory, but is explained by a detailed study of the transportequations for a gas m i ~ t u r e . ~ According to Enskog's theory theheavier molecules tend, in general, to diffuse towards the regionof lower temperature, and a transport of these molecules continuesl5 H. S. Booth and F. C. MericoIa, J. Amer. Chem. Soc., 1940, 62, 640.l6 0. Schmitz-Dumont and H. Kulkens, 2. anorg. Chem., 1938, 238, 189.' 7 0. Schmitz-Dumont and A. Braschos, ibid., 1939, 243, 113.P. 140.See ref. (2), and L. J.Gillespie, J . Chem. Physics, 1939, 7, 530; S. P.Physikal. Z., 1911, 12, 56, 533.Frankel, Physical Rev., 19M, [ii], 57, 661154 INORGANIC CHEMISTRY.until the purely thermal diffusion process is balanced out by theeffect of ordinary diffusion, which tends to re-establish uniformityof concentrations throughout the mixture. A similar processproduces a corresponding enrichment of the lighter molecules in theregion of higher temperature. The existence of thermal diffusionwas predicted independently by S. Chapman,* and confirmedexperimentally with mixtures of carbon dioxide and hydrogen byChapman and F. W. D o ~ t s o n . ~ The mixtures were confined intwo bulbs maintained at different temperatures (T,, T,) and joinedby a short connecting tube, and after attainment of equilibrium thechanges of concentration due to thermal diffusion were determinedby isolation and analysis of the contents of each bulb.Othermixtures have been studied by similar methods.6 In the case of abinary mixture the difference of relative concentrations in the twobulbs may be represented by the difference in mo1.-fraction ofeither constituent, which is given by. .kT is the coeficient of thermal separation, which in an isotopic mixturecontaining two species of molecule (mo1.-fractions A,, A ~ ) is equal toah,^, ; a is the thermal diffusion constant. If the gas molecules wereperfect elastic spheres of molecular weight M,, $I2, c( would be105(M2 - M1)/llS(M2 + M,) (cf. Enskog's theory '), but owingto the imperfect behaviour of real gas molecules the experimentalvalues of a are usually rather less than half the values calculatedfrom the Enskog expression. Equation (1) also gives the differenceof the concentrations established a t the surfaces of parallel platesenclosing a binary gas mixture, the plates being held at the tem-peratures T1 and T2.Inspection of the expressions given showsthat kT is relatively small in isotopic mixtures, particularly if theabundance of one of the isotopic species is low; the values of AAobtained by simple thermal diffusion are consequently too small tobe of practical value for efficient isotope separations. It has beencalculated * that the separation of 20% 13CH, from normalmethane (which contains 1.1% of l3CH,) would require a group of810 cells connected in series, each cell consisting of a horizontalplate heated to 500" K.situated 1 em. above a parallel plate cooledPhi,?. Trans., 1917, A, 217, 115.For bibliography, see T. L. Ibbs, Physica, 1937, 4, 1133; also B. E.Atkins, R. E. Bastick, and T. L. Ibbs, Proc. Roy. SOC., 1939, A, 172, 142;N. G. Schmahl and J. Schewe, 2. Elektrochern., 1940, 46, 203. The separationof isotopic mixtures (methane and neon) has been studied by this method byA. 0. Nier, Physical Rev., 1939, [ii], 56, 1009; 1940, [ii], 57, 338.See W. H. Furry, R. C. Jones, and L. Onsager, a i d . , 1939, [ii], 55, 1083.PhiE. Mag., 1917, 33, 248.* J. W. Westhaver and A. K. Brewer, J . Chern. Phy8k8, 1940, 8, 314WELCH: THE SEPARATION OF ISOTOPES. 155to 300" K.; the power consumption of such a system, all losses beingneglected, would be 880 kilowatt hours per g. of 13CH4 recovered a t20% concentration. It is clear from this example that some meansof increasing the effectiveness of the thermal diffusion processmust be employed if reasonable isotope separations are to beobtained.A simple means of enhancing the effect of thermal diffusion hasbeen described by K. Clusius and G. D i ~ k e l , ~ who discovered themethod while tracing the cause of some unexpected demixingeffects in gases. I n the parallel-plate thermal diffusion cellpictured, the concentration changes occur without intervention ofconvection currents, which are deliberately avoided by maintainingthe upper plate a t the higher temperature.angle, bringing the plates into a verticalplane as shown in section in the figure, a 8convection current develops in the manner aindicated by the arrows.This current does tnot prevent the normal thermal diffusion 2process from establishing a concentrationgradient in the direction AB; the gas a t A , 3 which is being carried upwards, therefore QQJ contains a higher concentration of the lighter $molecules than the descending current of gasa t B. The result of the combined thermaldiffusion and convection processes is con-sequently an upward transport of the lighter species of molecule,with a corresponding downward transport of the heavier species ;this causes a second concentration gradient to develop in thedirection CD.The transport must clearly continue as long asa difference of concentrations is maintained between A and B, anda considerable concentration gradient may ultimately developalong CD, even though the changes due to pure thermal dizusionalong AB are very small; the factor limiting the magnitude of thevertical gradient is not the smallness of the thermal diffusion effect,but the tendency for ordinary diffusion, with some remixing of thegas by the convection currents, to re-establish uniform concen-trations. If the height of the apparatus is large in comparison withthe distance between the plates, the opposing effects are relativelysmall, and considerable enrichments of the respective constituentsof the gas mixture are produced a t the top and the bottom of theapparatus.See also A.Bramley and A. K.Brewer, Physical Rev., 1939, [ii], 55, 590; J . Chern. Physics, 1939, 7, 553.If the apparatus is turned through a rightNaturwbs., 1938, 26, 546, and ref. (13)156 INORGANIC CHEMISTRY.The mathematical theory of the combined thermal diffusion-convection process, although complex, has been studied in somedetail.lo The results confirm that large separations of gas mixturesshould be obtainable, provided that the gas pressure, temperaturedifference, and distance between the heated and cooled surfacesare suitably chosen; optimum values, which vary with the par-ticular mixture, exist for these three variables. The separationultimately attained when the transport process reaches equilibriumwith the competing effects increases with the height of the apparatus,but the time required to reach the equilibrium separation alsoincreases.The apparatus used by Clusius and Dickel to study the processdescribed above consisted of a vertical glass tube 65-290 em.inlength and about 1 em. in diameter, which was cooled on the outsideby running water; an electrically heated platinum or nichromewire was run along the axis of the tube. This method of con-struction reproduces the conditions of the figure in a very simplemanner, with the parallel surfaces in the form of concentriccylinders. The same design has been adopted for all the '( thermaldiffusion columns " used for gas mixtures, although some workersprefer a rigid, internally heated, inner tube to an axial heatingwire; the former arrangement has the advantage that a greatervolume of gas can be accommodated in the column for a givendistance between the surfaces.demonstratedthe effectiveness of their method by rapid and nearly completeseparations of bromine vapour-helium and carbon dioxide-hydrogenmixtures.Definite isotope separations were then obtained by usingcolumns 2-6-2-9 m. in height, with a temperature difference (AT)of 600" between the wire and the cooled tube. Pure neon (atomicweight 20.18) gave a " heavy " fraction, for which a gas-densitybalance gave an atomic weight of 20.68, and a mass-spectrographicanalysis by J. Mattauch showed that the abundances of *ONe, W e ,and 22Ne had been changed from 90.0, 0.3, and 9.7% to 68.4, 0.6,and 31 .O%, respectively. A heavy fraction obtained from hydrogenchloride contained chlorine of atomic weight 35.56, i.e., 0.1 unitabove the normal value; this atomic weight was determinedchemically by 0.Honigschmid and (Frau) F. Hirschbold-Wittner.l4Many previous attempts have been made to separate the chlorineIn preliminary experiments Clusius and Dickel10 See ref. ( 7 ) ; L. Waldmann, Naturwiss., 1939, 27, 230; 2. Physik, 1939,114, 53 ; W. van der Grinten, Naturwiss., 1939, 2'7¶ 317 ; J. Bardeen, PhysicalRev., 1940, [ii], 57, 35; 58, 94. The following papers deal with the calculationof a : H. Brown, Physical Rev., 1940, [ii], 57, 242; 58, 661; R. C. Jones,ibid., p. 111; R. C. Jones and W. H. Furry, ibid., 57, 547WELCH : THE SEPARATION OF ISOTOPES.157isotopes, but atomic weight displacements greater than approxim-ately 0.05 unit (reached by W. D. Harkins by diffusion of hydrogenchloride through porous diaphragms 11) have not hitherto beenobtained. The Clusius and Dickel method is thus remarkablyefficient in view of the simplicity of the apparatus.The preliminary results just described led Clusius and Dickel toattempt a complete separation of the chlorine isotopes,12 using asystem of diffusion columns of much greater length; hydrogenchloride was again used as the " carrier gas." I n order to avoidthe constructional difficulties and lack of flexibility associated witha single long column, the diffusion apparatus was built up from anumber of units 6-9 m.in length. The bottom of each unit wasconnected to the top of the next by a closed loop of tubing heatedelectrically on one side so that the gas in the joined ends of thecolumns was continuously mixed by a convection current. Clusiusand Dickel l3 have considered in some detail the arrangement anduse of apparatus consisting of a succession of diffusion columns. I nsuch systems a progressive reduction in the diameters of the columnsresults in increased efficiency, since by this means the rate of trans-port of the constituent undergoing enrichment in the mixture canbe made uniform throughout the apparatus. I n the work onhydrogen chloride five different arrangements of units were tried.The highest contents of H3'Cl were obtainedin a system consistingof columns 7, 9, 6, 6, and 8 m.in length, connected in series; an18-1. reservoir was inserted in the convective mixing tube betweenthe 7-m. and the 9-m. column to provide a suitably large supply ofhydrogen chloride a t atmospheric pressure. The columns had aninternal diameter of 8.4 mm. (9-m. tube, 12.8 mm.), and heatingwas effected with 0.4-mm. platinum wires spaced from the columnwalls by perforated platinum washers a t 60-cm. intervals. Thewires were heated to approximately 690". To hasten the attain-ment of equilibrium some hydrogen chloride enriched in H37Cl inprevious experiments was introduced into the " heavy " end of theapparatus a t the commencement of the run. After continuousoperation for 17 days, 8 C.C. of the heavy hydrogen chloride wereremoved daily in a gas pipette, and after a further 20 days gas-density measurements on the same daily yield showed its chlorinecontent to have an atomic weight of 36.98.A concordant value(36-956) was subsequently obtained by chemical atomic-weightdeterminations of greater accuracy.14 The atomic weight of 37Cll1 W. D. Harkins and F. A. Jenkins, J. Amer. Chem. SOC., 1926, 48, 58;l2 2. physikal. Chem., 1939, B, 44, 451.l4 0. Honigschmid and (Frau) F. Hirschbold-Wittner, 8. anorg. Chem., 1939,W. D. Harkins and C. E. Broeker, 8. Physik, 1928, 50, 537.l3 Ibid., p. 397.242, 222158 INORGANIC CHEMISTRY.being assumed to be 36-968, the enriched gas contained 99.4% ofAs normal hydrogen chloride contains 75.7% of H3YX, theseparation of light hydrogen chloride was relatively easy, and itwas found that the “light ” end of the apparatus just describedyielded 96% H35Cl. The enrichment was increased further by usingthe 6, 6, and 8-m.columns in series, with an 8-1. reservoir of normalhydrogen chloride at the “ heavy ” end of the group; the columnsproper were filled initially with partly enriched gas (atomic weightof chlorine, 35.17) from earlier experiments. After four days thelight hydrogen chloride gave a densimetric atomic weight of 34.98,and in a further 14 days the aggregate of 28-C.C. portions of gaswithdrawn daily gave a chemical atomic weight of 35.021 for itschlorine content; on reducing the daily withdrawal of gas to 16 C.C.the atomic weight fell to 34.979, corresponding to H35Cl of99.6% purity.l4The separation of the isotopes of chlorine in a practically purestate has thus been effected by an efficient and reasonably economicalmethod. The energy required for the separation of the isotopicspecies in 1 g. of normal hydrogen chloride, in the apparatusdescribed, is 3.7 x 1O1O g.-cals., representing an efficiency of9 x The efficiency of the electrolytic separation of deuteriumoxide (which admittedly has to be enriched from a much lowerabundance) is of the same order, viz., 2 x The apparatusused by Clusius and Dickel could easily be rendered entirely con-tinuous in operation, and large quantities of the enriched gases couldbe accumulated in a moderate length of time.Clusius and Dickel, indicating the practical possibility of separ-ating 35Cl free from the heavier isotope, have suggested an interestinguse for the separated product ; l3 a mass-spectrographic comparisonof the atomic weights of 1*0 and 35Cl, together with a chemicaldetermination of the silver equivalent of 35Cl, would afford a valuefor the conversion factor €or the chemical and the physical atomicweight scale. Previous values of this ratio have depended onaccurate measurement of the oxygen isotope abundance ratio.A 10-c.c.sample of hydrogen chloride containing about 95% ofH35Cl has also been obtained by J. W. Kennedy and G. T.Seaborg,l5 using a single 7-5-m. column with a 12-1. reservoir atthe lower end. The sample was used to prove that the neutron-induced radioactivity of chlorine is due to 38Cl.The use ofseparated isotopes in this manner for the precise identification ofartificially radioactive elements opens up another field of use forpractical methods of separation.l5 Physical Rev., 1940, [ii], 57, 843.~ 3 7 ~ 1 WELCH THE SEPARATION OF ISOTOPES. 159The experiments with chlorine have been described first becausethey have afforded the only reasonably complete isotope separationyet secured by the thermal diffusion method. An obvious applic-ation of the method is to the separation of deuterium from ordinaryhydrogen. This is relatively difficult, however, because the easeof separation of an isotopic mixture is roughly proportional to kT(see above), and kT becomes small when the abundance of theisotope to be enriched is low.G. 2'. Seaborg, A. C. Wahl, and J. W.Kennedy l6 have carried out separations of hydrogen, deuterium,and the new radioactive hydrogen isotope, 3H, with a simple 7.5-m.column. With hydrogen already partly enriched in deuterium,the equilibrium separation was reached in three days with the heatedwire at 250"; gas a t the top and the bottom of the column thencontained 18 and 87% of deuterium, respectively, the separationfactor being 30. (The separation factor is .the concentration ratioof the molecular species at the bottom of the column divided bythe corresponding ratio a t the top.) The separation factor fordeuterium containing 10-11 mo1.-fraction of 3H, under identicalconditions was 5, the value rising t o 9 with the wire at 800".Ifallowance is made, according to theory, for the dependence of theseparation factor on the masses of the separated molecules, thevalues given above for hydrogen-deuterium and de~terium-~H areconcordant; this shows that the separation factor of a thermaldiffusion column is the same for moderate and extremely low con-centrations of either molecular species, a fact of considerablepractical importance in the separation of isotopes of low abundance.With two columns connected in series by the Clusius and Dickelmethod the total separation factor approximated to the product ofthe values for the individual columns, but the convective mixingdevice between the tubes was found to increase the time requiredto reach equilibrium.Seaborg, Wahl, and Kennedy calculate thatinstallations of one, two, or three 7.5-m. columns would produce1.1, 37, or 97% deuterium, respectively, from normal hydrogen.Clearly the rate of production of deuterium would be small, as itsabundance in the normal mixture is so low.Considerable attention is being directed to the separation of heavycarbon (13C) by the thermal diffusion method. So far all workershave used methane as the carrier gas for t'his purpose. W. W.Watson l7 has attempted this separation with two 1-m. columnsfollowed by a 3-m. unit, working continuously and yielding 3 C.C.of heavy gas per hour; fresh normal methane was supplied con-tinuously to the top of the first tube to prevent depletion of thesupply of I3CH,. With the heated surfaces a t 330" this apparatusl6 J. Chenz.Physics, 1940, 8, 639. l7 Physical Bev., 1939, [ii], 56,703160 INORGANIC CHEMISTRY.doubled the normal 13CH4 concentration (l.lyo). I n later work 18Watson has used standardised 2-m. units constructed round com-mercial heating elements; two such units gave a 2.77-fold con-centration of 13CH4, and a series of six should produce a 21-foldenrichment. A. 0. Nier,lg with a 5.5-m. single column filled withmethane at the optimum pressure of 46 cm., has obtained 4.57%13CH4, with a possible yield of 148 mg. of the enriched gas per hour.The same column would afford 376 mg. per hour of 304% 13CH4,using gas a t 61 cm. pressure. T. I. Taylor and G. Glockler 2O havedescribed a 10.9-m. column constructed from iron tubing of standardsizes, with which 3.6% 13CH4 has been obtained; the time requiredto reach equilibrium was about 180 hrs., with gas a t 28 cm.pressureand AT = 310". This column would yield 9% heavy methane iffresh normal methane were circulated a t the top. H. S. Taylor,21in describing the results of some preliminary experiments withdiffusion columns, states that a 12-m. column used under specifiedconditions should produce 20% 13CH4. The isotopic compositionof the methane obtained by all the above workers was determinedin a mass spectrometer. No doubt the application of the diffusionmethod to the carbon isotopes will be further developed in the nearfuture ; although the experimental technique is simpler, the resultsso far obtained are scarcely comparable with those secured by thechemical exchange method,22 which has yielded 2.5 g.of 25%Na13CN per day in an apparatus containing two exchange units.R. Fleischmann 23 has obtained enrichments of heavy nitrogen(15N) by the thermal diffusion method. Two 12-m. columns wereused separately to produce a quantity of gas containing 9.2% of15N14N from normal nitrogen (043% 15NI4N), and in a second stagethis enriched gas, in one of the columns, gave 110 C.C. of 18%15N14N. Unfortunately no attempt was made to separate the heavynitrogen from accompanying argon, and its concentration wasestimated by an indirect method. A marked enrichment was,however, confirmed by band-spectrum photographs. In the caseof nitrogen, the chemical exchange method has afforded muchbetter results, for H.G. Thode and H. C. Urey24 have obtainedmaterial containing up to 70% of 15N.No work has yet been published on the separation of l80 fromoxygen by thermal diffusion, although the method should beapplicable in this case. If the process were applied to oxygen gas thel8 Physical Rev., 1940, [ii], 57, 899.2o J . Chem. Physics, 1939, 7 , 851; 1940, 8, 843.21 Nature, 1939, 144, 8.22 C. A. Hutchinson, D. W. Stewart, and H. C. Urey, J . Chem. Physics, 1940,2s Physikal. Z., 1940, 41, 14.Ibid., p. 30.8, 532.24 J . Chem. Physics, 1939, 7 , 34WELCH : THE SEPARATION OF ISOTOPES. 161l a 0 would appear as la0l6O molecules, owing to the low abundance.Clusius and DickelZ5 have pointed out, however, that if the wiretemperature were sufficiently high the equilibrium 2180160 +la0, + 1602, established in the high-temperature region of thecolumn, would enable 1802 molecules to be separated.The partial separation of the neon isotopes obtained by Clusiusand Dickel has been described above; a somewhat smaller separ-ation has been secured by W.W. Watson.la Partial separationshave also been effected with krypton and xenon; 26 krypton gavelight fractions with atomic weights up to 1.74 units less than thenormal value, while heavy and light xenon fractions differing inatomic weight by 1.57 units were obtained. In both separationsrelatively short columns (5.5 and 2.5 m.), with unusually high wiretemperatures (1500", 1750"), were successfully used. The atomic-weight displacements were determined by a thermal conductivitymethod, which in the case of xenon is claimed to give results correctto 0.01% with 1 C.C.of gas. The separation of xenon is of specialinterest in that a parallel enrichment was effected in a series of 12Hertz diffusion pumps, which were found to be approximatelyequivalent to a thermal diffusion column only 1 m. long.Thermal diffusion separations of the isotopes of other elementshave not so far been described, although W. Krasny-Ergen2' hasgiven a specification for a column designed to separate 235U, usinguranium hexafluoride as the carrier gas. Considerable interestattaches to the enrichment of this isotope in view of its possibler6le in nuclear chain reactions in uranium. Probably the thermaldiffusion column provides the only means yet discovered of enrichingsuch a heavy isotope in appreciable quantities ; i t appears possible,however, that more success will be achieved in this separation ifi t can be carried out in the liquid phase (see below).Although the thermal diffusion constant, a, for a gas mixture ismainly determined by the masses of the molecules, other factors arealso involved.This is shown by successful separations of mixturesof gases of approximately equal molecular weight.2a In carbondioxide-propane, carbon monoxide-ethylene, and nitrogen-ethylenemixtures the larger species of molecule is enriched a t the bottom ofthe diffusion column, but in carbon dioxide-nitrous oxide thelatter gas separates at the bottom.No measurable enrichmentswere obtained with carbon monoxide and nitrogen. The separationss5 See ref. (13), p. 449.26 W. Groth and P. Harteck, Naturwiss., 1940, 28, 47; W. Groth, ibid.,27 Nature, 1940, 145, 742.28 F. T. Wall and C. E. Holley, jun., J. Chem. Physics, 1940, 8, 348.REP.-VOT,. XXXVII. F1939, 27, 260162 ZNORGmIC CHEMISTRY.recorded were too large to be accounted for by the fractionaldifferences in molecular weight of the two components .of eachmixture.Gases vary in their relative efficiency as “ carriers ” in isotopeseparations by thermal diffusion, and calculated values of cx obtainedfrom viscosity data 29 may be used as a rough guide to the probablesuitability of a given gaseous compound for this purpose.Theresults show, for example, that nitrogen should give better resultsthan ammonia, nitric oxide, or nitrous oxide in the separation ofnitrogen isotopes.Clusius and Dickel l2 have found that the efficiency of a thermaldiffusion column is increased by the washers used to space theheating wire from the walls of the tube. A. Brainley and A. K.Brewer 3O have also found that ( ( baffle plates ” inserted at intervalsalong a column increase its separation factor to a considerableextent. These results suggest that improved efficiency may beobtained by changes in the mechanical design of diffusion columns.Clusiua and Dickel l3 have pointed out several applications of thethermal diffusion method outside the field of isotope separations.Certain gases, such as xenon and helium, can be rapidly purifiedfrom substances of appreciably different molecular weight, Theformation of association compounds in the gas phase can be studied,and suspended particles in fogs and smokes can be separated veryeffectively. The vapours of azeotropic liquid mixtures should alsobe separable. It should be emphasised that the thermal diffusioncolumn is an experimental tool of general utility, which should bevaluable in effecting separations iii which the commoner physicalmethods of distillation, fractional adsorption, etc., are inefficient orhave failed.Thermal Diffusion in Liquids.An effect similar to the more recently discovered thermal diffusionprocess in gases has long been known to exist in solutions and€iquid mixtures, in which a temperature gradient produces a smallcorresponding gradient i n the relative concentrations (Ludwig-Soret effect 31) ; this effect is now regarded as a consequence ofthermal diffusion.With suitable modifications in experimentaltechnique the convective action utilised in the thermal diffusioncolumn for gases should therefore produce correspondingly largeSeparations in liquids. As the thermal diffusion constants forliquids are very much smaller than those obtaining in the gas phase,29 H. Brown, Physical Rev., 1940, [ii], 57, 242.30 J . Chem. Physics, 1939, 7, 972.31 C. Ludwig, Wien. Akad. Ber., 1856, 20, 539; C. Soret, Ann. Chim. Phys.,1881, 22,295WELCH: THE SEPARATION OF ISOTOPES. 163the separation process is slower and the optimum distance betweenthe heated and cooled surfaces is smaller (of the order of 0.1-0.25mm.). Preliminary attempts to apply the new process to liquidswere made by H. Korsching and K. Wirtz and by Clusius andDi~ke1,3~ who successfully separated several solutions and mixturesof organic liquids. A simple glass diffusion column in whichvisible sedimentation of aqueous copper sulphate and cobalt chloridesolutions can be obtained has been described by D. Taylor and M.R i t ~ h i e . ~ ~ The first actual isotope separation effected by thermaldiffusion in a liquid was obtained by Clusius and nickel 32 with heavywater ; in an apparatus consisting of heated and cooled parallel flatplates enclosing a diffusion space 150 em. long, 2 em. wide, and 0.1mm. thick, a sample of heavy water was separated to concentrationsof 62-9 and 644% deuterium oxide, respectively, a t the top and thebottom, after 8 hours' operation with a temperature difference of80" between the plates. Korsching and Wirtz 34 have also separateddeuterium oxide-water mixtures, in one case in a column only10 em. long. Although these separations were obtained underfavourable conditions in partly enriched material, they demonstratethe practical possibility of isotope separations in liquids. In spiteof the fact that the process is slow, the necessary apparatus is notcumbersome and reasonable quantities of enriched product shouldbe obtainable. In a more detailed paper,35 Korsching and Wirtzhave described suitable types of diffusion column, and indicatedmethods by which these may be used in successive stages to effectlarge separations.The theory of the thermal diffusion column for liquids has beendeveloped by P. D e b ~ e , ~ ~ and extended by J. W. Hiby and K.Wirtz 37 to practical types of apparatus. At present the theory isnot sufficiently developed to predict whether an isotope separationcan be effected in a given liquid; Wirtz 38 has discussed methods ofcalculating the thermal diffusion constant for a liquid mixture andthe possibility of isotope effects, but the problem is complex becausevarious factors connected with the structure of the liquid state areinvolved.Dissolved salts containing different isotopes should clearlysediment out at slightly different rates in a diffusion column, and apartial separation of the isotopes should result. A claim to havechanged the zinc isotope abundances in zinc sulphate solution by32 Naturwiss., 1939, 27, 110, 148.=4 Naturwiss., 1939, 27, 367.36 Ann. Physik, 1939, [v], 36, 284.37 Physikal. Z., 1940, 41, 77.38 Naturwiss., 1939, 27, 369; Ann. Physik, 1939, [v], 36, 295.33 Nature, 1940, 145, 670.35 Ber., 1940, 73, 249164 INORGANIC CHEMISTRY.this method remains uns~bstantiated.~~ Hiby and Wirtz 37 havediscussed the conditions necessary for this type of separation; itsdisadvantage is the large change of total concentration whichaccompanies the selective sedimentation effect. It may be possibleto avoid this in some degree by a suitable choice of solvent.No doubt the further development of the liquid diffusion columnwill indicate numerous possible applications, both in isotope separ-ations and in general chemistry. Several such applications arenow in course of e~amination.4~A. J. E. W.H. J. EMEL~US.A. J. E. WELCH.39 See ref. (34), and 2. Elektrochem., 1939, 45, 662.4') A. J. E. Welch, unpublished

ISSN:0365-6217    

DOI:10.1039/AR9403700125

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

CRYSTALLOGRAPHY.1. INTRODUCTION.THIS Report has been affected by present circumstances because thecontributors are all engaged to a considerable extent on unusualduties which make it difficult for them to collect the necessarymaterial, especially when access to many of the scientific journalsis difficult in any case. I n spite of this, however, the ground hasbeen covered fairly well, and it is encouraging to note that pureresearch work is being continued without very much abatement.As before, the first part of this Report deals with the morephysical aspects of X-ray analysis and crystallography. A newphenomenon has recently been engaging a great deal of attentionin this field. When X-rays are passed through a stationary crystal,a well-defined series of sharp spots are usually obtained on a photo-graphic plate placed behind the crystal-the well-known Lauepattern ; but in certain circumstances other effects can be observed.These are spots of a more diffuse character and disposed in a differentway.Their interpretation forms an interesting problem which isbeing vigorously discussed with a considerable amount of disagree-ment between different observers. There is no doubt, however,that the phenomena are of importance, and even in the absence oftheoretical explanation they can be utilised as an aid t o structuralwork in some cases. The matter is briefly dealt with in Section 2,as well as recent work on melting and crystal structure, and generallyon the transition from the liquid to the solid and other states ofmatter.Section 3 gives a brief account of recent work on hinderedrotation about single bonds, which has considerable bearing onthese as well as on the more structural topics in this Report.A proper review of recent work on structural chemistry is alwayshampered by the somewhat fictitious title of this Report. Liquidscan sometimes be disguised as crystals but gases tend to escape,and this year the structural work carried out by electron-diffractionstudies will be found in another part of the volume. In Section 4,however, the other main advances in the field of inorganic structuresare set out. Two very general works of great importance haveappeared during the year. The first of these 1 deals with the generalproblem of directed valency from the standpoint of group theoryand contains a valuable tabulation of spatial arrangements derivedG.E. Kimball, J. Chem. Physics, 1940, 8, 188166 CRYSTALLOGRAPHY.from all possible configurations of the valeiicy electrons on thecentral atom. In the second work there is a most valuablecollection of all the experimental data relating to the spatial arrange-ments of covalencies, with nearly 400 references to original papers,mainly in the fields of X-ray and electron-diffraction work. It isobserved that these arrangements tend to conform to quite a limitednumber of types, and the attempt is made to relate these types tosome familiar property of the atom. This property is the sizeof the valency group and the number of shared electrons which itcontains, together with the number of (unshared) electrons in thepreceding group.These works are reviewed in Section 4, and are followed by anaccount of recent experimental work on new structures in theinorganic field.These are not as numerous as usual, but containa number of interesting examples.New work in the field of organic structures is discussed in Section 5 .Circumstances have made it necessary to confine this account chieflyto those new structures which have been completely determinedduring the year, and although not very numerous they containsome very interesting results. Some of the macrocyclic pigmentswhich were expected to be isomorphous with the phthalocyanineshave been shown to possess a different type of crystal structure,but this has not been fully worked out.The detailed analysisof the platinum phthalocyanine structure, obtained in a verydirect manner from the intensity data, has been published, and hereagain the type differs from that of the other phthalocyanines.The difference, however, is one of molecular packing and not, to anyappreciable extent, of molecular structure.Some account is given of two other structures that have beenfully determined, the rhombohedra1 form of acetamide and thedimer of cyanamide, dicyandiamide. In both cases the moleculesare relatively simple ones, but crystallographically the structuresare complex, and their complete analyses represent a considerableachievement. The interatomic distances can be determined ineach case with sufficient accuracy to enable the state of the resonancehybrid to be ascertained.In each case the molecules are heldtogether by an interesting arrangement of hydrogen bonds whichhave a considerable effect on the stability of the structure.Much other work on hydrogen bonds has been published duringthe year, but as this is largely in the form of a general discussion 3it has not been thought necessary to review the work again here.N. V. Sidgwick and H. M. Powell (Bakerian Lecture), Proc. Roy. SOC.,J. M. Robertson, Trans. Paraday SOC., 1940, 56, 913.A , 1040, 1'76, 153UBBELOHDE CRYSTAL PHYSICS. 167There has also been a certain amount of work on complex polymers,rubber, and proteins, but an account of these subjects must bedeferred to a later Report.J. M. R.2. CRYSTAL F%YSICS : THERMODYNAMICS AND STRUCTURE.In more normal times, topics for the Annual Reports are usuallyselected so as to give a critical account of recent work, over a periodof years. The history of science during the war of 1914-1918shows that, although the advancement of knowledge may be thrownviolently out of balance, important developments do not wait forpeace. The choice of subjects in the following pages has thereforebeen directed to those fields whose development has been rapid,even when a critical review is not yet possible.Temperature Ejfects in the Rejection of X-Rays from Crystals.-As described in a previous Report,l the normal effect of raising thetemperature of a crystal is to increase the amplitude of its latticevibrations, and hence to decrease the intensity of X-rays reflectedby various lattice planes.Renewed attention has recently beengiven to certain " diffuse spot " reflections exhibited by somecrystals, when set up so as to give the familiar Laue reflections.These diffuse reflections are additional to the normal Laue pattern.It was furthermore shown by G. D. Preston that, contrary to thebehaviour of Laue spots, they are obtained with monochromaticX-rays, and that their intensity increases with rise in temperature.Diffuse reflections have been reported for single crystals of suchvarying nature as rock salt, sylvine, diamond, zinc, aluminium,3sodium nitrate: b e n ~ i l , ~ urea nitrate, urea oxalate, sorbic acid,h exame thylbenzene, and a-resorcinol.The theoretical interpretation of the origin of these diffusereflections is still under discussion.According to Preston, althoughthe reflections might be due to two-dimensional gratings (i.e.,sheets of atoms particularised in some way in the crystal), the in-tensities are in better agreement, at least in the case of aluminium,with the assumption that the crystals are broken up into groups,probably consisting of an atom and twelve neighbours. TheAnn. Reports, 1939, 36, 155.A. Guinier, Compt. rend., 1938, 206, 1641 ; A. P. Wadlund, PhysicaE Rev.,Ibid., 1939, 143, 76; Proc. Roy. Soc., 1939, 172, A , 116.(Sir) C. V. Raman and P. Nilakantan, Nature, 1940, 145, 667; Proc.5 (Miss) I. E. Knaggs, (Mrs.) K.Lonsdale, A. Miiller and A. R. Ubbelohde,(Miss) I. E. Knaggs, (Mrs.) K. Lonsdale, and H. Smith, ibid., 146, 332.1938, 53, 843; A. Taylor and D. Laidler, Nature, 1940, 146, 130.Indian Acad. Sci., 1940, 11, A , 379.Nature, 1940, 145, 820168 CRYSTALLOGRAPHY.interatomic distances are supposed to vary slightly from one groupto another, and this effect is ascribed to thermal vibrations. Analternative explanation is advanced by W. Zachariasen and hiscolleague^,^ following on the work of Faxen. According to theseauthors, the standing waves in a crystal, formed by the thermalvibrations, give rise to new regularities in the density distribution,and hence to new reflections a t higher temperatures. Yet anotherphysical explanation, analogous to the Raman effect for simplemolecules, is advanced by Raman and his colleagues.*This diversity of theories makes it clear that with the presentexperimental information more than one physical cause might beresponsible for diffuse reflections in single crystal^.^^ Indeed,since crystals can have more than one source of energy, whoseincrease with rise in temperature will lead to minor modificationsin structure,l the origin of diffuse spot reflections may not beunique.A geometrical examination of possible groupings of atomswhich may be responsible for diffuse reflections has been put forwardby (Sir) William Bragg,g and has the advantage that the reflectionscan be discussed without requiring any assumption about thephysical cause for such groupings.The marked temperature dependence of the intensity of diffusereflections suggests that further experiments will have considerableinfluence in developing the thermodynamics of crystals.The Transition from Liquid to Solid.-Owing to the fact thatmelting is conditioned by the equilibrium between a solid and a liquidstate of matter, an account of the transition should properly includesome description of recent work on the structure of both the solidand the liquid state.Within the last few years, fresh theories aboutthe liquid state have been almost more abundant than fresh experi-mental information. A comprehensive description of work on theliquid state is outside the scope of this report, but reference may bemade to two recent summaries.lO. l1It would not be possible, however, to give an account of modernwork on melting, without alluding to two broad lines of experi-mental research on liquids. The first of these has established thatmany properties of liquids can be accounted for much better by theassumption that their structure is quasi-crystalline, a t any rate in7 A.P. Wadlund and W. Zachariasen, Physical Rev., 1938,53, 843; Nature,1940, 145, 1019; W. Zachariasen and S. Siegel, Physical Rev., 1940, 57,597, 795.8 Proc. Roy. SOC., 1938, 168, A, 302. Nature, 1940, November.10 J. A. V. Butler, Ann. Reports, 1937, 34, 75.11 N. F. Mott and R. W. Gurney, Ann. Reports Physical SOC., 1938, 5,46; J . Chern. Physics, 1938, 6, 222; C. A. Benz and G. W. Stewart, PhysicalRev., 1934, 46, 102UBBELOHDE : CRYSTAL PHYSICS.169the neighbourhood of the melting point, than by the assumption,incorporated in van der Waals's famous theory, that the structureof liquids is quasi-gaseous. One of the simplest illustrations of thisconclusion is provided by the specific heats, in the neighbourhoodof the melting point. For solid mercury, C, = 6-72, and for liquidmercury, C, = 5.90 cals./mol.12 This behaviour is not restrictedto liquids with simple molecules, since the specific heats of solidand liquid long-chain paraffins are also closely ~imi1ar.l~~ l4Other information on the quasi-crystalline structure of liquidsincludes the Brillouin scattering of monochromatic light by stationarywaves in the medium, which gives a broadened central line forgases, a doublet for solids, and a doublet + a central line forliquids.12 The absorption of ultrasonic waves and high-frequencyelectromagnetic waves provides further evidence.15 X-Ray studieson liquids continue to show the roughly similar packing of themolecules, compared with the corresponding solids.16 Miscellaneousinvestigations cover work on liquid carbon dioxide,17 on thestructure of molten salts,18 on the viscosity of liquid metals nearthe melting point,lg on thermodynamic evidence for the restrictedrotation of molecules in certain and on some mechanicalmodels for liquids.21 The general conclusion 22 is that the packingof molecules in most liquids, near their freezing points, is roughlysimilar to the packing in the corresponding solids, but that anymicro-crystals in the liquid do not as a rule comprise more thanvery few molecules.An alternative description, referred to againbelow, is that a liquid, near the melting point of the solid, correspondswith a polycrystalline conglomerate in which the individuals havesuch a large number of defects that its crystalline character is notapparent except in special tests.l2 P. Debye, 2. Elektrochem., 1939, 45, 174.l3 A. R. Ubbelohde, Trans. Paraday SOC., 1938, 34, 289.l4 L. Brillouin, J . Phys. Radium, 1936, 7, 153 (see, especially, theoreticall5 P. Debye and W. Ramm, Ann. Physik, 1937, [v], 28, 28.l6 E.g., W. Pierce, J . Chem. Physics, 1935, 3, 266; W. Pierce and D.Macmillan, J . Amer. Chem. Soc., 1938,60,779; H. Sirk, 2.Physik, 1934,89,129.l7 L. H. Borchert, Physikal. Z . , 1938, 39, 156.l 8 E. Miller, Physical Rev., 1940, 57, 61.Is Y. S. Chiong, Proc. Roy. Soc., 1936, 157, A , 264.2o A. R. Ubbelohde, Trans. Paraday SOC., 1939, 35, 843; R. S. Halford,J . Chem. Physics, 1940, 8, 496; D. Osborne, R. Doescher, and D. Yost,ibid., p. 506; E. Fischer and G. Klages, Physikal. Z . , 1939, 40, 721.21 H. Rehaag and H. A. Stuart, ibid., 1937, 38, 1027; W. Kast and H. A.Stuart, ibid., 1939, 40, 714; R. Fiirth, L. S. Ornstein, and J. M. W. Milatz,Proc. K . Akad. Wetensch. Amsterdam, 1939, 42, 107.22 Cf. A. Magat, Trans. Paraday SOC., 1937, 33, 114; J. Malsch, 2. Elektro-chem., 1939, 45, 813; F. Frank, Physikal. Z., 1938, 39, 530.discussion)170 CRYSTALLOGRAPHY.A second broad line of experimental research on liquids refersto work on the formation of crystal nuclei.Evidence for the quasi-crystalline structure of liquids, near their freezing points, emphasisesthe similarity in packing of molecules in the liquid and the solidstate of matter. On the other hand, evidence about the spontaneousformation of crystal nuclei in melts or solutions makes it quite clearthat there must also be a marked difference in the arrangement ofmolecules in the two states. Numerous cases are known of thesupercooling of a melt or solution, without any separation of crystals.On the other hand, definitely established cases of the superheatingof crystals above their melting point are so rare 23 that they might bedescribed as non-existent.A further peculiarity of spontaneouscrystallisation is the tendency for the least stable state to separatepreferentially, according to the so-called rule of successive states.Recent examples of the formation of metastable states includezinc sulphideY24 the precipitation from certain solid solutions oncooling,25 and the crystallisation of some long-chain compoundsfrom their melts.26A theoretical interpretation of the rules governing meltingand crystallisation has been suggested in terms of the thermalfluctuations in the systems.27Experimental evidence for the persistence of crystallisationnuclei above the melting point of the solid,28 and for variousmechanisms which may apply in the spontaneous formation ofcrystal nuclei,29 has hardly done more than emphasise the difficultiesin studying nucleus formation.The recently discovered influenceof an electric field on the location of nuclei in supercooled meltsor solutions 30 may help to solve some of the experimental difficul-23 M. Volmer and 0. Schmidt, 2. physiknl. Chem., 1937,35, B, 467.24 S. Madigp, Physical Rev., 1937, 51, 61.2 5 A. J. Bradley, Proc. Physical SOC., 1940, 52, 80.36 J. W. H. Oldham and A. R. Ubbelohde, Proc. Roy. SOC., 1940,176, A, 50,65.27 A. R. Ubbelohde, Trans. Faraday Soc., 1937, 33,1203 ; G. Borelius, Ann.Physik, 1938, [v], 33,517 ; but see F. C . Frank, Proc. Roy. SOC., 1939,170, A, 182.28 V. Danilov and J. Radschenko, PAysikal. 2. Sovietunion, 1937, 12, 745;P. Anderson, Phyeicul Rev., 1936, 50, 386; R.Kaischew, Ann. Physik, 1937,[v], 30, 184; W. L. Webster, Proc. Roy. SOC., 1933, 140, A , 653.2) A. T. Jensen, 2. physikal. Chem., 1937, 180, A , 93; A. T. Wahramianand S. A. Alemian, Acta Physicochim. U.R.S.X., 1937, 7 , 95; E. Androni-kaschvili, ibid., 6, 689; G. L. Michnevitsch et al., PhysikaZ. 2. Sovietunion,1938, 13, 103; V. Danilov and W. Neumark, ibid., 1937, 12, 313; L. Hornand G. Masing, 2. Elektrochem., 1940,46,109; L. Hamburger, C'hem. WeekbZad,1938, 35, 886; I. Sokolov, Tech. Phys. U.R.S.S., 5, 619; G. Schmid andA. Roll, 2. Elektrochem., 1939, 45, 769.3O W. Rix, 2. Krist., 1937, 96, 155; C. Hammer, Ann. Physik, 1938, [v],53,445 ; H. C. Hamaker, Trans. B'aruday SOC., 1940,36,279 ; A. R. Ubbelohde,ibid., p. 863UBBELOHDE : CRYSTAL PHYSICS.171ties. In connection with the effect of chance dust particles onnucleation, mention may be made of studies on the melting ofadsorbed substances 31 and of two-dimensional layers of molecules.32Factors influencing the steady velocity of growth of crystals, oncethese have reached a sufficient size, are also of importance in de-termining the transition from liquid to solid. Recent experimentshave made some progress in this field.33Melting and CrystaZ Structure.-The heat of fusion and themelting point of crystals range from very low values, for helium,to very high values, such as are exhibited by certain carbides, andby graphite (m. p. ca. 4000” ~ . ) . 3 ~ The ratio of the heat of fusionto the melting point, measured on the’absolute scale of temperature,gives the entropy of fusion A#,, which is a far more useful quantityin correlating melting with crystal structure. Inspection of theBoltzmann equation for the entropy of fusion,35 vix., AS, =R log, Wl/W,, suggests that the number of ways, W, and WI, ofrealising the solid and the liquid state might be the same for crystalsof similar structure, irrespective of the actual melting point.Inparallel with Trouton’s rule, which states that the entropy ofvaporisation a t constant pressure is independent of the boiling point,Richards’s rule for solids states that for crystals with similarstructure, the entropy of melting is a constant, independent of theactual melting point.36In practice, Richards’s rule for the melting of solids is found tohave a far narrower scope than Trouton’s rule for liquids.Thereason for this is that as soon as the number of atoms per moleculeincreases, the number of degrees of freedom of the molecule likewiseincreases. Owing to the constraints imposed by the crystal structure,these degrees of freedom are frequently not fully excited in the solid,but may become active on melting, thereby increasing the entropyof fusion. A simple illustration of how this can upset a smoothsequence of melting points can be obtained from a comparison of aW. T. Richards, J . Amer. Chem. SOC., 1932, 54, 479.31 W. A. Patrick and W. A, Kemper, J . Physical Chem., 1938, 42, 369;32 D. Dervichian, J . Phys. Radium, 1939, 10, 333.j 3 J. Michel, BUZZ. Xoc.chim. Belg., 1939, 48, 105; J. Dubourg and R.Saunier, Bull. Xoc. chim., 1938, [v], 6, 1196; M. Pahl, 2. physikal. Chem.,1938, 184, 245; D. Mavroska and G. Valensi, Compt. rend., 1939, 208, 1648,1727; T. Erdey-Grbz and R. Kardos, 2. physikal. Chem., 1936, 178, A, 255,266; W. F. Berg, Proc. Roy. Xoc., 1938, 164, A, 79; U. Hoffmann, Ber.,1939, 72, 754; H. E. Buckley, Mem. Manchester L i t . Phil. Soc., 1939, 83, 32.34 W. Hume-Rothery, J. Physical Chem., 1940,44, 808; J. Bassett, J. Phys.Radium, 1939, 10, 217.35 Ann. Reports, 1939, 36, 163.36 Cf. A. Eucken, “ Energie and Warmoiiihalt ”; also tabulated data,Ann. Reports, 1939, 36, 159172 CRYSTALLOGRAPHY,range of similar crystals, in which rotation of the molecules occursin some examples, but not in others.37 Abnormally low meltingpoints are usually found in the sequence, when the entropy ofmelting is high, Le., when the entropy increase for rotation is includedin the entropy of fusion.When a wider range of crystals is considered, particularly inthe case of complicated organic molecules, it is found that the entropyof melting can have almost any value.38 Systematic investigationshave been carried out chiefly in the case of long-chain polymethylenemolecules. So far as is known,39 there is no upper limit to theentropy of melting. Empirical formulae for the freezing point ofvarious homologous series can be interpreted on the assumptionthat successive methylene groups make a constant addition to boththe heat of fusion and the entropy of fusion.If n is the number ofCH, groups, and H and X the respective increments of heat andentropy of fusion, these can be writtenAHf= H , -+ nHASf = Xo + nSwhere H, and So are constants characteristic of the compound(long-chain paraffin, alcohol, ketone, iodide, etc.). When theseformulae apply, the melting point T, = AHf/AS, reaches a limit-ing ‘‘ convergence temperature ” for very high values of n.40Some attempt has been made to give a physical interpretationof these equations connecting the entropy and heat of fusion withthe number of methylene groups,41 kinetic arguments being used todescribe the melting of chain molecules. There is evidence thattorsional vibrations of the chains play a considerable part in themelting of polymethylene compounds,26* 42 and rotation may notbe entirely free even in the liquid phase.43 It is interesting to notethat the stiffening of a long chain by conjugation, as in the p~lyenes,~Pleads to a quite different behaviour on lengthening the chain. So37 W.0. Baker and C. P. Smyth, J . Amer. Chem. SOC., 1939, 61, 1695.38 J. Timmermans, Bull. SOC. chim. Belg., 1935, 44, B, 17 ; J . Chim. physique,1938, 35, 335; Bull. Acad. roy. Belg., 1939, 25, 417; A. van de Vloed, Bull.Soc. chim. Belg., 1939, 48, 229.39 A. R. Ubbelohde, ref. (20).40 J. Timmermans, Bull. SOC. chim. Belg., 1919, 28, 392.4 1 (Miss) A. M. King and W. E. Garner, J., 1936, 1368.42 J. W. H. Oldham and A. R. Ubbelohde, Trans. Paraday Soc., 1939,35, 332; A. Muller, Proc. Roy.SOC., 1940, 174, A , 137; A. Eucken andE. Schroder, Ann. Physik, 1939, [v], 36, 609.43 J. S. Koehler and D. M. Dennison, Physical Rev., 1940, 57, 1006; W. B.Bridgman, J . Amer. Chem. SOC., 1938, 60, 530.44 R. Kuhn and C. Grundmann, Ber., 1936, 69, 224UBBELOHDE : CRYSTAL PHYSICS. 173far as has been determined, the melting point continues to risesteeply as the number of carbon atoms increases.Attempts to derive a statistical theory of melting depend onthe devising of a successful model on which to base calculations ofthe partition functions of the solid and the liquid state. Quitenaturally, theories have so far been restricted to the melting ofsimple atomic lattices, where no degrees of freedom need be con-sidered in the liquid or solid, apart from those connected withthe positions of the centres of gravity of the atoms.Considerablesuccess has been achieved with the carefully worked out model dueto Lennard- Jones and his c0lleagues.~5 According to this model,the crystalline state a t low temperatures corresponds with a state ofcomplete order of the atoms, all being situated a t the specifiedpoints or ct sites of the crystal lattice. With rise in temperature,some of the atoms can occupy intermediate p sites in the crystal,and the liquid corresponds with an arrangement in which theoccupation of p sites preponderates. The presence of empty usites in this model explains the easy migration of atoms in the liquid,and accounts for its mechanical properties. Although the modelis necessarily formalised for the purpose of calculation, successfulestimates have been made by its means of the entropy and volumechanges on melting the lattices of argon and nitrogen.The heatof fusion has also been calculated. Molecules are treated asequivalent to atoms, so that rotational transitions are neglected.Oxygen does not give satisfactory results with this model , possiblyowing to the r61e of non-van der Waals forces in the crystals. F. A.Lindemann’s semi-empirical correlation between the melting pointand the Einstein frequency of the crystals is also derived from thetheory. The possibility of a continuous transition between solidand liquid above a critical pressure and temperature 46 has also beendiscussed, in view of its importance for geophysical theories.With the present experimental evidence, the particular type oflattice disorder assumed by Lennard-Jones may be regarded asconvenient for calculation, without necessarily excluding other typesof lattice defects.From the behaviour of solid solutions, there issome experimental evidence for co-operative defects, in which theenergy of p sites is lowered when these occur in a suitable relation-s hip. 26Theories of melting in which the emphasis is laid on the changein mechanical properties, associated with the phase change, have45 J. E. Lennard-Jones and A. F. Devonshire, Proc. Roy. SOC., 1939, 169, A,317; 170, A , 464; A. F. Devonshire, ibid., 1940,174, A , 102; J. E. Lennard.Jones, Proc. Physical SOC., 1940, 52, 38.4 6 P.W. Bridgman, Physical Rev., 1934, 46, 930; Proc. Amer. Acad. ArtsSci., 1935, 70, 1174 CRYSTALLOGRAPHY.not been worked out in the same detail.47 One of the obstaclesis the want of a completely satisfactory theory of the mechanicalproperties of crystals, in relation to their structure.** Attentionmay be drawn to a semi-empirical approach 49 based on Sutherland'sempirical formula for the rigidity modulus p : 5OThe radiation pressure due to standing waves in the crystal, i.e., tothe thermal oscillations, can be expressed in terms of this rigiditymodulus and the temperature, and automatically leads to a phasewith no macroscopic rigidity above the melting point. For verysmall crystals, however, which only contain standing waves of highfrequency, the microscopic rigidity modulus may remain finite atthe melting p0int.~1 Hence very small crystals may persist tosomewhat higher temperatures.Although these views are notyet well defined, they contain a number of features which anymechanical theory of melting must take into account.Finally, brief mention may be made of the topic of premelting.No experimental means has yet been devised which decides whetherthe transition from solid to liquid really occurs a t a single temper-ature, or over a more or less narrow but finite interval, dependingon the crystal structure.26 Actual measurements of the volumechange or heat change have always shown a transition over anarrow range of temperatures : for instance, for very pure mercurya volume change was observed extending over an interval of 0.024°.52This behaviour has hitherto been explained on the assumption ofimpurities insoluble in the solid phase, but there is little theoreticaldifficulty in modifying the thermodynamic equations for a phasetransition to account for transitions over a narrow temperatureinterval, instead of a melting One view is that thefluctuations may be unusually large near the melting point,54 sothat it would be difficult to define the temperature with anyaccuracy-say, to 10-4".Though the significance of premelting must be regarded as un-decided at present, the topic is of some interest in view of its con-nection with the abnormally low mechanical strength of ~rysta1s.l~47 M.Born, Nature, 1940, 145, 741; J .Chem. Physics, 1939, 7, 591;48 Cf. discussion in Proc. Physical SOC., 1940.48 R. Luc&s, Compt. r e d . , 1938, 207, 1408.50 W. Sutherland, Phil. Mag., 1891, [v], 32, 31. 215, 524.51 L. Brillouin, Physical Rev., 1938, 54, 916.52 A. Smits and G. J. Muller, 2. physikal. Chem., 1937, 36, By 288.53 A. Rutgers and S. Wouthuysen, Physica, 1937, 4, 235, 515.54 J. Frenkel, J . Chern. Physics, 1939, '7, 538; E. Brody, ibid., p. 538.R. Furth, Nature, 1940, 145, 742UBBELOHDE : CRYSTAL PHYSICS. 175Transition f r m the Crystulline to 8pecial States of MatEer.--Inaddition to the transition from the crystalline to the anisotrophliquid state, a number of related transitions are known, whichthrow further light on the modes of entropy increase in crptals.T4e most familiar of these is the transition to the liquid crystalstate. Since the previous report on liquid crystals,55 work has beenrather spasmodic. References to recent discussions include thosegiven below,56 but a fuller discussion must be left for a subsequentreport.A large amount of work has recently been published on thetransition from the crystalline to the rubber-like state of matter.57In addition to natural rubber, it number of polymerised hydrocarbonshave been discovered, which normally exist in what is best describedas a new stake of matter.Two main properties of the rubber state,which throw light on its statistical make-up, are its behaviourcooling and on stretching. X-Ray photographs of unstretchedrubber a t ordinary temperatures show only diffuse rings, eom-parable t o those of a liquid or a gas.When the rubber is cooledto a fairly low temperature, the photographs change to the sharprings given by a plycrystalline material, so that in this respectrubber resembles the more familiar liquids. As was first discoveredby J. R. Katz,5* when rubber is skretched at room temperature,X-ray photographs show that crystallisation likewise takes place. Ithas long been known that heat is evolved on stretching rubber. Thethermodynamic consequence, that Young’s modulus €or rubber hasa, positive temperature coefficient, over a range of temperature, isin contrast with the majority of crystalline solids. The liberationof heat and the crystallisation which occur on stretching show thatthe entropy of the molecules is decrcased in the process.5 5 Ann.Reports, 1931, 28, 280.66 F. C. Frank and K. Wirtz, Naturwiss., 1938, 26, 688, 697; E. Omstein,Proc. K. Akad. Wetensck. Amsterdam, 1938, 41, 1046; R. Furth and K. Sitte,Ann. Physik, 1337, [v], 30, 388; W. Kast, 2. Elekt?-ochm., 1939, 45, $84;C. Weygand and R. Gabler, 2. physikal. Chem., 1939,44, B, 69; D. Vorlsnder,2. Krist., 1931, 79, 61; Ber., 1938, 71, 1688.5 7 E.g., E. Guth and H. Mark, 8. Elektrochem., 1937, 43, 683 ; F. Dart andE. Guth, Physical Rev., 1938, 53, 327; H. Mark, Nature, 1938, 141, 670;2. physikal. Chem., 1948, 38, B, 395; F. Misch and A. van der Wyk, J . Chern.Physics, 1940,8, 127; Helv. Chim. Acta, 1939, 22, 1358, 1362; W. Smith aadC.Saylor, J . Res. Nut. Bur. Stand., 1938, 21, 257; G. Clark, ibid., 22, 105;S. Bresler and J. Frenkel, Actu Physicochim. U.R.S.X., 1939, 11, 485; J.Frenkel, ibid., 1938, 9, 235; W. Kuhn, 2. Elektrochem., 1939, 45, 335; E.Wohlisch, Kolloid-Z., 1939, $9, 239; 8. physikal. Chern., 1939, 184, 416;E. Sa.uter, ibid., 1937, 36, B, 405; P. Tkiessen and W. Kirsch, ibid., 1939,43, B, 292; 1938, 41, B, 33.58 K01lozd-Z., 1925, 33, 300; 37, 19176 CRYSTALLOGRAPHY,These and other facts about the transition from the crystallineto the rubber state have been interpreted by assuming that the verylong hydrocarbon molecules can increase their entropy by coilingup, through rotations about the C-C bonds. The large number ofpossible states of coiling results in a considerable increase in entropy,a,s a result of the transition, and this explains why the equilibriumstate of unstretched rubber has the majority of the molecules in thecoiled state, even a t room temperature.Details of the statistical treatment will be found in the originalpapers.Although not all the possible states of the moleculesare known in such a complex mixture of polymers as rubber, thistype of change of state has its importance, even outside rubberchemistry. Other complex organic molecules, such as the proteins,may have modes of entropy increase, connected with changes of'' crystal structure " in the wider sense, which are intimately con-nected with such problems as the thermodynamics of muscularwork. A. R. U.3. HINDERED ROTATION ABOUT SINGLE BONDS.Where it is applicable, X-ray crystal analysis is the methodcapable of furnishing most detailed information regarding molecularstructure.Evidence as to how far crystal forces may alkr theconfiguration of a molecule is therefore of some interest.Great importance consequently attaches to the recent discoveryof high-energy barriers hindering so-called " free rotation " aboutsingle covalency bonds, even in quite simple compounds. Forexample in molecules such as s-dichloroethane in the gaseous stateevidence from electron-diffractionindicates that rotation about the C-C bond is hindered by a potentialbarrier almost certainly higher than 4 kg.-cals. /mol. More remark-able is the suggestion, first made by J. D. Kemp and K.s. P i t ~ e r , ~and subsequently confirmed by many ~ t h e r s , ~ that even in ethaneand its homologues the hindering potential is of the order of 3kg.-cals./mol. This hypothesis is based mainly upon a comparisonof the observed entropy (or heat capacity-a related quantity) ofthe gas with those calculated by statistical mechanics on the basis of1 This vol., p. 64; J. Y. Beach and K. 3. Palmer, J. Chenz. Physics, 1938,6, 639.2 E.g., J. Y. Beach and D. P. Stevenson, ibid., p. 635; this voI., p. 59.* G. B. Kistiakowsky and F. Nazmi, J . Chem. Physics, 1938, 6, 18;K. Schafer, 2. physikal. Chena., 1938,40, B, 357; E. B. Wilson, junr., J . Chem.Physics, 1938, 6, 740; G. B. Kistiakowsky, J. R. Lacher, and F. Stitt, ibid.,1939, 7, 289.and from dipole momentsIbid., 1936, 4, 749; J .Amer. Chern. Soc., 1937, 59, 276SPEAKMAN : HINDERED ROTATION ABOUT SINGLE BONDS. 177molecular models corresponding to various barriers; but it is sup-ported by other data.5 Barriers of the same order of magnitudehave been detected in many other simple molecules.6 Steric effects,calculated in a manner suggested by H. Eyring,' are possibly ade-quate to account for the barriers observed in substituted ethanes(e.g., CH,C1*CH,C18), but not for that in ethane itself; and herethe only explanations so far put forward would require the stableconfiguration to be the " opposed," or " eclipsed '' (of symmetryD,*), rather than the " staggered " one (D3J, which on generalgrounds is regarded as much the more probable.1°These high barriers to internal rotation, being considerably largerthan the estimated threshold energies for the rotation of the moleculeas a whole in the crysta1,ll will oppose the assumption by themolecule of any configuration different from the mean configurationoccurring in the gas.Certainly with small compact molecules,therefore, rotation of the molecule as a whole will take place ratherthan any considerable permanent distortion of its parts. As anexample, K. S. Pitzer l2 concludes from a study of the thermalproperties of solid s-dichloro- and -dibromo-ethane that, althoughtotal molecular rotation occurs when the temperature is sufficientlyhigh, there is no internal rotation about the C-C bond in the crystal ;and R. W. Sillars l3 and D.R. Pelmore l4 conclude that the dielectricbehaviour of solutions in solid waxes of certain long-chain estersindicates that the ester molecules rotate rigidly. On the other hand,with long polymethylene chains an accumulation of small oscillationsE.g., J. B. Howard, J. Chem. Physics, 1937, 5, 451 ; B. L. Crawford, junr.,W. H. Avery, and J. W. Linnett, ibid., 1938, 6, 682.J. G. Aston, Chem. Reviews, 1940,27,63 (for other references) ; C. R. Bailey,S. C. Carson, and E. F. Daly, Proc. Roy. SOC., 1940, 173, A, 347; G. B. Kistia-kowsky and W. W. Rice, J. Chem. Physics, 1940, 8, 610; J. S. Koehler andD. M. Dennison, PhysicaZRev., 1940,57,1006; D. W. Osborne, R. N. Doescher,and D. M. Yost, J. Chem. Physics, 1940, 8, 506; L. R. Zumwalt and R.M.Badger, J. Amer. Chem. SOC., 1940, 62, 305; J. G. Aston and A. M. Kennedy,ibid., p. 2567; B. L. Crawford, jun., J. Chem. Physics., 1940, 8, 744.J . Arner. Chem. Soc., 1932, 54, 3191.J. Y. Beach and K. J. Palmer, J. Chem. Physics, 1938, 6, 642.@ E. Gorin, J. Walter, and H. Eyring, J. Amer. Chem. Soc., 1939, 61, 1876;A. Eucken and K. Schiifer, Naturwiss., 1939, 27, 122.lo B. L. Crawford, junr., W. H. Avery, and J. W. Linnett, J. Chem. Physics,1938,6,682 ; J. B. Conn, G. B. Kistiakowsky, and E. A. Smith, J. Amer. Chew,.SOC., 1939,61, 1872; K. S. Pitzer, Chem. Reviews, 1940, 27,44; V. Schomakerand D. P. Stevenson, J. Chem. Physics, 1940, 8, 637.l1 E.g., L. Pauling, Physical Rev., 1930, 36, 430; Alex Muller, Helv. PhysicuActa, 1936, 9, 626; J.G. Kirkwood, J. Chem. Physics, 1940, 8, 205.l2 J . Amer. Chem. Soc., 1940,62,331; see also A. H. White and W. S. Bishop,ibid., p. 16.l3 Proc. Roy. Soc., 1938, 169, A , 66. l4 Ibid., 1939, 172, A , 502178 CRYSTALLOGRAPHY.about each C-C bond may give rise to a considerable torsionalflexibility between the ends of the chain.15 J. M. Robertson andhis couaborators l6 have found that crystals of stilbeae and of tram-azobenzene contain two kinds of molecule, differing with respect tothe orientations of the two benzene rings; but the difference inorientation is very small-about 15". (It may be noted that, if theopposed were the stable configuration in ethane and its homologues,then the flat zigzag shape normally assumed by polymethylenechains in crystals or in monolayers-a shape requiring the staggeredconfiguration round each C-C bond-could not be formed withoutgreat strain.)In some other instances, where crystal analysis has shown groupsto be greatly distorted from their normal orientations, the distortioncan be attributed to large steric effects, which would presumablyoperate similarly were the molecules in the gaseous sfate. Thisapplies to the conclusion of C.J. B. Clews and K. Lonsdale l7 thatin solid o-diphenylbenzene the phenyls are twisted through s w e50" from the plane of the C,W, nucleus; to that of J. M. Robert-~0n17a that in solid cis-azobenzene the rings are also twisted throughsome 50" from the plane of the central C-NIN-C atoms [a displace-ment no doubt contributing towards the (algebraically) lower energyof formation of the cis-isomer I*] ; to that of G.Huse and H. M.Powell that the o-nitro-groups in picryl iodide ere rotatedthrough 80" from the plane of the ring; and to that of J. D.McCullough 2o (obtained by an interesting extension of the radialdistribution method to X-ray powder photographs) that, whilst meso-stilbene dibromide has a trans-configuration, the molecule in theracemic form is twisted by 90" (in which direction is not certain)from that configuration in which the two bromine atoms are transto one another.A somewhat different aspect of the case is implied in the work ofS. Mizushima, Y. Morino,and their collaboratorsY2land of L. Kahovecand K. W. F. Kohlrausch,22 who show that a number of frequenciespresent in the Raman spectra of liquid or dissolved ethy€ene di-l5 This vol., p- 172.l6 J.M. Robertson and (Miss) I. Woodward, Proc. Roy. Xoc., 1937, 162, A ,568; J. J. de Lange, J. M. Robertson, and (Miss) I. Woodward, 2bid., 1939,171, A , 398.l7 Proc. Roy. Xoc., 1937, A , 161, 493; see also K. Lonsdale, 2. Krist., 1937,97, 91.lia J., 1939, 232.l8 R. J. Corruccini and E. C. Gilbert, J . Amr. Chem. Soc., 1939, 61, 2925.J . , 1940, 1398. 2o J . Amer. Chem. SOC., 1940, 62, 480.21 Sci. Papers Inst. Phys. Chem. Res. Tokyo, 1936, 29, 63; 1939, 36, 281 ;22 Ber., 1940, 73, 169.1940, 37, 205HAMPSON INORGANIC STRUCTURES. 179halides and related compounds are absent from the spectra of thecorresponding solids. Broadly speaking, both groups of workersagree in attributing this to a restriction of the molecules in thecrystal to a single configuration (in the dihalides, for instance, trans) ;in the ‘‘ mobile ” states the restriction is relaxed, so that anotherform may exist also (in the dihalides probably a cis-form).Asmentioned above 1* the evidence leads to the conclusion that thereis no appreciable proportion of cis-molecules present in the vapoursof these dihalides. The work of E. Pischer 23 on dielectric propertiesmakes it appear possible that the rotation of groups may be morenearly free in the liquid than in the gaseous or the solid state, thoughthere is evidence that barriers to internal rotation do occurin dissolved molecules.24 J. C. S.4. INORGANIC STRUCTURES.Inevitably the number of structure determinations which havebeen carried out during the past year is less than in previous years,and as no obvious general problem has been pursued, a brief mentiononly will be made in the second sub-section of this report of some ofthe more interesting individual cases.Two papers, however, on thegeneral problem of directed valency have recently appeared and areof considerable interest. In the first of these the problem has beentreated from the group-theory point of view ; it is really an exten-sion of the earlier work of Pauling and Mulliken,2 and in it are dis-cussed the spatial arrangements of all possible electron co~gurationsfrom two to eight s, p , or d valency electrons on a central atom. Themethod does not predict directly the type of bond arrangementformed by any given electron configuration, but merely tells whetheror not a given arrangement is possible.In some cases it is foundthat several arrangements are possible for a single configuration ofelectrons; for example, where we have a central atom covalentlylinked to two other atoms in a molecule AX,, the electron configura-tions sp and d p can lead to either a linear or an angular arrangement ;the configurations ds, d2, and p2, on the other hand, must be angular.In those cases where there is ambiguity, the arrangement which isadopted can only be predicted by applying other criteria to deter-mine the relative stability .of the various arrangements. Suchcriteria might be the “ strength ” of the bonds as determined by23 E.g., Physikal. Z ., 1939, 40, 645; see also T. Y. Wu, J . Chem. Physics,1939, 7, 967.24 E.g., G. L. Lewis and C. P. Smyth, ibid., p. 1085.1 G, E. Kimball, J . Chem. Physics, 1940, 8, 188.2 L. Pauling, J. Amer. Chem. Xoc., 1931, 53, 1367; R. S. Mullikon, J . Chem.Physics, 1933, 1, 492180 CRYSTALLOGRAPHY.the degree of overlapping of the bond orbitals,3 or considerations ofrepulsion between non- bonded atoms, or the possibility of utilisingorbitals for the formation of ( ( resonating " double bonds. I nmercuric chloride, for example, where the electron configuration i8sp, a linear rather than an angular arrangement is favoured bothbecause of the repulsive forces and also because of the possibility ofresonance with CltHgf-Cl.It must be emphasised that grouptheory cannot tell us anything about the relative stabilities of thevarious electron configurations, any more than it can tell us of thelikelihood of occurrence of any of the 230 space groups ; if, however,an atom is kno'wn to have a certain electron configuration, its possiblespatial arrangements can be deduced.It is because of the difficulty which many chemists experience ofknowing what electron configurations are involved in bond forma-tion, that N. V. Sidgwick and H. M. Powell have collected togetherthe experimental evidence, and correlated the stereochemistry of amultivalent atom with the size of its valency group and the numberof shared electrons which it contains, together with the number ofelectrons (unshared) in the penultimate electronic group.X-Rayand electron-diffraction experiments have shown that stereo-chemical configurations tend to conform to quite a limited numberof types, and the conditions under which they are formed can bestated as follows :I. When the valency group is less than an octet, the electronstake up positions of highest symmetry, namely, a linear arrange-ment for a group of 4, as in mercuric chloride, and plane symmetricalfor a group of 6, as in boron trifluoride. This also appears to be thearrangement when the sextet is not fully shared, as in covalentstannous chloride and lead bromide.11. With a complete octet the arrangement can be either tetra-hedral or planar. When the covalency is less than 4 it is alwaysderived from the tetrahedron as in the triangular water and thepyramidal ammonia.* The fully shared octet is always tetrahedral* A linear arrangement, however, with a bicovalent octet has been observedin certain crystal structures, e.q., the P-0-P group in the pyrophosphateion [P,0,]-4,5 and the S-S-S group in the trithionate ion [S,0,li2.." This- -0 0- I +._I -may be due to the contribution of structures such as 0-P=O--Y-O.0 - ' 0 -L. Pauling, " The Nature of the Chemical Bond," p. 76.* Bakerian Lecture, Proc. Roy. SOC., 1940, 176, A , 153.G . R. Leri and G. Peyronel, 2. Krist., 1935, 92, 190; G . Peyronel, ibid.,1936, 94, 511.13 W. H. Zachariason, zbid., 1934, 89, 529HAMPSON : INORGANIC STRUCTURES.181when the penultimate group is completed (2, 8, or 18 electrons).*I n the transitional elements where the number in the penultimategroup lies between 8 and 18, the arrangement is tetrahedral whenthe penultimate group contains not much more than 8, and planarwhen it contains not much less than 18 electrons, the two series over-lapping in the middle. The square planar arrangement may beregarded as that of an octahedron with the two trans-positionsunoccupied, and Sidgwick and Powell have suggested that it mayinvolve a 4-covalent duodecet 4, 8, as we have in the planar [ICl,]-i0n.t If this is so, the electronic configuration should be written,not as (n), 8, but as (n - 4), 4,8, and since the penultimate group can-not be reduced below 8, the planar structure would only becomepossible when n is a t least 12, which is roughly where it first appears.111.With a valency group of 10 electrons, whether they are fullyshared or not, the arrangement would always appear to be thatderived from a trigonal bipyramid. Thus when only 4 are sharedas in the ions [IClJ-, [I3]-, etc., which are known to be linear,8 theconfiguration can be considered that of a trigonal bipyramid withhalogen atoms a t the two pyramidal apices and the three unsharedelectron pairs in the equatorial plane. The structure of the covalentiodochlorides ArICl,, where six of the ten electrons are shared, hasnot been examined, but from the foregoing one would expect a planararrangement with the two unshared electron pairs now a t thepyramidal apices.Two interesting structures involving a decet with eight sharedelectrons have recently been determined, the crystal structure ofpotassium fluoroiodate, KI02F2, being investigated by means ofX-rays and the structure of tellurium tetrachloride in the vapourphase by the method of electron diffraction.10 In the former the[IO,F,]- ion is composed of an iodine atom forming bonds a t approxi-mately 100" with two oxygen atoms, and, perpendicular to the planeof these three atoms, two opposed bonds at 180" to the two fluorineatoms ; the third position in the 10, equatorial plane of the trigonalbipyramid is presumably occupied by the unshared pair.A similar(somewhat distorted) trigonal bipyramidal configuration has beenH.J. Dothie, F. 5. Llewellyn, W. Wardlaw, and A. J. E. Welch, J., 1939,426.R. W. G. Wyckoff, J . Amer. Chem. SOC., 1920, 42, 1100; R. C. L. Mooney,2. Krist., 1935, 90, 143; 1938, 98, 324; 1939, 100, 519.51 L. Helmholz and M. T. Rogers, J . Amer. Chem. SOC., 1940, 62, 1537.lo D. P. Stevenson and V. Schomaker, ibid., p. 1267.* A possible exception is K[Au(CN),dipy] which appears to be planar 'although the gold atom has the structure (18) 8.It is noteworthy that atoms such as Pd and Pt which can exhibit bothsquare and octahedral co-ordination have the same covalent radii in boththese forms182 CRYSTALLOGRAPHY.proposed for the tellurium halide as being in best accord with theelectron-diffraction measurements, the unshared pair again occupy-ing one of the equatorial positions.* The configuration is presum-ably that due to (incomplete) sp3d hybridisation, and not to p3d withthe unshared pair occupying a pure s orbital as was assumed byKimbal1,l since this would lead to an irregular tetrahedral con-figuration of symmetry C3v.There are, however, a number of cases known where there isapparently a valency group of ten of which eight are shared, whichdo not conform to this stereochemical type derived from a trigonalbipyramid. These include the 4- covalent complexes of univalentthallium, such as [T1(SC(NH,)2}4]N03, and of bivalent lead, such asK2[Pb(C204),].The structures of these compounds have not beeninvestigated in detail but enough has been done to establish theirplanar character.12 It is by no means certain, however, that thesecases involve a decet a t all.If we consider the complex plumbousoxalate, the usual way of writing this ion is (I), giving the metal theelectron configuration (78), 2,8 and a formal charge of -2, which issurprising when we remember that lead is an electropositive element. p=o~pb,o-""]-- po,G,/ \o-c -o,c=ofl - '0-c \O &C-0 \O(1.) (11.)An alternative structure would be (11), in which the lead has theelectronic configuration (76), 8. The inner 76 electrons can then beplaced in the orbitals 1s2s2p33s3p33d54s4p34d54f '5s5p35d4 leavingthe fifth 5d orbital as well as the 6s and 6 p orbitals available for bondformation. The 8 shared electrons would then occupy these orbitalsto give square 5d6s6p2 bonds.A number of cases have been examined of compounds containinga, fully shared decet (e.g., phosphorus pentafluoride) and they haveall been shown to be of the trigonal bipyramid type; further recentexamples include niobium and tantalum chlorides and bromides.l3It is noteworthy that this is the arrangement which has been foundin all 5-covalent compounds, even with an incomplete penultimategroup as in MoC1, and Fe(CO),. It is unfortunate that, owing to* This is also in agreement with the large dipole moment of 2-54 D. foundl1 C. P. Smyth, A. J. Grossman, and S. R. Ginsburg, J . Amer. Chem. Xoc.,la E. G. Cox, A. J. Shorter, and W. Wardlaw, Nature, 1937, 139, 71; J.,l3 H. A. Skinner and L. E. Sutton, Trans. Paraday Soc., 1940, 36, 668.for TeC1, in benzene solution.1l1940, 62, 192.1938, 1886HAMPSON : INORGANIC STRUCTURES.183the large difference between the atomic numbers of the atoms, thespatial configuration of iodine pentafluoride has not been deter-mined. Here the iodine has 12 valency electrons of which 10 areshared, and, by analogy with [IF,]-, one might expect a squarepyramid structure (Le., an octahedron with one of the cornersoccupied by the unshared pair).IV. With a covalency of six we should expect three possiblearrangements : (a) a trigonal prism, ( b ) a trigonal antiprism, (c) aregular octahedron ; ( c ) is simply a special case of ( b ) in which allthe edges and angles are equal. Octahedral bonds are formed bythe configuration d2sp3 and no other,14 and hence Kimball arguesthat the commonness of this arrangement is due to the fact that theconfiguration d2sp3 is the usual configuration of six valency electronsrather than to any particular virtue of the octahedral arrangement.Sidgwick and Powell, on the other hand, consider that the octa-hedron is to be preferred for geometrical reasons, since it is in thisarrangement that the strain due to repulsion between non-bondedatoms is a, minimum. Certainly, octahedral bonds have beenobserved in practically every type of complex AB,, it being only ingiant molecules of the molybdenum disulphide and nickel arsenidetype, where the complexes are subject to strong valency forces,that the trigonal prismatic form 0 c ~ u r s .l ~It should be pointed out, however, that the octahedral d2sp3configuration can arise in more than one way.With the large classof 6-co-ordinated ions of the transition elements, it is the d orbitalsof the penultimate group, having about the same energy as the sand p orbitals of the valency group, which take part in bond form-ation ; e.g., with [Fe( CN),]-* the hybridisation is 3d2494p3, with[PtC1,]-2 it is 5d26s6p3. I n those cases, however, either where thereare no d orbitals in the penultimate group (e.g., sulphur hexafluoride),or where these orbitals are fully occupied by unshared electron pairs,e.g., [SnC1,]-2, use must be made of the unstable d orbitals of thevalency shell itself, the hybridisation in the case of the hexafluoridebeing 3s3p33d2, and in the case of the chlorostannate ion 5s5p35d2.Still another case is where we have the phenomenon of the " inertpair," as in [SeBr,]-2.Here the fourteen electrons (12 shared +the inert pair) might occupy the orbitals 48, three 4p, and three 4d,and if, as might be supposed, the unshared pair occupied the orbitalof greatest coulornbic stability, vix., the 48, the configuration of thebonds would be 4p34d3, giving an antiprismatic arrangement. Thearrangement, however, is observed to be octahedral, and henceKimball suggests that the unshared pair occupies a 4d rather than14 G. E. Kimball, Zoc. cit., ref. (1).l5 R. Hultgren, Physical Rev., 1932, 40, 891184 CRYSTALLOGRAPHY.the 4s orbital, leaving the bond configuration 4s4p34d2. A morelikely explanation is that the inert pair does occupy the stable 4sorbital and that the 5s orbital is utilised in bond formation, theconfiguration being 4p34d25s; this view is in agreement with thelarge octahedral radius of selenium observed in these complexes.V.A covalency of 7 is extremely rare, but two different typeshave been observed; l7 the [ZrF7]-3 structure which may be derivedfrom an octahedron by adding a fluorine atom a t the centre of oneface, and the [NbF7]-2 or [TaF7]-2 structure which may be obtainedfrom a trigonal prism by adding a fluorine atom a t the centre of aprism face. The central atoms in these complexes are isoelectronic,and there is no obvious explanation for this surprising difference instereochemical arrangement.VI. Only one 8-covalent complex has been examined, namely,[Mo(CN),]-*, and it has been shown that the arrangement is thatof a dodecahedron.The configuration therefore, according toKimball, is d4sp3, the dodecahedron being more stable than eitherthe cubic or the antiprismatic structure. For the ion [TaF,]-3and the molecule OsF, with the configuration d5sp2, Kimball predictsa face-centred prismatic structure.Recent Xtructures.Metallic holmium has a hexagonal closest-packed structure, withsix nearest neighbours at a distance of 3-48 A. and six a t 3.557 ~ . 1 9Of the oxides, the high-temperature form of potassium hydroxidehas a rock-salt structure, the radius of the OH- ion being 1.53 A.**The sesquioxides of rubidium and cEsium, Rb,O, and Cs203, do notcontain the ion 0,- -, but should be formulated as Rb202,2Rb02, themagnetic susceptibilities being consistent with a proportion of oneper-oxide ion : 0-0 : to two superoxide ions : 0-=O :, the latter havinga three-electron bond.2l A study of the uranium-sulphur systemhas revealed the existence of a subsulphide U4S3 having a sodiumchloride structure in which one-quarter of the anion positions areleft vacant ; 22 and the subphosphides of iridium and rhodium, Ir2Pand Rh2P, are of the anti-calcium fluoride type with the distancesIr-P=2.40 A.and Rh-P = 2-38 A. An interesting structure whichhas been worked out completely by Patterson and Fourier analysis.. .. -*. .. .. * .18 J. Y . Beach, quoted by L. Pauling, op. cit., p. 171.1 7 See Ann. Reports, 1939, 36, 168, 170.18 J.L. Hoard and H. H. Nordsieck, J . Amer. Chem. Soc., 1939, 61, 2853.l9 H. Bommer, 2. amorg. Chem., 1939, 242, 277.2" W. Teichert apd W. Klemm, ibid., 243, 138.2 1 A. Helms and W. Klemm, ibid., 242, 201 ; L. Pauling, op. cit., p. 252.22 E. F. Stroteer, 0. Schneider, and W. Biltz, 2. anorg. Chem., 1940, 243,307 ; M. Zumbusch, ibid., p. 322HAMPSON : INORGANIC STRUCTURES. 185is that of tetraphosphonitrile chloride, P4N4C1,.23 The moleculeconsists of a puckered %membered ring of alternate nitrogen andphosphorus atoms, each phosphorus carrying two chlorine atoms andhaving approximately a tetrahedral arrangement of its four linkages.The angles within the ring are roughly 122" for nitrogen and 117"for phosphorus and the P-N distances are all approximately 1.67 A.Since the distance calculated for a single bond is 1-80 A.and for adouble bond 1.61 A., the eight-membered ring appears to possess an" aromatic " structure of resonating single and double bonds.Although phosphorus pentachloride has been shown by electrondiffraction to have a trigonal bipyramidal configuration in thevapour phase, the crystal structure of this substance and also thatof the bromide are entirely different. Crystalline phosphorus penta-chloride is ionic, being composed of tetrahedral [PCl,] + ions and octa-hedral [PC16]- ions,24 whilst crystalline phosphorus pentabromide isbuilt up of tetrahedral [PBr4]+ ions and Br- ions, the latter beingtwice as far removed from the phosphorus as the four covalentlylinked af0ms.2~ The difference between the two crystalline struc-tures may be due to the instability of an octahedral [PBr,]- complex,arising from the larger size of the bromine atoms.I n a metastablemodification of anhydrous sodium sulphate, Na,SO,-111, an S-0distance of 1.52 & 0.03 A. in the SO4-- ion has been reported z6;on the other hand, a careful examination of potassium sulphamate,NH2*S03K, has given the values S-0 = 1.44 0 . 0 3 ~ . and S-N =1.57 It 0.03 A. for the tetrahedral NH2*S0,- ion? The explanationfor the shorter S-0 distance may be that the sulphamate ion hasonly a single negative charge whereas the sulphate ion is doublycharged. These distances are even shorter than those to be expectedfor double bonds.I n sodium formate there is complete resonance between the twoC-0 bonds of the formate ion, the bond angle of 124" and C-0 dis-tance of 1-27 A.agreeing with those (125" and 1.29 A.) found in formicacid dimer by electron diffraction.28 A redetermination of theparameters in NH4HF2 29 has given an F----H----F distance of 2.32 -J--0.03 A. This is greater than that (2.26 0.01 A.) found in KHF2,3023 J. A. A. Ketelaar and T. A. de Vries, Rec. Trav. china., 1939, 58, 1081.2* H. M. Powell, D. Clark, and A. F. Wells, Nature, 1940, 145, 149.s5 H. M. Powell and D. Clark, ibid., p. 971.26 L. K. Frevel, J. Chem. Physics, 1940, 8, 290.27 C. J, Brown and E. G. Cox, J., 1940, 1.2* W. H. Zachariasen, J . Amer. Chem. SOC., 1940, 62, 1011 ; L. Pauling and29 M.T. Rogers and L. Helmholz, J . Amer. Chem. SOC., 1940, 62, 1533.30 L. Helmholz and M. T. Rogers, ibid., 1939, 61, 2590; Ann. Reports, 1939,L. 0. Brockway, Proc. Nut. Acad. Sci., 1934, 20, 340.36, 165186 CRYSTALLOGRAPHY.showing that in the ammonium compound the bonds are weakenedby the formation of the two additional hydrogen bondsA series of hexafluogermanates, M2GeF6, which have beenexamined 31 show an interesting change of structure according asM = NH,, K, Rb, or Cs. As is to be expected, the Gel?,-- ion hasan octahedral configuration with Ge-F = 1.77 A., but the packingof the GeF,-- ions and the alkali ions changes through the series.The ammonium and the potassium salt are isomorphous, thestructures corresponding very approximately to hexagonal closestpacking of alkali and fluorine atoms, as in the low-temperaturemodification of ammonium silicofluoride ; 32 each cation is sur-rounded by nine nearly equidistant fluorine atoms and three othersat a somewhat greater distance.With the rubidium salt the struc-ture is still hexagonal, but here the Rb+ ions appear to have twelveequidistant fluorine neighbours. The caesium salt, on the otherhand,33 has the well-known cubic structure of the ammoniumchloroplatinate type, the co-ordination number of the cations againbeing twelve. This same structure has also been observed forcaesium hexachlorogermanate CS,G~C~,.~~In an extensive investigation of the complexes formed betweentertiary phosphines and arsines and the halides of cadmium andmerc~ry,3~ five different classes of compound have been isolated.These can be represented as (a).(R3P)&tX2, (b) (R3P)2(MX2)2,halides forming complexes of all five types whilst cadmium halidesform only complexes of types (a), (b), and (e). In most cases only apreliminary structure determination has been carried out, but it isconsidered that complexes of the type (a) probably consist of discretemolecules having a tetrahedral arrangement [cf., on the other hand,the structures of (NH,),CdX2 and (NH3)2HgX2].36 Complexes ofthe type (b) are centrosymmetric andf hX/ \x ture (111) similar to the 4-covalent dipalla-dium 37 and diauric 38 complexes, exceptthat in the cases of the cadmium andH N-H ____ F ____ H ____ F ____ H-NH3 3'(c) (R3P)2(MX2)3, (') (R3P)2(MX2)4, and (e) (R3P)3(MX2)2, mercuricx x \M/ kMgPR3 presumably have the trans-bridged struc-R3P(111.)31 J.L. Hoard and W. B. Vincent, J . Amer. Chem. SOC., 1939, 61, 2849.32 B. Gossner and 0. Kraus, 2. Krist., 1934, 88, 223.33 R. W. G. Wyckoff and J. H. Miiller, Amer. J . Sci., 1927, 13, 347.34 A. W. Laubengayer, 0. B. Billings, and A. E. Newkirk, J . Amer. Chem.35 R. C. Evans, F. G. Mann, H. S . Peiser, and D. Purdie, J., 1940, 1209.38 C. H. MacGillavry and J. M. Bijvoet, 2. Krist., 1936, 94, 231, 249.37 F. G. Mann and co-workers, J., 1938, 702, 1949, 2086; 1939, 1622.3 8 A. Burawoy, C. S. Gibson, G. C. Hampson, and H. &I. Powell, J., 1937,1690.SOC., 1940, 62, 546HAMPSON : INORGANIC STRUCTURES. 187mercuric complexes, X-ray evidence indicates a tetrahedral ratherthan a planar configuration.With complexes of the type ( c ) (givenonly by mercury) there appear to be two distinct possibilities.(Bu,As),(HgBr,),, whose structure has been worked out in somedetail, is essentially a molecular compound[the same as type (b)] and HgBr,, as shown inI ',3*2I3.1 II \I \IHg As B rdistances and angles are given. The distortion in the bridged partof the complex is probably due to attractions between mercury andbromine atoms shown joined by the broken lines, so that thesemercury atoms become surrounded by a distorted octahedron of sixbromine atoms, two being directly bound a t 2.25 A., and the remain-ing four contributed by neighbouring bridged molecules at thegreater distances of 3.1 and 3.2 A.A similar distorted octahedronhas been observed in NH,HgCl, 39 and in HgBr,.*O (Et,As),(Hgl,),s9 H. Harmsen, 2. Km'st., 1939, 100, 208.40 H. J. Verweel and J. M. Bijvoet, ibid., 1931, 77, 122188 CRYSTALLOGRAPHY.is quite different, and it is possible that this complex has the bridgedstructure (IV).The structures of complexes of the type (d) and ( e ) are alsouncertain, but a space-group determination in the case of(Bu3P),(CdBr,), shows that it cannot be a molecular lattice ofZ(Bu,P),CdBr, and (Bu,P),(CdBr,), and the structure (V) has beensuggested. Here one of the cadmium atoms is 4-covalent a t thecentre of a tetrahedron, and the other 6-covalent at the centre ofan octahedron, the two polyhedra having a triangular face in common.G.C. H.5. ORGANIC STRUCTURES.Phthulocyanines and Reluted Structures.-X-Ray studies have beencarried out on a number of new compounds in this series duringthe year. Very detailed analyses have already been described forphthalocyanine itself and its nickel derivativeY2 and cell dimensionsand space group determinations are available for the beryllium,manganese, iron, cobalt, and copper compounds.3 The figures showthat all these crystals are very closely isomorphous, and hence themain conclusions regarding molecular structure and packing whichhave been deduced for the parent substance must apply to all thesederivatives .The new structures described are those of platinum phthalo-cyanine 4 (complete structure determination, see below), tetra-benztriazaporphin (I),5 tetrabenzmonazaporphin (11) 6 and tetra-benzporphin (111) (cell dimensions and space groups).The lastthree compounds are of interest because they represent a gradualapproach from the phthalocyanine to the natural porphyrin struc-tures, the linking nitrogen atoms of the great ring being replacedby one, three, and finally four methin groups. The synthesis ofthese compounds has been carried out, and their homogeneityestablished by the work of R. P. Linstead and others.5* 71 J. M. Robertson, J., 1936, 1195.2 J. M. Robertson and (Miss) I. Woodward, J . , 1937, 219.3 R. P. Linstead and J. M. Robertson, J., 1936, 1736.4 J. M. Robertson and (Miss) I. Woodward, J., 1940, 36.5 P.A. Barrett, R. P. Linstead, G. A. P. Tuey, and J. M. Robertson, J . ,6 (Miss) I. Woodward, J., 1940, 601.7 P. A. Barrett, R. P. Linstead, F. G. Rundall, and G. A. P. Tuey, J.,1935, 1809.1940, 1079; see dso Ann. Reporb, 1937, 34, 309ROBERTSON : ORGANIC STRUCTURES. 189The crystal structures of these new macrocyclic pigments are ofspecial interest, because only in the case of (I) is the phthalocyanineCCtype of structure maintained. The further replacement of the nitro-gen atoms in (11) and (111) has apparently sufficient effect on themolecular packing completely to alter the arrangement-a surprisingconclusion in view of the great stability of the normal phthalo-cyanine structure. There is little likelihood of any appreciablechange in the structures of the molecular frameworks, the chemicalevidence being fairly conclusive in this connection. The same ap-proximate molecular dimensions being therefore assumed, the celldimensions indicate that in (11) and (111) the molecular planes mustbe inclined at about 60" to the (010) crystal plane, as against 44" inphthalocyanine itself.The intensity data indicate that certainother changes in orientation probably occur as well. The molecules(I) and (11), like the symmetrical compounds, are found to display acentre of symmetry in the crystal. However, this is no doubt merelya statistical centre of symmetry, due to a random distribution of themolecules with respect to the unique group^.^The detailed structure analysis of platinum phthalocyanine 4is of interest in several connections.Here again the structure isfound to differ from the normal phthalocyanine type, but in theopposite sense to that of (11) and (111). The molecules are now lesssteeply inclined to the (010) plane, the angle being only 26.5" againstthe normal one of 44". All these different structures are probablygoverned to some extent by the amount of clearance between neigh190 CRYSTALLOGRAPHY.bouring atoms in overlying molecules, and the three distinct struc-tures probably represent three distinct maxima in this connection.As the molecular structures themselves are so closely similar, it ispossible that these variations are all potential polymorphic modifica-tions of one fundamental structure, although no actual transition forany one compound has yet been observed.,_-.. .Platinum phthalocyanine.projection plane.where the increment is 20 electrons per line.The planar molecule is inclined at 26.5’ to theContour lines at one electron per A.a except at the centre,The one-electron line is dotted.The results of the Fourier analysis of the platinum phthalocyaninestructure are shown in the contour map. The dimensions of themolecule are found to be very closely similar to those alreadydescribed for the other phthalocyanines, except that the nitrogen-metal distances have now increased to about 2.01 A. as comparedwith 1.92 A. and 1.83 A. in the metal-free and the nickel compounROBERTSON : ORGANIC STRUCTURES. 191respectively. This is no doubt due to the increased radius of theplatinum atom.The chief interest of this determination, however, lies in the methodof analysis employed.It depends upon the presence of an extremelyheavy atom in the molecule, whose contribution to the structurefactor for every plane is greater than that of all the other atomscombined. Briefly, unknown differences in phase are in this wayconverted into differences in amplitude which can be measured.The resulting Fourier synthesis thus becomes equivalent to the well-known Patterson synthesis, but the result is a direct picture of thestructure instead of its vector representation.It is not necessary that the atomic number of the heavy atomshould exceed the sum of the atomic numbers of all the other atoms,as might a t first be supposed, for the swamping effect on all thestructure factors to be complete.The contributions from theseother atoms tend to cancel each other out, and only a compar-atively small residual factor is finally left to be balanced by theheavy-atom contribution. It thus becomes possible to foresee theapplication of the method to very large molecules indeed, suchas the natural porphyrins, or even to some of the natural proteinstructures.*The method is thus of wide application, and the paper underdiscussion is a critical test of it and a study of the difficulties whicharise in its application. These are chiefly concerned with thenecessity for accurate intensity measurement, and with false detailwhich arises owing to non-convergence of the P series; but withsuitable precautions the resolution of even the light carbon atoms isfairly good, as the figure shows, and the maximum uncertainty intheir positions is about & 0.1 A.Even this figure can be consider-ably improved in most parts of the structure when calculated allow-ances are made for false diffraction effects, etc.Another interesting point about this work is the amount of materialused in the analysis. It consisted of a single crystal weighing0.0095 mg.Acetamide.-A very full investigation has been carried out byF. Senti and D. Harker of the crystal structure of the rhombohedra1form of acetamide. It is a complicated structure in the sense thatthere are 18 molecules in the unit cell (space group C!,-B3c), and itscomplete solution with a determination of all the atomic positionsrepresents a, considerable achievement.It is a pity, however, thatin such a thorough analysis the intensities cannot be determinedmore quantitatively. In the present instance they are determinedJ. M. Robertson, Nature, 1939, 143, 75.J . Arne?. Chem. Sw., 1940, 62, 2008192 CRYSTALLOGRAPHY.by the oscillation photographic method and visual estimation fromselected standards.The analysis is carried through by a, careful application of the well-known Patterson and Harker vector methods, and it includes aninteresting study of the refinement of parameters by Fourier seriesmethods in a structure which does not contain centres of symmetry,and in which the phase constants may, therefore, have any fractionalvalues.(Another such structure which has been analysed quantita-tively is resorcinol.lO) It is found that convergence to the truestructure is, nevertheless, quite rapid.The acetamide molecules are found to be planar, as would berequired from either of the resonance structures (IV) or (V). Theaccuracy claimed is considerable, and it is stated that the deviationof any atom from the median plane through the molecule must beless than 0-01 A.The interatomic distances are those that might be expected for thestructure : C-0 = 1.28 A., where the resonance is of the same type asin the carboxylic acids, C-CH, = 1.51 A., and C-NH, = 1.38 A. Thepossible uncertainty in all these values is given as about & 0.05 A.The resonance effect has certainly not equalised the valency angles,the CH,-CNH, angle being 109" and the other two 122-129".In the crystal the planar molecules of acetamide are arranged inrings of six, adjacent molecules in the ring being held together byN-H * * * * 0 bridges whose mean length is about 2-86 A. This re-latively close binding, of course, accounts for the high meltingpoint and boiling point of the substance, as compared with acetone,for example. From the directions of these hydrogen bridges it canbe concluded that the N-H bonds lie in the plane of the molecule.Other molecular contacts are of the residual or van der Waals kind.Dicyandiamide.-A very interesting and complete structuredetermination has now been carried out by E. W. Hughes on thedimer of cyanamide.11 The work clearly eliminates the possibilityof cyclic structures, and it is shown that the positions of the atomsNEC-N -N=C=N -N=C=N 3 1 45 Y-&Hz H2N>=NHl kHa>-NH2(VII.) (VIII.)H2N(VI.)correspond to a molecule which is a resonance hybrid, chiefly betweenthe structures (VI), (VII), and (VIII). Thus in the guanidine group10 J. M. Robertson, Proc. Roy. SOC., 1936, 157, A, 79.11 J. Amer. Chern. SOC., 1940, 62, 1258ROBERTSON : ORGANIC STRUCTURES. 193the double bond resonates to a certain extent into all possiblepositions.The planarity of the guanidine group and the linearity of thecyanimino-group have been established within narrow limits, theprobable error in atomic positions in this analysis being given as0.015 A., and the maximum possible error in bond lengthsestimated at about -& 0.04 A. The cyanimino-group appears todeviate only slightly from the plane of the guanidine group, themovement corresponding to a rotation about the =N-C--” bond of 2”.The interatomic distances are as follows, the numbers referringto the atoms as shown in (VI) : 1-3 = 1.22 A., 1 4 = 1-28 A.,2 4 = 1-36 A., 2-5 = 1.37 A., 2-6 = 1.34 A.The packing of the molecules is found to be dominated by hydro-gen-bond formation, one of these bonds being of the rare bifurcatedtype previously noted in the case of glycine.12 The five distincthydrogen bonds in the present structure have lengths varying from2-94 to 3.15 A., and are thus of a rather weak type, but they representsomething definitely stronger than the residual van der Waalsattractions, and they undoubtedly govern the orientations of themolecules relative to each other.Some of these bonds are directed towards the terminal nitrogenof the cyanimino-group. Now the structure (IX), with only one..“=C=N, (X.)unshared electron pair on the terminal nitrogen, could only give riseto very weak hydrogen bonds, as in the case of ammonia, where asimilar situation prevails ; but the alternative structure with twounshared pairs (X) would permit much stronger hydrogen bonds tobe formed, and it is thought that this structure consequently makesa larger contribution to the normal state, the final result being thatthe loss in resonance energy is more than made up by the gain inenergy of the hydrogen bonds. J. M. R.G. C. HAMPSON.J. M. ROBERTSON.J. c. SPEAKMAN.A. R. UBBELOHDE.la Arm. Reports, 1939, 36, 181.REP.-VOL. XXXVII.

ISSN:0365-6217    

DOI:10.1039/AR9403700165

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

ORGANIC CHEMISTRY.1. INTRODUCTION.THE present trend of aliphatic chemistry is being greatly influencedby developments in the petroleum industry. Branched-chainparaflfins are now being synthesised and studied with the object ofidentiffing the constituents of fuels and also of discovering com-pounds of high '' anti-knock value." The great industrial im-portance of polymerisation (e.g., of ethylene to isobutene) and of" alkanation " (e.g., addition of isobutane to isobutene in presenceof sulphuric acid) has initiated many researches. Advances in thetechnique of fractional distillation have made pure paraffins readilyaccessible, and these are a rich source of aliphatic compounds.Olefma are even more useful; propylene hydrates to isopropylalcohol, which on catalytic oxidation gives acetone ; a t 400-600"halogens substitute, propylene thus giving ally1 bromide or chloride,and by this route glycerol is synthesised.Higher paraffi fractions(10 to 16 carbon atoms) are a, source of detergents. The sulphonicacids may be obtained by direct sulphonation; halogenation andhydrolysis lead to alcohols (mainly secondary), which are convertedinto sulphuric esters ; both sulphonic acids and sulphuric esters havereceived considerable attention.Reference is also made in the aliphatic section of this Report torecent progress in the isolation and study of individual fatty acidsfrom natural fats. Many facts bearing on the problem of the bio-synthesis of fatty acids have emerged, and perhaps the most import-ant development is the synthesis of stearic acid from acetaldehyde(via crotonaldehyde). One of the plant wound-hormones,l trau-matic acid, is a decenedicarboxylic acid, and has been synthesised.The preparation of ctcc-dialkyl long-chain acids whose surface filmsresemble those of phthioic acid (from tubercle bacilli) is reported.Pantothenic acid,2 the " chick anti-dermatitis factor " present inliver, has been degraded to p-alanine and the lactone of aydihydroxy-pp-dimethylbutyric acid ; the vitamin has been synthesised.The evidence relating to the two mechanisms (unimolecular andbimolecular) of aliphatic substitution, up to the end of 1938, wassummarised in the Report for that year. The correctness of theview there outlined has now been confirmed.In cases where onereagent is the solvent (e.g., hydrolysis in an aqueous medium), theproof of the unimolecular mechanism depended upon observations1 See Ann. Reports, 1939, 36,368. Ibid., p. 344INTRODUCTION. 195of the effects of changes in the structure of the compound in whichsubstitution occurs and in the nature of the substituting agent andof the solvent ; in certain instances the influence of solvent changesupon the product, and the stereochemical course of the substitution,also provided a guide to the reaction mechanism. This evidencehas now been supplemented by a direct kinetic demonstration of theunimolecular character of a number of solvolytic substitutions. Amore detailed consideration of the unimolecular mechanism indi-cated that certain deviations from accurately first-order kineticsshould appear when the mechanism is unimolecular ; these havebeen observed, and the effects of added salts upon the reaction rateand upon the products are in harmony with the requirements of theunimolecular mechanism.In the course of this work, evidence of theconjugation of alkyl groups with the aromatic nucleus has beenfound.The mechanisms of esterification and ester hydrolysis constitutean old problem. It has now been shown, by four methods, thathydroxyl from the acid, or alkoxyl from the ester, appears in the wateror alcohol formed ; i.e., fission occurs in accordance with the schemesOther mechanistic details have also been studied ; several kineticinvestigations of the esterification of the saturated aliphatic acids andof the hydrolysis of their esters have been recorded and the resultsinterpreted.Condensations of carbonyl compounds were considered,in last year’s Report, and have therefore not been dealt with again.It is interesting to record, however, that several exaqples of thesecondensations in presence of boron fluoride or aluminium chloridehave been found.3 Their mechanisms have also been further dis-cussed; 4 the mechanism of the reaction of benzaldehyde withacetophenone has been considered in the light of new kinetic measure-ments which point to the conclusion that substituents in thealdehyde molecule influence mainly the energy of activation, whilegroups in the acetophenone change the P term of the kinetic equationk = PZe-ElRT.Although the triphenylmethyl radical was first prepared fortyyears ago, many points in the chemistry of such free radicals haveremained obscure, and notable advances have recently been made inthe elucidation of the reactions of alkyl-substituted triphenyl-D.S. Breslow and C. R. Hauser, J. Amer. Chem. Soc., 1940, 62, 2385,2389, 2611.* C. R. Hauser et al., ibid., pp. 62, 593, 1763.(Miss) E. Coombs and D. P. Evans, J., 1940, 1295196 ORGANIC CHEMISTRY.methyls which show unexpected instability. It has been found thatp-dkyltriphenylmethyls readily undergo disproportionation to givethe corresponding methane (I) and a quinonoid hydrocarbon (11) :2( CH2R*C6H4)3C- --+ ( CH2R*C6H4),CH +(1.)CHR=/=‘\ - C( C,H,CH,R), \=/-(11.)The p-tert.- butyl radical, where no such disproportionation ispossible, is stable. Halogenated triphenylmethyls have presentedspecial difficulties owing to the reactivity of the nuclear halogen withthe metal used in the preparation of the radical, which gives rise toa secondary radical such as (111). Contrary to expectation, thenuclear halogen in p-fluorotriphenylmethyl is more reactive thanthat in triphenylfluoromethane, whereas the reverse is the case forthe corresponding bromo-compound. Results of considerableinterest have also been obtained by the study of the penta-aryl-ethanes, which can dissociate, giving a mixture of triarylmethyl anddiarylmethyl radicals. Work upon the magnetic susceptibilities offree radicals has shown that, when two tervalent carbon atoms areattached to the 4 : 4’-positions in a diphenyl system as in (IV), thecompound is diamagnetic and exists in the quinonoid form (V), butthe corresponding 3 : 3’-disubstituted derivative (VI) is paramag-netic like the triarylmethyl radicals; if the radical is non-planarh 2 C -\/ hAI I1\/Hal(111.)I I L\/ y2owing to the presence of ortho-substituents, paramagnetic propertiesare shown even when the tervalent carbon atoms are attached t o the4 : 4’-positions.Gaseous thermal and photochemical reactions which are knownto involve the primary formation of free radicals have been thesubject of numerous studies ; observations of the gaseous nitrationof aliphatic hydrocarbons are of both theoretical and practicalinterest, and are in good agreement with the free-radical view.The presence of free alkyl radicals in the liquid phase was demonINTRODUCTION.197strated some time ango in the photolysis of aldehydes and ketones inhydrocarbon solvents, and aryl radicals are formed in the decom-positions of diazobenzene chloride, diazobenzene hydroxide, syn-diazobenzene cyanide, nitrosoacetanilide, benzeneazotriphenyl-methane, dibenzoyl peroxide, and their derivatives.6 The alkyl andaryl radicals are found to display many features in common, notablyin their reactions with hydrocarbon solvents ; examples have beenrecorded of attack by radicals on metals, and evidence has beenobtained of the addition of radicals to double bonds. There is alsoevidence of the participation of free radicals in certain Claisenrearrangements and in the Wurtz and Fittig reactions, and it is likelythat certain reactions of Grignard compounds also involve freeradicals. Atomic chlorination by means of the peroxide-catalysedreaction with sulphuryl chloride provides examples of abnormalaliphatic substitution comparable with the abnormal aromaticsubstitution reactions previously attributed to aryl radicals.One of the most interesting developments in carotenoid chemistryduring the past five years has been the discovery and examination ofthe spontaneous isomerisation which occurs in solutions of all caro-tenoids (especially at elevated temperatures or in the presence ofiodine), a phenomenon which has important repercussions on boththe identification and the technique of working with these pigments ;although conclusive evidence is not yet available, the phenomenonwould appear to be due to geometrical isomerism.By oxidationwith alkaline permanganate under carefully controlled conditionsthe stepwise degradation of several carotenoids has been achieved,the aldehydes obtained providing valuable spectroscopic data andalso information concerning the relationship between constitutionand provitamin A activity. Considerable interest has been arousedby the discovery that the polyene acid crocetin plays a fundamentalpart in the copulation of the male and the female gametes of Chlamy-domonas eugametos. A feature of the work on the identification ofpigments from natural sources has been a tendency to undertake asystematic study of particular groups, with the result that certainregularities in pigmentation have become apparent ; in particular,much attention has been devoted to such studies of the algB and ofthe leaf xanthophylls.Recent researches have shown that astaxan-thin, and not astacene, is the characteristic pigment of the Crus-taceae, the latter pigment being formed from the labile astaxanthinby oxidation during hydrolysis.Several authors have discussed the precise mode of linking of theimino-hydrogen atoms in the free porphyrins, and of the metallicatoms in their “ salts.” The catalytic properties of the latter haveAnn. Reports, 3937, 34, 282198 ORGANIC CHEMISTRY.been compared with those of the oxidases and catalase.-4 series oftetrabenzazaporphins-transitional between the phthalocyanine andporphyrin types-has been thoroughly investigated, and thesynthesis placed on a more rational basis by a study of isoindoles asintermediate stages. H. Fischer has proposed for chlorophyll amodified formula with the ‘‘ extra hydrogens ” in ring IVY based onthe results of oxidative degradation ; a corresponding structure isassigned to bacteriochlorophyll, now recognised as a tetrahydro-porphin. Substantial progress has been made in the syntheticapproach to chlorophyll. There is now some evidence that naturalhaemin is not homogeneous, but may contain 25% of hzmin 11,which has the same orientation as copro- and uro-porphyrins I.Since the bile pigments were last reviewed in 1933, notableadvances have been made in methods of synthesis, especially ofunsymmetrically constituted materials. The degradation of por-phyrins to bilirubinoids has been effected by several purely chemicalas well as biological methods ; hydroxylated dipyrrylmethenes havebeen recognised as further stages of the breakdown, and as respons-ible for the “ pentdyopent ” reaction.A reduced bilirubinoidstructure is assigned to stercobilin.The most spectacular achievement in steroid chemistry during theperiod under review is the total synthesis of the naturally occurringsex hormone d-equilenin and its three stereoisomerides. Greatprogress has been made in the development of methods for theconversion of androstane into pregnane derivatives ; new partialsynthesies of progesterone and deoxycorticosterone from dehydro-androsterone have been reported.One of the more promisingmethods for the preparation of 20-ketopregnane derivatives (e.g.,progesterone) appeared to be the hydration of the readily available17-hydroxy- 17-ethinylandrostane intermediates (I) by the mercuricoxide-boron fluoride method. Although hydration occurs, it hasbeen established that the reaction is accompanied by ring-D enlarge-ment with formation of the 17-keto-D-homoandrostane derivativeImportant advances have been made in the chemistry of thesteroid sapogenins, the side chain of which has been shown t ocontain a protected (ketal) carbonyl group (111). The ready iso-merisation of the sapogenins to +-sapogenins by means of acetiINTRODOUTION.199anhydride, and the facile oxidation of +-sapogenins to 2O-keto-AI6-pregnene intermediates (IV), have opened up a remarkably simpleroute to the pregnane (or allopregnane) series. The structures ofseveral of the less abundant sterols such as brassicasterol, a-spina-sterol and zymosterol have been determined.Two independent sets of phenomena have been correlated with therelative basic or acidic characters of heterocylic rings. Benzopyransin spiro-condensation with heterocyclic nuclei can undergo reversibleintramolecular ionisation with very different degrees of facility,yielding phenol-betaines in which the pyran ring is ruptured and thesecond nucleus now contains an 'onium atom.The intense colour ofcyanine dyes is attributed to resonance between structures in whichthe positive ionic charge is associated with different nitrogenatoms. In the bases of which the dyes are quaternary salts, onecanonical structure contains negatively-charged bicovalent nitrogen,and here the degree of resonance and observed depth of colour areusually, but with significant exceptions, diminished relatively tothose of the salts. The dimensions of several heterocyclic nuclei havebeen determined by electron-diff raction measurements ; indicationsof the participation of states in which a sulphur atom is associatedwith an electron decet agree with other purely chemical evidence.Natural products containing oxygen rings have received muchattention.Cannabinol has been synthesised by two differentmethods, and the constitution of homopterocarpin from red sandal-wood has been elucidated. Usnic acid probably does not containthe many-membered ring previously ascribed to it, and is representedby a more conventional dibenzofuran formula. Primetin methylether has been synthesised, and pedicin is regarded as a benzylidene-coumaranone derived from pentahydroxybenzene.Syntheses of the glyoxaline alkaloid pilosinine and of r-lupinineare recorded. Further progress has been made with the diiso-quinoline bases of curare, and several alkaloids from various sourcesprove to be simple benzylisoquinoline derivatives of familiar type.The nitrogenous rings of cevine and the Xolanurn alkaloids have beenexplored, and cassaine also is a steroid or polyterpenoid derivativewhich on saponification yields its nitrogen as dimethyIrtminoethano1.Recent work on the Xtrychnos bases is summarised.T. HENSHALL.D.H. HEY.E. R. H. JONES.J. C. SMITH.T. S. STEVENS.H. B. WATSON.F. s. SPRINcf200 ORGANIC CHEMISTRY,2. ALIPHATIC COMPOUNDS.The last three or four years have seen big changes in aliphaticchemistry, and these have been brought about mainly by the con-version of the petroleum industry into a major chemical industry.Petroleum gases (natural gas, and the gases from ‘( cracked ”petroleum) are, taken together, a rich source of most of the lowerparaffins and olefins. These hydrocarbons, of which many millionsof tons are handled each year, are becoming the raw materials for thepreparation of the lower aliphatic compounds, just as benzene,toluene, and naphthalene have been the starting points in otherseries.Thus ethyl alcohol has for many years been made cheaplyfrom ethylene ; acetone and glycerol are now made from propylene,methyl ethyl ketone from butylene, and the amyl alcohols from thepentanes; fatty acids (made by oxidising solid paraffins) can beconverted into synthetic fats.These industrial developments are not only causing a revolutionin chemical industry, but are creating many problems for solutionby academic chemists. As a further result of this progress manysubstances previously obtainable only in small quantities are nowavailable by the ton, and in a high state of purity.All that is intended in this Report is to record some of the more im-portant advances, and to indicate to organic chemists that the rangeof reagents available in quantity has been considerably widened.Full accounts can be obtained in the technical publications.1,Branched Chain and cycloPara$ins.-Investigations into thecomposition of petroleum and of ‘( cracked ” petroleum fractionshave required the preparation of pure hydrocarbons for use in theidentification of constituents.At the same time the search for fuelsof high “ antiknock rating ” has stimulated research on branchedchain and cycloparaffins. In several laboratories, university andindustrial, mainly in Holland and America, these hydrocarbons havebeen synthesised on a large scale by unequivocal methods, and thephysical constants have been recorded.The methods of synthesis are not in themselves new, the mostimportant series of reactions used being represented by the scheme :ketone, ester, aldehyde (or olefin oxide) -+ alcohol -+ olefin-> paraffin. Much use has of course been made of the Grignardreaction in the preparation of the alcohols, but apart from the useof gaseous methyl and ethyl chlorides (from cylinders) the usual1 Annual Review of Petroleum Technology, 1939, 5, 184 (F.H. Braybrook) ;Reports of the Progress of Applied Chemistry.2 “ The Science of Petroleum ” (A. E. Dunstan), Oxford, 1938; “TheChemistry of Petroleum Derivatives” (C. Ellis), New York (Vol. I, 1934;Vol. 11, 1937).R-MgCI - HZO+ HHENSHALL AND SMITH : ALIPHATIC COMPOUNDS.201rather tedious Grignard technique has had to be retained. Reduc-tion of the olefins has been much simplified by the introduction ofthe Raney nickel catalyst and of high-pressure hydrogenation.Other methods used in these researches are (a) the Wurtz-Fittigreaction, applicable to symmetrical hydrocarbons ; (b) interactionof a Grignard reagent with an alkyl chloride in presence of mercuricchloride; (c) reduction of halides with zinc dust and acetic acid;(d) interaction of halides and zinc dialkyls.For purification the almost universal method is fractional distill-ation ; great advances have in recent years been made in the designof efficient laboratory columns.There are at least two lists of hydrocarbons and their physicalproperties ; 3 but the compounds recently prepared for the purposesindicated above are as follows :Pentanes : n-pentane 4* 2-methylbutaneHexanes : n-hexane 4 9 5 2-methylpentane 4~ l82 : 2-dimethylbutane 6$ laHeptanes : n-heptane 7 0 5 3-methylhexane * 2-methylhexane d l 7v82 : 4-dimethylpentane2 : 2 : 3-trimethylbutane 4, 6# la3-methylpentane2 : 3-dimethylbutane l82 : 3-dimethylpentane 4 ~ 83 : 3-dimethylpentane 4 3 83-ethylpentane 2 : 2-dimethylpentane 82 : 3-dimethylhexane 49 D* 1 0 2 : 4-dimethylhexane lo2 : 5-dimethylhexane 4 p 1 0 s 11 3 : 4-dimethylhexane 4110.183 -methyl-3 -ethylpentme 4* 1 0 2-methylheptane lo2 : 2 : 3-trimethylpentane 4 ~ 6 3 1% 18 2 : 3 : 4-trimethylpentane l 8Nonanes : n-n0nane,s~1~ 2-methyloctane 71 1 4 2 : 3-dimethylheptane lo2 : 6-dimethylheptane l5 3 -eth ylhep tane lo2 : 2 : 4 : 4-tetramethylpentane 14, l7Decanes : n-decane 5p l6 and 2-, 3-, 4- and 5-methy1n0nane.l~Octanes : n-octane 49 5t lo 3-methylheptane 103 A.L. Ward and S. S. Kurtz (jun.), Ind. Eng. Chem. (Anal.), 1938,10, 559;4 J. P. Wibaut, H. Hoog, S. L. Langedijk, J. Overhoff, and J. Smittenberg,5 B. J. Mair, J. Res. Nut. Bur. Stand., 1932, 9, 457.G. Egloff, “ Physical Constants of Hydrocarbons,” New Pork, 1939.Rec. Trav. chim., 1939, 58, 329.D. B. Brooks, F. L. Howard, H. C. Crafton, ibid., 1940, 24, 33.F. C. Whitmore and H. P. Drem, J. Amer. Chem. SOC., 1938, 60, 2573.* G. Edgar, G. Calingaert, and R. E. Marker, ibid., 1929, 51, 1483.9 F. C.Whitmore and W. L. Evers, ibid., 1933, 55, 812.lo A. Maman, Compt. rend., 1937, 205, 319.l1 M. J. Timmermans and (Mme.) Hennant-Roland, J. Chim. physique,l2 F. C. Whitmore and K. C. Laughlin, J. Amer. Chem. SOC., 1932,54, 4011.l3 L. Clarke and R. Adams, ibid., 1915, 37, 2536.l4 F. C. Whitmore, and H. A. Southgate, ibid., 1938, 60, 2571.l5 J. D. White, F. W. Rose (jun.), G. Calingaert, and H. Soroos, J. Res.l6 G. Calingaert and H. Soroos, J. Amer. Chem. SOC., 1936, 58, 636.l7 F. L. Howard, J. Res. Nat. Bur. Stand., 1940, 24, 677.1932, 29, 529.Nut. Bur. Stand., 1939, 22, 315.L. Schmerling, B. S. Friedman, and V. N. Ipatieff, J. Amer. Chem. Soc.,1940, 62, 2446802 ORGANIC CHEMISTRY.A synthesis which shows both the general method and a specialdifficulty owing to an isomerisation is that of 2: 2 : S-trimethyl-pentane : yepMe*$F--$WH:CH,Etmoi / H MeMe*$+CO -+ Me*vC*EtMe Me b H \ P M e y eMe$%!XCHMe -+ Me*j=-CHEtThe carbinol resulting from the action of ethylmagnesium chlorideon pinacolone has been dehydrated in several ways, but in every caseisomerisation also occurred.Refluxing with naphthalene- p-sulphonicacid gave the best results, and the resulting mixture of olefins waspartly separated by fractional distillation.4e 1, After hydrogenationof the olefin, the paraffin was subjected to thorough fractionation.416P. L. Cramer and M. J. Mulligan,lg however, have shown that in thepreparation of 2 : 2-dimethylbutane from pinacolyl alcohol, the iso-merisation can be avoided by use of the acetate, which when passedover glass-wool at 400" yields only 3 : 3-dimethyl-Al-butene :CMe,*CHMe*OH ---+ CMe,*CHMe*OAc 40% CMe,*CH:CH,The mechanisms of the pinacol-pinacolone and the Wagner-Meerwein transformations have recently been discussed by H.B.Watson.,OEthylcyclobutane, cyclopen t ane, me thylcyclopent ane, cyclohexane,methylcyclohexane and isopropylcyclohexane have been preparedby Wibaut? n-octyl- and n-octadecyl-cyclohexane by Waterman.21The preparation of tert. -butyl- and tert.-amyl-cyclopentanes hasbeen reported.22 Failure of tert.-butylmagnesium chloride andcyclopentanone to yield tert. -butylcyclopentanol (compare the actionof isopropylmagnesium iodide23) led to a search for anotherintermediate.This was found in the commercially available p-tert.-bu ty lp henol . Hydrogenat ion gave 4 - tert . - butylcyclohexan- 1 - 01,which on oxidation with nitric acid and ammonium vanadate 24gave p-tert. -butyladipic acid. Cyclisation of the adipic acid (bariumsalt) yielded 2-tert.-butylcyclopentanone ; hydrogenation to thecarbinol, dehydration and reduction led to tert.-butylcycEopentane.2O Ann. Repwts, 1939, 36, 196.Me Mel9 J . Amer. Chem. SOC., 1936, 58, 373.2l H. I. Waterman, J. J. Leenderste, and D. W. van Krevelen, J. Inst.22 H. Pines and V. N. Ipatieff, J. Amer. Chem. SOC., 1939, 61, 2728.as H. Meerwein, Annalen, 1914, 405, 155.24 J. Niederl and R. A. Smith, J . Amer. Chem. SOC., 1937, 59, 715.Petroleum, 1939, 25, 801HENSHALL AND SMITH : ALIPHATIC COMPOUNDS.203Olefins.-Ethylene is oxidised (with oxygen at 350" in presence offinely divided silver) to ethylene oxide,25 from which glycol chlor-hydrin, glycol and glycol ethers can be made, thus reversing theprevious order. In recent years the reaction between ethyleneoxide and hydrogen sulphide with the formation of p p'-dihydroxydi-ethyl sulphide (" thiodiglycol ") has not been overlooked.26 Besidesits value in organic synthesis ethylene oxide is effective as an in-secticide. Other oxides are used in industry as intermediates:propylene oxide is isomerised by acids to propaldehyde, and iso-butylene oxide similarly yields is~butyraldehyde.~, P. 2784Debydrogenation of olefins is receiving a good deal of attention.If the butenes are passed over the oxide of chromium, molybdenumor vanadium on alumina a t 600-650" (at 0.25 atmosphere pressure)an 18% yield of butadiene is obtained.3-Methyl- Al-butene similarlygives 21% of isoprene 27 (the butadiene content of the butene frac-tion of " cracked " petroleum amounts t o 12-14%). With thehigher members of the series the reaction goes further. Ringclosure and aromatisation of paraffins and olefins (above C,) takeplace when the aliphatic compound is passed over amorphouschromium oxide a t 400" : 28paraffin + olefin + hydrogenaromatic + hydrogenn-Hexane yields benzene, n-heptane or 2-methylhexane yieldstoluene, n-octane mainly o-xylene, and n-nonane yields mainlyo-methylethylbenzene.29 This reaction may prove of great valueboth as a source of aromatic hydrocarbons and as a means of raisingthe " antiknock value " of fuels.Polymerisation of propylene with dilute phosphoric acid is adecidedly stepwise process, the dimeride, trimeride, and tetrameridehaving been Butenes have been polymerised in the coldto materials with rubber-like physical properties, but with much less2 6 Brit.Pat. (1935), 462,487.26 A. E. Chichibabin, French Pat. (1934), 769,216; D. F. Othmer and D. Q.Kern, Id. Eng. Chem., 1940, 32, 160.27 A. V. Grosse, J. C. Morrell, and J. V . Mavity, Ind. Eng. Chem., 1940, 32,309.2 8 B. Moldawski and H. Kamusclier, Compt. rend. Acad. Sci. U.S.S.R., 1936,1, 355; Brit. Pat., 409,312; 466,609.2B H. Hoog, J. Verheus, and F. J. Zuiderweg, Trans.Paraday SOC., 1939,35, 993; K. C. Pitkethly and H. Steiner, ibid., p. 979; A. V. Grosse, J. C.Morrell, and W. J. Mattox, I d . Eng. Chem., 1940, 32, 528.30 L. A. Monroe and E. R. Gilliland, Ind. Eng. Chem., 1938, 30, 58204 ORGANIC CHEMISTRY.chemical reactivity. Owing to the absence of unsaturation thesematerials are resistant to the action of ozone.31Large quantities of isobutene arise from the cracking of petroleumand can be converted into liquid hydrocarbons of high “antiknockrating ’,. The C, fraction is separated by distillation; treatmentwith 65% sulphuric acid removes the isobutene more than 100 timesas fast as it does the A1- or A2-butene. Heating the sulphuric acidlayer to about 100” causes polymerisation (dimerisation) of theisobutene; the product separates as an oily layer and the acid canbe used again.32 Instead of sulphuric acid, copper cadmium phos-phate can be used for this p~lymerisation.~~ The product, diiso-butene, is then reduced catalytically to “ isooctane ”, 2 : 2 : 4-trimethylpentane.This substance,2CMe2:CH2 + CMe,*CH,*CMe:CH, + CMe3*CH2*CHMe2because of its exceptionally small tendency to “ knock ” in engines,is taken as the standard and given the antiknock rating or ‘‘ octanenumber ” of 100. (Another aliphatic compound of high “ octanenumber ” is diisopropyl ether.)A process which makes use of both isobutene and isobutane andavoids the necessity for a hydrogenation of the product is theremarkable “ alkanation ” * which occurs in presence of coldsulphuric acid.When isobutene, isobutane and 97% sulphuric acidare stirred at 20°, the pressure falls during 90 minutes from 50 Ib.per sq. in. to atmospheric, and a high yield of “isooctane” r e s u l t ~ , ~ ~ ~ ~ CHMe, + CH2:CMe2 + CMe,*CH,-CHMe,A similar reaction has been shown to occur with, on the one hand,the simpler isoparaffins, isobutane, isopentane, 2-methylpentane, and,on the other hand, with the olefins propene, A1- and A2-butene, iso-butene, and the di- and tri-merides of isobutene. No convincingexplanation of the reaction has yet appeared.The action of sulphuric acid on olefins and on branched chainparaffins is complex. W. R. Ormandy and E. C. Craven 34a have31 W. J. Sparks, I. E. Lightbown, L. B. Turner, P. K. Frohlich, and C.A.Klebsattel, I n d . Eng. Chem., 1940, 32, 731 ; R. M. Thomas, I. E. Lightbown,W. J. Sparks, P. K. Frohlich, and E. V. Murphree, ibid., p. 1283.32 W. J. Sparks, R. Rosen, and P. K. Frohlich, Trans. Faraday SOC., 1939,35, 1040; F. C. Whitmore, I d . Eng. Chem., 1934, 26, 94.38 A. E. Dunstan, Nature, 1940, 146, 186; S. F. Birch and A. E. Dunstan,Trans. Paraday SOC., 1939, 35, 1013.34 Brit. Pat. (1936-1938) 479,345; S. F. Birch, A. E. Dunstan, F. A. Fidler,F. B. Pim, and T. Tait, J . In&. Petroleum, 1938, 24, 308; I n d . Eng. Chem.,1939, 31, 1079.340 J . Inst. Petroleum, 1927, 13, 311, 844. But see V. N. Ipatieff andH. Pines, J . Org. Chem., 1936, 1, 464.* Drs. A. E. Dunstan and 5. F. Birch (private communication) prefer thisterm to “ alkylation,” now being used by American writ,ersHENSHALL AND SMITH : ALIPHATIC COMPOUNDS. 205observed a kind of disproportionation (which they term " hydropoly-merisation") when an olefin and sulphuric acid give rise to aparaffin and a highly unsaturated product.Because of the great demand for esters as solvents for lacquersthe preparation of esters from olefins has been investigated. When anolefin is stirred and heated under pressure with acetic and sulphuricacids a t 60" for two hours, an acetate results; propylene yieldsisopropyl acetate, A1- and A2-butene give sec.-butyl acetate ; iso-butene yields tert.-butyl acetate.35 (In practice the butene-butanefraction is used in the preparation of sec.-butyl acetate.)The tertiary olefin, trimethylethylene, reacts readily with phenolin presence of sulphuric acid to yield p-tert.-amylphenol (which withformaldehyde gives a resin suitable for incorporation in drying oils).Under the same conditions secondary olefins, such as A2-pentene,yield stable phenol ethers which rearrange only with P.2799Halogenation of 0leJins.-An obvious way to use the large quantityof propylene available would be to convert it into glycerol. It wassoon realised, however, that the essential intermediate was ally1chloride, and many ways of obtaining this from propylene weretried.36 isoButylene, at the ordinary temperature, does not addchlorine appreciably, but by substitution yields methallyl chloride(P-methylallyl chloride, CH2:CMe*CH2Cl). This reaction of iso-butylene at a low temperature is exceptional, however, and reallymisled the earlier investigators. It was discovered finally that,when propylene was chlorinated at 400-600", almost exclusivesubstitution occurred.CH,*CH:CH, %> CH,Cl*CH:CH, + CH,*CCI:CH, + CH3*CH:CHC1.3796% 3% 1 %Despite the attendant dangers of this high-temperature exothermicreaction the process is being successfully worked. Ally1 chloride(or bromide) and allyl alcohol are now readily available; methallylchloride and methallyl alcohol are easily obtained from isobutene.By various methods, such as the addition of hypochlorite, followedby hydrolysis, allyl chloride is converted into glycerol of high purity,and methallyl chloride into P-methylglycerol.Chlorination of Pcarafins.-According to H.B. Hass, E. T. McBee,and P. Weber 38 the hydrogen atoms in a paraffin are chlorinated inthe vapour phase (at 300") in the ratio, primary : secondary : tertiary35 T. W. Evans, K. R. Edlund, and M. D. Taylor, I d . E n g . Chem., 1938, so, 55.36 E. C. Williams, I d . Eng. Chem., News Edition, 1938, 16, 630; B. T.37 H. P. A. Groll and G. Hearne, Ind. Eng. Chem., 1939, 31, 1530; W. E.88 I d . E n g . Chem., 1935, 27, 1190; 1936, 28, 333.Brooks, I d . E n g . Chem., 1939, 31, 515.Vaughan and F. F. Riist. J . Org. Chem., 1940, 5, 449, 412206 ORGANIC CHEMISTRY.= 1 : 3.25 : 4-43. These chlorinations are now carried out in-dustrially on a huge scale. To take an example, even in 1937 oneplant chlorinated per day 100,000 gallons of the pentane-isopentanefraction from natural gas.For this, 22 tons of chlorine were passedinto a 60 mile-per-hour stream of hot pentane vapour, only 3 gallonsof pentane and 8 oz. of chlorine being in the reaction zone at onetime. Absorption of the hydrogen chloride produced was one of themain p. 2795The chlorides so obtained may isomerise immediately after form-ation and the mixture is difficult to analyse by fractional distillationbecause of the loss of hydrogen chloride from the tertiary halides.Conversion into the alcohols by heating with sodium oleate andsodium hydroxide solution '(emulsifying hydrolysis) is also accom-panied by olefin formation. The 50/50 mixture of n- and iso-pentane used yields a product containing 50-60% of primaryalcohols ; these are technically more valuable than the secondaryalcohols.2¶ p.2796 The aliphatic alcohols thus become available inlarge quantities not only from the olefins (by hydration) but alsofrom the paraffins; in the c6 series, n-hexyl alcohol, hexaldehydeand hexoic acid are commercial products. Study of the physicalproperties of the hexyl alcohols has been continued by H a v ~ r k a . ~ ~Attention is being paid to the halogenation of higher paraffinfractions mainly with a view to obtaining alcohols whose sulphuricesters may serve as detergents.4O For example, the paraffin fraction,b. p. 95-100"/15 mm., on chlorination gives a product containing80-90% of monochlorides. These can be converted into alcoholsin 30% yield by heating with sodium oleate solution a t 170" for 9h0urs.~1 By the action of chlorosulphonic acid the alcohols areconverted into sulphuric esters and these into the sodium salts.Unfortunately the value of these salts from the mixture of alcoholsis much less than that of primary alcohol derivatives such as sodiumlauryl (n-dodecyl) sulphate.A. R. Padgett and E. F. Degering*lsynthesised the sulphuric esters of the dodecan-2-, -3-; -4-, -5- and-6-01s; they found that the " foam value " and the ability tolower interfacial tension decreased as the hydroxyl (or ester) groupwas moved towards the middle of the chain. Solutions of sodiumcetyl (GIG) and lauryl sulphate show on standing a rapid and thena slow fall in surfaceThe preparation of aliphatic sulphonic acids from the halides and39 F.Havorka, H. P. Lankelme, and I. Schneider, J. Amer. Chem. Xoc.,40 C. F. Reed, U.S. Pat. (1936) 2,046,090; Ann. Reports, 1939, 36, 233.4 1 Ind. Eng. Chem., 1940, 32, 204.42 G. C. Nutting, F. A. Long, and W. D. Harkins, J . Amer. Chem. SOC.,1940, 62, 1096. See also S. C. Stanford and W. Gordy, ibid., p. 1247.1940, 62, 1496HENSHALL AND SMITH : ALIPHATIC COMPOUNDS. 207sulphites has been simplified by S. Zuffanti.43 As the series isascended and the alkyl chain becomes longer, the sulphonic acids(in solution) gradually take on the properties of colloidal electro-lytes. In order to explain the properties of these solutions McBainpostulates the existence of (a) neutral micelle and ( b ) ionic m i ~ e l l e .~ ~The Nitration of Parafins.-Apparently working on the plan ofmaking compounds first and finding a market for them afterwards,the enterprising American chemists have devised a successful pro-cess for the manufacture of nitroparaffins. The n-paraffins arescarcely attacked by nitric acid in the liquid phase, but in thevapour phase at 420" a mixture of nitric acid and paraffin (1 mole ofacid to 2 moles of paraffin) reacts in less than one second.45 (Nitricacid is completely dissociated at 250".) The products obtained arethose theoretically derivable from the free radicals formed by loss ofhydrogen or fission of a C-C bond. Ethane yields nitroethane (73%)and nitromethane (27%) ; 46 n-butane yields 2-nitrobutane (50%),l-nitrobutane (27%), l-nitropropane (5%), nitroethane (Ex), andnitromethane (6%).46 Pentane 47 and isopentane 48 have also beennitrated.The mechanism of the nitration is discussed by R.F. McClearyand E. F. Degering ; 49 the application of the nitroparaffins to organicsyntheses is illustrated by C. L. Gabriel.50The nitroparaffins have found uses as solvents for nitrocelluloseand lacquers,45 but their value lies in their varied reactivity and theyare becoming most important synthetic intermediates. Reductionwith iron and hydrochloric acid, or catalytically with Raney nickeland hydrogen, gives 90% yield of amine; reduction with stannouschloride or with zinc dust and acetic acid gives the oxime, which onhydrolysis with dilute sulphuric acid gives a 43% yield of the alde-h ~ d e .~ ~ [A series of n-primary amines (C6H,*NH, to C18H,,*NH2)has been prepared and the boiling points measured under various43 J. Amer. Chem. SOC., 1940, 62, 1044. Compare D. L. Vivian and E. E.Reid, ibid., 1935, 67, 2559; C. R. Noller and J. J. Gordon, ibid., 1933, 55,1090.44 E. L. McBain, W. B. Dye, and S. A. Johnston, ibid., 1939, 61, 3210;E. L. McBain, Proc. Roy. SOC., 1939, A, 170, 415; P. van Rysselberghe,J. Physical Chem., 1939, 43, 1049.45 H. B. Hass, E. B. Hodge, and B. M. Vanderbilt, I d . Eng. Chem., 1936,28, 339.46 H. J. Hibshman, E. H. Pierson, and H. B. Hass, ibid., 1940, 32,427.47 H. B. Hass and J. A. Patterson, ibid., 1938, 30, 67.48 L. B. Seigle and H. B. Hass, ibid., 1939, 31, 648.4B Ibid., 1938, 30, 64; see also P.G. Stevens and R. W. Schiessler, J. Amer.Chem. SOC., 1940, 62, 2886.Ind. Eng. Chem., 1940, 32, 887.61 K. Johnson and E. F. Degering, J . Amer. Chem. SOC., 1939, 61, 3194208 ORGANIC CHEMlSTRY.pressure^.^, The salts of the amines with organic acids are usefulemulsifying agents.]By the action of concentrated sulphuric acid a t 60" the nitro-paraffins yield hydroxamic acids :Heating with 85% sulphuric acid takes the reaction further,R*CGNoH OH + R*C<gH + NH,*OH. Thus nitroethane and 1-nitroisobutane give 90% yields of acetic and isobutyric acid respec-t i ~ e l y . ~ ~ As a by-product of this reaction hydroxylamine becomesa cheap reagent. The reactions with alkali and with chlorine willlead to important products,50 but the most promising is the con-densation with aldehyde~.~~a In presence of sodium carbonate form-aldehyde may react with all the a-hydrogen atoms : CH,*NO, -+CH2(CH2*OH)*N02 + CH(CH,*OH),*NO, + C(CH,*OH),*NO,.The trinitrate of this trishydroxymethylnitromethane is an excellentexplosive.Similarly nitroethane yields the bishydroxymethylderivative, CH,*C(CH,*OH),*NO,, the dinitrate of which is also anexplosive. Other aldehydes (and ketones) condense with onlyone hydrogen atom of the nitroparaffin. Reduction of the nitro-hydroxy- gives the aminohydroxy-compounds such as 2-amino-2-methylpropane-1 : 3-dio1, CMe( CH,*OH),*NH, ; these are solublebases which yield with higher fatty acids valuable emulsifying soaps.50Diketen.-The pyrolysis of acetone to keten (CH,*CO*CH, --+CH, + CH,:C:O), followed by reaction with acetic acid to formacetic anhydride, is a well-established process [CH,*CO,H +CH,:C:O+ (CH,*CO),O.] Keten is much used in industry as anacetylating agent.%Wil~more,~~ who first prepared keten, found that it polymerisedto a lachrymatory liquid, diketen (b. p.127"), which had manyinteresting reactions. The substance is now available commerciallyin America; although the liquid polymerises on keeping, the solid(m. p. -6.5") can be kept indefinitely in aluminium containerscooled with solid carbon di0xide.~6 One of its industrial uses is inthe manufacture of acetoacetic esters. A. L. Wilson (quoted byBoese 56) proposes the formula (I).52 A. W. Ralston, W. M. Selby, W. 0. Pool, and R.H. Potts, I d . Eng.Chern., 1940, 32, 1093.53 S. B. Lippincott and H. B. Hass, ibid., 1939, 31, 118.530 L. Henry, Bull. SOC. chim., 1895, 13, 999.54 G. H. Morey, ibid., p. 1129.5 5 N. T. M. Wilsmore, J., 1907, 91, 1938; F. Chick and N. T. M. TVilsmore,5 8 A. B. Boese (jun.), I n d . Rng. Chenz., 1940, 32, 16.J . , 1908, 93, 946; 1910, 97, 1978HENSHALL AND SMITH : ALIPHATIC COMPOUNDS. 209A few of the reactions are :CH,:$F--vH, + R-OH + acid --+ CH,*CO*CH,*CO,R0-co(14 + Ni/H -+ CH,*yH--yH, p-butyrolactone0-co + Ar*NH, + CH,*CO*CH,-CO*NHPh + C6H6 -f AlCI, + CH,.CO*CH,COPhThe Peroxide Eflect.Since last year's Report,S7 work on the now widely recognisedeffects of oxygen and peroxides on organic reactions has been con-tinued by Kharasch, Mayo, and their collaborators at Chicago, and acomprehensive review has appeared.58Although in no case has any effect of oxygen or peroxides on theorientation of addition of hydrogen iodide (or chloride) been observed,it has been found by M. S. Kharasch and C. Hannum 59 that peroxidesgreatly accelerate the addition of hydrogen iodide to ally1 bromide.This is now shown to be due to the liberation of iodine from hydrogeniodide by the peroxide. In the presence of 0.005 mole either ofascaridole (menthene peroxide) or of iodine the addition of hydrogeniodide to propylene is complete in less than 1 minute at - 80".Iodine thus has an enormous effect on the rate of addition ofhydrogen iodide to olefins . 6oThe view that the orientation of addition of hydrogen iodide wasinfluenced by the internal pressure of the solvent used was put for-ward by Ingold; it was stated that up to 24% of n-propyl iodidecould be obtained from propylene in propane solution.61 A carefuland extensive repetition of the work has failed to detect the presenceof n-propyl iodide in the product.60It is sometimes suggested that the presence of traces of iodine inthe hydrogen bromide may greatly influence its addition to an olefin.Even if this were so, i t would not alter the fact that in nearly everycase the presence of oxygen or peroxide is necessary before abnormalorientation can occur.F. R. Mayo and C. Walling 589 p- 375 onthermochemical grounds explain the absence of abnormal addi-tions of hydrogen fluoride, chloride and P.226 Furtherresults bearing on the mechanisms of the peroxide-catalysed and themetal-catalysed reactions have appeared. 6257 Ann. Reports, 1939, 36, 219.58 F. R. Mayo and C. Walling, Chem. Reviews, 1940, 27, 351.59 J . Amer. Chem. Soc., 1934, 56, 1782.6o M. S. Kharasch, J. A. Norton, and F. R. Mayo, ibid., 1940, 62, 81.6 1 C. K. Ingold and (Miss) E. Ramsden, J . , 1931, 2746.62 M. Takebayashi, Bull. Chem. SOC. Japan, 1939,14,290 ; 1940,15,113,116 ;0. Simamura,ibid., 1940,15,292; M. S . Kharasch and W. R. Haefele, J. Amer.Chem. Soc., 1940,62,2047 ; A. Michael and N. Weiner, J . Org. Chem., 1940,5,389210 ORGANIC CHEMISTRY.The use of sulphuryl chloride in the chlorination of hydrocarbons(e.g., in the preparation of benzyl and benzylidene chlorides, orcyclohexyl chloride) has already been reported.63 It has now beenshown 64 that, although sulphuryl chloride in the dark and inabsence pf catalysts does not attack aliphatic acids or acid chlorides,it will in presence of traces of peroxides (and in the dark) reactvigorously with most aliphatic acids and acid chlorides.Theimportance of the reaction lies in the fact that the usually reactivea-position is least affected. n-Butyric acid yields 15% of a-, 55% ofp-, and 30% of y-chlorobutyric acid; a second chlorine atom tendsto enter the chain at a point remote from the first chlorine atom.Acetic acid does not react.If a halogen carrier such as iodine is added (in place of the peroxide),the reaction is slow, even at 70°, and the product is exclusively theor-chloro-acid or a-chloro-acid chloride.64In the presence of light a photochemically catalysed sulphonationoccurs; propionic acid and sulphuryl chloride, refluxed for 2 hourswhilst illuminated with a 300 watt lamp (at 5 em.), give a 37% yieldof p-sulphopropionic acid (H03S*CH,*CH2-C02H).Again aceticacid is unreactive.An interesting application of the use of peroxides is recorded byC. D. Hurd and W. A. H0ffman.6~ If hydrogen bromide is addedto o-allylphenyl acetate in the presence of peroxide the bromine addsterminally and then by elimination (apparently) of acetyl bromidechroman is formed; in the presence of quinol the hydrogen addsterminally and the final product is 2-methylcoumaran. From theunacetylated allylphenol, only 2-methylcoumaran could be obtainedin the presence or absence of peroxide, as the effect of the hydroxylgroup neutralises that of the peroxide.OAc-+ChromanAnn.Reports, 1939, 36, 233.64 M. S. Kharasch and H. C. Brown, J . Arner. Chern. Xoc., 1940, 62, 925.65 J. Org. Ohem., 1940, 5, 212; L. I. Smith, Chern. Reviews, 1940, 27, 287HENSHALL AND SMlTH ALIPHATIC! COMPOUNDS. 211Higher Aliphatic Compounds.It has been possible to establish the constitution of the commonersaturated and unsaturated long-chain acids without completelyfreeing them from accompanying acids. But before the physicalproperties could be studied, methods of purification had to beimproved. For the saturated compounds distillation at low tem-perature through a long column (electrically heated66 or wellinsulated 6') separated the homologues sufficiently for crystallisationto become effective ; repeated crystallisation then removed theremaining homologues and the unsaturated imp~rities.~' An accountof studies on the saturated acids (and also on esters, alcohols, andparaffins) has already been given; 6* further examples of the nowwidely recognised polymorphism of these compounds have beenreported.69Unsaturated C,, Acids.-Because of their lower melting pointsand their instability the olefinic acids are difficult to purify.Theolder methods involve the rather tedious use of lead and lithiumsalts; in some cases success has now been achieved by the methodof low-temperature crystallisation worked out by J.B. Brown andhis pupils. When a dilute (5-10%) acetone solution of mixedfatty acids from natural glycerides is cooled to between - 20" and- 30", the saturated acids are almost completely removed; furthercooling to between - 40" and - 80" precipitates the unsaturatedacids. The apparatus and technique are simple, and should begenerally applicable to the purification of low-melting substances. 70p '1Okic acid (cis-CH,*[CH,],*CH:CH[CH,],*C02H). The most per-sistent impurity in oleic acid is not palmitic but stearic acid, whichalso has 18 carbon atoms. Crystallisation at - 20" to - 25",however, gives a sharp separation from the saturated acids.72, 71In the binary systems oleic-palmitic and oleic-stearic acids theeutectics occur at 95 and 98% of oleic acid respectively.As 1% ofstearic acid lowers the melting point of oleic acid by only 0-13", the6 6 E. Jantzen and C. Tiedcke, J . pr. Chem., 1930, 127, 277.67 J. C. Smith, J., 1931, 802.68 Ann. Reports, 1938, 35, 251.Acids and esters : F. Francis and S. H. Piper, J. Amer. Chem. SOC., 1930,61, 577; J. B. Guy and J. C. Smith, J., 1939, 615; P. E. Verkade andJ. Coops, Biochem. Z., 1929, 206, 468. Acetates and ethyl esters: R. vanBellinghen, Bull. SOC. chim. Belg., 1038, 47, 640. Alcohols: K. Higasi andM. Kimbo, Sci. Papers Inst. Phys. Chem. Res., Tokyo, 1939, 36, 286. Glycer-ides: M. G. R. Carter and T. Malkin, J., 1939, 1518. p-Alkoxycinnami6acids : G. M. Bennett and Brynmor Jones, J., 1939, 420.'O J. B. Brown and G.G. Stoner, J . Amer. Chem. Soc., 1937, 59, 3.71 J. C. Smith, J., 1939, 974.72 J. B. Brown and G. Y. Shinowara, J . Amer. Chern. SOC., 1937, 59, 6212 ORGANIC CHEMISTRY.inadequacy of melting points taken in capillary tubes should beevident. Pure oleic acid 719 729 73 has nz'' 1.4597 and melts a t 13.36"or 16.25" & 0.04" (dimorphous), slowly oxidising in the air.Ebidic acid. This trans-isomeride of oleic acid is even moredifficult to free from stearic acid; it has consequently been madefrom the purest oleic acid, and the binary systems elaidic-palmiticand elaidic-stearic acid are recorded. An X-ray spacing of 48.95 A.for pure elaidic acid corresponds to a double molecule with a verticalchain (the longest spacing observed for stearic acid is 46.2 A.).71Linoleic acid ( Ag : l2-octudeccadienoic acid).74 J. B. Brown andJ. Frankel 75 have shown that the a-linoleic acid, m. p. - 6.8", isolatedfrom corn oil by conversion into the tetrabromide, followed by de-bromination with zinc,76 is identical with the main acid obtained bylow-temperature crystallisation from acetone. I n this case thebromination-debromination procedure does not cause a shift of thedouble bonds, and it actually yields a purer product than does thecrystallisation process.Linoleicacid from natural sources is probably cis-cis ; elaidinisation withoxides of nitrogen or with selenium yields a solid acid (m. p. 28-29") and also a liquid acid which has not yet been purified.77 Lino-leic acid, which appears to be essential for the diet of animals,contains one :CH*CH,*CH: unit.78Linolenic acid (As : l2 : 15-octudecatrienoic acid).This has beenobtained in 83--88% purity, by crystallisation first from acetoneand then from light petroleum, of the acids of linseed or perilla oil.The specimen melted a t - 11.5". In this case the acid obtained bybromination-debromination seems different from that obtained bycrystalli~ation.~~ T. Moore 8o has shown that prolonged heating oflinseed oil during saponification causes conjugation of the doublebonds. The acid formed (m. p. 77-79") is named $-elaeostearicacid because of the similarity of its absorption band to that ofelzeostearic acid.73 D. H. Wheeler and R. W. Riemenschneider, OiE and Soap, 1939, 16, 207;P.J. Hartsuch, J . Amer. Chem. SOC., 1939, 61, 1142.7 4 R. D. Haworth, J., 1929, 1456.7 5 J. Amer. Chem. SOC., 1938, 60, 54.713 A. Rollett, 2. physiol. Ohm., 1909, 62, 410.7 7 J. P. Kass and G. 0. Burr, J . Amer. Chem. SOC., 1939, 61, 1062.7 8 E. M. Hume, L. C. A. NUM, I. Smedley-Maclean, and H. A. Smith,79 G. Y. Shinowara and J. B. Brown, J . Arner. Chem. SOC., 1938, 60, 2734;80 Biochern. J., 1937, 31, 142; J. P. Kass and G. 0. Burr, J . Aner. Chem.The isomerism of the linoleic acids is not yet worked out.Compare R. W. Riemenschneider,D. H. Wheeler, and C. E. Sando, J . Biol. Chem., 1939, 127, 391.Biochem. J., 1938, 32, 2162.J. W. McCutcheon, Canadian J . Res., 1940, 18~, 231.SOC., 1939, 61, 3292HENSHALL AND SMITH : ALIPHATIC COMPOUNDS.213Linoleyl and linolenyl alcohols can be prepared by reduction ofthe corresponding esters with sodium in ethyl alcohol; use of butylalcohol brings about conjugation of the double bonds.s1Castor oil is a mixture of several glycerides andthe acid obtained by hydrolysis contains S0-S6~0 of ricinoleic acid,12 - hydroxy-cis- A9-octadecenoic acid,CH,*[CH,],-CH(OH)*CH,*CH:CH*[CH,],*CO,HAlthough many preparations are described, i t is unlikely that theacid has ever been obtained pure. By low-temperature crystallis-ation (- 50" and - 65") J. B. Brown and (Miss) N. D. Greenprepared methyl ricinoleate of above 99% purity, but hydrolysis ofthis with hot alkali caused dehydration and polymerisation. Partialhydrolysis at 0 4 " , followed by low-temperature crystallisation,gave an acid of m. p.5.5" and approximately 96% purity.A new series of oxidation experiments a3 shows that both ricinoleicacid and its trans-isomeride, ricinelaidic acid, yield mixtures oftrihydroxystearic acids (m. p. 110" and 141"). These trihydroxy-acids on oxidation with periodic acid yield p-hydroxypelargonicaldehyde; CH3*[CH2],*CH(OH)*CH2*CH0 and aldehydoazelaic acid,CHO*[CH,],*CO,H, confirming the accepted structure of ricinoleicand ricinelaidic acids.In a comprehensive study of the thermal decompositions of (crude)castor oil by A. Barbot 84 conditions are indicated for obtainingmaximum yields of heptaldehyde and AlO-undecenoic acid (500 g.of oil give 115 g. of aldehyde and 66 g. of acid). The bulky residuefrom the thermal decomposition has received a good deal of attention.Barbot notes that dehydration of the ricinoleic acid to A9 : 11- andA9 : 12-octadecadienoic acids precedes the polymerisation, andsuggests that the residue is largely formed by a Diels addition of theA9 : ll-acid to the other unsaturated acids.The tetrahydrobenzenederivatives with long side chains thus formed immediately change tobenzene derivatives, the hydrogen atoms reducing some of the sidechain double bonds. In support of this view he claims to haveobtained from the residue (by nitration and reduction) a diazotisableamine.From the examples already given it is evident that good progresshas been made in the isolation and purification of the octadecenoicacids. That they should be obtained lOOyo pure is of special im-portance in biochemical studies ; by modification of the existingRicinoleic acid.J. P.Kass and G. 0. Burr, J . Amer. Chem. SOC., 1940, 62, 1796.82 Ibid., p. 738.83 St. E. Brady, ibid., 1939, 61, 3464. Compare J. T. Scanlan and84 Ann. Chirn.. 1939, 11, 519.D. Swern, ibid., 1940, 62, 2305, 2309.This paper contains a full bibliography214 ORGANIC CHEMISTRY.methods (low-temperature crystallisation, fractional distillation ofesters, lead, mercury and lithium salt separation, bromination-debromination) and by use of chromatographic absorption 85 homo-geneous products should eventually be obtained.Elceostearic acid (Ag : l1 : 13-octadecatrienoic acid). The a-form(m. p. 49") of this acid is the chief acid of tung oil, which is used asa " drying " oil.On standing or on irradiation with ultra-violet lightit changes to a @form (m. p. 72"). The two forms give differentaddition products with maleic anhydride.86s 87 The molecularrefraction and the parachor have been determined with speciallypurified specimens .88Although the isolation of severalisomerides of elsostearic acid has been reported, usually the sub-stances have been insufficiently characterised. Pomegranate seed(Punica grandurn) yields an acid, punicic acid,89 which has a highexaltation of molecular refraction (6.9 units), and is reducible (6atoms of hydrogen) to stearic acid. It is oxidised to n-valeric,oxalic and azelaic acids (the methyl ester yields methyl azelate);i t isomerises to P-elx?ostearic acid, identified as the maleic anhydridederivative; it depresses the melting point of the a- or p-elaeostearicacids to a point below the eutectic of the a-P-system.Finally, X-rayanalysis shows it to be distinct, and not merely a mixture of a- andp-elaeostearic a ~ i d s . 8 ~ ~ The isolation of a fourth isomeride, tri-chosanic acid (m. p. 35"), from the seeds of Trichosanthes cur-cumeroides is ~laimed.8~The first conjugated acid of the tetraene seriesis parinaric acid from the kernels of Parimrium Ecsurinum. Thisacid, which crystallises in large plates, m. p. 83.5", and oxidisesrapidly in the air, was a t first regarded by M. Tsujimoto andH. Koyanagi 91 as a triene. Later it was shown to absorb four molecularproportions of hydrogen to yield stearic a ~ i d .~ 2 Oxidation withpermanganate yielded azelaic, propionic, and oxalic acids, all care-fully identified.92 Four structures fit these facts ; E. H. Farmer andE. Sutherland 92 prefer CH,*CH2*[CH:CH],*[CH2],*C02H, as itcontains the widely occurring system, :CH*[CH2],*C02H.Punicic and trichosanic acids.Parilzarric acid.85 C. Manunta, Helw. Chim. Acta, 1939, 22, 1156.86 R. S. Morrell and H. Samuels, J., 1932, 2251; H. P. Kaufmann andE. H. Farmer and E. S. Paice, J., 1935, 1630.S. W. W'an and M. C. Chen, J. Amer. Chem. SOC., 1939, 61, 2283.Y. Toyama and T. Tsuchiya, J. Soc. Chem. Ind. Japan, 1935, 30, 182 B ;J. Baltes, Fette und Seifen, 1936, 43, 93.Y. Toyama and K. Vozaki, ibid., 1937, 40, 249 B.3 O E.H. Farmer and F. A. Van den Heuvel, J., 1936, 1809.s1 J . SOC. Chem. Ind. Japan, 1933, 36, 110, 673 B.92 J . , 1935, 759HENSHBLL AND SMITH : ALIPHATIC COMPOUNDS. 215Limnic acid. Another unusual type of acid is licanic acid, fromBrazilian oiticica oil. It was originally thought 93 to be an isomerideof elaeostearic acid. W. B. Brown and E. H. Farmer 94 found that itwas reduced catalytically to hexahydrolicanic acid, which proved tobe a ketostearic acid. Chromic acid oxidation of this keto-acidgave succinic and probably myristic acid, CH,*[CH2]1,*C0,H, whichindicate a y-ketostearic acid. The hexahydrolicanic acid was thenfound 94 to be identical with a specimen of synthetic y-ketostearicacid.95 On permanganate oxidation licanic acid yields valeric acidand y-ketoazelaic acid.From these facts licanic acid is seen t o be4-keto-A9 : 11: 13-octadecatrienoic acid (m. p. 74-75"). During thecommercial preparation of oiticica oil the licanic acid is partlyconverted by the heat treatment into an iso-acid, probably ageometrical is0rneride.9~Higher Unsaturated AcidS.-Ara;chicEonic acid (C20H3,02). Thisacid, which is obtained from 'the phosphatides of ox suprarenals,has been reduced to arachidic acid (eicosanoic acid, C,,Ha02) ; thenormal iodine number shows that the four double bonds are notc~njugated.~~By fractional crystallisation from acetone of the methyl esters ofthe mixed acids, methyl arachidonate (liquid even at - 80') is leftin the filtrate. It is obtained in 75% purity by evaporation; ifthe product is fractionally distilled or subjected to the bromination-debromination process, the purity (as estimated by iodine or brominenumber) reaches 90-100% .97Neither carbon dioxide nor oxalic acid is produced by ozonolysisor by oxidation with permanganate in acetone solution, and theabsence of the :CH*CH,*CH: unit is inferred.From the identi-fication among the oxidation products of acetaldehyde, succinicand adipic acids the authors tentatively suggest that arachidonicacid is A6 : lo : : 18-eicosatetraenoic acid.97Preliminary results of another investigation 98 suggest adifferent formula. Arachidonic acid purified by the bromination-debromination process and oxidised with aqueous alkaline per-manganate certainly gave oxalic acid ; valeric and hexoic, succinicand glutaric acids were probably present.As the :CH*CH,*CH:unit is not excluded, and because of the biosyntheses of arachidonics3 J. Van Loon and A. Steger, Rec. Trav. chim., 1931, 50, 936.94 Biochem. J . , 1935, 29, 631.O 5 P. W. Clutterbuck and R. Raper, ibid., 1925, 19, 385.s6 A. W. Bosworth and E. W. Simon, J . BioZ. Chern., 1934, 107, 489.97 G. Y. Shinowara and J. B. Brown, ibid., 1940, 134, 331.98 D. E. Dolby, L. C. A. Nunn, and I. Smedley-Maclean, Biochem. J . , 1940,34, 1422216 ORGANIC CHEMISTRY.from linoleic acid,s9 i t is suggested that arachidonic is A5 : ' 11 : 14-eicosatetraenoic acid :CH3*[CH2],*CH:CH*CH,*CH:CH [CH ,],*CO ,H, linoleic acid ;CH3*[CH,],*CH:CH*CH2*CH.CH,.CH*CH2.CH]2:CH*[CH2]3~CO~H, ara-chidonic acid.These divergent results cannot be reconciled even by makingallowance for the different courses which oxidations with per-manganate may take under different conditions.l? 74Many highly unsaturated acidshave been reported to occur in the fish oils, but in most cases it isdoubtful whether the acids are homogeneous ; the wider applicationof a recent advance in technique seems likely to clarify the position.E.H. Farmer and F. A. Van den Heuvel2 have subjected to mole-cular distillation the mixed methyl esters from fish oils (after pre-liminary removal of the more saturated acids). This distillationinvolves heating a film of the ester for only 15 seconds t o temper-atures below 100" a t a pressure of mm.or lower, and condens-ation of the vapour on a cold surface separated from the heatedfilm by less than the mean free path of the molecules. The authorspoint out that the degree of separation of oleic and erucic acids(Cls and C,,) should be 3.5 times as great at 60" as a t 264" andcalculate that a t a given temperature the relative number of mole-cules of various acids distilling from an equimolecular mixture will be :Docosahexaenoic acid (C22H3202).acids ..................... CM c,, c,, c,,molecules ............... 100 21.5 4.2 0-85Such a procedure separates the acids (or esters) sharply accordingto chain length, and the yields are high: hundreds of grams ofthese fractions now become available. The C16, CIS, and C,, acidfractions each contain acids of different degrees of unsaturation,but the C,, fraction is homogeneous docosahexaenoic acid, andattention has been concentrated on this.The molecular refractionshows that there is no conjugation of the double bonds. Ozonisationof the acid gave acetaldehyde, acetic acid, carbon dioxide andsuccinic acid; the methyl ester on oxidation gave, besides these,methyl hydrogen succinate. No formaldehyde or oxalic acid wasdetected. It follows that the end groups are CH3*CH: and:CH*CH,*CH,*CO,H (or :CH*CH,*CH,*CO,Me) and that four:CH*CH,*CH: units are interpolated in some way not yet determined ;five structures are possible (omitting cis- and trans-forms).There was no sign of " clupanodonic acid " (C22H3402), isolatedg9 L.C. A. Nunn and I. Smedley-Maclean, Biochem. J., 1938, 32, 3178.A. Lapworth and E. N. Mottram, J., 1925, 127, 1987.?J. SOC. Chem. Tnd.. 1938, 57, 24; ,J., 1938, 427HENSHALL AND SMITH : ALIPHATIC COMPOUNDS. 217by the older methods in very small yield from Japanese sardineand there was good reason to believe that other docosapentaenoicacids were absent. E’aTmer and Van den Heuvel, find that thehydrogen value/refractivity relationship is a straight line for frac-tions obtained by molecular distillation (up to the C,, acid), butthat the products of ordinary distillation (0.1-5 mm.) and clupan-odonic acid do not fall on this line. The heat treatment has causedthe loss of a double bond; it is difficult to avoid the conclusionthat much of the earlier work on the highly unsaturated acids is inneed of revision.Biosynthesis.-Support for the view thatfatty acids in Nature are synthesised from sugars comes from theobservation that an increase in fat content is often accompanied bya decrease in sugar.That the synthesis was via direct union ofhexoses seemed to be supported by the widespread occurrence ofacids with 12, 18, 24 and 30 carbon atoms.4 In the last few years,however, it has become obvious that fatty acids occur in a moreevenly graded sequence than was formerly supposed and theoriesof synthesis by addition of 2 carbon atoms a t a time, from acet-aldehyde or pyruvic acid, are now more favoured.5 Lignoceric andcerebronic acids were hitherto regarded as having 24 carbon atoms,and the disproof of this view is worth discussing at length.The cerebrosides of the brain are usually divided into phrenosinand kerasin.Phrenosin on hydrolysis yields galactose, sphingosin(C1,H,,02N) and cerebronic (phrenosinic) acid ; kerasin similarlyaffords galactose, sphingosin and lignoceric acid.6 Lignoceric acid,which has approximately the composition C24H4S02, is obtainedalso from beech-wood tar and peanut oil; its history, like that ofcerebronic acid, is very confused and only the recent work canusefully be reported. By careful fractional distillation of lignocericesters from peanut oil E. Jantzen and C. Tiedcke66 obtainedn-eicosanoic, docosanoic and tetracosanoic acids, and a residue ofhigher acids. F. Francis, S. H. Piper, and T.Malkin 7 showed byX-ray analysis of the crystals that the acid from beech-wood tarwas mainly n-tetracosanoic acid and that branched-chain acidswere absent. Kerasin was carefully purified by A. C. Chibnall,S. H. Piper, and E. F. Williams,a and when the lignoceric acid fromit was examined by X-ray methods it was found to be a mixtureM. Tsujimoto, J . SOC. Chem. Ind. Japan, 1920, 23, 1007.E. Fischer, ‘. Untersuchungen uber Kohlenhydrate und Fermenten,”I. Smedley-Maclean, Ergebn. Enxymforsch., 1936, 5, 285 (a review).Ann. Reports. 1929, 26, 223.Proc. Roy. SOC., 1930, A , 128, 242.Biochem. J . , 1936, 30, 100.Higher Saturated Acids.Berlin, 1909, p. 110218 ORGANIC CHEMISTRY.of approximately 10% of C,,, 80% of C,, and loo/, of C,, acids;there must therefore be at least three kerasins.There was a prolonged controversy between A.Klenk and P. A.Levene over the constitution of cerebronic acid. Klenk consideredthat cerebronic acid was a-hydroxytetracosanoic acid, since onoxidation (with chromic acid) it yielded tricosanoic acid. Accord-ing to Levene the oxidation product was tetracosanoic acid, andcerebronic acid was therefore a-hydroxypentacosanoic acid.9 In1929 10 and again in 1933 11 Levene stated that the cerebronic acidwas really a mixture of acids. Klenk disagreed with this view andclaimed an 85% yield of pure tricosanoic acid on oxidation of cere-bronic acid.12 Since the confusion seemed in part to be due todifficulty in identifying the oxidation products, tricosanoic andtetracosanoic acids were prepared 13 from pure undecoic anddodecoic acids respectively by the Robinson synthesis.14From the binary system tricosanoic-tetracosanoic acid it wasevident that the identification of the acids by melting points incapillary tubes would be unsatisfactory, there being less than 0.5"difference in melting point between mixtures containing from 0 to40% of tetracosanoic acid. A small amount of the oxidationproduct supplied by Professor Klenk was tested by a micro melting-point method and appeared to melt slightly below any point in theC23-C24 system.13 (Miss) D.M. Crowfoot 15 examined the syntheticacids and the oxidation product by X-ray methods and concludedthat the oxidation product was mainly tricosanoic acid, mixedprobably with pentacosanoic acid.In the meantime cerebronic acid (from highly purified phrenosin)and its oxidation product were prepared by A.C. Chibnall, S. H.Piper, and E. F. Williams.8 After a comprehensive examinationby mixed melting-point and by X-ray methods they concludedthat the oxidation product was a mixture of C21, C23, and C25 acidsand therefore that cerebronic acid was a mixture of a-hydroxy-docosanoic, -tetracosanoic and -hexacosanoic acids. It is nowobvious that the composition of cerebronic acid and of the oxidationproduct will depend on the purification to which they are subjected :recrystallisation will reduce the proportion of the shorter-chainacids.Ann. Reports, 1929, 26, 224.10 F. A. Taylor and P. A. Levene, J.Biot. Chem., 1929, 84, 23.l1 Idem, {bid., 1933,102, 535.l2 E. Klenk and W. Diebold, 2. phy&oL Chem., 1933, 215, 70.1s R. Aehton, R. Robinson, and J. C. Smith, J., 1936, 283.14 (Mrs.) G. M. Robinson and R. Robinson, J., 1925,127,175 ; (Mrs.) G. M.l6 J., 1936, 716.Robinson, J., 1930, 745; Ann. Reports, 1938, 85, 261HENSHALL AND SMITH: ALIPHATIC COMPOUNDS. 219Cerebronic acid (d-) melts at 100~5-101°,16 or at 102.3-102.6°,with [a]Y + 3-33O.89 13 Synthetic (dE) - a - hydroxytetracosanoicacid,l3, 17 melts at 99.7", and synthetic Z-a-hydroxytetracosanoicacid a t 99", with [a]Y - 3.13".17 The significance of these melting-point differences is not clear, and unfortunately crystals of thea-hydroxy-acid suitable for determination of the chain length byX-ray methods have not yet been obtained.Chibnall's anrtlyticalfigures for cerebronic acid indicate an average composition ofC,,H,,O, and the X-ray measurements on the oxidation productsmake it almost certain that cerebronic acid is a mixture of oc-hydroxy-docosanoic, -tetracosanoic, and -hexacosanoic acids. *When these results are considered in the light of recent work onwaxes, it would appear exceptional if lignoceric and cerebronicacids were homogeneous compounds.In a most important series of papers, which unfortunately cannotbe fully reviewed here, Chibnall, Piper and their collaboratorsexamined about twenty plant and insect waxes.19 They developedmethods for the isolation of the acids, primary and secondaryalcohols, ketones and A great number of these com-pounds were also synthesised as standards,21 and the binary andternary systems were examined by a special melting-point tech-nique.22 Another feature of the work was the application of X-rayanalysis to films of the pure substances and of mixtures of homo-logues.Most of the leaf waxes consisted principally of primaryalcohols, with lesser amounts of fatty acids, paraffins andsecondary alcohols, but tobacco leaf wax is made up exclusively ofl6 E. Klenk and L. Clarenz, 2. physiol. Chem., 1939, 257, 268.17 A. Muller and I. Binzer, Ber., 1939, 72, 615.Is H. Mendel and J. Coops, Rec. Trav. chim., 1939, 58, 1133.lo Summarising papers: A. C. Chibnall, S. H. Piper, A. Pollard, E. F.Williams, and P. N. Sahai, Biochem.J., 1934, 28, 2189; A. C. Chibnall andS. H. Piper, ibid., p. 2209.2o A. C. Chibnall, S. H. Piper, A. Pollard, J. A. B. Smith, and E. F. Williams,ibid., 1931, 25, 2095.21 S. H. Piper, A. C. Chibnall, and E. F. Williams, ibid., 1934, 28,2 175.22 S. H. Piper, A. C. Chibnall, S. J. Hopkins, A. Pollard, J. A. B. Smith,and E. F. Williams, ibid., 1931, 25, 2072; Ann. Reports, 1938, 35, 260.* In Klenk's latest paper l6 cerebronic acid is oxidised with lead tetra-acetate, and the resulting aldehyde converted via the oxime and nitrile intoan acid, m. p. 77.7-78.1°. If this is a melting point K a capillary tube, i tis 1-2' lower than the value for synthetic tricosanoic acid.A very useful method of degrading a fatty acid t o the next lower homologuevia the a-hydroxy-acid (which is oxidised with lead tetra-acetate and oxygenin 89% yield) has been carefully worked out.l* Despite the six stages involvedthe process is said to give an 84% yield of pure pentadecoic acid frompalmitic acid220 ORGANIC CHEMISTRY.paraffins.19 The insect waxes vary from cochineal wax (from whichtriacontanoic acid, 15-keto-n-tetratriacontanol and 13-keto-n-dotria-contanoic acid were isolated 23) to beeswax, which yielded the 24,26, 28, 30, 32, and 34 carbon acids, the primary alcohols C24 to C34,and the paraffins C25, ,,, 29, 31.19 These researches have swept awaymany conflicting views previously held on the waxes; they havedefinitely established that all the acids and primary alcohols havean even number of carbon atoms * and that all the paraffins areodd-numbered.Most of the wax acids and wax alcohols formerlyregarded as single substances have been shown to be mixtures ofhomologues.Assuming that the saturated acids, ketonic acids, and hydroxy-acids are formed from unsaturated acids synthesised from shorterunits, Chibnall and Piper 19 discuss in detail the metabolism ofwaxes. They suggest that the saturated acid, by direct reduction,gives the primary alcohol ; p-oxidation of the acid gives a keto-acid,which then yields either the next lower even-numbered acid, or, byloss of carbon dioxide, a methyl ketone. Reductlion of this ketonethen gives an odd-numbered paraffin :R*CH,*CH,*CH,*OH + R*CH2*CH,*C02HR*CO*CH,*CO,H --+ R*CO*CH,R*CH,*OH +- R*C02H R*CH,*CH,The interesting w-hydroxy-acids (OH*CH2*[CH,],,*CO2H, sabinic ;OH*CH,fCH,],,*CO,H, juniperic ; andOH*CH,*[ CH,] ,*CH :CH*[CH,] ,*CO,H ,ambrettolic) are considered by Chibnall and Piper 19 as hydroxyl-ated products of unsaturated acids such as AS-decenoic 24 (in thiscase the addition of water to the double bond would have to beabnormally oriented).Alternatively these authors suggest that the hydroxy-acids arise23 A.C. Chibnall, A. L. Latner, E. F. Williams, and C. A. Ayre, Biochem.24 A. Grun and T. Wirth, Ber., 1922, 55, 2197.* No long-chain n-fatty acid with an odd number of carbon atoms has yetbeen found in Nature. Probably the most convincing claim to have isolated" daturic " acid (margaric, heptadecyclic acid, from thorn apple seeds) wasmade by H.Meyer and R. Beer (Monatsh., 1912, 33, 311), but this was com-pletely disproved by P. E. Verkade and J. Coops, jun. (Biochmn. Z., 1929, 206,468).J., 1934, 28, 313HENSHALL AND SMlTH : ALIYHATlC COMPOUNDS. 221by the w-oxidation 25 of n-fatty acids; palmitic acid would yieldthapsic, and this on reduction, juniperic acid : CH3*[CH2]14*C02H +HO,CfCH,],,*CO,H --+ OH*CH2*[CH2]14*C02H.A different view is taken by P. C. Mitter and P. N. Bagchi,26who consider that these acids are derived from terpenes (whichsometimes occur in the same oil). Farnesol, by reduction of thedouble bonds, removal of the side-chain methyl groups, and oxid-ation of the terminal methyl, would yield sabinic acid :CH3~CMe:CH*CH,*CH2-CMe:CH*CH2~CH2*CMe:CH*CH,*OHCO2H~CH2~CH2~CH2~CH,~CH2~CH2*CH,~CH,~CH2~CH2~Similarly a hypothetical diterpene alcohol would yield junipericacid.R.Kuhn,Z7 making use of the reactivity of the methyl group incrotonaldehyde,28 condensed crotonaldehyde with itself in presenceof piperidine acetate, and obtained a series of polyene aldehydes :2CH3*CH:CH*CH0 --+ CK,*[CH:CH],*CHO2CH,*[CH:CH],*CHO + CH,*[CH:CH] ,CHOThe hexadecaheptaenal on catalytic reduction gave hexadecyl(cetyl) alcohol ; condensation with malonic acid gave hexadeca-heptaenylidenemalonic acid (deep violet) which on decarboxylationand catalytic reduction yielded stearic acid.These syntheses indicate one way in which fatty acids andalcohols may be built up in Nature from acetaldehyde, croton-aldehyde or p p v i c acid.According to Kuhn, reduction in Naturewould have to occur a t an early stage, since highly unsaturatedacids with conjugated double bonds would be coloured, and thesecolours have not yet been observed in the fats of plants or animals.It is interesting to speculate on the way in which various unsatur-ated acids could arise from Kuhn's polyene acids, but recentwork has shown that in animals (if not in plants) the fatty acidscan be interconverted with great ease. When esters of acids con-taining deuterium are fed to rats, the metabolism of the acidscan be investigated by estimating the deuterium content of the25 P. E. Verkade and J. Van der Lee, Biochem. J . , 1934, 28, 31; P. E.Verkade, Chem. and Ind., 1938, 16, 704.26 J .Indian Chem. SOC., 1939, 16, 402. Compare R. Kuhn, F. Kohler, andL. Kohler, 2. physwl. Chem., 1936, 242, 171.27 Pedler Lecture, J . , 1938, 608. Compare K. S. Raper, J., 1907, 91,1831.28 A. Lapworth, J., 1901, 79, 1273; I. Sm dley, J . , 1911, 99, 1627222 ORGANIC CHJZMISTRY.acids isolated, after a Iapse of time, from various parts of thebody.29 The following changes have been shown to occur :11 11Oleic PalmitoleicCH,*[CH,],*CH:CH*[CH,],CO,H CH3*[CH2],*CH:CH*[CH,],-C02HT. H.J. C. S.Traumatic Acid.-It has long been recognised that there areformed or liberated at the injured surfaces of plant tissues, water-soluble substances which are capable of promoting renewed growthactivity in mature uninjured cells or tissues; such substances arecalled wound-hormones.The isolation of a wound-hormone from the aqueous extract ofground bean-pods has been reported,30 and the constitution of thissubstance, now called traumatic acid, determined.31 Analysis andmolecular weight determinations gave the formula CI2H2,O4, andan equivalent weight determination by direct titration gave a valueof 118, corresponding to a dibasic acid.Catalytic hydrogenationfurnished decane-1 : 10-dicarboxylic acid, identified by its meltingpoint and its mixed melting point with a synthetic specimen.32Oxidation of traumatic acid with permanganate in acetone fur-nished sebacic acid, thus proving the +-position of the doublebond. Hence traumatic acid appeared to be Al-decene-1 : 10-di-carboxylic acid, and this constitution was proved by ~ynthesis.~lMethyl undecylenate, ozonised according to the method ofNoller and Adams,33 yielded the half aldehyde-ester of sebacicacid.This condensed readily with malonic acid in pyridine, andcarbon dioxide was evolved. Hydrolysis of the product yieldedthe required A1-decene-1 : 10-dicarboxylic acid, identical with thenatural product :CH,:CH*[CH,],*CO,Me -% CHO*[CH,],*CO,Me ---+ CH,(CO,H),- colCO,H*CH :CH*[CH,] 8*CO,HTraumatic acid29 D. W. Stetten and R. Schoenheimer, J. Bid. Chem., 1940, 133, 347;30 J. English (jun.), J. Bonner, and A. J. Haagen-Smit, Proc. Nat. Acad.31 Idem, J. Amer. Chem. SOC., 1939, 61, 3434.32 P. Chuit, Helv. Chint. Acta, 1926, 9, 264.33 C. R. Noller and R. Adams, J. Amer. Chem.SOC., 1926, 48, 1074.compare Ann. Reports, 1938, 35, 346.Sci., 1939, 25, 323HENSHALL : ALIPHATIC COMPOUNDS. 223In the come of the investigations 30931 a large number of organiccompounds were tested for similar activity ; these included indole-a-acetic acid, vitamins B,, B,, and B, and other plant growth sub-stances, but the only active substances were certain homologuesand anaIogues of decene-1 : 10-dicarboxylic acid. The hydrogen-ation product, decane-1 : 10-dicarboxylic acid, possesses approxim-ately half the activity of traumatic acid; hence the double bond,whilst not essential, nevertheless enhances activity. The saturateddibasic acids containing 7 or fewer carbon atoms are withoutsignificant activity. On the other hand, suberic and azelaic acidswith 8 and 9 carbon atoms respectively possess slight activity, andsebacic acid of 10 carbon atoms possesses half the activity oftraumatic acid.Phthioic and Tuberculostearic Acids.-In an investigation of thelipides of B.tuberculosis, R. J. Anderson34 isolated two acids,which were separated by high vacuum distillation of their methylesters.35 Hydrolysis of the lower-boiling ester furnished tuberculo-stearic acid, C19H3802, an acid of only slight physiological activity.Hydrolysis of the higher- boiling ester, however, yielded phthioicacid, C26H6202, an acid of great importance, showing almost aIl thetoxic properties of the bacillus itself.Determination of the constitution of tuberculostearic acid (m. p.10-11O) was undertaken by M.A. SpielmanF6 who considered itto be 10-methylstearic acid; for on oxidation with chromic acid ityielded methyl n-octyl ketone and azelaic acid together with a littleoctoic acid. Oxidative rupture of the molecule thus proceedsaccording to the scheme :CH3-[CH2]7*CHMe*[CH,]8*C0,H --+CH,*[CH,],*COMe +.1 CH,*[CH,],*CO,HHO,C*[CH,] ,*CO,HIn an endeavour to prove conclusively this supposition, Spielmansynthesised 10-methylstearic acid as follows : Interaction of n-octyl-magnesium bromide (after removal of the ether) and fused zincchloride in dry benzene provided the zinc alkyl chloride (I). Treat-ment of this with o-cazbethoxynonyl chloride (11) 3, gave 10-keto-stearic acid (111), which was converted into its barium salt, and astirred suspension of the latter in ether treated with methylmag-nesium iodide.The 10-hydroxy-10-methylstearic acid (IV) thus84 J. Biol. Chem., 1929, 83, 169.36 R. J. Anderson and E. Chargaff, ibid., 1930, 85, 77.97 C. R. Fordyce and J. It. Johnson, J . Amer. Chem. SOC., 1933,55, 3368.Ibid., 1934, 106, 87284 ORGANIC CIIEMIIISTRY.produced was dehydrated, the olefinic acid esterified arid hydro-genated (PtO), and the resulting saturated ester saponified :CH,*[CH2],*ZnC1 + Cl*CO*[CH,],*CO,Et --+(1.1 (11.)CH3*[CH2]7*CO*[CH2],*C0,H---+ CH3*[CH,],*CMe(OH)*[CH2],*C0,HI (111.) (IV.)4CH,*[CH,],*CHMe*[CH,],*CO,HIt was found, however, that this synthetic acid melted a t 20-21",whereas tuberculostearic acid melted a t 10-ll", and the mixedmelting point was intermediate.The melting points of the amidesand of the 2 : 4 : 6-tribromoanilides, the densities and the refractiveindices of the two acids agreed exactly. Spielman attributed thedifference in melting point of the acids either to the presence of atrace of impurity such as 9-methylstearic acid in the natural pro-duct, or to the possibility of the natural acid being a d- or E-form(with negligible specific rotation). Resolution of the synthetic acidhas not yet been reported. Owing to the interest attaching to theoccurrence of branched-chain, odd-numbered acids in Nature it isunfortunate that the tuberculostearic acid has not been submittedto thorough X-ray examination.The determination of the constitution of phthioic acid, on theother hand, has proved more difficult, and indeed the problem stillawaits solution. Only a small measure of success has attendedordinary chemical methods of attack.M. A. Spielman and R. J.Anderson 38 obtained on chromic acid oxidation a small quantityof an acid, which gave analytical figures closely corresponding toC,,H,,02 but was not n-undecoic acid; indeed it was a liquid acid,and both its p-bromophenacyl ester and its 2 : 4 : 6-tribromoanilidediffered greatly from those of n-undecoic acid : the conclusionwas drawn that the acid must have a branched chain. E. Chargaff 39attempted to throw light on the structure of phthioic acid by analternative route, that of synthesising isomerides of hexacosanoicacid. By application of the malonic ester synthesis with long-chain halides, he prepared a-ethyl-n-tetracosanoic, a-n-butyl-n-docosanoic, a-n-hexyl-n-eicosanoic, or-n-octyl-n-octadecanoic, u-n-decyl-n-hexadecanoic, and u-n-dodecyl-n-tetradecanoic acids.Theinteresting fact emerged that introduction of an a-substituent intohexacosanoic acid (m. p. €48") lowered its melting point by 20-30°,and since phthioic acid has a melting point of 28", this indicatesthe presence of at least two branchings in the chain.Application of surface-film measurements has strengthened thisSB J . Biol. Chem., 1936, 112, 759. 39 Ber., 1932, 65, 745HENSHALL : ALIPHATIC COMPOUNDS. 225belief. E. Stenhagen40 carried out film measurements on somedisubstituted acetic acids, and with decyldodecylacetic acid, thefilm collapsed at 60 ~ .2 With phthioic acid,* on the other hand,the film was much more compressed and collapsed at 38 A . ~ Thisindicated to Stenhagen that in phthioic acid ther3 was a smalla-substituent which brought about a closer packing of the chains.Accordingly he expressed the opinion that the phthioic acid mole-cule is of the type CH,*[CH ]CH,*[CH2ly-C*CO2H x\CH,*[CH,I/where z and y are of the order of 12 and x = 0 or 1, and that themost probable formula is ethyldecyldodecylacetic acid.The synthesis of such trisubstituted acetic acids has been attempteda t Oxford9 Reichstein 42 has shown that trisubstituted aceticacids can be obtained by application of the Friedel-Crafts reactionwith tertiary halides to methyl furoate, followed by oxidation of thefuran ring :Using methyldioctylcarbinyl chloride in this procedure, there wa6obtained a small quantity of methyldioctylacetic acid,41 whichformed surface films similar to those of phthioic acid, and collapsingat 38 ~ .2 Further, when injected into rabbits, it produced toxiccell reactions, though these were different from those produced byphthioic acid.41 Progress in this field has, however, been unavoid-ably slow owing to the great difficulties encountered in manipulatinglong-chain compounds with their attendant steric effects.In the field of synthetic bactericides for the treatment of tuber-culosis and the analogous pathological condition of leprosy muchhas been accomplished by Adams and his co-workers.43 Adamsprepared a series of di-n-alkylacetic acids and showed that, whereasmaximum activity against B.Zeprce occurred in those acids with16 carbon atoms, they possessed only slight activity towards B.tuberculosis. A more potent factor against this bacterium has beenprepared by Birch and Robinson.41 Application of the Guareschi,a0 Trans. Faraday SOC., 1940, 36, 597.Sir R. Robinson, Presidential Address, J., 1040, 505; A. J. Birch,42 T. Reichstein, H. R. Rosenberg, and R. Eberhardt, Helv. Chim. Acta,43 C. R. Noller and R. Adams, J. Amer. Chem. SOC., 1926, 48, 1080, and* Stenhagen, unpublished result.REP.-VOL. XXXVII. HD. Phil. Thesis, Oxford, 1940.1935, 18, 721.subsequent papers; see Ann. Reports, 1930, 27, 240.See ref. (41)226 OMANI(3 UHEMISTRY.reaction * to methyl n-octyl ketone led to P-methyl-p-n-octyl-glutaric acid.This dibasic acid was converted into its ester chloride,which was condensed with ethyl sodio-a-acetyl-n-heptoate, and theproduct hydrolysed to 5-keto-3-methyl-3-n-octyl-n-undecoic acid ;Clemmensen reduction of this gave 3-methyl-3-n-octyl-n-undecoicacid :Me.aO-UNa.CO,E t IMe CH,-CO,Et ,CO*CH, bpi, 4 C8H1 7>c<CH2-CO-~,H11 CHC0,Et ~ 8 ~ 1 7 ' C<CH2*Co2Et CH,COCIHydrolysis I4.( p 1 7 YsH17Me*[CH,I,*CO*CH,-~CH,*CO,H Zn/HgyHCi Me*[CH,],-$XH,-CO,HMe Me&?-Di-n-octylbutyric acid(3-Methyl-3-n-octyl-n-undecoic acid.)This acid, too, forms very compressed flms of the phthioic acidtype : it appears that the molecule in water is immersed fromthe carboxyl up to the branch in the chain.This fact togetherwith the non-toxicity of the compound further indicates thatphthioic acid is a trisubstituted acetic acid. When tested againstB. tuberculosis in witro, it has shown greater bactericidal power thanany of the acids of the same molecular weight prepared by Adams.Pantothenic Acid.-The structure of pantothenic acid has beenelucidated and its synthesis completed. This acid, of which byfar the richest source is liver (in which there are 40 parts permillion), was first recognised as a factor capable of promoting thegrowth of yeast and of bacteria.45 It has since been identifiedwith the " chick anti-dermatitis factor " 46 and is a vitamin ofimportance in animal nutrition; the name has been chosen toindicate its wide occurrence.Its isolation by the original process 47was exceedingly laborious, the vitamin being finally obtained as itscalcium salt. Even in 1939 48 the calcium salt then available gave44 I. Guareschi, Atti R. Accad. Sci. Torino, 1901, 36, 443.46 R. J. Williams, C. M. Lyman, G. H. Goodyear, J. H. Truesdail, and46 T. Jukes, ibid., 1939,61, 975; D. W. Woolley, H. A. Waisman, and C. A.47 R. J. Williams, J. H. Truesdail, H. H. Weinstock, E. Rohrmann, C. M.48 R. J. Williams, H. H. Weinstock, E. Rohrmann, J. H. Truesdail, H. I(.D. Holaday, J . Arner. Chem. SOC., 1933, 55, 2912.Elvehjem, ibid., p. 977.Lyman, and C. H. McBurney, ;bid., 1938, 60, 2719.Mitchell, a d C. E. Meyer, ibid., 1939, 61, 454HENSHBLL : ALIPHATIC COMPOUNDS.227an analysis indicating a formula subsequently shown to be wrong[( C,H,,05N),Ca instead of (C,H,,O,N),Ca].The determination of the structure of pantothenic acid is theoutcome of a series of brilliant micro-chemical investigations. Acarboxyl and at least one hydroxyl group were shown to be present,the substance did not react with nitrous acid, had no basic pro-perties and did not evolve ammonia on hydrolysis with alkali:the -NH,, the NH, and -CO*NH, groups were absent, but this didnot preclude the possibility of a substituted amide.48 Hydrolysisof 3 mg. of pantobhenic acid with dilute hydrochloric acid gave afraction (0.2 mg.), b. p. 140-160"/0~06 mm., which was proved tobe p-alanine hydrochloride (identified by preparation of p-naphthal-enesulpho- p-alanine) .49 Attention was then focused on the " non-p-alanine " portion of the molecule, which appeared to be ana-hydroxy-lactone,50 for after pantothenic acid had been hydrolysedwith alkali (but not after hydrolysis with acid) the solution gavethe ferric chloride test (clear yellow) characteristic of a-hydroxy-acids .51In confirmation of this a micro-method of determining a-hydroxy-acids was applied : 5O the carbon monoxide evolved in the reactionR*CH(OH)*CO,H -> RCHO + CO + H,O was measured, and thevolume agreed well with that obtained from other a-hydroxy-acids.Further, a micro-method which had given consistent results withp-hydroxy-acids [permanganate titration of the product from thereaction R*CH( OH)*CH,*CO,H ___t R*CH:CH*CO,H] gave a zerovalue with pantothenic acid (and its hydrolysis product).As thehydroxy-lactone (or acid) condensed readily with aldehydes, it wasassumed that a six-membered ring had been formed and thereforethat the hydroxyls were in the a- and the y-po~ition.~~Still under the impression that the lactone contained five carbonatoms, Williams synthesised p-alanine derivatives of cc-hydroxy-y -n-valer olac t one, a - h ydr oxy- p -met hyl-y- but y r olactone and a-hydr -oxy-a-methyl-y-butyrolactone : these derivatives had only slightphysiological activity.50 Then came a preliminary announcement 52that the lactone had been characterised as a-hydroxy- pp-dimethyl-y-butyrolactone (C6Hl0O3 and not C5H80,).In the meantime (in 1939) R.J. Williams and his co-workershad begun to co-operate with the Merck laboratories (U.S.A.).49 H. H. Weinstock, H. K. Mitchell, E. I?. Pratt, and R. J. Williams,J. Amer. Chern. SOC., 1939, 61, 1421.60 H. K. Mitchell, H. H. Weinstock, E. E. Snell, S. R. Stanberry, and R. J.Williams, ibid., 1940, 62, 1776.51 M. A. Berg, Bull. SOC. chim., 1884, 11, 882.6% R. J. Williams and R. T. Major, Science, 1940, 81, 246.25% &SO228 ORGANIC CHEMISTRY.The workers in these laboratorie~,~3 having first devised a rapidextraction of pantothenic acid from sheep’s liver, began furtherinvestigations on the “ lactone half.’’ Analysis and a cryoscopicmolecular weight determination indicated the formula C,HI,O,.Titration showed the absence of a free carboxyl group, but onheating with alkali one equivalent of alkali was consumed.Further,the rate of lactonisation indicated a y- and not a S-lactone ; the pre-sence of a free hydroxyl group in the lactone was confirmed by thepreparation of a monoacetate and a 3 : 5-dinitrobenzoate. A Kuhn-Roth determination of C-methyl gave a value corresponding to26% of one C-methyl, and such a result would be expected fromthe presence of a gem-dimethyl group. - -CH,*CMe,-CH( OH)*vO, which These facts suggested a formulaWATSON : REACTION MECHANISMS. 22999-100% pure. Better yields were obtained by heating the drylactone with the dry sodium salt of p-alanine.55 *The only analogous compound showing activity is hydroxy-pant ot henic acid,56 C Me (CH,*OH) ,*CH (OH) *CO*NH*CH,*CH,-CO,H.For the synthesis of this the necessary aldehyde CMe(CH2*OH),*CH0was prepared by the action of formalin on propaldehyde.Hydroxy-pantothenic acid has strong but relatively specific biologicalactivity.T. H.3. REACTION MECHANISMS.Esterification and Hydrolysis.-It was suggested in 1912 byJ. Ferns and A. Lapworth that, whereas the group -SO,*OAlkreacts with fission of the O-Alk bond (e.g., in alkylation by an alkylsulphate), the corresponding bond of the ester grouping -CO*OAlknormally remains intact; the points of division in the two casesare indicated by the dotted lines in the formuh-SO,-O-iAlk and -COi-O-AlkE. E. Reid2 had already shown that the reactions of thiolbenzoicacid with ethyl alcohol and of benzoic acid with ethyl mercaptanwere as follows :Ph*COi*OH + HiSEt + Ph*CO*SEt + HeOH 1 - - - - - - - - - - - - I(the second reaction is reversible, but the first is not; in fact theaction of hydrogen sulphide upon ethyl benzoate gave benzoic acidand mercaptan).It has now been demonstrated by four methodsthat, in the esterification of a carboxylic acid, hydroxyl from theacid unites with hydrogen from the alcohol to form water (oxygenfrom the alcohol passing into the ester), whereas in ester hydrolysis5 5 R. J. Williams, H. K. Mitchell, H. H. Weinstock, and E. E. Snell, J . A m ~ r .5 6 H. K. Mitchell, E. E. Snell, and R. J. Williams, ibid., p. 1791.a Amer. Chem. J . , 1910, 43, 489.Chem. SOC., 1940, 62, 1784.J., 1912, 101, 373.* Prior to the announcement of the American synthesis 54 T.Reichsteinand A. Griissner (HeZv. Chim. Acta, 1940, 23, 650; paper received May loth,1940) reported the preparation of methyl dl-pantothenate from the racemiclactone, and also methyl d-pantothenate from the 2-lactone ; biological assayswere not recorded. Following the publication of Stiller’s paper, 5* A. GIriissner,M. Gatzi-Fichter, and T. Reichstein (Helv. Chim. Acta, 1940, 23, 1276) repeatedthe synthesis of methyl d-pantothenate and showed that it was biologicallyactive in doses of 10 y230 ORGANIC CHEMISTRY.the alkoxyl group of the ester is transferred intact to the alcohol.The processes are to be representedRCOi-OH I - - - - - - - - - - - - - - I + HI-OAlk+ RGO-OAlk + H*OHR-COj-OAlk ,_-_-__________-_ + HI-OH + R*CO*OH + H-OAlk *,--------------,--- ___----- - - - --The proof is as follows :If the group Alk in the ester R-C0,Alk is linked tooxygen by an asymmetric carbon atom, it will not retain its con-figuration completely if a t any stage it becomes free; preservationof asymmetry therefore gives proof that the 0-Alk bond is notbroken.This was demonstrated for the alkaline hydrolysis ofZ-acetylmalic acid by B. HolmbergY4 and E. D. Hughes, C . K. Ingold,and S. Masterman have now found a full retention of the con-figuration of the p-n-octyl radical in the esterification of P-n-octylalcohol by acetic acid.If the positive alkyl ion is mesomeric, a mechanismof hydrolysis in which this ion becomes free at any stage wouldlead to a mixture of alcohols.It has been shown that crotylacetate and a-methylallyl acetate on hydrolysis by alkali or acid 7yield crotyl and a-methylallyl alcohol respectively, and it may beconcluded that the mesomeric ion MeCH-CH-CH2 never becomesfree. Since acid hydrolysis is a reversible process, and muchesterification must therefore occur if such a reaction is followedalmost to equilibrium, the same conclusion is rendered inevitablefor esterification also, for otherwise an indefinite result would havebeen obtained for acid hydrolysis.8A re-examination of sarsasapogenoic acid led R. E. Marker andE. RohrmanngO to formulate this acid as a hydroxy-diketo-mono-carboxylic acid (XV). A further point advanced in favour of theketal side-chain structure (XIII) or (XIV) for the sapogenins is theobservation 77 that oxidation of digitogenin and of gitogenin givesmethylsuccinic and a-methylglutaric acids.The formation of thelatter had previously to be attributed to the nucleus.The further investigation of the acid isomerisation of the neutralsapogenins together with the application of a method for the con-version of a 3-hydroxycoprostane derivative into the corresponding3( p)-hydroxycholestane derivative led to the establishment of therelationships between tigogenin, neotigogenin, sarsasapogenin, andmi lag en in.^^ Oxidation of sarsasapogenin to the corresponding(&-ketone, sarsasapogenone, followed by bromination of the latter,gives a dibromo-ketone, one halogen atom being attached to C4 andthe other to the side chain.Treatment of this with pyridine yieldsa bromo-A4-dehydrosarsasapogenone, reduction of which withsodium and alcohol effects replacement of the side-chain bromineatom by hydrogen and addition of four hydrogens to the ap-unsatur:ated ketone group. The product proved to be identical with nqo-Sarsasapogenone. (+ 1 Br in side (+ 1 Br in side neo-tigogenin, a sapogenin isolated by L. H. Goodson and C. R. Noller 'J2from Chlorogalum pomeridianum. In exactly the same way smila-J. Amer. Chern. SOC., 1939, 61, 2072, 3477.Dl R. E. Marker and E. Rohrmann, ibid., 1940, 82, 647; It. E. Marker,E. Rohrmann, and E. M. Joneb, ibid., p. 1162.O2 Ibid., 1939, 61, 2420.chain) chain) TigogeninSPRING : STEROIDS. 351genin was converted into tigogenin and, as expected, treatment ofneotigogenin with mineral acid yielded tigogenin, thus showing thatthis pair of sapogenins is related in the same way as sarsasapogeninand smilagenin. Sarsasapogenin and neotigogenin have the" normal " side-chain structure and differ from one another in theconfiguration of C, ; smilagenin and tigogenin have the " is0 "- side-chain structure and likewise differ in the orientation around C,.The Conversion of Neutral Xapogenins into Pregmne Derivatives .-#-Sapogenins.Late in 1939 it was shown by R. E. Marker andE. Rohrmann 93 that, when a sapogenin is heated to 200" with aceticanhydride, it is isomerised to a $-sapogenin, which on mild oxidationyields a pregnane derivative. Thus sarsasapogenin (XIII) 93 andsmilagenin (isosarsasapogenin) (XIV) ,94 when heated with aceticanhydride (followed by hydrolysis), give the same +-sarsasapogenin(XVI),95 which when mildly oxidised gives A16-pregnen-3 : 20-dione(XVII) 96 together with a small quantity of 3-ketoztiobilianic acid.In the same way tigogenin gives +tigogenin 97 and thence Als-allo-pregnen-3 : 2O-di0ne.~~ That the hydroxyl group in tigogenin hasthe p-configuration is also confirmed by a similar conversion of thegenin into 3( ~)-hydroxyaZEopregnan-20-one.R. E.Marker and his collaborators conclude that the availableevidence indicates that the $-genins are oxygenated a t C,, andcontain a c,6-c,7 ethylenic The $-genins have beenformulated as (XVI) and, although evidence has been adduced infavour of this formulati~n,~~ it cannot be considered to be rigidlyestablished.(XIII) + (XIV)Me 1J OH COMeMe I\A/ L(XVI.)The isomerisation of sarsasapogenin into $-sarsasapogenin isreversible ; on treatment with mineral acids under mild conditions,98 J .Amer. Chem. Soc., 1939, 61, 3592; 1940, 62, 518.94 R. E. Marker, E. Rohrmann, and E. M. Jones, ibid., 1940, 62, 648.Q5 R. E. Marker and E. Rohrmann, ibid., p. 521.96 Ber., 1939, '72, 1614.9 7 R. E. Marker and E. Rohrmann, J. Amer. C'hem. Soc., 1940, 62, 898.9 B Idem, ibid., p. 896.g9 R. E. Marker, E. M. Jones, and J. Kreuger, ibid., p. 2632352 ORG ANlC CII EMISTRY.the $-genin regenerates sarsasap~genin.~* Thus the $-genin of thecoprostane configuration re-forms the " normal " genin side chain.On the other hand, $-tigogenin reacts with mineral acids to givetigogenin and not neotigogenin; in this case the $-genin of thecholestane configuration re-forms the " is0 "-genin side chain.The ease of conversion of a sapogenin into a pregnane derivativehas been further demonstrated by R.E. Marker, E. Rohrmann,H. M. Crooks, E. L. Wittle, E. M. Jones, and D. L. Turner.1 Oxid-ation of sarsasapogenin acetate with potassium persulphate, followedby hydrolysis, gave 3 : 16 : 20-trihydroxypregnane, the reactionbeing formulated as follows :/I-Sarsasapogenin has been coiiverted into testosterone by a modific-ation of this reaction ; A16-pregnenedione (XVII) (obtained-fromsarsasapogenin) was reduced to pregnane-3 : 20-dione (XVIII), the4-bromo-derivative of which (XIX) 3 was oxidised with persulphuricacid to yield (XX), which on treatment with pyridine, followed byhydrolysis, gave testosterone (XXI) in small yield.COMe COMeH BrOCOMe1 J .Amer. Chern. SOG., 1940, 62, 525. R. E. Marker, ibitl., p. 2543.A. Butenandt and J. Schmidt, Ber., 1034, 67, 1901SPRING : STEROIDS. 353parison(XXII.HODiosgenin.-Diosgenin was isolated by T. Tsukamoto, Y. Ueno,and T. Ota4 from the Dioscorea and characterised as an un-saturated neutral sapogenin.5 It was later isolated from therhizomes of Trillium erectum and Aletris farinosa (L).7 A com-of the constants of dihydrodiosgenin and tigogenin led(XXVI. )Me I I(XXV.)R. E. Marker and E. Rohrinann * to suggest that diosgenin is a,dehydrotigogenin, a suggestion substantiated by later investigations.Diosgenin was shown to have the structure (XXII) ; $-diosgeninon oxidation (the 3-hydroxyl group and the A5-linkage beingprotected in the usual manner) yields 3-hydroxy-A5 : 16-pregnadien-20-one (XXIII) and thence progesterone. Oxidation (Oppenauer) ofdiosgenin gives A4-tigogenone * (XXIV), reduction of which (a) withsodium and alcohol gave tigogenin (XXV) and ( b ) with a palladiumcatalyst gave srnilagenone (XXVI).More recently R. E. Marker lohas converted diosgenin into dihydroandrosterone.Ch1orogenin.-From the rhizomes of Chlorogcalurn pomeridicsnumP. Liang and C. R. Noller l1 isolated, in addition to tigogenin, a newsapogenin, chlorogenin, C27H4404 ; from the same source, gitogenin l24 J .Plmrrn. SOC. Japan, 1936, 56, 135; 1937, 57, 9.6 T. Tsukamoto, Y. Ueno, T. Ota, and R. Tschesche, ibid., 1937, 57, 283.6 R. E. Marker, D. L. Turner, and P. R. Ulshafer, J. Amer. Chem. SOC.,7 R. E. Marker, ibid., p. 2620.* Ibid., 1939, 61, 1516.Q R. E. Marker, T. Tsukamoto, and 1). L. Turner, ibid., 1940, 62, 2525.10 R. E. Marker, ibid., p. 2621.11 Ibid., 1935, 57, 525.12 C. R. Noller, L. H. Goodson, and M. Synerholm, ibid., 1939, 61, 1707.* The nomenclature is confusing; this derivative would be better namedA4-diosgenone to differentiate it from tigogenone, which is a saturated ketone.REP.-VOL. XXXVII. M1940, 62, 2542; R. E. Marker and J. Krueger, ibid., p. 2548354 ORGANIC CHEMISTRY.and neotigogenin l3 have subsequently been isolated.C. K. Noller 14provisionally formulated chlorogenin as a 3( a) : 12-dihydroxy-sapogenin. R. E. Marker and E. Rohrmann15 showed that thegenin gives a digitonide (see, however, the comments of C. R.Noller l6 on digitonide formation) and that on oxidation it gives adiketone, chlorogenone, which yields a pyridazine derivative ; onthis evidence Marker and Rohrmann suggested that chlorogenin is a3( p) : 6-dihydroxy-steroid derivative. The characterisation of a3 : 6-diketone by the formation of a " pyridazine " derivative hasbeen criticised by C. R. N0ller.l'R. E. Marker, E. M. Jones, and D. L. Turner l8 have confirmed the3 : 6-dihydroxycholestane structure for chlorogenin by the conver-sion of diosgenin into chlorogenone by the following route :/?'y cfis /qA/ Zn + A c O 5 /qT) - O b \ ) o:h/ H i jChlorogenone0 "oc\/DiosgeninThese reactions are comparable with the conversion of cholesterolinto 3 : 6-diketo-A4-cholestene,1s and of the latter into 3 : 6-diketo-cholestane .20Sterols.Oxidation of Chobstero2.-In addition to dehydroandrosterone andacidic products, oxidation of cholesterol (as acetate dibromide)yields 3( ~)-hydroxy-A5-norcholesten-25-one (I),21 a hydroxy-lactoneto which has been ascribed a structure of the type (11),22 A5-pregnen-lone,^^ and the unsaturated keto-alcohol (111) ; 24 in the formationof the last, the removal of the group originally attached to CI3 isprobably to be ascribed to a post-oxidative change (cf.the oxidationof @-ergostenol).25 Partial removal of the sterol side chain withformation of neutral products has also been achieved in the case of13 L.H. Goodson and C. R. Noller, J. Amer. Cheire. SOC., 1939, 61, 2420.l4 Ibid., 1937, 59, 1092; 1938, 60, 1629.l6 Ibid., 1939, 61, 946. l6 Ibid., p. 2717.l7 Ibid., p. 2976. l8 Ibid., 1940, 62, 2537.lQ H. Mauthner and W. Suida, Monatsh., 1896, 17, 579; A. Windaus, Ber.,2O A. Windaus, Ber., 1903, 36, 3755; 1906, 39, 2252.21 L. Ruzicka and W. H. Fischer, Helv. Chim. Acta, 1937, 20, 1291; J.22 K. Miescher and W. H. Fischer, Helv. Chim. Acta, 1939, 22, 155.e3 K. Fujii and T. Matsukawa, J . Pharm. SOC. Japan, 1936, 56, 24.2* H. Koster and W. Logemann, Ber., 1940, 73, 298.%. Achtermann, 2. physiol.Chern., 1934, 225, 141.1907, 40, 257.Hattori, J. Pharm. SOC. Japan, 1938, 58, 150SPRING : STEROIDS. 355epicholestanol (as acetate), oxidation of which gives androsteronetogether with the aldehyde (IV). At a lower reaction temperature,the hydroxy-ketone (V) is obtained in relatively high yield, theformation of androsterone being suppressed.26 The previouslyreported oxidation of 5 : 6-dibromocholestanone to progesterone 27could not be repeated by M. A. Spielman and R. K. Meyer,2* who,however, obtained a 0.2% yield of progesterone by oxidation ofcholesterol dibromide.0. Rosenheim and W. W. Starling29 showed that oxidation of chole-sterol with selenium dioxide gives cis-3 : 4-dihydroxy- A5-cholestene ;/ \ I C1H #[ CH,] ,*COMeoxidation of cholesteryl esters with the same reagent gives mono-esters of the cis-3 : 4-di01,~~* 30 together with esters of 3 : 6-dihydroxy-A4-~holestene.~~, 31 Monoesters of the cis-3 : 4-diol have also beenobtained by the action of potassium acetate on cholesteryl acetatedibromide 32 and by the action of the same reagent on 6-chloro-3-benzoylo~y-A~-cholestene.~~ In contradistinction to the cis-3 : 4-diol, it has been shown by F.S. Spring and G. Swain33 that the4-monoesters do not give insoluble digitonides in spite of thepresence of a 3( p)-hydroxyl group.At various times four different substances have been allocated26 L. Ruzicka, M. Oberlin, H. Wirz, and J. Rleyer, Helv. Chim. Acta, 1937,20, 1283; M. I. Ouchakov, P. F. Epifanski, and A. D. Tshinasva, Bull.SOC.chirn., 1937, 4, 1390.27 N. I. Tavastsherna, Arch. Sci. biol., 1936, 40, 141.2 8 J . Amer. Chem. Soc., 1939, 61, 893.29 J., 1937, 377.30 R. E. Marker and E. Rohrmann, J. Arner. Chem. SOC., 1939, 61, 3022.31 A. Butenandt, and E. Heusmann, Ber., 1937, 70, 1154.32 V. A. Petrow. J., 1937, 1077.33 F. S. Spring and G. Swain, J., 1939, 1356; Nature, 1940, 146, 718356 ORGANIC CHEMISTRY.\/A 73 \1/ SOCI, KO I\/ I 2 H.JO HO \,id OH4<>/ OR SPRING : STEROIDS. 357O b \ <I --+ ():1'0\ /\I/, \/\ 1 + ":(/"C OAc OH(XIII), hydrolysis of which gives the dilactone (XIV). Thusbromination of the 3-ketone (XI) gives a 2-bromo-derivative and39 J . , 1939, 1078.40 R. H. Pickard and J. Yates, J . , 1908, 93, 1678; T. Westphalen, Ber.,4 1 Ber., 1932, 65, 1770.42 A.Windaus, Ber., 1907,40,259; see also R. E. Marker and E. Rohrmann,43 M. I. Ushakow and A. I . Lieutenberg, Nature, 1937, 140, 466; J . Gen.1915, 48, 1064.J . Amer. Chem. SOC., 1940, 62, 516.Chem. Russia, 1939, 9, 69358 ORGANIC CHEMISTRY.oxidation leads to C,IIC3 rupture, a behaviour known to be char-acteristic of C&-keto-steroids in which rings A and B are trans-fused.44 Ellis and Petrow conclude that this orientation pertainsin the 3-keto-5 : 6-diol monoacetate and consequently in the parenttriol-I, which is therefore represented by (XV) and triol-I1 by (XVI).Unfortunately these decisions are dependent upon the assumptionthat the course of bromination and oxidation of saturated 3-keto-steroids is not altered by the replacement of the C,-hydrogen by ahydroxyl group."Sulphomtion of Steroid Ketones.-The sulphonation of a series ofsteroid ketones has been studied by A.Windaus,46 one object of thestudy being to sulphonate the methyl group attached to Clo; thistype of sulphonation has not been observed. Sulphonation of3-keto-A4-cholestene gives the corresponding 6-sulphonic acid(XVII), which is soluble in water; its sodium salt yields water-soluble complexes with cholesterol, cholestene, 3-keto- A*-cholestene,vitamin-D, methylcholanthrene and benzpyrene. Sulphonation ofcholestanone gives the 2-sulphonic acid and sulphonation of copro-stanone gives a mixture of the 2- (XVIII) and the 4-sulphonic acid.The formation of the former is surprising in view of the tendency for(XVIII.) (XIX.)coprostanone to brominate in the 4-position and to suffer C,llC,oxidation. Oxidation of coprostanone-2-sulphonic acid gives thehitherto unknown dicarboxylic acid (XIX) .t44 A. Butenandt and A. Wolff, Ber., 1935, 68, 2091.45 J. Org. Ghem., 1939, 4, 506.46 A. Windaus and E. Kuhr, Anrzalen, 1937,532,52 ; A. Windaus and K. H.Mielke, ibid., 1938, 536, 116; E. Kuhr, Ber., 1939, 72, 929.47 R. E. Marker, E. L. Wittle, L. Plambeck, E. Rohrmann, J. Krueger, andP. R. Ulshafer, J . Amer. Chem. SOC., 1939, 61, 3317.* The configurations (XV) and (XVI) adopted for triols-I and -11, respect-ively, appear to be independent of the assumption made by Ellis and Petrow.First, the immediate lactonisation of the (not isolated) C,l IC, dicarboxylicacid to the lactonic acid (XIII) and of the (not isolated) hydroxy-lactone-acidt o the dilactone (XIV) requires that rings A and B in triol-I be trans-fused.Secondly, the oft-observed fact that cis-hydroxylation of 3(P)-hydroxy-A5-steroids by means of osmic acid is inhibited by acetylation fhds an explanationif triol-I1 is represented by (XVI) (cf. M.Ehren~tein).~5-f A serious confusion between the dicarboxylic acids derived from chole-stanol and coprostanol has arisen in the literature. R. E. Marker et aL47olaim that the dicarboxylic acid, m. p. 247", obtained by oxidation of coproSPRING : STEROIDS. 369Photoisomers of Ergosterol and Related Sterols.-The four stereo-isomers ergosterol, lumisterol, pyrovitamin-D, (pyrocalciferol) andisopyrovitamin-D, (isopyrocalciferol) have been further correlatedby T.Kennedy and F. S. Spring,51 who find that the acetate ofpyrovitamin-D,, like ergosterol and unlike lumisterol and iso-pyrovitamin-D, acetate, yields a bimolecular oxidation product-" pinacol diacetate "-pyrolysis of which yields neoergosterylacetate. The '' pinacol " of dehydroergosterol has been examinedby T. Ando 52 and by A. Windaus and C. Ro~sen-Runge,~~ theconclusion being that it has the structure (XX).Following upon the isolation of vitamin-D4 (obtained by irradi-ation of 22-dihydroergosterol) 54 A. Windaus and B. Giintzel55 haveisolated the intermediate photoisomers lumisterol, and ta~hysterol~.The pyrolysis of vitamin-D3 has been investigated by A.Windaus,M. Deppe, and C. Roosen-Runge,66 who find that, as in the case ofvitamin-D,, a mixture of two isomers is formed. These two isomers,p yr ovit amin-D, and isop yr ovi t amin -D, , together with 7 - deh y dr o -cholesterol and lumistero13 represent the four stereoisomers ofstructure (XXI) differing in orientation around C, and Clo. Thestanol or copro~tanone,~s is the C,IIC,-diacid (XIX) on the grounds that A.Windaus and T. Riemann *9 further oxidised the dibasic acid to the tribasiciaolithobilianic acid (which without doubt is a C,llC, acid). The structureascribed by R. E. Marker et al. to the dicarboxylic acid from coprostanone isincorrect, since this acid gives the same pyroketone as is obtained from dihydro-Diels acid, thus showing that the two acids must be stereoisomers differingin the orientation of an asymmetric centre in the a-position to one carboxylgroup, i.e., they are C,l IC,-dibasic acids differing in orientation around C,.b0Concerning the Windaus-Riemann transformation of the coprostane diacidinto isolithobilianic acid, A.Windaus says that this " scheint unrichtig zusein."The new coprostanedicarboxylic acid obtained by R. E. Marker et al. bythe oxidation of " coprostanediol-3 : 4 " and formulated as a C,l IC,-diacidmay be identical with the coprostane C,I IC,-diacid obtained by A. Windausand K. I€. Mielke46 and mentioned above. It is possible that during thepreparation of the coprostanediol from 4-bromocoprostanone, migration fromC, to C, has occurred and that the coprostane-3 : 4-diol is in reality a copro-stane-2 : 3-diol.The conclusion of R. E.Marker et al. that " The greater part of the oxidativefission of 3-substituted derivatives of the coprostane configuration (exceptthose of the bile acid series) takes place a t the 2, 3 bond rather than a t the 3, 4bond as has been generally assumed," is not sound.48 J. A. Gardner and W. Godden, Biochem. J., 1913, 7, 588; A. Windaus,Ber., 1916, 49, 1724.40 2. physiol. Chem., 1923, 126, 277. 50 A. Windaus, ibid., 1932, 213, 185.61 J., 1939, 250.63 Ber., 1940, 73, 321.64 A. Windaus and G. Trautmann, Z. phpiol. Chem., 1937, 247, 185.5 5 Annalen, 1939, 538, 120.52 Bull. Chern. Soc. Japan, 1939, 14, 482.66 Ibid., 1939, 537, 1360 ORGANIC CHEMISTRY.stereorelationships of these isomers were established in the same wayas in the case of the corresponding ergosterol derivatives.7-Dehydroepicholesterol [3( ~)-hydroxy-A~ ' 7-cholestadiene] hasbeen prepared by A.Windaus and J. nag gat^.^^ Although thechanges in the absorption spectrum during irradiation with ultra-violet light show that this compound suffers the same type of chemi-cal transformations, the product has only one-tenth the antirachiticactivity of a similar product from 7-dehydrocholesterol [3( p)-hydroxy-A5 : 7-cholestadiene], Photopyrocalciferol and photoiso-pyrocalciferol, obtained by ultra-violet irradiation of pyrocalciferoland isopyrocalciferol respectively (changes which are reversed onheating),58 have been further examined by A.Windaus, K. Dimroth,and W. Breywisch. 59 The photoisomers contain two ethyleniclinkages ; during the photoisomerisation one of the nuclear ethyleniclinkages of the pyrosterols is saturated with simultaneous formationof an unstable bridge. The irradiation of As 8-cholestadienol hasbeen further examined by G. Zuhlsdorff.soReduction of tachysterol (XXlI) with sodium and propyl alcoholgives a mixture of dihydrovitamin-D2 I (XXIII) and dihydro-tachysterol (XXIV).sl The latter is of interest in that it possessesin a high degree the property of raising the blood serum calcium58 K. Dimroth, Be?., 1937, 70, 1631.6o Ber., 1940, 73, 328.5 7 Annalen, 1939, 542, 204.59 Annalen, 1940, 543, 240.F. v. Werckr, Z. physiol.Chem.. 1039, 260, 119SPRING : STEROIDS. 361level; it is employed (A.T. 10) in the treatment of ideopathic andpostoperative tetanies. The property of increasing the blood serumcalcium level appears to be less specific than that of antirachiticactivity, the former property appearing in tachysterol, tachysterol,,dihydrotachysterol, vitamin-D,, vitamin-D,, and dihydrovitamin-D,11.conclude that7-dehydrocholesterol is the significant provitamin-D of the skin.An isomer of 7-dehydrocholesterol, A4 : 6-cholestadienol, has beenprepared by V. A. Petrow ; 63 the product obtained from this dienolby irradiation with ultra-violet light does not show vitamin-Dactivity.Other Sterols.-Brassicasterol, a diethenoid sterol isolated fromrape-seed has been shown to be 3(P)-hydroxy-A5 : 22-ergosta-diene (XXV) by E.Fernholz and H. E. S t a ~ e l y . ~ ~ Tetrahydro-brassicasterol is identical with ergostanol ; ozonolysis of brassica-sterol gives methylisopropylacetaldehyde and oxidation of brassica-steryl acetate 5 : 6-dibromide gives 3( ~)-hydroxybisnor-A5-cholenicacid (XXVI).J. M. Bunker, R. S. Harris, and L. M. MosherCHMe*CH:CH*CHMe*CHMe, CHMe*CO,HMe I Me I(XXVI . )a-Spinasterol,66 which has recently been isolated from alfalfa,67has been ascribed the structure (XXVII) by E. Fernholz and W. L.Ruigh.68 Ozonolysis of a-spinasterol gives ethylisopropylacetalde-hyde, thus establishing the presence of a A2,-linkage. Dihydro-a-spinasterol (a-spinastenol) (XXVIII) is identical with the hydrogen-ation product of 7-dehydrostigmasterol (XXIX) .69The diethenoid sterol zymosterol isolated from yeast byI.Smedley-MacLean 70 has been examined by B. Heath-Brown,I. M. Heilbron, and E. R. H. J0nes.~1 The sterol contains a readily62 J . Amer. Chem. SOC., 1940, 62, 1760.6 5 J . Amer. Chem. SOC., 1939, 61, 142; 1940, 62, 428, 1875.6 6 M. C. Hart and F. W. Heyl, J. Biol. Chem., 1932, 95, 311; C. D. Larsenand F. W. Heyl, J. Amer. Chem. SOC., 1934, 56, 942; J. C. E. Simpson, J.,1937, 730; C. D. Larsen, J. Amer. Chem. SOC., 1938, 60, 2431.13' E. Fernholz and M. L. Moore, J. Amer. Chem. SOC., 1939, 61, 2467;L. C. King and C. D. Ball, ibid., p. 2910.68 Ibid., 1940, 62, 2341.69 0. Linsert, 2. physiol. Chem., 1936, 241, 125.70 Biochem. J., 1920, 14, 484; 1928.22, 28, 980.J., 1940, 66. 64 Ber., 1909, 42, 612.71 J . , 1940, 1482362 ORGANIC CHEMISTRY.reducible ethenoid linkage and an inert ethenoid linkage. The fullysaturated zymostanol is identical with cholestanol and the reactiveCHMe*[CH&*CHEt*CHMe,CHMe *CH:CH CHE t CHMe,(XXVIII.)(XXVII.)CHMe*CH:CH*CHEtCHMezCHMe*[CHJ,*CH:CMe,(XXIX. )ethylenic linkage is situated in the side chain as an isopropylidenegroup, since ozonolysis of zymosterol gives acetone ; similar treat-ment of dihydrozymosterol (a-zymostenol) does not give acetone.The last-mentioned derivative is almost certainly identical witha-cholestenol 72 obtained by partial reduction of 7-dehydrochole-sterol ; it is concluded that zymosterol is represented by the structureTreatment of the acetate-tetrabromides of both stigmasterol 7q75and brassicasterol 74 with sodium iodide gives the correspondingsterol 22 : 23-dibromides.Generd-The reduction of steroid ketones by the Wolff-Kischnermethod has been examined by J.D. Dutcher and 0. Wintersteiner.78Whereas the semicarbazones of steroid 7- and 12-ketones behavenormally, those of steroid 3-ketones give mainly the corresponding3( a)-carbinols, this abnormal reaction being independent of theconfiguration at C5. The abnormal reaction is completely suppressedby the addition of hydrazine hydrate; under these conditionscholestanone semicarbazone gives a quantitative yield of cholestane.An interesting modification of the Oppenauer reaction 79 is72 F.Schenck, I(. Buchholz, and 0. Wiese, Ber., 1936, 69, 2696.7 3 E. Fernholz and H. E. Stavely, J . Amer. Chem. SOC., 1939, 61, 2956.74 Idem, ibid., 1940, 62, 428, 1875.75 E. Fernholz, W. L. Ruigh, and H. E. Stavely, $bid., p. 1554.7 6 S. Bernstein and E. S . Wallis, J . Org. Chem., 1937, 2, 341 ; R. E. Marlierand E. L. Wittle, J . Amer. Chem. SOC., 1937, 59, 2704.7 7 W. Discherl, 2. physiol. Chem., 1939, 257, 239.78 J . Amer. Chem. Soc., 1939, 61, 1992.79 R. V. Oppenauer, Rec. Trav. chim., 1937, 56, 137.(XXX)SPRING : STEROIDS. 363described by A. Wettstein.80 Using benzoquinone instead of acetoneor cyclohexanone as hydrogen acceptor, he oxidised 3-hydroxy- A5-steroid derivatives (XXXI) with aluminium tert.-butoxide to thecorresponding A4: 6-ketodienes (XXXII). It is shown that theA4-3-ketones (XXXIII) which are the products of a normal Oppenaueroxidation of (XXXI) 79 cannot be intermediates in the new reaction,since they are unaffected when treated with aluminium tert.-butoxidewith quinone as hydrogen acceptor.On the other hand, the A6-3-(XXXI. ) (XXXII. )ketones (XXXIV) 81 are readily oxidised to the A4: 6-3-ketones bythe modified method.By the action of sodium iodide on the p-toluenesulphonate ormethanesulphonate of cholesterol, stigmasterol or sitosterol, B.Helferich and E. Gunther 82 have obtained the corresponding steryliodides. J. H. Beynon, I. M. Heilbron, and F. S. Spring 83 hadpreviously obtained cholesteryl iodide by the action of hydriodicacid on i-cholesteryl methyl ether.6-Iodocholesterol has beenobtained by R. H. Levin and M. A. S ~ i e l m a n . ~ ~It has been shown by R. P. Linstead 85 that, contrary to thesuggestion of K. Miescher and W. H. FischerYs6 there is no connectionbetween the C3-configuration of a sterol and its capacity to form aglucoside. W. Bergmann and F. Hirschmann 87 have formulatedrules applicable in determining whether a steroid conjugated dienehas its unsaturated system restricted to one ring or whether thesystem extends over two rings. A2 : 4-Cholestadiene has been shownto possess many of the chemical properties associated with theconjugated system of 7-dehydro~holesterol.~~F. s. s.80 Helv. Chim. Acta, 1940. 23, 388.81 A. Butenandt, Ber., 1936, 69, 882, 889, 2773.82 Ibid., 1939, 72, 338, 932.84 J .Amer. Chem. SOC., 1940,62, 920.8 6 Helv. Chim. Acta, 1938, 21, 336.8 8 W. Bergmann, F. Hirschmann, and E. L. Skau, J . Org. Chem., 1939, 4,29; R. P. Jacobsen and C. Z. Nawrocki, J . Amer. Chem. SOC., 1940, 62, 2612.83 J., 1936, 907.8 5 Ibid., p. 1766.J . Org. Chem., 1939, 4, 40364 ORGANIC CHEMISTRY,8. HETEROCYCLIC COMPOUNDS.General.Structure.-V. Schomaker and L. Pauling have determined thedimensions of several heterocyclic rings by electron-diffraction.The C-N distances in pyridine and pyrazine are greater than thatcalculated for 50% double-bond character, as if structures such as(I) and (11) were in resonance with the Kekul6 forms, which wouldaccord2 with the chemical character of the nuclei and especiallytheir resistance to ordinary substitution. In furan, thiophen, andpyrrole, the corresponding distances are in defect, relatively to thesingle-bond values, to an extent in accordance with contributionsof very roughly 10, 34, and 24% from structures of the types (111)and (IV).The dipole moments, determined by H. de V. R ~ b l e s , ~indicate the values 15, 16, and 23% respectively. It is possible toattribute the discrepancy in the case of thiophen to the participationof structures such as (V, a , b, c ) in which the sulphur octet hasexpanded to a decet.i5H(1.1 (11.) (111.) (IV.)The resonance energy of the 5-rings has been re-assessed : furan,23 ; thiophen, 31 ; pyrrole, 31 kg.-cals.The peculiarity of rings containing sulphur has been explored fromanother dire~tion.~ Under standard conditions the following com-pounds exchanged for deuterium theitalicised methyl groups only :hydrogen atoms of theM e - NHO,C(S)lNe(VII.)I n general the hydrogen atoms of methyl side chains in nitrogen ringsbecome mobile only when they are part of a system *N:&CH, or its1 J .Arner. Chem. SOC., 1939, 61, 1769.2 C. Naegeli, W. Kundig, and H. Brandenburger, Helv. Chim. Acta, 1939,3 Rec. Trav. chim., 1939, 58, 111 ; 1940, 59, 184.4 H. Erlenmeyer and H. M. Weber, Helv. Chim. Acta, 1938, 21, 863;See also H.22, 912.H. Erlenmeyer, H. M. Weber, and P. Wiessmer, ibid., p. 1017.Erlenmeyer and H. Ueberwrtsser, ibid., 1940, 23, 1268STEVENS : HETEROCYCLIC COMPOUNDS. 365vinylogue ; the reactivity of the 4-methyl radical in (VI) and (VII)becomes intelligible in terms of structures analogous to (V), but thedata are too few as yet, and the inference is not the only possible one.Basic Character.-The intense colour of cyanine dyes is attributedto resonance in the cation :and the absorption of a dye having two different heterocyclic ringsis commonly intermediate between those of the correspondingsymmetrical pigments. I n the case formulated, owing to the feeblebasicity of the indole nucleus, the first form is subordinated, theresonance therefore limited, and the intensity of colour is less thanthat of either related symmetrical cyanine. In the bases of whichthe cyanines are quaternary salts, the structure involving a bicovalentanionic nitrogen atom will be in general disfavoured, and the basesare lighter in colour than their alkyliodides or their salts with acids :I n this particular case, however, the character of the indole nucleusthat hindered the resonance of the salt favours that of the base,which is now the more highly coloured of the two.On the otherhand the base (11), from dealkylation of the thiazole ring, is colour-less. The series (111), (IIIa), (IV) shows a similar relationship.(111.) Bluish-purple.(IIIa.) Rather paler. (IV.) Pale yellow.ti L. G. S. Brooker, R. H. Sprague, C. P. Smyth, and G. L. Lewis, J . Amer.Chem. SOC., 1940, 62, 1116366 ORGANIC CHEMISTRY.This interpretation is reinforced by the measured dipole moments ofthe bases: contrast (I) ( 7 - 6 8 ~ .) with (11) (4-06) and (111) (10-6)with (IV) (5-43); the values for the non-ionic structures would be2-26 D., and for the betaines 20-30 D.has thoroughly investigated the methodsavailable for the preparation of symmetrical and unsymmetricalbases of this kind, and finds that substances of the type (V) areyellow and deepen in colour in presence of acids, whereas those ofthe type (VI) are orange and unaffected by acidification. Thedistinction is in the direction to be expected, but is surprisinglyemphatic.(Miss) F. M. HamerWhen di-p-naphthaspiropyran (VII) is heated in inert solvents,an intense colour develops, attributed to an intramolecular ionisationwhich is reversed on cooling :CH'CH CH'CHc10H6<O;__)C<~)C10H6 *The process may be regarded as a function of the basic character ofthe central carbon atom or of the left-hand ring as a whole.Of thefollowing naphthapyransYs the first two show no sign of ionisation,the last three are coloured even in the solid state, and the othersdevelop colour in solution at progressively lower temperatures :(VIII. )Valency Angles and Ring CEosure.-Luttringhaus I) uses facility ofring closure in order to assess the distance between the oxygen atomsJ., 1940, 799.7 See, e.g., R. Dickinson and I. M. Heilbron, J., 1927, 1699.8 R. Wizinger and H. Wenning, Helv. Chim. Acta, 1940, 23, 247.* A. Luttringhaus, Annalen, 1937, 528, 211, 223; Ber., 1939, 72, 887;R. Kohlhaas and A. Luttringhaus, ibid., pp. 897, 907 ; A. Luttringhaus andK.Buchholz, Ber., 1939, 72, 2057; 1940, 73, 134STEVENS : HETEROCYCLIC COMPOUNDS. 367and therefore the C-X-C valency angles in compounds of the typep-HO*C6H4*X*C6H4*OH. Under standard conditions the sub-stances HO *c6H4-x*C6H4*O fCH,],Br give the following percentageyields of monomeric ether O~C6H4~Xoc~H4.0~[(H~]~ :I IX. n. X. n.0 - 0 - 0 36 - CH, - - 5 27 68 - s - 0 - 16 51 - CMe, - 0 - 24 54 -5 6 7 8 1 0 1 2 5 6 7 8 1 0 1 2SO, 6 10 - - 24 - co - o - - o 11.6The indication of a small valency angle for S in the sulphone isnotable, and also the absence of any evidence for the Thorpe-Ingoldeffect in the case of CMe2. An X-ray study- --- of the cyclic ether, X = S ; n = 10, gave theC-S-C angle 112.4" & 1-5" and this value hasbeen combined with the chemical data toestimate the angle in other cases.Thus forX = CH,; n = 8, the yield is 27% ; byinterpolation, the same yield corresponds to +o,c -- -- the imaginary case X = S; n = 8.6. Giventhe GS-C angle of 1124", the distance c (I)will be 9 ~ . , and the corresponding distance for the methylenecompound is taken to be 9 A. x 818.6; whence the C-GH2-C angleis 110" & 3", as it should be. A similar calculation gives 133" forC - 0 4 , corrected on somewhat subjective grounds to 129"The possibility of isornorphous replacement of 0, S, and CH, intheir diary1 derivatives and in the cyclic structures just mentionedappears to depend on similarity of valency angle; 10 but whensimilar angles are constrained, as in fluorene and diphenylene oxideand sulphide, complete miscibility is observed.N. M. Cullinaneand W. T. Rees l1 come to similar conclusions in a study of systems/- ffo\c t i'r i(1.)4".c6&<$>C6H4 (X, Y = 0, S, NH).isoIndoles and isoBen2furans.-Their relation to phthalocyanines(see p. 316) has stimulated work on isoindoles and led incidentallyto the demonstration 12 that the only simple unreduced isoindole onrecord is really a dihydro-derivative, further emphasising thecontrast with the readily formed indole system. Phthalonitrile isthe starting point for isoindole syntheses of a new type; withmethyl-lithium or methylmagnesium iodide l3 the strongly basiclo A. Luttringhaus and K. Hauschild, Bey., 1940, 73, 145.11 Trans. Paraday SOC., 1940,36,507 ; N.M. Cullinane and C. A. J. Plummer,l2 R. P. Linstead and E. G. Noble, J . , 1937, 933.J . , 1938, 63.P. A. Barrett, R. P. Linstead, and G. A. P. Tuey, J., 1939, 1809; idemand F. G. Rundall, J . , 1940, 1079368 ORGANIC CHEMISTRY.I-amino-3 : 3-dimethyl-+-isoindole is formed, probably by thefollowing stages :c6H4(cN C6H4<CN ‘<C:NLiCN CMe:NLi CMe+ C H > N +CMe,c~H4<‘3Li 42 C:NLiAcetylation, followed by hydrolysis, yielded dimethylisoindolinone.Phthalonitrile reacts in an analogous fashion with the sodio-deriv-atives of phenylacetonitrile or ethyl cyanoacetate or malonate, thefirst stage being a Thorpe reaction : l4 1 C* CH ( C0,E t ) ,C H / c N - - +4\CNY,C:CH-CO,H ,C:C( CO,Et),C: H >NH CJ34\ >NH (11.14 \ ~ ~ coThe ester (11) can also be prepared from iminophthalimidine,C6H4<’F:H, and ethyl malonate ; l5 ethyl acetoacetate behaves cosimilarly.With hydrogen sulphide, phthalonitrile affords l6 thebinuclear isoindole derivative (111),which condenses readily with reactive Cmethylene compounds with elimin-(111.) ation of hydrogen sulphide.Apparently contradictory results in the investigation of methylene-and carboxymethylene-isoindolinones (-phthalimidines) were ex-plained by the unravelling l5 of the following series of preparativelyuseful transformations :c6H4<c20 C6H4’ \CO*NH,C H / > N N2>6H44\C*SK HS-CC:CH*CO,H CO*CH,*CO,NH,I I<CO*CH,*CO,H CCH,‘6*4 CO-NH, a c6H4<c2NH I acid(roomtemp.)C:CH*CO,H CH*CH,*CO,HPhthalimide C6H4<CZNH .Na-Hgr C,H,<CzNH14 P.A. Barrett, R. P. Linstead, J. J. Leavitt, and G. A. Rowe, J . , 1940,1076.l5 R. P. Linstesd and G. A. Rowe, J . , 1940, 1070.l6 Imperial Chem. Ind., B, 1940, 192, 349, 434STEVENS : HETEROCYCLIC COMPOUNDS. 369In contrast to isoindoles, 3 : 4-benzfurans (isobenzfurans) areknown, at least when arylated in the furan ring; theyiare yellowand fairly stable, but polymerise easily, and i t is suggested on thebasis of optical data that free radicals (V) may be c0ncerned.l’The relatively ready formation of this system is illustrated by theproduction of isobenzfurans from phenolphthalins (triphenyl-methane-o-carboxylic acids) and sulphuric acid, a reaction involvingmigration of an aryl radical.18 A new method for their synthesis isbased on the Diels-Alder reaction : l9Butadiene -+ \trans-CHBz:CHBzjCH,/ \CH vH*COPhCH CH-COPh -+- ”CH, CPh/ \ / \\ / \ / 1 Zn-NaOHCH, CPh + CPh/\/l\CH*COHCl [ J 0 1 >O rnakic fYF0 - \/\l,CH*CO f,h_ \\ CPhCPh (IV-1(two stereoisomerides)HOAc I + 1 : 4-DiphenylnaphthaleneThe isobenzfurans (IV) themselves undergo the diene-synthesiswith ethyl cinnamate, acraldehyde, and especially maleic anhydride,giving derivatives of 1 : 4-diphenylnaphthalene 19* 20 not readilyobtained otherwise.The analogous addition of ct-naphthaquinoneto (IV) affords diphenylnap ht hacenequinone .2 With phosphoruspentasulphide, (IV) yields the related isobenzthiophen,22 which doesnot add maleic anhydride, and is identical with the supposed“ mesothioanthracen-dihydrid ” of A.Bistrzycki and B. Brenken.2317 R. Adams and M. H. Gold, J . Amer. Chem. SOC., 1940, 62, 2038.18 F. F. Blicke and R. A. Patelski, J . Amer. Chem. SOC., 1936, 58, 276, 559.19 R. Adams and T . A. Geissman, ibid., 1939,61,2083 ; R. Adams and M. H.2o R. Weiss and A. Beller, Monatsh., 1932, 61, 143; R. Weiss and A. Abeles,21 C. Dufraisse and P. Compagnon, Compt. rend., 1938, 207, 585.22 C. Dufraisse and D. Daniel, Bull. SOC. chim., 1937, 4, 2063.23 Helu. Chim. Acta, 1922, 5, 20.Gold, ibid., 1940, 62, 5 6 ; R. Adams and R. B. M7earn, ibid., p. 1233.ibid., p. 162 ; C. Dufraisse and R. Priou, Bull. SOC. chim., 1938, 5, 502370 ORGANIC CHEMISTRY.W. Dilthey, E.Graef, H. Dierichs, and W. Josten2* describe thecorresponding phenanthrathiophen.Natural Products Containing Oxygen Rings.Coumaranones and Plarvones.-Didymocarrpus pedicellata affordsderivatives of pentahydroxybenzene.25 The constitution of pedi-cellin (I; R = Me) is well established; pedicin, of which pedicellinis the dimethyl ether, is probably (I ; R = H) ; and isopedicin and4-isopedicin are possibly the active and the racemic form of therelated coumaranone (11).RO OR HO OL(I.) MeO$>O*CH:CHPh MeOC-)CO*CH*CH,Ph (II.)The carmine-red constituent pedicinin, also obtained by oxidationand hydrolysis of pedicellin, was formulated as the benzylidene-coumaranone (111), but P. K. Bose and P. Dutt 26 give cogent argu-ments in favour of the structure (IV).Me OMe Me0 OMeHO /o\ HO /o@-(CO*CH:CHEh (Iv.) (111.) MeOQO*C:CHPh HO HOH \-/The view that primetin is 5 : 8-dihydroxyflavone has now beenconfirmed by synthesis2’ of the $-monomethyl ether (VI). Oxid-ation of 2 : 6-dihydroxyacetophenone or its monobenzyl ether withpersulphate, followed by methylation and partial hydrolysis,afforded 2-hydroxy-3 : 6-dimethoxyacetophenone (V) :Friedel-Crafts acetylation of 1 : 2 : 3 : 5-tetramethoxybenzene isaccompanied by demethylation, and the resulting Z-hydroxy-3 : 4 : 6-trimethoxyacetophenone affords, via its anisylidene deriv-ative, 5 : 7 : 8 : 4’-tetramethoxyflavanone (carthamidin tetramethylether).28 Synthetic 5 : 8 : 4’-trimethoxflavone is not identical withS.Siddiqui, J .Indian Chern. SOC., 1937, 14, 703; V. Sharma and24 J . pr. Chem., 1939, 151, 185.S. Siddiqui, ibicl., 1939, 16, 1 ; S. Warsi and S. Siddiqui, ibzd., p. 519.26 Ibid., 1940, 17, 499.W. Baker, N. C. Brown, and J. A. Scott, J., 1939, 1922; Z. Horii,J . Pharm. SOC. Japan, 1939, 59, 209.2e G. Bargellini, Atti X Congr. Internaz. Chim., 1938, 111, 32STEVENS : HETEROCYCLIC COMPOUNDS. 371the dimethyl ether of ginkgetin, from the leaves of the Maidenhairtree ; exclusion of other alternatives and reconsideration of theanalytical data suggest that ginkgetin may contain three extracarbon atoms and an additional ring.29The glucose residue of gossypitrin is in position 7,30 since theglucoside on methylation and hydrolysis affords 7-hydroxy-3 : 5 : 8 : 3‘ : 4’-pentamethoxyflavone agreeing in properties withsynthetic materiaL31 Lespedin, isolated from Lespexa cryptobotrya,and doubtfully identical with campheritrin, is a dirhamnoside ofcampherol.With diazoniethane it yields a monomethyl etherreadily and a dimethyl ether with difficulty ; hydrolysis of the formeraffords campherol 4’-monomethyl ether, and i t is inferred that therhamnose residues are in the 3- and 7-po~itions.~~Pterocarpin and homopterocarpin, colourless constituents of redsandalwood, are chromanocoumarans,33 the latter, more abundant,substance being (VII; R, = OMe, R, = H).0V ’ ‘c6 Me??- ‘&The four oxygen atoms of homopterocarpin are indifferent, butreduction yields a phenolic dihydrohomopterocarpin (VIII) whichcan be oxidised to a p-quinone and also to the acid (X).Oxidationof the methyl ether of (VIII) affords the isoflavanone (IX), whichhas been synthesised as follows [R = 2 : 4-(MeO),C,H3] :The isoflavone is reduced catalytically to the methyl ether of(VIII) and then oxidised to (IX). Pterocarpin contains a methylene-29 W. Baker and W. H. C. Simmonds, J., 1940, 1370.31 W. Baker, R. Nodzu, and (Sir) R. Robinson, J., 1929, 74.aa S. Hattori and M. Hasegawa, Proc. Imp. Acad. Tokyo, 1940, 16, 9.33 E. SpLth and J. Schliiger, Ber., 1940,73, 1 ; A. McGookin, A. Robertson,P. S. Rao and T. R. Seshadri, Proc. Indian, Acad. Sci., 1939, A, 9, 177.and W. B. Whalley, J., 1940, 787372 ORGANIC CHEMISTRY.dioxy-group and shows similar reactions to those of its associate ;it is formulated as (VII; R,R, = CH,O,), the orientation followingfrom the production of a p-quinone from d&jdropterocarpin.Fuurano- and Pyrano-compounds.-Nodakpnin, from Peucedanumdecursivum, is the glucoside of nodakenetin,$rom which it has beenresynthesised.Nodakenetin by dehydration and reduction affordsthe known lactone (I), the relative stabilitg,to dehydration showingthat the hydroxyl group is not in the furan ring ; and the productionof acetone on oxidation defines $be structure as ( II).34It has now been shown that the linearly constituted xanthyletin(111) accompanies its angular isomeride seselin in the synthesis ofthe latter from methylbutinol and umbelliferone : 35The formation of methyl isopropyl ketone by oxidation of dun-nione establishes the structure (IV) for the latter.36 Migration ofa methyl group from the P- to the a-position in the furan ringprobably takes place in the conversion into P-isodunnione, whichyields acetone on oxidation.34 E.Spiith and P. Kainrath, Ber., 1936, 89, 2062; E. Spiith and (Frl.35 E. Spiith and R. Hillel, Ber., 1939, 72, 963, 2093; compare Ann. Reports,36 J. R. Price and (Sir) R. Robinson, J., 1940, 1493.E. Tyray, Ber., 1939, 72, 2089.1939, 38, 316STEVENS : HETEROCYCLIC COMPOUNDS. 373A phenanthrafuran system seems to be present in tanshinone I, aconstituent of the Chinese drug tan-shin. The substance is ano-quinone, not easily hy+ogenated beyond the stage of quinol, andyields on oxidation the ghhydride (VI). The third oxygen atom isindifferent, two C-methyl groups are present, and tanshinone I isalmost certainly (V) or a position i~omeride.~'Usnic A~id.~~-The nature of the side chain in decarbousnic acid(I ; R = Me) is conkmed by the.demonstration that the relatedacetousnetic ester ( I ; R = OEt) i$ a y- and not an a-substitutedacetoacetic ester.39Usnic " acid " (which is not a carboxylic acid) yields decarbousnicacid by simple heating with alcohol at 150"-~,,H,,O, + H,O -+C,,H,,O, + CO,; since its optical activity persists in alkalinesolution, the additional carbon atom is probably linked in such a wayas to give rise to an asymmetric centre in the coumarone nucleus,with disturbance of the hromatic conjugation of one or both rings.F. H. Curd and A.Robertson,40 and R. T. Foster, A. Robertson, andT. V. Healy 41 propose the formula (11), which accounts for the pro-duction of an a-coumaranone and oxaloacetic acid on ozonisati~n.~~Many reactions of usnic acid then imply facile opening of thedihydrobenzene ring,'&ributed to mobility caused by the quaternarycarbon atom and the tendency to form a normal coumarone system.Sulphuric acid converts usnic and decarbousnic acids respectivelyinto usnolic acid (a true carboxylic acid) and decarbousnol. On thebasis of the formula (11) the dihydrobenzene ring opens and re-closesto (I11 ; R = CO,H), decarbousnol being (TI1 ; R = H).This has been confirmed by model experiments in which (IV;R = H or Me) gave on treatment with sulphuric acid productsclosely resembling usnolic acid, in the formation of which, therefore,neither the 7-carbon atom nor the attached acetyl group is involved.97 M.Nakao and T. Fukushima, J . Pharm. SOC. Japan, 1934, 54, 844;38 Ann. Reports, 1938, 35, 314.40 J., 1937, 894.41 J., 1939, 1594.42 C. Schopff and F. Ross, Naturwiss., 1938, 26, 772.F. von Wessely and S. Wang, Ber., 1940, 73, 19.Y . Asahina and M. Yanagita, Ber., 1939, 72, 1140374 ORGANIC CHEMISTRY.Y. Asahina 43 has discussed the transformations of usnic acid andproposes among others tlhe structure (V).0 CHeCOMe/\/\/\/\//c:O f7 c:CH,*CO*qH*COMe \ /Me C02Et co (V.)MeG CO 70C 0 CHMeCH L C H M eCannubinoL-Earlier work 44 established the structure (I), apartfrom the orientation of the substituents in the right-hand nucleus.It is now found 45 that both in hashish and in American wild hemp(marihuana) cannabinol is accompanied by the dihydric phenolcannabidiol, the structure of which has been largely elucidated.46* 47The phenol can be decomposed into p-cyniene and olivetol (5-n-amylresorcinol), and it is hydrogenated to tetrahydrocannabidio1,which yields menthane-3-carboxylic acid on oxidation.The absorp-tion spectrum indicates that no double bond is conjugat.ed with thearomatic nucleus; and, by comparison with those of syntheticmaterials, that cannabidiol is a 2-substituted resorcinol. Canna-bidiol, which yields formaldehyde in quantity on ozonolysis anddoes not combine with maleic anhydride, is provisionally formulatedas (11).Acids convert cannabidiol into a variable mixture of isomerictetrahydrocannabinols, dehydrogenated to cannabinol itself$'* 48The orientation (I) for cannabinol, already rendered probable by43 Proc. Imp.Acad. Tokyo, 1939, 15, 311.44 Ann. Repwts, 1932, 29, 191.45 (Miss) A. Jacob and A. R. Todd, J . , 1940, 649; R. Adams, D. C. Pease,and 3. H. Clark, J. Amer. Chem. SOC., 1940, 62, 2194.46 R. Adams, C. K. Cain, J. H. Clark, M. Hunt, and H. Wolff, ibid., pp. 196,732, 735, 1770, 2215.4 7 R. Adams et al., ibid., p. 2566.48 R. Adams, D. C. Pease, C. K. Cain, and J. H. Clark, J. Amer. Chem. Soc.,1940, 62, 2402; idem, B. R. Baker, H. WOW, and R. B. Wearn, ibid.,p. 2245STEVENS : HETEROCYCLIC COMPOUNDS. 375its absorption spectrum and colour reactions, is thus confirmed.Following much exploratory work, cannabinol was synthesisedindependently by both groups of investigators :A 49 (AcO) ~~Me,&O ypCbH1lA iEaioias.Pyrrolidine and Piperidine Bases.-The minor bases of Aobelicainflab have been further ~tudied.~l dl-Lelobanidine has beendegraded to the diketone Bz*[CH,],*COEt, independently synthesised,and to 1 -methylpiperidine-2-carboxylic-6-acetic acid, whence it isformulated as (I).The other alkaloids are related to it in much thesame way as the previously studied C,, and C,, bases to lobelanidine(I, with Ph for Et).4Q R. Ghosh, A. R. Todd, and S. Wilkinson, J., 1940, 1121, 1393; comparealso R. Adams, and B. R. Baker, J . Amr. Chern. SOC., 1940, 62, 2401;G. Powell and T.H. Bembry, ibid., p. 2568.5O R. Adams, B. R. Baker, and R. B. Wearn, $bid., p. 2204.51 0. Thorn&, Artnalen, 1939, 540, 99; H, Wieland, W. Kosthara, (Frl.)E. Dane, J. Renz, W. Schwarze, and W. Linde, ibid., p. 103376 ORGANIC CHEMISTRY.Senecio species h a ~ e furnished a number of new alkaloids, severalof which are esters of retrone~ine.~~ Monocrotaline, the alkaloid ofCrotalaria spectabilis and C. retusa, affords retronecine, monocroticacid, and carbon dioxide on hydrolysis, and retronecanol and mono-crotalic acid on reduction. The former acid (11) has been syn-thesised, and i t is concluded that the latter is (II?).= Heliotric acidfrom heliotrine has been shown to be (IV).54OH cO,H OH OMe0 - I C0,H(111.1 Me*C*CHMe-CMe*CO CHMe,*g-CHMe (IV.)N-Methylpyrrolidine has been isolated from tobacco, and it hasbeen shown that isonicoteine is 2 : 3’-dipyrid~I.~~ A new synthesisof r-lupinine (VI) is recorded ; picolinoyl chloride, by the successiveaction of diaLomethane and acid, yields picolinoylcarbinol, whichwith y-ethoxypropylmagnesium bromide affords (V) .56OH CH, CH-CH,*OHC-CH,=OH /\/\H,-PtO, v H 2 yH v H 2(v-) p p F .2 -zzi-+ CH, N CH, (VI.)\/N P H 2 \,A/CH,*OEt CH, CH,Renzylisoquinoline AZkaZ~ids.-CheiIanthifoline,~~ obtained fromCorydalis species, is (I), since its methyl ether is identical withsinactine, and its ethyl ether can be oxidised to the known lactam(11)-M e 0 ~ C ~ ~ ~ 2 CH2/N\ mA’ \?H2 (11.)Vo)CH2HOEta()\ / N HCH CH,co (1.) C’H, /\/\OCryptocavine yields, by Emde reduction of the methosulphate,followed by dehydration and oxidation of the resulting carbinol, thesame products as are obtained by similar treatment of the isomeric62 R.H. F. Manske, Canadian J. Res., 1939, 17, B, 1, 8 ; H. L. de Waal,53 R. Adams, R. S. Long, C. F. Rogers, and F. J. Sprules, J . Amer. Chem.54 G. P. Menschikov, J. Gen. Chem. Russia, 1939, 9, 1851.5 5 E. SpZith and S. Biniecki, Ber., 1939, 72, 1809.6 6 K. Winterfeld and H. von Cosel, A T C ~ . Pharm., 1940, 278, 70.67 R. H. F. Manske, Canadian J . Res., 1940, 18, B, 100.Nature, 1940, 146, 777.SOC., 1939, 61, 2815, 2819, 2822; 1940, 62, 2289STEVENS HETEROCYCLIC COMPOUNDS. 377cryptopine; and it is regarded asthe adjacent methylene group of cCH,H,:H2(111), in which the carbonyl a,ryptopine are interchanged.58mdAnonaine, from Anona reticulata, is (IV).59 The inactive base hasbeen synthesised by standard methods, and, although i t could notbe resolved, i t yielded on Hofmann degradation the same nitrogen-free product as the natural alkaloid. Rameria refracta affords thebase rmmerine, which appears to be AT-methylanonaine ; there isfair agreement as to the properties of common transformationproducts. The demethylenated base is not identical with apo-morphine (the 5 : 6-dihydroxy-analogue) ; its degradation byexhaustive methylation proceeds as usual in this series and affordsultimately the known 3 : 4-dimethoxyphenanthrene, thus locatingthe methylenedioxy-group of the originalThe relative positions of the hydroxyl and methoxyl groups inbebeerine (V) have now been established by Hofmann degradationand oxidation of bebeerine diethyl ether, which afforded the acids(VI) and (VII), identified by synthesis.61The alkaloids of several Chondrodendron species have beenthoroughly examined, primarily in order to establish the botanicalprovenance of the drug radix pareirce bravce.62 Protocuridine andisochondrodendrine give the same product on methylation and areisomeric dimethyl ethers of (VIII), as is also neoprotocuridine, whichaffords the same O-methylmethine as isochondrodendrine.The firsttwo active bases belong to the same stereochemical series; and thelast is internally compensated. The new alkaloid chondrofoline isa trimethyl ether related to (V), since it gives an O-methylmethinemethiodide, also obtained from bebeerine.A group of bisbenzyl-isoquinoline alkaloids may be classified as follows :isochondrodendrine type (VIII) Bebeerine type (V).&Bebeerhe d-isochondrodendrineAsymmetric centres O-Methylisochondrodendrined-ProtocuridineOpposite sign .i-neoProtocuridine d-Tubocurarine5 8 R. H. F. Manske and L. Marion, J . Amer. Chem. SOC., 1940, 62, 2042.6o S. Junusov, R. A. Konovalova, and A. P. OrBkhov, J. Qen. Chem. Russia,61 H. King, J., 1939, 1157.G. Barger and 0. Weitnauer, Helv. Chim. Acta, 1939, 22, 1036.1939, 9, 1356, 1507, 1868.62 H. King, J., 1940, 737378 ORGANIC CHEMISTRY.Radix parreirct! bravs also contains the simpler base d-isococlaurine,which is (IX), since it yields the same dimethyl ether methiodide asthe known coclaurine, and gives no catechol reaction.0 CO,HCHoTetrandrine, in which the benzylisoquinoline systems are joinedhead to head and tail to tail, is the methyl ether of the new alkaloidfangchinoline ; both occur in the Chinese drug Han-fang-chi.Oxid-ation of tetrandrine or of fangchinoline ethyl ether yields 2-meth-oxydiphenyl ether-5 : 4'-dicarboxylic acid ; the free hydroxyl groupof fangchinoline is therefore in an isoquinoline ring.63Steroid and Polyterpenoid Bases.-H. Rochelmeyer 64 proposes torename solanines t and s solatunine and solasonine, and the respec-tive aglycones solatubine and solasodine ; solancarpidine is identicalwith the last-named substance.Sola~odine,~~ now found to beC2,H430,N, yields cycbpentenophenanthene on fusion with selen-63 C. K. Chuang, C. Y. Hsing, Y. S. Kao, and K. J. Chang, Bey., 1939, 72,t14 H. Rochelmeyer, Awh. Pharm., 1937, 275, 336; H. Rochelmeyer and519.H. Chen, ibid., 1939, 277, 329; L. H. Briggs, Nature, 1939, 144, 247STEVENS : HETEROCYCLIC COMPOUNDS. 379ium, gives sterol colour reactions, and is precipitated by digitonin.It thus contains a steroid system with hydroxyl in the cis-3-position ;Medehydration affords the solanosodine obtained along with solasodinefrom solasonine by previous workers. Two active hydrogen atomsare present, the nitrogen is tertiary and unmethylated; and solaso-dine is provisionally formulated as (I). A very similar formula hadpreviously been proposed for s o l a t ~ b i n e , ~ ~ which undergoes trans-formations analogous to those of sterols.66Cevadine from Veratrurn sabadilla is the tiglic ester of cevine, thedegradation of which has been studied by W. A. Jacobs and L. C.Craig.67 Energetic decompositions afforded a series of volatile bases,including 5-methyl-2-ethyl-pyridine and -piperidine, and a dicyclictertiary base regarded as (11); coniine could not be identified.68Cevine methiodide yields a strongly alkaline base now believed to bea phenol- or enol-betaine. By distillation with soda-lime this gavea substance C8H15(OH)(NMe) probably derived from 5-methyl-2-ethylpiperidine ; Hofmann degradation appeared to give a cyclicether. It is suggested that cevine may have the partial structureOH -?*OH GO,H gO--CO(111).CH, CH, CH CH CH CQH2 ' \ O H 2 v p 2 /\/p y p 3 2 d,\C;;; yo MeCH N CH, MeCH N CH, CH2 C CH \ / I \ //H,C Me I: \ / \ / CH, CH,(11.) (111.) (IV.) OHThe oxidation of cevine throws some light on the nature of theremainder of the molecule, in which a benzphenanthrene system isalready indicated. An acid C14H1406 is obtained, termed decevinicacid, which is dehydrogenated by selenium to 2-hydroxynaphthalicanhydride. It is a lactonic acid giving the ferric chloride reaction61 0. R. Clemo, W. McG. Morgan, and R. Raper, J., 1936, 1299.66 Ann. Reports, 1936, 33, 378; H. Rochelmeyer, C. S. Shah, and E. Geyer,67 J . Biol. Chem., 1937, 119, 141; 120, 447; 1938, 124, 659; 125, 625;8 8 Compare A. K. Macbeth and (Sir) R. Robinson, J . , 1922, 121, 1571.\ / \ /CH, CH,Arch. Pharm., 1939, 277, 340.1939,129, 79; 1940, 134, 123; J . Amer. Chem. SOC., 1939, 61, 2252.Ann. Reports, 1936, 33, 376380 ORGANIC CHEMISTRY.of an enol ; aqueous alkali adds a molecule of water and successivelyremoves two of carbon dioxide, leading to a keto-lactone C,,H,,03.These and other reactions are interpreted by the tentative formula(IV) for decevinic acid.Cassaine, isolated from Erythophleum guineense, is the dimethyl-aminoethyl ester of the mono-olefinic keto-hydroxy-acid cassaic acid,C20H,004. By reduction, followed by fusion with selenium, this gave1 : 7 : 8-trimethylphenanthrene, and is hence a diterpenoid butprobably not a steroid derivative.70Xtrychnos Alkaloids.-The principal point against the mostattractive variant (I) of Robinson's formula for strychnine was theready bromination of diketonucidine (11) ; as such bromination doesnot after all take place, it is no longer necessary to attribute thegrouping *CO*eH- to diketonucidine. 71, 72 The transformations ofstrychnine and brucine have been discussed in detail in terms of theformula (I) .71CO CH\/\/?- FH\/\CH,CH, CHClH2-CH/ \-QH FH-CH CH\/\/ \N CH I(111.)Several papers have been devoted to the exhaustive methylationand Emde degradation of +-strychnine (I, with OH for H*) and+-brucine.73 The results, which could be accommodated on thebasis of (I) or the similar formula (111) of Leuchs, included dis-concerting but intelligible transferences of methyl groups betweenoxygen and nitrogen :J/ . I . :CH*C--% / :C=q ;yMeOMe OMe7O F. Faltis and L. Holzinger, Ber., 1939, 72, 1443 ; G. Dalma, H e h . C'him.7 1 H. L. Holmes and (Sir) R. Robinson, J . , 1939, 603.72 H. Leuchs and H. Grunow, Ber., 1939, 72, 679; H. Leuchs, ibid., p. 1588.7 3 H. Leuchs, Ber., 1937, 70, 2455; 1938, 71, 660; H. Leuchs andActu, 1939, 22, 1497; L. Ruzicka and G. Dalma, ibid., p. 1516.K. Tessmar, Ber., 1939, 72, 965STEVENS : HETEROCYCLIC COMPOUNDS. 381The Hofmann degradation of dihydrostrychnidine A (dihydro-I,with CH, for CO) has been carried to the point of elimination ofN(b) .74 The complications arising from stereoisomerism andmigration of double bonds can be interpreted by Robinson's or lesseasily by Leuchs's formula. Similar degradations have been effectedin the brucidine and vomicidine series.75+Strychnine has been degraded to @-indolylethylamine deriv-a t i v e ~ , ~ ~ and Leuchs has oxidised brucinonic acid to an acid for-mulated as (IV), in which only three of the alkaloid's seven ringssurvive./C,H*CH \NHCH CH*CO,H/ \ /HO,C*vH TH\ /CHVomipyrine, the degradation product CI5Hl6N2 of ~ o i i i i c i n e , ~ ~ isnot identical with either of the alternatives (V), both of which havebeen synthesised. Comparison of the absorption spectrum of vomi-pyrine with those of synthetic materials indicates, however, that theformer must be a pyrroquinoline derivati~e.'~PiZosinine.-N. A. Preobrashenski, A. M. Poljakova, and V. A.Preobrashenski have synthesised d2-isopilocarpine and -pilocarpine *Oby the method previously used for d-pilocarpine.81 The sameauthors have prepared pilosinine (11) in the same wayB2 frompilosinic acid (I), synthesised by heating ethyl itaconate :>CHQH,-G*CO,Et - ~H,-$lH*CO,Et ?H,--CH*CH,*$-NMeC0,Et CH, CO*O*CH, CO*O*CH, CH-N(11.)T. S. S.74 0. Achmatowicz and (Sir) R. Robinson, J., 1938, 1467 ; 0. Achmatowicz,ibid., p. 1472 ; 0. Achmatowicz and C. Dybowski, ibid., p. 1483.7 5 Idem, ibid., p. 1488; 0. Achmatowicz and B. Racinski, Rocz. Chem.,1938, 18, 315.'13 M. Kotake, T. Sakan, and S. Kusumoto, Sci. Papers Inst. Phys. Chem.Res. Tokyo, 1939, 35, 415.7 7 Ber., 1938, 71, 2237.79 H. Wieland and L. Hornor, Annalen. 1938, 536, 89; L. Horner, ibid.,*O Ber., 1936, 69, 1314, 1835.7 8 Ann. Reports, 1937, 34, 367-8.1939, 540, 73.81 Ann, Reports, 1935, 32, 331.J. Gen. Chem. Russia, 1939, 9, 1402

ISSN:0365-6217    

DOI:10.1039/AR9403700194

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

BIOCHEMISTRY.1. INTRODUCTION.IN the course of comparatively a few years biochemistry has sodeveloped that, from being a little-explored side-line of chemistryor physiology, i t has become itself a vast new tract for investigation,with numerous sub-divisions of its own, each of which in turn formsthe subject for intensive and specialised study by teams of re-searchers. Indeed so specialised have the various branches becomethat i t is felt that the time is ripe to make a new departure with thisissue of the Annual Reports, and to call in the aid of experts to dealwith recent developments in their own lines of work.Nutrition still accounts for by far the largest output of papers-no fewer than 5,127 abstracts are included in the year’s list inNutrition Abstracts and Reviews-and, following the custom of theseReports for a good many years past, i t is accordingly placed first inorder of attention.The increased importance of this subject intime of war needs no emphasis. Next follows an account, by J. R.Marrack, of some of the more remarkable of the recent findings inanother field of equally great practical importance, namely immuno-chemistry. Still considering “ the animal as a whole,’’ we turn thento a discussion of the chemical control of the body through theagency of the secretions of the ductless glands. The number of suchhormones already known exceeds that of the vitamins, and so withthe restricted space at his disposal E. Kodicek has wisely limited hisattention to one gland only, or a portion of it, namely, the anteriorpituitary.Turning now, in a sense, from the needs of the wholeanimal t o the chemistry of single components of its tissues, A.Neuberger deals with some significant recent advances in the studyof nitrogenous substances, and J. F. Danielli reviews certain under-lying physicochemical principles of concern to the biochemist. Theenzymes are substances of such universal significance for all livingtissues that an adequate understanding of their nature and theirmode of action is a prime necessity for the advance of all aspects ofbiochemical knowledge; this field is covered by D. J. Bell in thesection on biochemical catalysis. Leaving the animal kingdombehind, P. W. Norris, as in several recent years, contributes a sectiondealing in a general way with the chemical changes which occur inthe living plant. A special aspect of this, glucoside formation, isthen touched by R.Hill. Finally, coming to micro-organismsHARRIS : NUTRITION AND VITAMINS. 383(Miss) M. Stephenson and E. F. Gale devote their section on bacterialbiochemistry to two topics which have been much to the fore inrecent literature, namely, " accessory food factors " and nitrogenmetabolism.L. J. H.2. NUTRITION ANI) VITAMINS.Vitamin A .Human Requirements.-Interest has centred largely on thedetection of deficiency by the dark-adaptation test and themeasurement of the daily human requirement for the vitamin.Estimates of the latter seem to work out in reasonably good agree-ment. I n one particularly convincing experiment 2 ten volunteerswere deprived of the vitamin for 188 days, until their visual sensi-tivity was 1/28 of the normal, and it was then ascertained that2000-2500 I.U.of vitamin A, or 5000 I.U. of carotene, daily werethe smallest amounts which would bring about a slow improvementin adaptation, No vitamin A was present in the blood until thedose of 2500 I.U. of vitamin A was reached or exceeded. The fore-going estimate errs perhaps in being on the low side, since thereseems no doubt that larger intakes would be needed for a moreprompt cure. L. E. Booher and E. C. Callison3 found the require-ment to be about 4000 and 7000 I.U. of carotene, given as cookedpeas and cooked spinach respectively (recalculated to 70 kg.ofbody-weight) ; and W. v. Drigalski considers that about 9800 I.U.of carotene daily are needed, increasing to about double or treblein lactation or late pregnan~y.~ H. R. Guilbert et aZ.,6 after review-ing past data and comparing different mammalian species, cite1400-4200 I.U. for vitamin A and 2800--14,000 I.U. for carotene,the former figure in each instance being the minimum for freedomfrom clinical symptoms, and the latter the optimum for normaladaptation, growth, and significant storage. Guilbert points outthat the requirement for vitamin A remains relatively constant for asurprisingly large number of species, including the ox, sheep, pig,horse, and rat-namely, about 25 I.U. per kg. of body-weight forvitamin A, and about 50 I.U. for carotene.The differences betweenthe vitamin and the " pro-vitamin " are of course due t o the lesscomplete absorption or assimilation of the latter, which may alsoCompare Ann. Reports, 1939, 36, 338.J . Nutrition, 1939, 18, 459.Klin. Woch., 1939, 18, 1269.W. v. Drigalski and H. Kung, ibid., p. 1318.H. R. Guilbert, C. E. Howell, and G. H. Hart, J . Nutrition, 1940, 19, 91.a I(. H. Wagner, 8. physiol. Chem., 1940, 264, 163384 BIOCHEMISTRY.vary considerably according to the digestibility or the method usedfor cooking the vegetable containing the carotene.Conditioned DeJiciency in Man.-Several workers have confirmedthat deficiency of vitamin A, as detected by the dark adaptation test,is commonly found in diseases of the liver (e.g., cirrhosis, parenchy-matous liver disease).'* ** The deficiency runs parallel with thedegree of damage of the liver,8 and may be due either to inadequateabsorption (as in obstructive jaundice and steatorrhcea), or todeficient storage caused by the hepatic dysfunction, and sometimesalso to a lowered intake.9 I n juvenile diabetes dysadaptation iscommon, and the explanation is thought to be an inability to convertcarotene into vitamin A.l0 Further indications have been givenalso of a frequent correlation between renal calculus and deficiencyof vitamin A.11Incidence of DeJicienc9.-Various investigators agree that darkadaptation tests give a reliable index of " partial deficiency," pro-vided that the method is used with the proper precautions.G.Sankaran,12 examining 600 convich in a jail in Calcutta, found that43% of the lower-class groups and 17% of the better-class groupsshowed a high degree of deficiency.This may be compared withthe results of L. J. Harris and M. A. Abbasy,13 who found a com-paratively mild degree of deficiency present in about 30% of poorworking-class children in England.Testsfor DeJ'iciency in Man.-O. D. Abbott, C. F. Ahmann, andM. R. Overstreet l4 in U.S.A. and K. H. Wagner in Germany pointout that a differential leucocyte count is of value in diagnosingdeficiency. Wagner found that after 188 days on a deficient diet thethrombocytes, too, were 70-85% below normal, and there was aleucopenia with a " right shift " of the differential count. Thus,four distinct criteria are now available for assessing deficiency inman : (1) dark adaptation, (2) the level of vitamin A in the blood, (3)presence of cornified cells in scrapings from the bulbar conjunctiva,(4) blood counts.Experimental Avitaminosis A .-H.Mellanby l5 has recordeddefects in the dental structure of young rats when their mothers'A. J. Patek and C. Haig, J. Clin. Invest., 1939, 18, 609.8 M. G. Wohl and J. B. Feldman, J. Lab. Clin. Med., 1940, 25, 485.9 M. Salah, J. Egypt. Med. ASSOC., 1940, 23, 153.lo J. G. Brazer and A. C. Curtis, Arch. intern. Med., 1940, 65, 90; cf'. alsol1 H. Long and L. N. Pyrah, Brit. J. Urol., 1939, 216.l2 Ann. Report All-India Inst. Hygiene and Pub. Hlth., Calcutta, 1939,l3 Lancet, 1939, ii, 1299, 1355. l4 Amer.J . Physiol., 1939, 128, 254.15 Brit. Dental J., 1939, 67, 187.J . Clin. Invest., 1939, 18, 495.p. 33HARRIS : NUTRITION AND VITAMINS. 385diets were lacking in vitamin A. These include degenerated enamelorgans, poorly calcified dentine, ossifying cells in pulp, abnormalitiesin the mandible. According to M. E. Sauer,16 the epithelialmetaplasia of avitaminosis A is independent of any nerve change.C. A. Baumann and T. Moore 1' have been unable to confirm thetheory of a specific antagonism between thyroxin and vitamin A.Workers in dietetics will have noted with interest that the so-called" light-white casein,' ' frequently used in nutritional experiments, iscontaminated with appreciable amounts of vitamin A ; l8 also thatthe disproportionately high values suggested for cheese have notbeen confirmed.lg Reference must be made to a valuable review 2ocovering both chemical and physiological aspects of carotene andrelated pigments.Vitamin B,.Vitamin B, and the Pyruvate Oxidation 8ystem.--It is becomingclear that the system by which pyruvic acid is oxidised, under theinfluence of vitamin B, (as its pyrophosphate ester, cocarboxylase),may vary in some of its details from one tissue to another.Inbrain, or kidney cortex, C, dicarboxylates (e.g., succinate andfumarate) seem to be of importance, and other components of thesystem include inorganic phosphate, Mg++ (or Mn++), adeninenucleotide and probably cozymase (pyridine nucleotide) .21 Inpigeon breast muscle22 or l i ~ e r , ~ 3 on the other hand, the reactionappears to proceed through the citric acid cycle-i.e., citric acid isformed as an intermediate-and among the principal productsformed may be acetoacetate and a-ketoglutarate.The last-mentioned substance has been identified in large amounts in theurine of deficient rats,24 and the quantity of citric acid is said to risesteeply shortly after a cure.25An interesting development has followed the discovery of H. G.Wood and C. H. Werkman 26 that " propionic-acid bacteria " absorbcarbon dioxide. They suggested that the mechanism might involvethe addition of carbon dioxide to pyruvate to give oxaloacetate.H. A. Krebs and L. V. Eggleston27 have now propounded thel6 Anat. Rec., 1939, 74, 223.l8 M. K. Maitra and T. Moore, ibid., p.1648.A. W. Davies and T. Moore, ibid., p. 1645.R. A. Morton, Ghem. and Ind., 1940, 59, 301.1 7 Biochem. J . , 1939, 33, 1639.21 I. Banga, S . Ochoa, and R. A. Peters, Biochem. J . , 1939, 33, 1980.22 H. A. Krebs and L. V. Eggleston, ibid., 1940, 34, 442.23 E. A. Evans, ibid., p. 829.24 P. E. Simola, Biochem. Z., 1939, 302, 84.25 H. A. Sober, M. A. Lipton, and C. A. Elvehjem, J . Biol. Chem., 1940,134,26 Biochem. J., 1938, 32, 1262; 1940, 34, 7, 129.605.2 7 Ibid., p. 1383.REP.-VOL. XXXVII. 386 BIOCHEMISTRY.hypothesis that in certain aninial tissues, e.g., in liver, and inbacteria, the vitamin is concerned not in the direct oxidation ofpyruvate as hitherto supposed but in a preliminary carboxylation,analogous with the decarboxylation known to OCCLU in yeast.Thiscarboxylation results in the formation of oxaloacetate, and thelatter then passes through the citric acid cycle, as follows :(1) Carbon dioxide + pyruvate + oxaloacetate ;(2) 20xaloacetate + pyruvate --+ citrate + (fumarate ++(3) Citrate + oxaloacetate --+ a-ketoglutarate + (fumar-the net result being the production of a-ketoglutarate plus fumaratein equilibrium with oxalate, according to the following equation :(4) 4Pyruvate + carbon dioxide --+ a-ketoglutarate +However, this scheme does not appear to be of universal applic-ability; for example, these reactions do not occur in brain.There can be no doubt that the interconnections between pyruvateoxidation and other related systems are most complicated.Forexample, P. J. G. Mann and J. H. Quaste128 have shown that theproduction of acetylcholine from pyruvate is increased whenvitamin B1 is added to an aerobic preparation of polyneuritic brain.A further example of the complexity of the system is the finding thatthe oxidafion of acetate by bacteria requires the participation of thevitamin.28a Again, the observation 29 that insulin, like vitamin B,,is able to lower the level of pyruvate in normal human blood suggeststhat the glycogen-glucose equilibrium and glucosemonophosphateformation are connected, if only indirectly, with the conversion ofpyruvic acid.Vitamin B, and Phosphorykating Mechanisms.--It is now clearlyestablished that a cycle of phosphorylation accompanies the enzymicoxidation of pyruvate. I n a preparation of brain, the r61e of adenylicacid seems to be to carry phosphate to hexosemonophosphate to formhexosediph~sphate.~~ Phosphoglyceric and phosphopyruvic acidsmay have similar functions.21 F.Lipmann’s experiments 31 withB. ctelbriickii suggest that the unstable acetyl phosphate may be anintermediate phosphorylated derivative in the oxidation of pyruvicacid. As is well known, vitamin B, itself becomes phosphorylated inmalate) + carbon dioxide ;ate f3 malate) + carbon dioxide;2(fumarate +-+ malate).28 Nature, 1910, 145, 850.38a J. H. Quastel and D. M. Miebley, ibid., 1939, 144, 633.H. v. Euler and B. Hogberg, Naturwiss., 1940, 28, 29.30 S. Ochoa, Nature, 1940, 145, 747; cf. also 146, 267.31 J. Biol.Ckein., 1940, 134, 463HARRIS : NUTRITION rWD VITAMINS. 387the organism to form the biologidy active cocarboxylase, andrtccording to H. Weil-Malherbe 32 it does so through the agency ofadenine pyrophosphate ; phosphoglyceric acid also may acceleratethe action.Activation of Comrboxyluse by Free Vitumin B,.-Some two yearsago S. Ochoa 34 made the surprising observation that cocarboxylaseactivity is stimulated by the presence of free vitamin B,. Anelegant piece of work by H. G. K. Westenbrink and D. A. vanDorp 35 has now proved that the aneurin does this by inhibiting thehydrolysis of cocarboxylase by the phosphatases present in thetissue juices-Le., as a product of the reaction, it slows the net rateof the reaction.Assessment of the Level of Ntxlrition.-The suggestion was made byG.G. Banerji and L. J. Harris 36 that the diminished tolerance to-wards the intermediates of carbohydrate metabolism observed invitamin B, deficiency could be used as a test for assessing the levelof nutrition. In rats the urinary excretion of pyruvate, or its levelin the blood, is proportional to the degree of the deficiency; and thisfinding has been c o n b e d by other workers.37 For detecting partialdeficiencies in man, the procedure is slightly less simple, since itinvolves loading with test doses of glucose or pyruvate or otherproducts of intermediate metabolism, followed by examination forthe accumulation of abnormal amounts of such metabolites aspyruvic acid or a-ketoglutaric acid : the precise experimental detailsare still in process of being worked out?6*38 An alternative tech-nique is the direct determination of the vitamin in urine,39 and thismethod has been carefully standardised by Y.L. Wang and J.Yudkin,*O who have also calculated from their data the probablehuman requirements. Alternatively, vitamin B,, or cocarboxylase,can be measured in the bl00d.~1* 4232 Biochem. J., 1939, 33, 1997.33 M. A. Lipton and C. A. Elvehjem, Nature, 1940, 145, 226.3r Ibid., 1938, 141, 831; see also S . Ochoa and R. A. Peters, Biochem. J.,35 Nature, 1940, 145, 465.36 Chem. and Ind., 1938, 57, 1190; Biochem. J., 1939, 33, 1346.37 M. Shils, H. G. Day, and E. V. McCollum, Science, 1940, 91, 341; H. v.Euler and B. Hogberg, Naturwiss., 1939, 27, 769.38 G.G. Banerji, Biochem. J., 1940, 34, 1329; seepp. 1332-1333; K. 0.Elsom, F. D. W. Lukens, E. H. Montgomery, and L. J o m , J . C h . Inwest.,1940, 19, 153; W. D. Robinson, D. Melnick., and H. Field, ibid., p. 483.3* I;. 5. Harris and P. C. Leong, Lancet, 1936, i, 886; L. J. Harris, P. C.Leong, and C. C. Ungley, ibid., 1938, i, 539; Y . L. Wang and L. J . Harris,Biochem. J., 1939, 33, 1356.4 0 Ibid., 1940, 34, 343.4 1 R. Goodhart and H. M. Sinelah, J . Biob. Chem., 1940, 132, 11.4 2 R. Goodhart, ibid., 1940, 135, 77.1938, 38, 1501388 BIOCHEMISTRY.Methods of Assay.-The bradycardia method 43 has been usedfor large-scale surveys of foodstuffs, in the East Indies, China, andelsewhere, and found to agree with various other methods (ratgrowth, rice-bird, colorimetric-diazo, e t ~ .) . ~ ~ R. Goodhart 42 hasrecorded modifications in Sinclair and Goodhart's method forcocarboxylase in blood, H. M. Sinclair 45 in the Schopfer-Meiklejohn" fungus " test for B, in blood, and H. G. K. Westenbrink 46 in theOchoa-Peters reaction for B, and cocarboxylase in yeast.Vitamin-B, Complex.Nicotinic Acid (Pellagrca-preventive Factor) .-E. Kodicek 47 hasmade a valuable study of the method for estimating nicotinic acidwith p-aminoa~etophenone.~~a Both he and W. R. Aykroyd andM. Swaminathan 48 have compared the distribution of the vitaminin various cereals and draw attention to the rather surprising factthat rice and millet may have little or no greater activity than maize,notwithstanding the well-knownassociation of the latter withpellagra.I n keeping with the known physiological function of nicotinic acidas a constituent of the pyridine coenzymes (phosphopyridine nucleo-tides), it has been proved that a deficiency, in either dogs or pigs,causes a fall in the amount of coenzyme I in their livers andmuscles.49 The so-called Factor V (needed for cultivation ofHcemophilus parccinJluenxce, and served by either the di- or tri-phosphopyridine nucleotides) is similarly said to be lowered in thelivers and muscles-but not in other tissues-of deficient dogs.50Human erythrocytes are able to synthesise Factor V in vitro whenincubated with the vitamin, or in vivo when the latter is ingested bymouth.51 The biochemical r6le of the pyridine and other coenzymeshas been much studied (see, for example, M. Dixon and L.G.Zerfas 52), and is discussed more fully elsewhere in this Report4 3 T. W. Birch and L. J. Harris, Biochem. J., 1934, 28,602 ; cf. Ann. Reports,1939, 36, 340.44 S. J. E. Pannekoek-Westenburg and A. G. van Veen, Geneeslc. Tijdschr.Neder1.-Indie, 1939, 79,2891 ; E. F. Yang and B. S. Platt, Chinese J . Physiol.,1939, 14, 259; A., 1940, 111, 233; cf. also D. G. H. MacDonald and E. WMcHenry, Amer. J . Physiol., 1940, 128, 608.46 Biochem. J., 1939, 33, 2027.4 6 Enzymologia, 1940, 8, 97.4 7 Biochem. J . , 1940, 34, 712, 724.47a L. J. Harris and W. D. Raymond, ibid., 1939, 33, 2037.4 a Indian J . Med. Res., 1940, 27, 667.4D A. E. Axelrod, R. J. Madden, and C. A.Elvehjem, J . Biol. Chem., 1939,50 H. I. Kohn, J. R. Klein, and W. J. Dann, Biochem. J., 1939, 33, 1432;61 H. I. Kohn and J. R. Klein, J . Biol. Chem., 1940, 135, 685.52 Biochem. J., 1940, 34, 371.131, 85.M. Pittman and H. F. Fraser, Publ. Health Reps., Wash., 1940, 55, 915HARRIS : NUTRITION AND VITAMINS. 389(see p. 415) and in a recent review by J. H. Q ~ a s t e l . ~ ~ a Of clinicalimportance is the suggested use of nicotinic acid for treatment ofVincent’s disease (trench mouth) ,s3 in sulphanilamide poisoning, 54and in the so-called “ encephalopathic syndrome.” 55Riboflavin.-Deficiency of riboflavin in man is marked not onlyby lesions of the lips and seborrhmic accumulations on the face,56but by ocular manifestations, particularly photophobia and keratitis.These corneal lesions can be cured or made to reappear a t will byadministering or withholding riboflavin.57 Similarly in rats, inaddition to the skin lesions,58 ocular symptoms are prominent :in order of onset these comprise conjunctivitis, blepharitis, cornealopacity, vascularisation and ulceration of the cornea, and finallycataract.All except the cataract are cured by riboflavin withoutany other treatment.59 Flavin is important, likewise, in the rearingof pigs 6o and poultry.61 A method for checking the biological activityof various flavins, using lactic acid bacteria, has been described.62Pantothenic Acid (Chick-pellagra Factor, Filtrate Factor) .-Thestructure of pantothenic acid has now been settled.63 The syntheticdextrorotatory compound 6 5 p 66 has been testcd on micro-organisms,on chicks and on rats and found to have the identical activity of thenaturally occurring substance : 65* 67 the racemic acid has 50% of the5 2 , ~ Lecture delivered before the Institute of Chemistry of Great Britain and53 J.D. King, Lancet, 1940, 2, 32.54 J. F. Doughty, J . Amer. Med. ASSOC., 1940, 114, 756; G . B. Cottini,Dermatologica, 1940, 81, 83.6 6 N. Jolliffe, Res. Publn. Ass. neru. ment. Dis., 1939, 19, 148; N. Jolliffe,K. M. Bowman, L. A. Rosenblum, and H. D. Fein, J . Amer. Med. ASSOC.,1940, 114, 307.5 6 W. H. Sebrell and R. E. Butler, Publ. Health Reps., Wash., 1939, 54,2121 ; N. Jolliffe, H. D. Fein, and L. A. Rosenblum, New Eng7. J . Med., 1939,5 7 H.D. Kruse, V. P. Sydenstricker, W. H. Sebrell, and H. M. Cleckley,5 8 H. Chick, T. F. Macras, and A. N. Worden, Biochem. J . , 1940, 34, 580.59 M. M. El-Sadr, Chem. and Ind., 1939, 58, 1020.6o E. H. Hughes, J . Nutrition, 1940, 20, 233.6 1 A. E. Schumacher and C. F. Heuser, Poultry Sci., 1939, 18, 369.62 E. E. Snell and F. M. Strong, EnzymoEogia, 1939, 6, 186; R. E. Feeney63 See this vol., p. 226.64 R. J. Williams, H. K. Mitchell, H. H. Weinstock, and E. E. Snell,6 5 E. T. Stiller, S. A. Harris, J. Pinkelstein, J. C. Keresztesy, and K. Folkers,6 6 D. W. Woolley, ibid., p. 2251.6 7 H. H. Weinstock, A. Arnold, E. L. May, and D. Price, Science, 1940, 91,411 ; S. H. Babcock and T. H. Jukes, J . Amer. Chem. SOC., 1940, 62, 1628.Ireland, 4th October, 1940.221, 921.Publ. Health Reps., Wash., 1940, 55, 157.and F.M. Strong, J . Biol. Chem., 1940, 133, proc. xxxi.J . Anaer. Chem. Soc., 1940, 62, 1784.ibid., p. 1785390 BIOCHEMISTRY.activity, and (-)pantothenic acid is inactive ; 65 the hydroxy-derivative 68 has a variable potency depending 011 conditions, butvarious other synthetic analogues are biologically inert.69The pathological and physiological relations of pantothenic acidare now receiving attention. In a deficiency results indegeneration of the nerve fibres of the spinal cord. I n rats,71 aswell as the better known symptoms 58 such as nose bleeding, stickyexudate on eyelids, and depilation about the nose, the adrenals aresaid to suffer hsmorrhage, atrophy, and necrosis.The tissues ofdeficient chicks have been shown to be low in the vitamin; 72 anda method has been suggested for estimating i t in human bl00d.73The conclusion given in last year’s Report about the identity ofthe “ filtrate factor,” needed by rats, with pantothenic acid has beenfurther confirmed.74 Associated with the filtrate factor are othersubstances also needed by rats, vix., the so-called fi- and y-factorsof the British workers74 and the Factor W of Elvehjem and hiscollaborators. 75’Vitarmin C.DeJiciency in Man.-Working details have been given of thesimplified procedure for assessing the level of nutrition of vitamin Cin human subjects, by means of the urine test : controls a t a homefor waifs and strays in Cambridge where an orange was provideddaily were found to be up to standard, whereas in poor-class homesin the same town some 40% of the children were “ below standard ”and 5% had a relatively severe defi~iency.~~ When it is merelydesired to distinguish in a rough qualitative way between childrenwith high and with low “ reserves,” it is possible to use a still simplermodification of the test.77 In polar regions, the large consumptionof fresh meat, and a t some seasons of arctic flora, generally keeps the8 8 H.K. Mitchell, E. E. Snell, and R. J. Williams, J. Amer. Chem. SOC.,69 H. H. Weinstock, E. L. May, A. Arnold, and D. Price, J. Biol. Cltem.,70 P. H. Phillips and R. W. Engel, J. Nutrition, 1939, 18, 227; quoted by71 F. S. Daft, W. H. Sebrell, S. H. Babcock, andT.H. Jukes, Publ. Health73 E. E. Snell, D. Pennington, and R. J. Williams, J . Biol. Chem., 1940,1940, 62, 1791.1940, 135, 343.Brit. Med. J . , 1940, 2, 230.Reps., Wash., 1940, 55, 1333; L. L. Ashburn, ibid., p. 1337.133, 559.S. R. Stanbery, E. E. Snell, andT. D. Spies, ibid., 1940, 135, 353.74 B. Lythgoe, T. F. Macrae, R. H. Stanley, A. R. Todd, and C. E. Work,76 S. Black, D. V. Frost, and C. A. EIvehjem, J . Biol. Chem., 1940, 132,76 L. J. Harris, Lancet, 1940, ii, 259.77 J. Pemberton, Brit. Med. J., 1940, 2, 217.Bwchem. J., 1940, 34, 1335.65; J. J. Oleson and S. Black, ibid., 1940, 133, proc. lxxiiiHARRIS : NUTRITION AND VITAMINS. 39 1people free from scurvy,78 but certain Eskimos on a diet devoid ofmeat were found to be deficient in the vitamin when tested either forthe ascorbic acid in their blood or for capillary resistance-bothcould be improved if orange juice were given.79 Other signs ofdeficiency included rhinitis, edematous nasal mucous membranes,epistaxis and increased severity of the cutaneous response totuberculin.79- 8o The increased importance of an adequate supplyof vitamin C: in war-time is illustrated by the fact that the healingof wounds and the formation of callus are delayed by partial de-ficiencies.81 Those working on the apparent connection betweenvitamin C and resistance to infection may receive further clues fromthe observation concerning the tuberculin reaction mentioned above,and from the finding of the workers a t Pittsburg that the administra-tion of hypnotics such as barbiturates or chloretone caused anincreased synthesis of the vitamin by rats, attributable, it is pre-sumed, to its connection with detoxication processes.82Properties and Distribution of Ascorbic Acid.-A valuable compil-ation has been made by M.Olliver g3 of the distribution of vitamin Cin numerous raw and cooked fruits and vegetables. C. L. Arcus andS. S. Zilva 84 have examined the photochemical decomposition ofascorbic acid and conclude that “ it is improbable that any ultra-violet light penetrating superficial tissues containing Z-ascorbic acidwould bring about the oxidation of the vitamin in v~vo.’’Other Vitamins.Owing to lack of space this year, a discussion of recent work onvitamins D and E will have to be postponed until the next volume.Vitamin K.-The formula given provisionally for vitamin K, in0 Vitamin K ,last year’s Report has been modified.Instead of two farnesylresidues, one each in the 2- and the 3-position, it is now thought that7 8 K. Rodahl, private communication.7Q V. E. Levine, J . Biol. Chem., 1940, 133, proc. lxi.8O A. Hoygaard, Nord. med. Tidsskr., 1938,16, 1647. For further referenceto the tuberculin reaction in vitamin C deficiency, see also K. M. Birkhaug,Actu tuberc. Scand., 1939, 13, 45; quoted by V. E. Levine (loc. cit.).81 E. W. Lexer, Arch. klin. Chir., 1939, 195, 611.82 H. E. Longenecker, H. H. Fricke, and C. G. King, J . Biol. Cliem., 1940,135, 497.Lancet, 1940, ii, 190.84 Biochem. J., 194-0, 34, 61392 BIOCHEMISTRY.the two are combined, joined head to tail, in the 3-position.85According to this view vitamin K, is 3-difarnesyl-2-methyl-1 : 4-naphthaquinone. This brings K, into line with all the more potentK-factors, in having a methyl substituent in the 2-position.“ Vitamin P.”-The claims for “ vitamin P ” have again beencontroverted, so far as concerns the guinea pig; 86 for the humansome clinical observers record positive and others negativeresults.88 We may reasonably hope for an explanation of thediscrepancy within a year or two.Vitamin F (Essential Unsaturated Fatty Acids) .-The previousindications given by T.W. Birch and his co-workers of inter-relationsbetween vitamin B, and the nutritionally essential fatty acids havebeen confirmed : symptoms of so-called acrodynia in the rat may,it seems, be cured by either factor independently.89 Possibly,however, the symptoms in the two deficiencies are similar but notidentical; indeed G .0. Burr, J. B. Brown, and W. 0. Lundbergsay that the unsaturated fatty acids themselves (linolenic, linoleic,arachidonic, etc.) should not be treated as an interchangeable group,since they differ appreciably in their effects on growth and on theskin lesions. Interesting studies of the relative potencies of variousesters of these fatty acids, and of their effects on fat metabolism,have been made by (Mrs.) I. Smedley-Maclean and her colleagues atthe Lister I n ~ t i t u t e . ~ ~Vitamin H.-Vitamin H is the substance which protects ratsagainst “ egg-white injury,’’ that is to say, against the symptomsproduced by consumption of large amounts of raw or of insufficientlycooked egg-~hite.~la Investigations on this vitamin were begunby P.Gyorgy in Heidelberg in 1927, were resumed with T. W. Birchin Cambridge from 1933-1935, and have since been continued inconjunction with others in America. A full resume of all theseextended experiments has appearedeg2 One of the most interesting8 6 S. B. Binkley, R. W. McKee, S. A. Thayer, and E. A. Doisy, J . Biol.88 L. E. Detrick, M. S. Dunn, W. L. McNamara, and N. E. Hubbard, J. Lab.8’ H. Scarborough, Biochem. J., 1939, 33, 1400; Lancet, 1940, ii, 644;8 8 E.g., E. Davis, ibid., p. 1062.*g H. Schneider, H. Steenbock, and B. R. Platz, J. Biol. Chem., 1940,132,539.QO Proc.SOC. Exp. Biol. Med., 1940, 44, 242.91 E. M. Hume, L. C. A. Nunn, I. Smedley-MacLean, and H. H. Smith,Biochem. J., 1940, 34, 879; I. Smedley-MacLean and L. C. A. Nunn, ibid.,p. 884; G. C. Hevesy and I. Smedley-MacLean, ibid., p. 903.Chem., 1940, 133, 721.Clin. Med., 1940, 25, 684.D. R. Gorrie, ibid., i, 1005.glc M. A. Boas, ibid., 1927, 21, 712.92 P. Gyorgy, J. Biol. Chem., 1939, 131, 733; P. Gyorgy, R. Kuhn, andE. Lederer, ibid., p, 745; T. W. Birch and P. Gyorgy, ibid., p. 761HARRIS : NUTRITION AND VITAMI". 393points is that vitamin H as it occurs in yeast or liver is insoluble inwater or fats, but can be made soluble in water by autolysis (in thecase of yeast) or by a suitable process of digestion (in the case ofliver).It is three to five times more effective when given parenter-ally than by mouth.93 Lastly the important suggestion has beenmade that vitamin H is probably identical 947 95 with biotin ("biosI1 B ,')-a factor needed for the growth of yeast and by certainstrains of X. aureus-and also with " Coenzyme R " 94-a growthfactor needed by nodular organisms in certain plant roots.New Vitamins.-The casual reader of current literature on vita-mins will probably be mystified by allusions to such novel terms asthe grass-juice factor, the anti-grey-hair factor, vitamins L and M,etc. Most of these newly described substances are as yet but poorlycharacterised chemically, but for enumeration and classificationreference may be had to a recent review.96Nutrition in War-time.In Britain the nutritional problem is, in essence, how to secureadequate nutrition while reducing the importation of food to theminimum.The solution lies, mainly, in an increased reliance oncommodities already produced in large quantities a t home. Potatoesand milk come foremost in this category. Together these two foodsconstitute a regimen which is not far from complete as regards mostof the vitamins, mineral salts, and protein; and enough of each isavailable to furnish a considerable proportion of the daily foodrequirements of the p o p ~ l a t i o n . ~ ~ Examples of two other home-produced foodstuffs which are being encouraged are carrots, valuableas a source of vitamin A, and oatmeal, also a native crop and goodfor cheap calories and for vitamin B,.The " nutrition front " mustbe of supreme interest to every chemist and biochemist just now, asan instance where applied chemical and biochemical knowledge hasbecome of fundamental importance to the national life. The theme,unfortunately, cannot be more than vaguely outlined here, and wemust content ourselves with a cross reference to some more detailedrecent writings.9*93 P. Gyorgy and C . S . Rose, Proc. SOC. Exp. Biol. Med., 1940, 43, 73.s4 P. Gyorgy, D. B. Melville, D. Burk, and V. du Vigneaud, Science, 1940,9 5 J. R. Porter and M. J. Pelczar, ibid., p. 576.O6 L. J. Harris, Post Grad. Med. J., 1941, 17, 34.9 7 For discussion on the potato as a war-time food, see R. N. Salaman andothers in contributions to a symposium reported in Chem. and Ind., 1940, 59,735 et seq.98 E.g., Sir J.B. Orr and D. Lubbock, " Feeding the People in War-Time,"1940; L. J. Harris, PTactitioner, 1940, 145, 105; J. C. Drummond, Sir R.91, 243394 BIOCHEMISTRY.White versus Brown Bread.-Nutrition workers are unanimous inrecognising wholemeal flour, or flour of a similar high degree of“ extraction,” as superior to white flour. The proposal to fortifywhite bread by adding to it crystalline aneurin plus a calcium salt,stated to be the carbonate, has been discussed by many a~thors.9~~ 100The addition of vitamin B, and calcium does not, however, makegood all the deficiencies of white flour as compared with wheatmeal,and this has been convincingly confirmed in growth tests on rats byH.Chick.lol L. J. H.3. IMMUNOCHEMISTRY.Bacterial Antigens.have comparedthe antigens isolated from members of the XaZmoneZZa group(Bact. Typhi-muriunz and Bact. Tgphosus) by the methods ofA. Boiviii and L. Mesrobeanu,2 and R. Raistrick and W. W. C. Topley3and by extraction with ethylene glycol and diethylene glycol.4 Theyalso prepared the antigens from bacteria grown on a syntheticmedium. J. Walker used a method of extraction with concentratedsolutions of urea. The four methods gave very similar products.G. M. Mackenzie, R. H. Pike, and R. E. Swinney ti also found thatsimilar products were obtained by the methods of Raistrick andTopley and of Boivin and Mesrobeanu. It may be supposed thatthe antigenic material obtained by these diverse methods exists as acomplex in the bacteria and is not a chance mixture of variouscontents of the bacterial cell and the culture made.extracted from Bcact.Dysentericc! (Shiga) an antigenG. G. Freeman, S. W. Challinor, and J. WilsonMorganMcCarrison, Sir J. B. Orr, Sir F. Keeble, L. H. Lampitt, V. H. Mottram, J. C.Spence, and F. Kidd, “ The Nation’s Larder . . .,” 1940; G. Bourne, “ Nutri-tion and the War,” 1940; Sir J. Russell, Nature, 1940, 145, 11. See also thevaluable tabulations by A. L. Bacharach and J. C. Drurnmond, Ghem. and Ind.,1940, 59, 37, gn the human requirements for vitamins, expressed both asmarginal and as optimal limits, and by A. L. Bachrrach, Food, 1940, 9, 110,on the contributions made by the more important foodstuffs towards theserequirements.gg X.g., Accessory Food Factors Committee, Lancet, 1940, ii, 143; W.C.McCulloch, Brit. Med., J., 1940, 2, 397; J. P. McGowan, ibid., p. 398; (Sir)E. Graham-Little, Lancet, 1940, ii, 311 ; A. L. Bacharach, Pood Manufacture,1940, 15, 220; E. R. Dawson, Chem. and Ind., 1940, 59, 784.100 T. Moran and J. C. Drummond, Nature, 1940, 146, 117.101 Lancet, 1940, ii, 511. Biochem. J., 1940, 34, 307.2 Compt. rend. SOC. Biol., 1933, 112, 76.3 Brit. J . Exp. Path., 1934, 15, 114.6 Ibid., 1940, 34, 325.W. T. J. Morgan, Biochem. J . , 1937, 31, 2503.J . Bact., 1940, 40, 197MARRACK IMMUNOCHEMISTRY. 395(ABC) composed of a phospholipin (A), a polysaccharide (B), and apolypeptide (C). considerthat the product is a homogeneous complex, which can induce theformation of antibodies like the intact organism.If A is removed,BC is still antigenic (induces formation of antibodies on injection),but A alone and AB are not. C on injection into rabbits gives riseto antibodies that react with C or BC ; however, these do not appearto be the same as those evoked by BC or the intact organism. BCcan be regenerated from isolated B and C. The purified poly-saccharide B forms precipitates with antisera to the bacteria.To obtain a bacterial antigen with the minimum disintegrationA. A. Miles and N. W. Pirie extracted Br. Melitensis with chloroformwater. They obtained a substance [PLAPS], from which a phospho-lipin PL, and a protein-like substance S could be successively re-moved, leaving AP.Hydrolysis with weak hydrochloric acid splitAP, liberating a phospholipin and an N-formylated amino-poly-hydroxide compound A, which may be a formylated amino-hexose.[PLAPS], which forms viscous opalescent solutions that showanisotropy of flow, is split on treatment with sodium dodecyl sulphateinto smaller particles, PLAPS, with molecular weights between lo5and lo6. AP has a molecular weight of about lo6, which can bereduced to between 1 and 2 x lo5. The amine, freed from formicacid, is fairly homogeneo~s,~ with a minimum particle weightof 3300.The ability to form antibodies on injection (antigenicity) is in theorder [PLAPS] > PLAPS > AP; the antigenicity of AP may bedue to contamination with traces of PLAPS. [PLAPS] forms moreprecipitate with antisera than PLAPS or AP.A is not antigenic andforms no precipitate with antisera. These substances, when addedin excess, inhibit the agglutination of Br. Melitensis by homologousantisera. The concentrations required for inhibition are in theorder Free amine > A > [PLAPS] > AP.These observations illustrate with single natural antigens thedependence of antigenicity and the formation of precipitates withantisera on the degree of aggregation and complexity of the complex,and of the specificity on a relatively simple fraction of the antigen.Natural Protein Antigens.W. T. J. Morgan and S. J. PartridgeCatalase.-Antisera to crystalline ox catalase have been obtainedDog and horse catalase also formed pre-Biochem.J., 1940, 34, 169.Brit. J . Exp. Path., 1939, 20, 83, 109, 278; Biochem. J . , 1939, 33, 1709,by immunising rabbits.1°1716.* J. St. L. Philpot, Biochem. J., 1939, 33, 728.lo D. A. Campbell and L. Fourt, J . Biol. Chern., 1939, 129,383396 BIOCHEMISTRY,cipitates with the antiserum ; but haemoglobin and haematin neithergormed precipitates with the antiserum nor inhibited the formationof a precipitate by catalase and antiserum. The specificity of thecatalase is, therefore, not determined by the iron-porphyrin group.When the catalase was precipitated in combination with antibody,its enzyme activity was unimpaired ; the specific groups that com-bine with antibody are not the same as those involved in the enzymeactivity.Bence Jones Proteins.-L.Hektoen and W. H. Welker l1 haveconfirmed conclusions from previous work that Bence Jones proteinsfall into two groups, which are immunologically distinct. Both typesmay be excreted by the same patient. They report that Medes hasfound the nitrogen distribution similar in all but one of their prepar-ations. The immunological behaviour of this one resembled thatof other preparations from which i t - differed chemically. Thissuggests that the specificity of these proteins is determined by asmall part of the protein molecule.L. Pillemer and others l2 have studied the specificity of keratinderivatives. They find a difference between feather and woolkeratin and consider that the S*C*CO,H group is of prime importancein determining specificity.Synthetic Antigens.W.F. Goebel l3 has continued his investigations 14 on the r61eof uronic acids in the immunity to pneumococci. He found thatsynthetic antigens containing gentiobiuronic and cellobiuronic acidslinked to horse serum globulin (GeA- and CA-globulin) evoked inrabbits antibodies which conferred passive immunity on miceagainst multiple doses of Type I1 pneumococci. However, antiserato GeA-globulin did not protect mice against Type I11 and TypeVIII pneumococci, as did antisera to CA-g10bulin.l~ The specificpolysaccharides of Type I11 and Type VIII pneuniococci cont,aincellobiuronic acids; the uronic acid constituent of Type I1 poly-saccharide is still unknown. The cross reactions of GeA- and CA-globulins and of similar compounds containing gentiobiose werestudied.The point of attachment of the glucose or glucuronic acidto the glucose seemed of more importance than the presence of acarboxyl group.Further investigations have been made into the effects of sub-stitution in the amino-group of proteins on their immunologicalbehaviour. If the amino-group of a protein P, (e.g., horse serumglobulin) is treated to form P,R, will (1) P,R still form a precipitatel1 Biochem. J., 1940, 34, 487.l3 Ibid., 1940, 72, 33.12 J . Exp. Med., 1939, 70, 287.14 Ibid., 1939, 69, 33MARRACK : IMMUNOCHEMISTRY. 397with antisera to the untreated protein P,, (2) P, form a precipitatewith antisera to P,R, (3) the product P2R, formed with anotherprotein P, (e.g., egg albumin), form a precipitate with antisera toP,R, and (4) the formation of a precipitate by P,R and antisera t oP,R be inhibited by the relatively simple compounds (AR) formedby treating amino-groups of amino-acids ‘1Formaldehyde Treatment (P,F).-L.J. Jacobs and S. C. Gommers l5compared the effects of the various methods of treatment withformaldehyde that have been used by previous workers. Theyfound that P,F formed precipitates with antisera to P,, and P, withantisera to P,F; P2F formed little or no precipitate with antiserato P,F.A. B. Kleczkowski,lG using quantitative methods in place of thetraditional inspection and rows of + signs, found that P,F pre-cipitated all the antibody from antisera to p,.Kleczkowski found that P,U alsoprecipitated all the antibody from antisera to P,.P,U precipitatedonly part of the antibody from antisera to P,U, which did not reactwith P,.Diethyl sulphide and diethylsulphone proteins (PMS and PMSO).L. Berenblum and A. Wormall 18 treated horse serum globulinwith pp’-dichlorodiethyl sulphide and pp’-dichlorodiethylsulphone,forming PIMS and P,MSO. P,MS formed a slight precipitate withantisera to P,, and P, with antisera to P,MS. P2MR and P,MSOreacted weakly with antisera to P,MS and PIMSO respectively.Carbobenxyloxyproteins (PC). P2C formed precipitates withantisera to P,C. This reaction was inhibited by carbobenzyloxy-glycine and the rather similar phenylureidoglycine .19G . C. Butler, C. R. Harington, and M. E.Yuill (working with Miles) 2O found that P,A reacted with antiserato P,A; the reaction was inhibited by aspirylglycine and to a lessdegree by salicylglycine.These experiments show that the specificity of proteins is littleaffected when the amino-group is changed by treatment with form-aldehyde ; but that when larger groups are attached a t this site, thespecificity is altered.They suggest that the amino-groups have littlerelation to the immunological behaviour of proteins and that when anew specificity is introduced i t is not due to loss of the free amino-group but to the presence of large new groups.Phenylureidoproteins (PU) .17Aspirylproteins (PA).l5 J . Immun., 1939, 36, 531.l7 S. J. Hopkins and A. Wormall, Biochem. J., 1933, 27, 740; 1934, 28,l 8 Ibid., 1939, 33, 75.l9 W.E. Gaunt and A. Wormall, ibid., p. 908.2o Ibid., 1940, 34, 838.l6 Brit. J . Exp. Path., 1940, 21, I.228398 BIOCHEMISTRY.According to A. B. Kleczkowski,21 in the course of iodination ofhorse serum globulin the ability to react with and the ability toevoke antibodies to native globulin disappear when all the tyrosinis substituted.Butler, Harington, and Yuill 2o also found that injection of anti-sera to P,A reduced the antipyretic action of aspirin (compare asimilar effect with thyroxin; Chem. Reviews, 1938, 356).Antibodies.The recognition that antibodies are modified serum globulins andthe new methods of characterising proteins have led to considerablework on the physical properties of antibodies. A. Tiselius 22 foundthat normal serum globulin consisted of three fractions, whosemobilities in an electric field a t pH between 7 and 8 were in the ordera > p > y.Horse antisera may contain either an excess of they-fraction or a new fraction, T, with a mobility between those of thep- and the y-fraction. When the antibody is removed by absorp-tion with the corresponding antigen, the y-fraction is reduced or they-fraction removed. In the sera of men, monkeys, and rabbits,antibodies were found mainly in the y-fraction.22s 23, 24According to J. van der Scheer, R. W. G. Wyckoff, and F. H.Clarke25 all horse antisera to toxins contain the new fraction T,whereas horse antisera to bacterial polysaccharides and proteinscontain abnormal amounts of the y-fraction.Some antitoxic seracontain both T-fraction and excess of y-fraction. These authorssuggest that these two fractions may cont'ain antibodies to twodifferent antigens present in the complex substance used for im-niunisation. Pappenheinier, Lundgren, and Williams 39 also foundthat diphtheria antitoxin formed a new serum globulin fraction.However, Tiselius and Kabat 23 found that the antibody of horseantisera to pneumococcal polysaccharide was in the T-fraction, andN. Fell, W. G. Stern, and R. D. Coghill 26 found no abnormal fractionin antisera to various bacterial toxins.Using the ultra-centrifuge, Biscoe, Hercik, and Wyckoff 24 andH. Heidelberger and K. 0. Pedersen2' showed that purified anti-bodies, from horse antisera to the pneumococcal polysaccharides,had very high sedimentation constants, whereas antibodies of rabbitantisera both to pneuinococcal polysaccharides and to proteins hadsedimentation constants close to those of normal serum globulin.E.A. Kabat 28 extended this work; he contrasted the antibodies21 Brit. J. Exp. Path., 1940, 21, 98.23 A. Tiselius and E. A. Kabat, J . Ezp. Med., 1939, 69, 119.24 J. Biscoe, F. Hercik, and R. W. G. Wyckoff, Science, 1936, 83, 602.25 J. Immun., 1940, 39, 65 (references are given to previous work).26 Ibid., p. 223.22 Biochem. J., 1937, 31, 1464.2 7 J . Exp. Med., 1937, 65, 393. 28 Ibid., 1939, 69, 103MARRACK : 1MMUNOCHEMISTRY. 399found in the sera of horses, cows, and pigs, which are very large(molecular weight about 990,000) and unsymmetrical, with those ofthe sera of men, monkeys, and rabbits which resemble normal serumglobulin (molecular weight 157,000 to 195,000).A. M. Pappen-heimer (jun,), H. P. Lundgren, and J. W. Williams,2g however,found that the molecular weight of diphtheria antitoxin in horseserum is about 155,000; and Fell, Stern, and Coghill,26 that horseantisera to various bacterial toxins contain no proteins of largemolecular weight. Kabat 28 also found that, on prolonged immnnis-ation of horses against pneumococci, smaller antibody moleculeswere formed. So that it seems that there is no inherent differencebetween antibodies either in molecular weight or in charge, dependenton the species in which they are formed or on the antigen with whichthey react.Enzymic Digestion.-Quantitative experiments were made byA.M. Pappenheimer (jun.) and E. S. Robinson 30 with the diphtheriaantitoxin prepared by the Parfentjev digestion process. Thesesuggest that the antibody molecules are split by the enzyme into anactive and an inactive part. Similar work by P. Grabar 31 indicatesthat the antibody to pneumococcal polysaccharides is split by pepsininto two approximately equal parts, one active and the otherinactive. C. G. Pope 32 studied the effect of varying conditions, p,,time of digestion, etc., on the purification of antibody by digestionwith proteolytic enzymes. He finds that antibodies are readily splitby digestion in a very short time into an inactive portion, whichis easily denatured by heat, and an active portion more resistantto heat.Ordinary antitoxin contains 60,000-90,000 units, andenzyme-treated antitoxin about 140,000 units, per gram of protein(also found by Pappenheimer and Robinson). Continuation of thiswork would shed light on the structure of proteins and the peculiari-ties of the structure of antibodies.The practical value of this method lies in the elimination of non-specific proteins, which make up some 90% of the proteins of anuntreated serum. The results can be judged by the number ofunits of antitoxin per gram of protein in the product, which may beincreased eight times. Various techniques depending on longdigestion have proved less sati~factory.~~, 34y 35 Since the digested2o J . Exp. Med., 1940, 71, 247.31 Ann.Inst. Pasteur, 1938, 61, 765.32 Brit. J. Exp. Path., 1939, 20, 132, 201, 213.33 F. Modern and G. Ruff, Compt. rend. Xoc. Biol., 1038,129, 851 ; Biochem.34 F. Hansen, Compt. rend. Xoc. Biol., 1938, 129, 216; Biochenz. Z., 1938,35 G. Sandor, Compt. r e d . SOC. Biol., 1939, 130, 840, 1l87.30 J . Irnrnun., 1937, 32, 291.Z., 1938, 299, 377.299, 363400 BIOCHEMISTRY.antibody appears not to be antigeni~,~’ injection of it should notcause serum sickness or anaphylaxis, which, however, is a very rareaccident in man. The method can also be applied to tetanusantitoxin.36The Reaction between Antibodies and Antigens,Knowing the molecular weights of antibody (A) and antigen (G),it is possible to calculate the composition of the aggregate formedby their interaction; this has been done by M.Heidelberger 3* andA. M. Pappenheimer (jun.), H. P. Lundgren, and J. W. Williams.39Theories of the reaction between A and G suppose that large solubleaggregates are formed when G is in great excess. The formation ofsuch aggregates has been shown by the ultra-centrifuge.The reactions of A and G when spread in films on surfaces havebeen studied.*O Measurements of the thickness of films on solidsurfaces have given objective demonstration of the specific union ofA and G, proof that a molecule of A can combine with more than onemolecule of G, and vice versa, and dimensions of A and G moleculesthat agree fairly well with those obtained by other methods. Somediscrepancies may be due to unequal spreading and to the use ofdifferent methods of spreading and different types of surface.When antibodies to the bacterial polysaccharides are spread onwater to form a film, one amino-acid thick, they do not combinewith the corresponding polysaccharides.In films on solid surfaces,the antibodies preserve a “globular” form and combine withpolysaccharide . J. R. M.4. HORMONES.Since the anterior pituitary hormones have not been discussed inthese Reports lately, this section will deal exclusively with the moreimportant investigations on this gland, the consideration of otherendocrine organs being kept over for later reports. Even so, it isobviously impossible, in one brief section, to mention all contributionsconcerning this particular gland.36 G.Sandor and R. Richou, Cornpt. rend. SOC. Biol., 1939, 131, 461.37 A. J. Weil, I. A. Parfentjev, and K. L. Bowman, J. Irnmun., 1938, 35,38 J . Arner. Chern. SOC., 1938, 69, 103.39 J . Exp, Med., 1940, 71, 247.4 O I. Langmuir and V. J. Schaefer, J. Amer. Chem. SOC., 1937, 59, 1406;J. F. Danielli, J. M. Danielli, and J. R. Marrack, Brit. J. Exp. Path., 1938,19, 393; M. E. Shaffer and 5. H. Dingle, Proc. SOC. Exp. BioZ. Med., 1938,38, 528; A. Rothan and K. Landsteiner, Science, 1939, 40, 6 5 ; E. F. Porterand A. M. Pappenheimer (jun.), J. Exp. Med., 1939, 09, 755.399KODICEK : HORMONES. 401Anterior Pituitary.Gonadotrophic Hormones.-Several workers have recently suc-ceeded in separating distinct follicle-stimulating (FSH) and luteinis-ing fractions (LH) from pituitary extracts, using the relative in-solubility of the luteinising hormone in ammonium sulphate solutiona t pn 4-2 or 5-6.l These results were confirmed by biologicalassays.2 The LH is apparently identical with the hormone of Evans,stimulating the interstitial cells of the testes and ovaries (ICSH).3The follicle-stimulating hormone (FSH) contained in unfractionedpituitary extracts is a protein complex whose activity is, accordingto several laboratories (with the exception of A.A. Abraniowitz andF. L. Hisaw *), relatively unaffected by digestion with commercialor crystalline trypsin, while the luteinising potency (LH) is de-stroyed.5 ESH was found to be very rich in carbohydrate (prin-cipally glucosamine) when compared with the ICSH (or LH).6From acetylation by keten of these sugar-rich proteins, C.H. Li,M. E. Simpson, and H. M. Evans conclude that the activity of theFSH and LH is dependent on the free arnino-gro~ps.~ H. Fraenkel-Conrat et al., contrary to recent results of F. Bischoff,8 found thatcysteine destroys all gonadotrophic activity of pituitary hormones,acting upon the supposed S-S linkings considered to be essentialfor their potency. The chorionic gonadotrophic hormone (CGH,prolan from human pregnancy urine), however, showed no loss ofa ~ t i v i t y . ~ For these and other reasons, CGH is regarded as differentfrom pituitary luteinising hormone, with which i t has some bio-logical effects in common. CGH (of placental origin) resemblesinsulin in that the phenolic hydroxyls rather than the free amino-groups are essential for its activity.’ CGH is apparently a gluco-H. L.Fevold, Endocrinology, 1939, 24, 435; H. Jensen, M. E. Simpson,S. Tolksdorf, and H. M. Evans, ibid., 1939, 25, 57; H. Jensen, S. Tolksdorf,and F. Bamman, J . Biol. Chem., 1940,135,791 ; R. 0. Greep, H. B. Van Dyke,and B. F. Chow, ibid., 1940, 133, 289; H. Rinderknecht and P. C . Williams,J . Endocrinol., 1939, 1, 117.H. M. Evans, M. E. Simpson, S. Tolksdorf, and H. Jensen, Endocrinology,1939, 25, 529.H. L. Fevold, J . Biol. Chem., 1939, 128, 83.Endocrinology, 1939, 25, 529.W. H. McShan and R. K. Meyer, J . Biol. Chem., 1938,126, 361 ; G. Chenand H. B. Van Dyke, Proc. SOC. Exp. Biol. Med., 1939, 42, 454 ; R.C. Li, ibid.,1940, 43, 598; B. F. Chow, R. 0. Greep, and H. B. Van Dyke, J . Endocrinol.,1939, 1, 440.H. M. Evans, H. Fraenkel-Conrat, M. E. Simpson, and C. H. Li, Science,1939, 89, 249; W. H. McShan and R. K. Meyer, J . Biol. Chem., 1940, 135,473.J . Biol. Chem., 1939, 131, 259.H. Fraenkel-Conrat, M. E. Simpson, and H. M. Evans, ibid., 1939, 130,Ibid., 1940, 134, 641.243; Science, 1940, 91, 363402 BIOCHEMISTRY.protein (4 RU per 1 pg.), the carbohydrate of which consists ofhexosamine-digalactose units. The amino-group of the hexosamineis probably acetylated, and another acetyl group is attached toanother part of the molecule. The minimal molecular weight liesbetween 60,000 and 80,000. The isoelectric point was found to beat pH 3.2-3.3.10The isolation of the interstitial cell stimulating (or luteinising)hormone in pure form (1 unit in 5 pg.) has recently beenclaimed by two laboratories.ll9 12 Although the isolated proteinshave similar biological properties, the results from the electro-phoretic study are quite different. The isoelectric point is stated tobe at pH 7.45 and 4.6, respectively, and the mobility of the protein6.36 x and 0.66 x respectively.The gonadotrophicprotein of C. H. Li et al. contained 4.45% of mannose, 5.86% ofglucosamine, and 14.2% of nitrogen with approximately 4.5% oftyrosine and 1% of tryptophan.Thyreotrophic Hormone.-Various methods for the evaluation ofthe thyreotrophic potency of pituitary extracts have been proposed,of which Q.K. Smelser’s test based on the increase of the thyroidweight of the one-day-old chicks may be mentioned.13 Recentreports of the isolation of nearly pure thyreotrophic hormone with anegligible luteinising potency l4 do not support the claim that thethyreotrophic is a property of the Iuteinising hormone rather thandue to a separate entity.15 J. Fraenkel-Conrat et cal. found that thisactive protein contains 13% of nitrogen, 3.5% of carbohydrate, and2.5% of glucosamine. Cysteine and keten treatment inactivates thehormone. Some metabolic effects of pituitary extracts, however,cannot be ascribed to the thyreotrophic hormone and seem to beindependent of the thyroid.16lo S. Gurin, C. Bachman, and D. W. Wilson, J . B i d . Chem., 1939, 128,proc.xxxvii, 525; 1940, 133, 467, 477; J . Amer. Chem. Soc., 1939, 61, 2251.l1 T. Shedlovsky, A. Rothen, R. 0. Creep, H. B. Van Dyke, and B. F.Chow, Science, 1940, 92, 178.l2 C. H. Li, M. E. Simpson, and H. M. Evans, ibid., p. 356.l3 Q . K. Smelser, Proc. SOC. Exp. Biol. Med., 1937, 38, 388; Endocrinology,1938, 23, 429.l4 C. G. Lamble and V. M. Trikojus, Biockem. J., 1937, 31, 843; B. F.Chow, R. 0. Greep, and H. B. Van Dyke, J . Edocrinol., 1939, 1, 440 ; H. L.Fevold, M. Lee, F. L. Hisaw, and E. 3. Cohn, Endocrinology, 1940, 26, 999;R. W. Bonsnes and A. White, i b d . , p.. 990 ; J. Fraenkel-Conrat, H. Fraenkel-Conrat, M. E. Simpson, andH. M. Evans, J. Bid. Chern., 1940,135, 235, 199.15 H. Jensen and J. F. Grattan, Amer. J. Phylsiol., 1937, 37, 388;H.Jensen and S . Tolksdorf, Endocrinology, 1939, 25, 429; Proc. SOC. Exp.Biol. Med., 1939, 42, 466.l6 D. K. O’Donovan and J. B. Collip, Canadicart Med. Assoc. J., 1938, 39,83; H. H. Neufeld and J. B. Collip, dbicE., p. 83; 0. F. Denstedt and J. B.Collip, ibid., p. 84KODICEK : HORMONES. 403Growth Hormone.-The growth of " plateaued " or hypophysecto-mised animals has been used for assays of the growth potency ofpituitary extracts.17 Recently J. Freud and L. H. Levie foundthat the growth of the tail and development of caudal vertebrai? ofhypophysectomised young rats ceased altogether. They use thiseffect upon the proliferating zone of the cartilage for assays by mak-ing X-ray pictures and histological examination of the tails.lgHowever, thymus extracts also give a positive result in this test.lSThe evidence that there is a pituitary hormone specifically concernedwith growth is convincing.Preparations of the growth hormonecan be almost freed from thyreotrophic (another growth-promotingfactor) and lactogenic hormone by treatment with cysteine.20Highly purified preparations were obtained (active in doses of3-10 pg.) having the elementary composition : C, 49.76; H,7.24; N, 14.27; S, 1.47% (in dithio-groups). They were labile inheat, acid, and alkali, and were destroyed by trypsin and pepsin.An absorption maximum was observed at 2830 A . ~ ~Lactogenic Hormone.-An extensive review of this hormone hasappeared.22 All evidence indicates that the factor which initiateslactation in mammals is identical with the pigeon crop stimulatingfactor.Upon this fact are based most methods of assay.23 Severalworkers have attempted a separation of pr~lactin.~* The crystallineproduct of protein nature, the lowest effective dose of which wasfound to be 0.1-0-2 pg. per pigeon, contained C, 51.11 ; H, 6-76 ;N, 14.38 ; S, 1.77% ; tryptophan, tyrosine, and phenylalanine wereidentified in the preparation.26 The isoelectric point was at pH5.7 and its mobility dp/dpH 4.5 x 10-5 in electrophoretic study.26l7 H. M. Evans, N. Uyei, Q. R. Bartz, and M. E. Simpson, Endocrinology,1938, 22, 483.l8 L. H. Levie, Acta brev. neerl. Physiol., 1937, 7, 119; J. Freud and L. H.Levie, Arch. int. Pharmacodyn., 1938, 59, 232; J. Freud, L. H. Levie, andD.B. Kroon, J . Endocrinol., 1939, 1, 56.l9 L. H. Levie, I. E. Uyldert, and E. Dingemannse, dcta brev. n e e d Physiol.,1939, 9, 50.2o H. M. Evans, M. E. Simpson, and R. I. Pencharz, Endocrinology, 1939,25, 175; D. L. Meamber, H. L. Fraenkel-Conrat, M. E. Simpson, and H. M.Evans, Science, 1939, 90, 19; A. E. Light, E. J. de Beer, and C. A. Cook,Proc. SOC. Exp. Biol. Med., 1940, 44, 189.21 E. Dingemanse, Proc. XVI Intern. Physiol. Congr., Zurich, 1938.22 S. J. Folley, Biol. Rev., 1940, 15, 421.23 A. J. Bergmann, J. Meites, and C. W. Turner, Endocrinology, 1940, 26,716.24 R. W. Bates and 0. Riddle, J . Pharm. Exp. Ther., 1935, 55, 365 ; J . Biol.Chem., 1938, 123, proc. v ; A. White, H. R. Catchpole, and C. N. H. Long,Science, 1937, 86, 82; W.R. Lyons, Proc. s'oc. Exp. Biol. Med., 1937, 35, 645.25 A. White and G. I. Levin, J . Biol. Chem., 1940, 132, 717.26 C. H. Li, W. R. Lyons, and H. M. Evans, Science, 1939, 90,622 ; J . Qen.Physiol., 1940, 23, 433404 BIOCHEMISTRY.The crop-stimulating activity of prolactin depends upon the presenceof free amino-groups in the molecule, as shown by its inactivationwith nitrous with phenyl isocyanate,28 and keten,29 anddepends upon the integrity of the tyrosine component, as indicatedby its inactivation by iodine.30Anterior Pituitary Hormones and Carbohydrate Metabolism.-Theinformation regarding this very complicated question is a t presentsomewhat confused and contradictory ; it is nevertheless evident thatpituitary extracts exert a profound influence upon metabolic pro-cesses in general (respiratory-quotient reducing substances 31) andcarbohydrate metabolism in particular. It is not certain, however,to what extent these effects can be attributed to specific hormonesrather than to already known active principles.Hypophysectomised animals tend to pass into hypoglyczmia.The glycotropic (glycostatic) pituitary principle, which counteractsthis tendency, is claimed to be identical with the adrenocortico-trophic principle.32 Certain pituitary extracts, named “ diabeto-genic,” produce hyperglyczmia, sometimes ketonemia, and evenpermanent diabetes in dogs.33 F.G. Young states that onlyglobulin and +-globulin fractions have diabetogenic activity.=Antihormones.-The appearance in the blood of principles antag-onistic to certain hormones after pretreatment with the respectivehormones seems to be established 35 beyond doubt. Whether thesesubstances are hormones or antibodies cannot be answered withcertainty as yet.E. K.5. PROTEINS.The great progress made during the last decade in our knowledgeof proteins is mainly due to the application of physical methods and27 C. H. Li, W. R. Lyons, M. E. Simpson, and H. M. Evans, Science, 1939,90, 376.28 A. C. Bottomley and S. J. Folley, Nature, 1940,145, 304.2D C. H. Li, M. E. Simpson, and H. M. Evans, Science, 1939, 90, 140.ao C. H. Li, W. R. Lyons, M. E. Simpson, and H. M. Evans, ibid., 1940,91, 530.31 H. M. Evans, J. M. Luck, R. J. Pencharz, and H. C . Stoner, Amer. J .Physiol., 1938, 122, 533; J.I). Greaves, J. K. Freiberg, and H. E. Johns,J . Biol. Chem., 1940, -133, 243 ; W. W. Billingsley, D. K. O’Donovan, andJ. B. Collip, Endocrinology, 1939, 24, 63.32 J. F. Grattan and H. Jensen, J . Biol. Chem., 1940, 135, 51 1 ; Amer. J .Physiol., 1940, 128, 270.33 J. Campbell, H. C. Keenan, and C. H. Best, Amer. J. Physiol., 1939, 126,455; F. G. Young, Brit. Med. J., 1939, 2, 393; C. H. Best, J. Campbell, andR. E. Haist, J . Physiol., 1939, 97, 200.34 J . Endocrinol., 1939, 1, 339.35 J. B. Collip, H. Seyle, and D. L. Thomson, Biol. Rev., 1940, 15, 1 NEUBERGER : PROTEINS. 405it is deemed desirable to review recent advances with a specialemphasis on physicochemical work.Molecular Xixe and Homogeneity.Mainly as a result of the work of the Svedherg school, which hasbeen reviewed in an earlier report,l we know that most of theproteins exist in solution as molecules of a well-defined size and thatmost of the usual protein preparations contain only one or fewmolecular species.The earlier chemical methods which claimed toisolate definite chemical individuals by making use of differences insolubility of proteins in salt solutions have thus been justified to alarge extent. During the last few years, however, i t has becomeincreasingly clearer that the fact that a single sharp boundary isobtained in the ultra-centrifuge is not infallible evidence for thehomogeneity of the preparation.2 A sharp boundary, especially ifi t has been observed by the refractive index method, certainly showsthat all the protein molecules in the solution have similar molecularweights, provided that the shapes of the molecules are also similarand are not too asymmetrical.But the fact that the molecularweights of most proteins tend to assume values which are multiplesof 17,600 produces a certain a priori probability that a mixture ofproteins will contain molecules of very similar size and thus give afortuitous appearance of homogeneity. In fact, several suchexamples can be quoted. Crystalline egg albumin, which had beenfound to be homogeneous by sedimentation methods, has now beenshown by the use of the cataphoresis method to contain two com-ponent~.~, * But even electrophoretic measurements, which, whenperformed a t different acidities, furnish one of the most reliablecriteria of purity, may fail to show inhomogeneities detectable byother means.Thus R. A. Kekwick was able to separate serumalbumin by fractional crystallisation into two crystalline andapparently homogeneous fractions, named A and B. The twofractions had the same electrophoretic mobilities between pH 4.0and 5.5 and identical diffusion and sedimentation constants.Fraction A contained 1 a950A of carbohydrate, whereas B was almostcarbohydrate-Iree. Similarly, several egg albumins obtained fromdifferent species of birds, which can be easily distinguished byserological reactions, show very close resemblance in electrophoresisAnn. Reports, 1937, 34, 302.N. W. Pirie, Biol. Rev., 1940, 15, 377.A.Tiselius and I. B. Eriksson-Quensel, Biochem. J., 1939, 33, 1752.L. G. Longsworth, R. K. Cannan, and D. A. MacInnes, J . Amer. Chern.SOC., 1940, 62, 2580.5 Biochem. J . , 1938, 32, 552406 BIOCHEMISTRY.experiments.6 It has also to be appreciated that crystallisation,which is still one of the most important methods in the purificationof proteins, is not a reliable indication of purity and as a criterionof homogeneity is definitely inferior to other physical or biologicalmet hods .2Another point of importance has emerged clearly from recentwork on sedimentation and electrophoresis of proteins. Althoughfor many proteins molecular size is independent of changes in thecomposition of the solution within rather wide limits, for otherproteins this is not the case. Syedberg had already observed in hisearlier work that hamlocyanin dissociates reversibly into smallercomponents if the pH of the solution is outside the “ stabilityrange.” But i t has now been shown that a salt concentration of amolarity of 0.5-1 -0 causes horse CO-haernoglobin to dissociate intosmaller molecule^.^ Similar effects are obtained on dilution.Bivalent ions seem to have a marked effect on molecular size evenin small concentrations.Thus, the sedimentation constant ofcaseinogen showed an increase from 6 x 10-13 to 10 x 10-13 onaddition of increasing amounts of calcium.* One of the globulincomponents of serum, @-globulin, is particularly sensitive to changesin salt con~entration.~ Other low-molecular substances such asurea, lysine, and ammonium chloride cause dissociation of someproteins and not of others.There is also a large amount of evidenceof an interaction between different proteins. It was shown byA. S. McFarlane lo that the sedimentation diagram of a mixture ofserum albumin and globulin is not an additive pattern of the isolatedproteins. Pedersen 9 has studied such interactions in a large numberof cases and found that the effect varied with different proteins, andhe considered that the protein-bound carbohydrate may play a partin these reactions. Electrophoretic experiment^,^ too, indicate aprotein-protein interaction in the case of egg-white.Serum Proteins.-These changes produced in the size and shape ofproteins by variations in their molecular environment assumeparticular importance in connection with the question, how farproteins isolated from cells and biological fluids correspond todefinite chemical entities in their natural state.Serum, one of thebest-studied protein systems, is a case in point. The sedimentationdiagram of native untreated horse serum shows in the ultra-centri-6 K. Landsteiner, L. G. Longsworth, and J. van der Scheer, Science, 1938,88, 83.7 “ The Ultracentrifuge,” by The Svedberg and K. 0. Pedersen, Oxford,1940.8 F. J. Philpot and J. St. L. Philpot, Proc. Roy. Soc., 1939, B, 127, 21.M. Jersild and K. 0. Pedersen, Acta path. microbial. Scand., 1938, 15,426NEUBERGER : PROTEINS. 407fuge two main peaks.1° A lighter fraction, named “albumin,”represents nearly 80% and the heavier “ globulin ” about 20% of thetotal protein.Chemical separation by half sauration with ammon-ium sulphate gives a ratio of albumin to globulin of nearly 1, andthe same ratio is approached from the sedimentation diagram if theserum is suitably diluted. Thus the “albumin” peak can beresolved into two components, of which one is due to albuminproper and the other one to a protein called X by McParlane andlater shown by Pedersen to be identical with a globulin, namedp-globulin by A. Tiselius. The variability of the sedimentation ofthis protein is caused by its extreme sensitivity to changes in the saltconcentration of the medium. More important information concern-ing the complex system of proteins in serum and plasma was, how-ever, obtained by the elegant electrophoretic method of Tiselius.It was shown that serum contains four components which could bedistinguished by their different mobilities in an electrical field; inaddition to serum albumin there are three different globulins,labelled a, p, and y,ll having different mobilities but nearly identicalmolecular weights.More recently I. A. Luetscher l2 has demon-strated in human serum two albumin peaks a t pH 4.0, and a similarobservation was made for mouse serum.l3 By improved opticalmethods it was shown that the p-component of human serum isactually composed of two substances, labelled p1 and @z.14, l5The values for the relative amounts of the various components innormal human serum given by different workers vary slightly, theaverage value for albumin being about 65%, which gives an albumin-globulin ratio of 2 : 1 ; the a-globulin is present in only smallamounts.Plasma, as might be expected, contains another electro-phoretically well-defined protein, fibrinogen. l6It is extremely difficult to correlate these electrochemical investig-ations, which study the protein complex of serum in its native state,with the results of workers who isolate proteins by chemical methods,such as salting-out or fractional crystallisation. It is certain thatthe classification of serum proteins on the basis of these old methodscannot now be accepted. It has been shown, e.g., that “pseudo-globulin’’ and “ euglobulin ” are mixtures and that “ albumin ” asusually prepared contains in fact a large proportion of g10bulin.l~10 Biochem.J., 1935, 29, 407.l1 A. Tiselius, ibid., 1937,31,313,1464; R. A. Kekwick, ibid., 1939,33, 1122.12 J . Clinical Invest., 1940, 19, 313.l3 J. Bourdillon and E. H. Lennette, J . Exp. Med., 1940, 72, 11.l4 H. I. Svemson, Kolloid-Z., 1939, 87, 181.l5 R. A. Kekwick, Biochem. J., 1940, 34, 1248.l6 E. Stenhagen, ibid., 1938, 32, 719.L. F. Hewitt, ibid., 1936, 30, 2229; 1938, 32, 26408 BIOCHEMISTRY,This does not mean that fractionation by chemical methods is useless ;on the contrary, the Reporter believes that important results by suchmethods can still be obtained if the separation is followed by physicalmethods, especially by electrophoretic measurements. A very goodexample for such a controlled separation is Kekwick’s isolation ofp-globulin by fractional precipitation with sodium sulphate.l5These requirements are not fulfilled in many recent papers dealingwith serum proteins; it seems desirable, however, to review therather complicated position at the present juncture. It was themerit of L. F. Hewitt l7 to reopen the problem of the homogeneity ofserum albumin and he was able to isolate three main fractions :(1 ) a crystalline fraction, named crystalbumin, with practically nocarbohydrate, [aD] - 71”, high tyrosine and cystine content, and acontent of tryptophan of 0.26% ; (2) an amorphous, carbohydrate-rich fraction, seroglycoid, with [aD] - 57”’ low cystine content anda tryptophan content of 1% ; (3) a fraction, globoglycoid, behavinglike a globulin after crystalbumin has been removed and containingcarbohydrate.Hewitt’s findings have recently been criticised byC. Rimington,l* who has also prepared a protein very rich in carbo-hydrate from ox-serum and named i t seromucoid. There is generalagreement that a orystalline fraction can be prepared from crudealbumin which is carbohydrate-free ; this fraction is probablyidentical with Kekwick’s serum albumin B.5 But even this prepar-ation is not homogeneous, as two boundaries of p= 4.0 were demon-strated by electrophoretic experiment.19 An apparently homo-geneous fraction was, however, obtained by crystallising the albuminsulphate a t p , 4-0 from water.20The albumin fraction contains proteins with a high content ofcarbohydrate which have not been examined by physical methods ;i t seems very likely that the preparations called “ seromucoid ” byRimington and “ seroglycoid ” by Hewitt are very similar and differonly in their respective “ impurities.” Whether Kekwick’s crystal-line serum albumin A is a chemical individual or a complex formedbetween crystalbumin and seromucoid remains doubtful.It seemsestablished that the albumin fraction contains an appreciablequantity of globulin-like material; but whether a new name shouldbe coined for this rather ill-defined preparation is less certain. Froma broader point of view the most important question in this connec-tion is, whether serum contains a small number of distinct proteinswhich are in every respect chemical individuals and in which thel 8 Bwchem.J., 1940, 34, 931 ; C. Rimington and M. van den Ende, ibid.,l9 J. A. Luetscher, J. Amer. Chern. SOC., 1939, 61, 3888.2o T. L. McMeekin, ibid., p. 2884.p. 041NEUBERGER : PROTEINS. 409difficulties of isolation are due to a lack of specificity of the analyticalmethods used and the interaction of proteins with each otherdiscussed above; or whether the proteins of serum contain anindefinite number of molecular species which fall into groups ofcertain physical and chemical similarity-a fact which would beresponsible for an apparent and deceptive uniformity in physical andbiological behaviour of different fractions. Physical experiments,such as electrophoresis, seem to favour the first, non-defeatist theory.Changes in Xerum Proteins in Disease.The examination of pathological sera by modern physical methodsis a very promising field, opened by A.S. McFarlane,21 using ultra-centrifugal methods. More recently the cataphoresis technique hasbeen applied and some interesting results have been obtained. Themost striking changes occur in multiple myelomatosis, which hadbeen known for some time to be associated with a hyperproteinemiaand a change in the albumin-globulin ratio. Prom the results ofdifferent workers (Kekwick; l5 Jersild and Pedersen; L. G. Longs-worth, Th. Shedlovsky, and D. A. R/lacInnes22) it appears that inall cases the relative proportion of globulin is very much increased.I n one group of cases this is due to an increase in the y-globulin peak ;in others an excess of @-globulin was apparent. Nephrosis leads alsoto a change in @-globulin which is associated with the presence of aprotein-lipoid complex ; this combination is apparently brokendown by ether e x t r a ~ t i o n .~ ~ Febrile patients, on the other hand,show an increase in ol-globulin.22Shape of Protein Holecules.Whereas the size of proteins can be determined by methodswhich rest on a secure thermodynamic basis such as sedimentationequilibrium measurements, exact information as to the shape ofthe molecules is more difficult to obtain. Broadly speaking, proteinsother than fibrous proteins can be divided into two classes. Thefirst group comprises highly asymmetrical types such as myosin andmany virus nucleoproteins which show double refraction of flow a tcomparatively small velocity gradients and " anomalous viscosity " ;i.e., the viscosity of their solutions depends markedly on the rate offlow.The second group is made up of the so-called " globular "proteins, which do not show any definite orientation if the usualrates of flow are applied. These molecules are either spherical ordo not deviate much from a symmetrical shape.The methods most extensively used to estimate the degree of21 Biochem. J . , 1935, 29, 1175.28 L. G. Longsworth and D. A. MacInnes, ibid., 1940, 71, 77.22 J . Exp. Med., 1939, 70, 399410 BIOCHEMISTRY.asymmetry of " globular " proteins are based either on a comparisonbetween observed and calculated diffusion or frictional constants oron viscosity measurements.The " observed ) ) frictional constant, f,can be calculated from the diffusion constant or from the molecularweight, M , partial specific volume and sedimentation constant ; thetwo methods give generally identical results. On the other hand, a" theoretical " frictional constant,fo, can be calculated if it is assumedthat the particle of the molecular weight nil. is rigid, spherical anddoes not combine with the solvent. In most cases f andfo are notidentical and the ratio f/fo, which is called the frictional ratio, isgreater than 1. This indicates that the particles either are solvatedor are not spherical. Attempts have been made to assess thedimensions of protein molecules in solution by assuming that hydr-ation can be neglected and that the shapes of the particles are thoseof ellipsoids of rotation.24 These assumptions are, however, to alarge extent arbitrary. Hydration cannot be neglected and mayaccount in many cases for the values of the frictional ratio.ThusG. S. Adair 25 has shown that the observed diffusion constants ofegg albumin, hzmoglobin, and serum albumin can be quantitativelyexplained in terms of spherical shapes, if i t is assumed that thesemolecules have a symmetrical shell of hydrated solvent equal inamount to that found for their crystals. It will be difficult to correctfor hydration in all cases; our knowledge of the forces operating inhydration and their magnitude is still incomplete and differentmethods have yielded different results. Moreover, it is quite possiblethat, if hydration is largely due to electrostatic interaction of chargedgroups of the protein surface with the dipolar solvent molecuIes,asymmetrical hydration will take place with symmetrical particlesprovided that the charge density is asymmetrical.In that case theincrease of the frictional ratio will be greater than could be accountedfor by mere increase in volume. But, even if solvation could beneglected, the use of an ellipsoidal model is somewhat arbitrary ;any marked irregularity of the molecular surface will cause anincrease in f.It is generally accepted that there exists a close relationship be-tween the shape of molecules and the viscosity of their solutions,although the quantitative significance of the different calculationsis less certain.W. Kuhn26 has given an equation correlating theobserved specSc viscosities of solutions containing molecules of theshape of long cylinders with their axial ratios and J. M. Burgers 2724 H. Neumth, J . Amer. Chern. SOC., 1939, 61, 1841.26 Proc. Roy. SOC., 1939, B, 127, 18.26 2. physikal. Chem., 1932, A, 161, 1.2 7 Second report on viscosity and plasticity, Amsterdam, 1938NEUBERGER : PROTEINS. 41 1made similar calculations for elongated ellipsoids ; A. Polson 28proposed a semi-empirical formula which has given quite satisfactoryresults. The last equation can be considered to be a modification ofKuhn’s formula, including, however, a correction for hydration.In all these calculations it is assumed that the Brownian movementis sufficient to suppress any orientation of the particles due to thehydrodynamic force applied.This assumption is almost certainlyjustified if the specific viscosity is independent of the rate of shear,that is, for most “ globular ” proteins. As pointed out by J. R.Robinson,29 the position becomes more complicated in the case ofhighly asymmetrical particles, like tobacco mosaic virus or myosin ;here work has to be done to rotate the molecules and this amount ofenergy becomes less as the velocity gradient is increased, comparedwith the energy required to maintain the rate of flow. Measiire-ments with ordinary capillary viscometers are in such cases open tograve objections.These equations have been used by severalauthors to deduce the shapes of globular proteinsY2*9 307 31 and theresults obtained would indicate that nearly all the proteins whichhad hitherto been considered spherical have axial ratios of 1 : 3-1 : 8. These figures must, however, be accepted with some reserve.The difficulties concerning the uncertainty of hydration and shapeof the particles, mentioned in connection with the frictional ratios,apply equally to the interpretation of viscosity measurements.J. W. Mehl, J. L. Oncley, and R. Simha32 have compared thevalues obtained from frictional ratios and from viscosity measure-ments and found that agreement between these methods is not satis-factory if the equations of Kuhn and Burgers are applied.Anequation derived by Simha33 gives better results. The wholeproblem of viscosity and shape has also recently been reviewed byJ. M. Burgers,s4 who calculated theoretical sedimentation constantsfrom the axial ratios obtained by viscosity measurements, assumingellipsoidal shape, and compared these values with those observedin sedimentation experiments. The agreement was not very good,even if allowance was made for hydration. It is felt that, althoughthe quantitative interpretation of results obtained by viscositymeasurements is doubtful, they may yield important information.Thus the increase of the viscosity of egg albumin 31 on denaturationcannot be explained by increased hydration and must be due to anincreased asymmetry of shape.28 Kolloid-Z., 1939, 88, 51.30 H.Neurath and G. R. Cooper, J . Amer. Chem. SOC., 1940, 62, 2248.31 H. B. Bull, J . Bwl. Chem., 1940, 133, 39.33 Science, 1940, 92, 132.34 Proc. K . Akad. Wetensch. Arnsterdmn, 19.40, 43, 307.29 Proc. Roy. SOC., 1939, A , 170,619.33 J . Physical Chem., 1940, 44, 25412 BIOCHEMISTRY.Interesting observations have also been recently reported onanomalous viscosities of proteins obtained from developing eggs andembryos.35 The results, which were obtained by the use of aCouette viscometer, indicate that these proteins are highlyasymmetrical.The most reliable criterion of a marked asymmetry of shape isdouble refraction of flow. By the use of this method it was shownthat myosin molecules are long and rod-shaped particles,36 an inter-pretation which is in accordance with all other physical properties ofthis protein.Later it was shown by several workers that manyother proteins show double refraction of flow at low velocity gradients.Thus a highly asymmetrical shape was proved for tobacco mosaicvirus,37 other plant viruses,38, 41 fibrin~gen,~~ antibody globulins,40hog thyrcoglobulin, different haem~cyanins,~~ and ovoglob~lin.~~The quantitative aspects of the relationship between double refrac-tion of flow and the dimensions of particles have recently beendiscussed by J. W. Meh1,43 who, using the theoretical treatment ofP. Boeder 44 and W. Kuhn,26 calculated the length of the myosinmolecule to be about 8500 A. Such calculations are based on certainsimplifications, as Mehl himself points out, and the calculated dimen-sions cannot be considered extremely accurate.The influence of different chemical and physical factors on theshapes of these highly asymmetrical particles has recently beeninvestigated. It was found, e.g., that the double refraction of flow ofmyosin is abolished by such substances as the chlorides of bivalentcations, guanidinium salts, and potassium iodide in very dilute solu-t i ~ n , ~ ~ and urea produced a similar effect in more concentratedsolution.The loss of double refraction of flow was associated witha decrease of viscosity, indicating that " denaturation " in this caseconsisted in a definite decrease of the asymmetry of the molecule;as mentioned above, globular proteins show the opposite behaviour.The recently developed electron microscope has also been used to35 A.S. C. Lawrence, J. Needham, and Shih-Chang Shen, Nature, 1940,36 A. L. von Muralt and 5. T. Edsall, J . Biol. Chem., 1930, 89, 315, 351.146, 104.F. C. Bawden, N. W. Pirie, J. D. Bernal, and I. Fankuchen, Nature,1936, 138, 1951; M. A. Lauffer, J . Physical Chem., 1938, 42, 935.38 G. A. Kausche, H. Guggisberg, and A. Wissler, Natuyw., 1939, 27, 303;H. S. Loring, J . Biol. Chem., 1938, 126, 455.39 G. Boehm and R. Signer, Klin. Woch., 1932, 11, 599.4 0 E. A. Kabat, J . Exp. Med., 1939, 69, 103.41 M. A. Lauffer and W. M. Stanley, J . Biol. Chem., 1938, 123, 507.p 2 G. Boehm and R. Signer, Helv. Chim. Acta, 1931, 14, 1370.4 3 Cold Spring Harbor Symposia Quantitat. Biol., 1938, 6, 218.4 4 2.Physik, 1932, 75, 258.4 5 J. T. Edsall and J. W. Mehl, J . Biol. Chem., 1940, 133, 409DANIELLI : PHYSICOCHEMICAL PHENOMENA. 413observe directly the shapes of colloidal particles. Thus M. vonArdenne 46 has recently published photographs of spherical HeEixhamocyanin molecules and two different viruses which appear in theform of long threads. Photographs of tobacco mosaic virus particlesobtained by an electron diffraction method have also been publishedby G. A. Kausche, E. Pfankuch, and H. R ~ s k a , ~ ' who state that theseparticles are about 150 A. in cross-section and about 3000 A. inlength. The dimensions obtained were of the same order of magni-tude as those obtained by other methods, such as viscosity.Theelectron microscope has also been used to study the reaction betweencolloidal gold particles and tobacco mosaic These photo-graphs were obtained on dried films, and the shapes and sizesobserved are therefore not necessarily those of native proteins orviruses. A. N.6. PHYSJCOCHEMICAL PHENOMENA.Mehlbporphyrins.W. M. Clark and his colleagues 1 have made an extensive surveyof the use of oxidation-reduction potentials and spectrophotometryin the analysis of the behaviour of metalloporphyrins in the presenceof substances capable of co-ordinating with the metal atom.Equations are obtained relating electrode potential, total concen-tration of metalloporphyrin, total concentration of base co-ordinatingwith the metal atom, the ratio of the concentrations of the oxidisedand the reduced complex, the number of electrochemical equivalentsinvolved in the oxidation-reduction process, constants describing thedissociation of the base metalloporphyrin complexes (both thereduced complex and the oxidised complex), the numbers of mole-cules of base co-ordinating with the metal atom (a) when reduced,( b ) when oxidised, and the degree of association of the oxidised andthe reduced metalloporphyrin molecules.Equations are also givenby which the various equilibrium constants may be derived fromspectrophotometric data. Sixteen propositions are given which areopen to experimental investigation, and graphical methods are out-lined for determining dissociation constants, etc., from potentialmeasurements.Thus the analysis of these results is placed on anobjective basis. This work has not so far led to many fundamentally4 6 Naturwiss., 1940, 28, 113,4 7 Ibid., 1939. 27, 292.4 8 G. A. Kausche and H. Ruska, Kolloid-Z., 1939, 89, 21.W. M. Clark, J. F. Taylor, T. H. Davies, and C. S. Vestling, J. Biol. Chem.,1940, 135, 543; J. F. Taylor, ibid., p. 569; T. H. Davies, ibid., p. 597; C. S.Vestling, ibid., p. 623; W. M. Clark and M. E. Perkins, ibid., p. 643414 BIOCHEMISTRY.new conclusions, but i t makes possible the drawing of objectiveconclusions on many hitherto controversial points.Protein Adsmption and the Suspended Fat of the Bbod.A. C. Frazer and his colleagues have investigated the factorsresponsible for the stability of the suspended plasma neutral fat.According to Frazer the fat is mainly contained in the chylomicrons,which are microscopic or submicroscopic fatty droplets protected byan adsorbed layer of protein. Towards precipitants such as am-monium sulphate, chyloniicrons behave like globulins, but whencentrifuged t'he chylomicrons move in the centripetal direction,unlike the serum globulins.Similar behaviour is shown by artificialfat emulsions in the presence of globulins. Albumins also protectthe artificial emulsions, but the precipitation reaction differs fromthat of chylomicrons and fat globules in the presence of globulin.Frazer concludes that the chylomicrons consist mainly of neutralfat, but that the outer layer which controls the precipitation re-actions is adsorbed globulin.A. Tiselius3 has shown that theopalescence of normal serum, which is probably due to fat droplets,migrates on electrophoresis as though the particles were coated withp-globulin, and L. G. Longsworth and D. A. McInnes have shownthat in serum from cases of lipoid nephrosis the lipoid behaveselectrophoretically like p-globulin, after cold ether extraction muchof it is removed from the serum, and the p-globulin peak of theelectrophoretic pattern is correspondingly reduced.Protein adsorption a t oil-water interfaces has been studied bymany authors,5 with conclusions compatible with those of Frazer.The first layer of protein adsorbed is denatured, as a t the air-waterinterface, but on this layer a second layer of globular protein maybe adsorbed and this layer will control the precipitation and electro-phoretic properties of the interface. The loss of toxicity of toxinsand venoms after mixing with oil emulsions is probably due mainlyto the denaturation of proteins adsorbed on the oil droplets.J. F.D.J. J. Ekes, A. C. Frazor, and H. C. Stewart, J. Physiol., 1030, 95, 6 s ;Kolloirl-Z., 1937, 31, 1464.J. F. Danielli and E. N. Harvey, J . Cell. Comp. Physwl., 1934, 5, 483;H. Devaux, Comp. rend., 1936, 202, 1957; I. Langmuir and D. F. Waugh,J. Cen. Physiol., 1938, 21, 745; J. F. Dunielli, Cold Spring Harbor Symposia,1935, 6, 190; A. E. Alexander and T. Teorell, Trans. Paraday Xoc., 1930, 35,727; F. A. Askew and J. F.Dnnielli, ibid., 1940, 36, 785.G. N. Myers, J . Hyg., 1934, 34, No. 2 ; A. C. Frazer and V. G. Walsh,Brit. Med. J., 1934, March; J. Oerskov and S. Schmidt, Rev. Immunol., 1935,1, No. 4; A. C. Frazer and V. G. Walsh, J. Phnrm. Exp. Ther., 1939, 67,476.A. C. Frazer and H. C. Stewart, ibid., 4 P, 5 P.4 J . Exp. Med., 1940, 71, 77BELL BIOLOGICAL CATALYSIS. 4157. BIOLOGICAL CATALYSIS.As befits the fundamental object of biochemical research, enzymicprocesses continue to be studied with ever-increasing revelation oftheir mechanisms. In particular, studies of oxidation, of C-C linkdisruption, and of reversible phosphorolytic breakdown of starchhave yielded important results.Oxidation Mechanisms.*A number of reconstructed biological oxidations can take placethrough three essentially similar stages, vix.:Substrate + coenzyme I(or 11)(a) Spec@c deiydrogenase ’ _3oxidised substrate + dihydrocoenzyme I( or 11)Dihydrocoenzyme I(or 11) + 2FeIII cytochromeDihydrocoenzyrne I( or 11) -dehydrogenase V ’+ ( b )coenzyme I(or 11) + 2FeII cytochrome2Fen cytochrome + &02Cytochrome oxzdase( c ) 2FeIn cytochrome + H20(a), ( b ) , and (c) each consist in the transfer of two hydrogen atoms(or their equivalent) from a hydrogen donator to a hydrogenacceptor under the influence of an enzyme highly specific with respectboth to donator and to acceptor. Modern nomenclature tends toreserving the term “ oxidase ” for enzymes directly catalysing thereduction of molecular oxygen. Reactions (a) and (b) are anaerobic,whereas (c) is aerobic.It is not yet clear whether, in vivo, thereoxidation of dihydrocoenzymes takes place through the cyto-chrome system or whether some alternative mechanism is concerned.It is, however, significant that it is generally believed that cyanide,known to inhibit cytochrome oxidase, blocks the greater part of mosttissue-respiration.I n this brief review it is convenient to regard both enzymes andcarriers as catalysts. The view has recently been put forward thatcoenzymes I and I1 (loosely termed “ phosphopyridine nucleotides ”)should not be regarded as individual catalysts, but as being dissociablylinked to ‘‘ specific proteins ” (the dehydrogenases) and thus formingthe prosthetic group of a catalytically-active conjugated protein.This idea has been suggested from the recognition of the nature ofthe cytochromes and of the so-called “ flavoproteins ” (Report for1939, p.353). M. Dixon and L. G. Zerfas,l however, have in-Biochem. J., 1940, 34, 371.* D. E. Green, “Mechanisms of Biological Oxidations” (Cambridge Unk.Press, 1940)416 BIOCHEMISTRY.geniously obtained evidence in direct contradiction to this concep-tion and in support of the older and more widely held view. Bychoosing appropriate substances to act as hydrogen-acceptors, theyshowed that both the alcohol and the maleic dehydrogenases ofyeast could oxidise their appropriate substrates in complete absenceof coenzyme I. They therefore regard the “protein” of thedehydrogenases as the complete enzyme and point out that co-enzyme I and the “ protein ” are in the relation of substrate toenzyme as indicated in ( a ) above.The cytochromes and the succinic dehydrogenase system have beenthe subject of new investigations by D.Keilin and E. F. Hartree.2They emphasise that attempts to dissect the system (which oxidisessuccinate zrobically) have failed. The activity of the systemdepends on the presence of insoluble particles to which the catalystsare bound. The generally accepted course of succinate oxidationmay be expressed as follows :(4 ‘Succinate + 2FeIII cytochrome-cSuccinic dehydrogenase2FeII cytochrome-c + i02(e) CytochroAe oxidase ’r J --+ fumarate + 2FeII cytochrome-c__f 2FeIII cytochrome-c + H,OBut Keilin and Hartree have found that with certain oxidasepreparations reaction ( d ) would not proceed.They suggest that thereduction of FeIII cytochrome-c may be an indirect one, requiring theintermediate intervention of cytochrome-b or some hitherto un-recognised substance. These authors have recognised, in heart -muscle, insect thoracic-muscle, baker’s yeast, and strictly axobicbacteria a new cytochrome component, termed a3. This has manyproperties identifying it with the enzyme, cytochrome oxidase ; e.g.,it is thermolabile and is, in the FeII state, very easily autoxidisable.Carbon monoxide combines with the FeII compound, giving astabilised derivative, whereas the FeIII form combines reversiblywith potassium cyanide, hydrogen sulphide, sodium azide, etc.Onthe other hand, it has so far not been possible to demonstrate eitherdirect or indirect reduction of Fe111a3 by Fe% under strictly anaxobicconditions and in complete absence of other reducing substances.The original papers should be consulted.Fission of the C-C Link.Remarkably little is known of biological processes involving thedegradation of a carbon chain. Carboxylase (which decarboxylatesa-keto-acids, CHR,R,*CO*CO,H) and zymohexase (which splitsProc. Roy. SOC., 1939, B, 127, 167; 1940, B, 129, 277BELL BIOLOGICAL CATALYSIS. 417fructofuranosc 1 : 6-diphosphate into glyceraldehyde 3-phosphateand dihydroxyacetone phosphate) are probably the only authenticmembers of this class to be identified as individuals.Pyruvate is known to undergo two different degradation processesin vivo.I n yeast, carboxylase in the presence of thiamine (aneurin)diphosphate, forms acetaldehyde and carbon dioxide from pyruvate.I n animal tissues, and in certain bacteria, pyruvate is oxidised toacetate and carbon dioxide, thiamine diphosphate again beingnecessary. D. E. Green, D. Herbert, and V. Subrahmanyan3describe the isolation from brewer’s yeast of a highly active, stablepreparation of carboxylase containing 0.46% of thiamine diphosphateand 0.13% of magnesium, no other metal being detected. Theauthors regard their material as containing the complete enzyme.In high concentrations of salts the three components, protein,magnesium, and thiamine diphosphate, are bound together ; ondilution, dissociation takes place.Other bivalent cations canreplace magnesium ; the following were found active in descendingorder : Mn, Mg, Pel1, Cu, Cd, Zn.The oxidative fission of pyruvate by preparations of Bcsct. del-briickii has been studied by F. Lipmann.4 The original papersshould be studied for details; for present purposes i t will suffice todiscuss Lipmann’s views on the part played by phosphate in thisreaction. Recognising the importance of enolpyruvic acid phos-phate in glycolysis (Report for 1939, p. 358), he sought for possiblephosphorylated intermediaries in the oxidation. He found (i) thatphosphate was transferred to adenylic acid, and (ii) that crude acetylphosphate in the presence of the bacterial material could serve asa source of phosphate for reaction (i).Lipmann has thereforesuggested the following scheme for this oxidation :Pyruvate + phosphate - 2 hydrogen --+ Acetyl phosphate + carbon dioxideAcetyl phosphate + adenylic acid + Acetate + adenylpyrophosphate.It should be noted that S. Ochoa, R. A. Peters, and L. A. Stocken 5areport that acetyl phosphate does not act as an intermediary inpyruvate oxidation by brain nor does i t act as a phosphate donorto adenylic acid in muscle extract.The widely distributed zymohexase (Report for 1939, p. 359) hasbeen obtained as a very active preparation by D. Herbert, H.Gordon, and V. Subrahmanyan from the water-soluble protein ofJ . Biol. Chem., 1940, 135, 795.* Nature, 1939, 144, 381 ; Cold Spring Harbor Symposium on Quantitative6 E.Kamerer and G. Carius, Annalera, 1864, 131, 165.6 a Nature, 1939, 144, 750.Biology, 1939, 7, 248; J . Biol. Chem., 1940, 134, 436.Biochem. J . , 1940, 34, 1108.REP.-VOL. XSXVII. 418 BIOCHEMISTRY.rabbit muscle. Some 4% of this fraction was isolated with anactivity 150 times greater than that of the original tissue. Thepreparation was non-crystalline and contained C, 50.6 ; H, 7-04 ;N, 15.8 ; S, 1.25% ; neither phosphorus nor iodine nor significantamounts of carbohydrate were detected. An exhaustive study of theenzyme was made; it was concluded that no oxidisible or reduciblegroup was concerned in its activity; heavy metals had a markedinhibitory effect,. The material was examined by E. C. Bate Smithin the Tiselius apparatus.The Reversible Phosphorolysis of Xtcarch : Phosphorylation ofGlucose.I n the Report for 1939 (pp.359, 361) mention was made of thereversible splitting of glycogens from liver, muscle, and yeast intoglucopyranose l-phosphate. Two papers by C. S. Hanes describethe extension of this work to starch, enzymic preparations havingbeen obtained from both pea-seeds and potatoes (see this vol., pp. 419,421). I n such systems the normal course undergone by the glucose1 -phosphate is its conversion into hexose-6-phosphates. Only ifthe latter reaction can be prevented does the 1 -phosphate accumulate.I n that event synthesis of the polysaccharide tends to result, as theenzymic equilibrium is on the side of synthesis. It would thereforebe of interest to demonstrate the production by enzymes of thel-phosphate from glucose and H,PO,-.This has not yet beendone.Although yeast preparations are believed to esterify carbohydratewith phosphoric acid, i t is only recently that the phosphorylation ofglucose by animal tissues has been demonstrated, although only withrespect to kidney extrack H. Kalckar * has shown that phosphoricesters accumulate in such ext,racts, under zrobic conditions, whenfluoride is added to inhibit phosphatase activity. The process isstimulated by the presence of alanine, glutamic, citric, and fumaricacids, all of which can be oxidised by the kidney. Later, S. P.Colowick, M. S. Welch, and C. F. Cori showed that fructose diphos-phate and glyceric acid phosphate are formed in aerobic kidneyextracts in the presence of fluoride.The authors consider that thephosphorylating activity is effected through energy derived from theoxidation processes involving dicarboxylates, e.g., succinate tofumarate. Adenylic acid, coenzyme I and Mg" are necessary. Thesame authors have further shown that, in the absence of fluoride,added fumarate catalyses the oxidation both of glucose and ofProc. Roy. Soc., 1940, By 128, 421 ; 129, 174.Enzymolog-ia, 1939, 6, 209.J. B i d Chem., 1940, 133, 359. 10 Ibid., p. 641NORRIS : PLANT BIOCHEMISTRY. 419pyruvate by dialysed kidney extract. They therefore consider thatfumarate is an essential link between the phosphorylation of glucoseand its subsequent oxidation.Carbonic Anhydrase and Zinc.D.Keilin and T. Mann 11 have isolated extremely active proteinpreparations from erythrocytes and gastric mucosa. The materialis remarkable in containing zinc, which appears to be necessary forits activity. It has further been shown l2 that sulphanilamide, andin general, sulphonamides, act as specific inhibitors for carbonicanhy drase.Crystalline Preparations of Enzymes.The following list gives those enzymes which to date have beenobtained in the form of crystalline protein preparations. (For adiscussion on the problem of the homogeneity of crystalline proteins,see this vol., p. 405.)Pepsin,13 pepsinogen, trypsin, trypsinogen, chymotrypsin, chymo-trypsinogen, papain,l*, carboxypeptidase, n ~ c l e a s e , ~ ~ urease, catal-ase, alcohol dehydrogenase, triosephosphate dehydrogenase,16 lacticdehydrogenase of heart,17 tyrosinase.la(References are given only to the most recent advances.)D.J. B.8. PLANT BIOCHEMISTRY.Xome Products and Enzymes of Plants.Starch and Amy7ases.-One of the outstanding achievements ofthe year in the field of carbohydrate biochemistry is due to C. S.Hanes,l who has published interesting and valuable papers on thebreakdown and synthesis of starch in the higher plants, and has beenable to effect the synthesis in vitro. The origin of the investigationsmay be said to arise from an attempt to discover whether thephosphorylated sugars play a similar r61e in the carbohydrate meta-boli~m of the higher plants, to that which they have been shown to11 Biochem.J., 1940, 34, 1163.le T. Mann and D. Keilin, Nature, 1940, 146, 164.la V. Desreux and R. M. Herriot, ibid., 1939, 144, 289.l4 A. K. Balls and H. Lineweaver, J . Bwl. Chem., 1939, 130, 669.1 5 M. Kunitz, J . Gen. Physiol., 1940, 24, 15.l 6 0. Warburg and W. Christian, Biochem. Z., 1939, 303, 40.1' I?. B. Straub, Biochem. J., 1940, 34, 483.1 8 H. R. Daltonand J. M. Nelson, J . Amer. Chem. Soc., 1939, 61,2946.1 Proc. Roy. SOC., 1940, B, 128, 421420 BIOCHEMISTRY.do in the case of yeast and some animal tissues. Necessary steppingstones to the proof of such a theory are the discovery of phosphoryl-ated sugars, and of enzymes capable of acting on them, in the planteconomy. A phosphorylating enzyme had been discovered byJ. Bodnar in 1925 in the flour from ground mature peas, andB.Tank6 in 1936 was able to show that such an enzyme convertedinorganic phosphate in a phosphate buffered mixture of the flourinto fructofuranose 1 : 6-diphosphate. Hexose monophosphatesalso were present. Hanes has found that the enzyme, termedphosphorylase, separated from the tissue, is able to phosphorylatestarch and a number of starch dextrins and that simple sugars withthe exception of maltose, which is attacked only slowly, are un-attacked. The first product in the phosphorylation of starch is thenon-reducing glucose l-phosphate, and this is found to be a reversiblereaction in that addition of glucose l-phosphate to an extract ofpeas involves the production of a certain proportion of starch withliberation of free phosphate.An alternative transformation ofglucose 1 -phosphate involving two distinct enzymic actions hasbeen observed : the glucose l-phosphate rapidly disappears fromthe system and a mixture of glucose-, fructose-, and mannose-6-phosphate is produced. Tank6’s observation that hexose diphos-phate is formed when pea flour suspensions act in presence ofphosphate is confirmed and i t is further found that addition of starchto the mixture greatly accelerates the esterification. If dialysedextracts are used, however, the diphosphate is not formed and itappears that such formation depends on the presence of a dialysableco-en zyme.The glucose 1 -phosphate appears to be identical with that obtainedas the first product of esterification in the action of muscle phos-phorylase on gly~ogen.~Observation shows that it is probable that the glucose 1 -phosphatearises by direct phosphorylation of the saccharide chains in starch.Thus starch and complex dextrins are esterified more rapidly andcompletely than the lower members of the starch-maltose series ;that the higher members are directly involved in the reaction isshown by the decrease in iodine colour under phosphorylase action;and the change in iodine colour corresponds with what might beanticipated from an endwise degradation of the chains.Glucoseitself is not esterified and from these facts i t is concluded thatterminal glucose units of the chains are phosphorylated and liberatedfrom non-aldehydic chain-ends.Biochern. Z., 1925, 165, 1.Biochem.J . , 1936, 30, 692.C. F. Cori and G. T. Cori, PTOC. SOC. Exp. Biol. Med., 1936, 34, 702NORRIS : PLANT BIOCHEMISTRY. 42 1Hanes summarises the preceding results as follows :I (Phosphorylase)+phosphate - - ' Starch , \L Glucose 1 -phosphateI - phosphate II11(Phosphoglucose-conversionenzymes) + (Amylase) Reducing hexose 6-phosphatesI Vphosphate (Enzyme + dialysable co-Fructose 1 : 6-diphosphatewater + I enzyme)J.Dextrins, maltose, glucoseAt a later stage, Hanes describes the preparation of phosphoryl-ase from potato tubers; the enzyme is obtained in highly activeform and is free from those enzymes which promote the alternativetransformations already referred to. The reversibility of the starch-glucose 1-phosphate reaction is shown, since the ratio of inorganicphosphate to ester reaches the same value in either direction.Moreover, the equilibrium point is not affected by wide variations inconcentrations of the reactants or enzyme.The position of equili-brium is notably altered by changes in pOH and it has been shown thatthis is due to the effect of varying hydrogen-ion concentration on thedissociation of inorganic and ester phosphate. The bivalent ionsdetermine the equilibrium, and between pH limits of 5 and 7 a constantvalue of 2-2 is found for the ratio [HPO,"]/[C,H,,O,*O*PO,"].Large-scale preparations of pure glucose 1 -phosphate as thecrystalline potassium salt are described, and by the action of purifiedphosphorylase on glucose l-phosphate large amounts of a poly-saccharide resembling potato starch have been obtained.Theproduct shows the typical granules of native starch when viewedunder the microscope ; but further examination shows certainpoints of difference from natural potato starch. For example, thesynthetic product is sparingly soluble in water and rapidly retro-grades when in solution; again, the iodine colour is much moreintense than that given by potato starch. A third point of differencewas obrjerved in relation to its degradation by the p-amylase ofungerminated barley. Natural starches normally degrade to about60% of maltose and the residue consists of resistant a-amylodextrin.The enzymically prepared starch, however, allows the action of thep-amylase to proceed almost to completion, 95-100% of the prepar-ation being converted into maltose.The synthetic product corre-Proc. Roy. SOC., 1940, B, 129, 174422 BIOCHEMISTRY.sponds thus most closely to the amyloamylose fraction of starch.Much interest thus attaches to further investigation of the enzymemechanisms involved in these remarkable transformations and toconstitutional studies on the synthetic polysaccharide, which arenow proceeding.X-Ray comparisons between the natural and the synthetic starchare discussed by W. T. Astbury, F. 0. Bell, and C. S. Hanes 7 ina recent note. The A-, B-, C-, and V-powder photographs corre-spond to wheat starch, potato starch, a mixture of the two, andto alcohol-precipitated starch respectively.8 The only differenceobservable between the photographs of potato starch and thesynthetic starch is that the former is very slightly sharper.Afurther curious point was discovered in that amyloamylose precipit-ated by alcohol after preparation by electrophoresis gave theV-photograph, whereas the synthetic starch still gave the B-photo-graph after precipitation with alcohol. A re-examination of theconditions governing the production of the different types of X-rayphotograph may throw further light on the relation between thevarious types of starch. It is not clear a t present whether differentphotograph types are due to differences in phosphorylases or, asseems more probable, in the method of crystallisation, ie., prepar-ation of the starches.The probability of the latter explanation isindicated by the fact that A-, B-, or C-photographs may beobtained from the same starch a t different temperatures ofdepo~ition.~A number of well-known factors influence the rate of breakdown ofstarch by enzymes and the nature of the products. It has recentlybeen shown that comparatively mild treatment of starch prior todiastatic action leads to a much greater degradation than is possiblewith untreated starch. Thus, R. H. Hopkins, E. G. Stopher, andD. E. Dolby1* describe the separation of starch into amylopectinand amyloamylose by electrophoresis, the starch having beenpreviously treated in one of two ways. I n some experiments thestarch was finely ground in order to rupture the granules ; in othersthe starch was made directly into a paste and dispersed a t 120".By alternate redispersion and electrophoresis the amylopectinfraction yields high proportions of amyloamylose, and the latteris far more completely degraded by barley diastase, but a t a slowerrate, than the untreated starch.The conversion into amyloamyloseM. Samec, 2. physiol. Ghem., 1936, 238, 103.Nature, 1940, 146, 558.J. R. Katz, " Die Rontgenspektrographie usw.," Berlin and Wien, 1934.J. R. Katz and J. C. Derksen, Z . physikal. Chem., 1933, A , 165, 228.l o J . Inst. Brew., 1940, 48, 426NORRIS : PLANT BIOCHEMISTRY. 423may reach yields as high as SOY0 if dispersion is carried out above100" ; the product reverts on standing to a substance which behavestowards diastase as does the starch from which i t is obtained.I n recent years some doubt and minor controversy have arisenwith reference to the effect of the particular buffer solution employedon the activity of enzymes.Not only has the nature of the bufferbeen held to affect the activity, but it has been suggested that ionicstrength is an important factor. Experiments designed to investig-ate these points in the case of taka-diastase are described by G. A.Ballou and J. M. Luck.ll Using a starch substrate, a temperatureof 30°, and a series of buffers of an ionic strength of 0.05~, they foundthat the optimum pE, in respect of saccharogenic action, was 5.1 forformate, acetate, propionate, butyrate, valerate, phenylacetate, andsuccinate buffers.A slight shift to pH 5-4 was observed for bufferscontaining phthalate or citrate. Where the enzyme was employedat the optimum pH or a t a reaction on the alkaline side of this value,variation in the anion of the buffer was without effect on the relativeactivity; but on the acid side of the optimum, marked differenceswere observed with varying anions.Employing the usual methods of methylation and end-group assay,E. G. E. Hawkins, J. K. N. Jones, and G. T. Young l2 have foundthat the starch present in unripe bananas conforms structurally tothe usual pattern in that the repeating unit consists of a chain ofabout 24 glucose residues. Similarity to rice starch l3 is furtherindicated by the close resemblance of the course of disaggregation inthe two cases.Disaggregation was discussed in last year's Report l4and is effected in methylated starches by hydrolysis with oxalicacid in a mixture of methyl alcohol and water. The hydrolysis takesplace a t the bonds between the repeating units, since, althoughproducts of lower molecular weight are obtained, the chain lengthof the repeating unit remains unchanged.A more rapid method of assay of the end group in methylatedstarch is reported by S. Peat and J. Whet~t0ne.l~ The methodinvolves acetolysis of methylated starch by acetyl bromide in coldchloroform. Treatment of the product with methyl alcohol gives amixture of methyl glucosides from which the tetramethyl glucosideis separated by distillation. The removal of the end group is selec-tive and rapid, the tetramethyl derivative separating within fiveminutes of the commencement of the reaction.Pectin, MuciEages, etc.-Constitutional studies on pectic acidscontinue and are involving the preparation and characterisation ofl1 J .Biol. Chern., 1940, 135, 111.l 3 E. L. Hirst and G. T. Young, J., 1939, 1471.l4 Ann. Reports, 1939, 36, 272.l2 J., 1940, 390.l6 J . , 1040, 276424 BIOCHEMISTRY.a number of derivatives of galacturonic acid exhibiting a fructo-furanose structure. Among these may be mentioned the methylester of 2 : 3 : 5-trimethyl P-methylgalacturonoside, which has beensynthesised by S. Luckett and F. Smith.16 In the course of investig-ations on citrus pectic acids, these authors l7 have also isolated themethyl ester of 2 : 3-dimethyl methylgalacturonoside and havesuggested that citrus pectic acid is built up of pyranose residues ofgalacturonic acid joined by 1 : 4 a-glycosidic linkages.Osmoticpressure measurements on the methyl ester of methylated pecticacid appear to indicate a small molecule of about 13 units.Using such sources as Tuso pith and sliced radishes, S. Onon l8has isolated pectic substances on the usual lines and with the usualproperties. A method of extraction is of interest, however, in thatthe extractive is boiling water containing copper sulphate. By thismeans it is claimed that pectins are obtained of a snow-white colour,and of a galacturonic acid and methoxyl content much higher thanthose normally obtained. The resulting pectin is considered to be apolymerised trimethyl tetragalacturonic acid and contains no arabanor galactan residues.Valuable contributions to the study of jelly formation by pectinhave been made by C.L. Hinton,20 who suggests that pectins maybe regarded as complex mixtures of carboxylic acids whose con-stituents cannot be separated a t present. However, they may bestudied from a physicochemical standpoint and much informationhas been gained in this way. Thus the effect of variation of anumber of factors, such as concentration of pectin and of othersoluble substances present, the pH of the mixture, the effect ofde-esterification of pectin by pectase or alkali, on the strength of thejelly has been studied. Electrolytic dissociation of pectin wasinvestigated and the " constant " was found to diminish as neutral-isation with alkali became more complete ; but there was no import-ant change in the constant for pectins boiled for one hour orde-esterified enzymically, or by citric acid. Jelly formation involvesonly those molecules of pectin which are in the un-ionised condition,and these must reach a certain solubility or saturation limit varyingwith the total solids concentration of the mixture.The strength ofa particular jelly with a given buffer salt was found to be propor-tional to the amount of non-ionised pectin above the solubility limit.The ratio of jelly strength to the amount of jellying substancediffered for different pectins. This theory of jelly formation hasbeen elaborated and has explained many of the phenomena observed,especially those relating to the effect of pH.Encouraging resultsl7 Ibid., p. 1106.2o Biochem. J . , 1940, 34, 1211.l6 J . , 1940, 1114.Bull. Sch. Agric., Taihoka, 1940, 1, 1NORRIS : PLANT BIOCHEMISTRY. 425may be expected to follow from this investigation, which is probablythe first attempt to treat in ordered and mathematical fashion themass of diffuse and often contradictory data which have tended toobscure rather than clarify the problem.The constitution of the mucilage of carrageen moss (Chondruscrispus) is the subject of investigation by T. Dillon and P. O’Colla,21who submit the mucilage to acetolysis in the usual manner andobtain, after removal of acetyl groups, two polysaccharides whichappear to be galactans. These correspond to those found by P.Haaset a1.,22 in that one is soluble in water and one soluble only in hotwater. The latter workers were unable to isolate the polysacchar-ides, which they showed to exist as calcium salts of ethereal sul-phates. The carbohydrate hydrolysis products appeared to bechiefly galactose and fructose, but Dillon and O’Colla were unableto confirm the latter, although indications were obtained thatfructose was present in solution on deacetylation.I n another communication on the same subject, E. G. V. Percivaland J. Buchanan 23 confirm the work of Haas and criticise the laterfindings of Dillon and O’Colla on the grounds that their method ofacetolysis involved considerable degradation of the polysaccharides,all constituents except galactose being lost.This indicates thatgalactose must be contained in the most resistant portions of themolecule, but the complexity of the latter is indicated by otherhydrolysis products, which appear to include glucose, a pentose, andpossibly a ketose. The probable configuration of the molecule isresponsible for the difficulties observed in acetolysis and methylation.A product of the nature of a polysaccharide and hydrolysinglargely to xylose is reported for the first time from the red alga,Rhodymenia palmcata, by V. C. Barry and T. Dil10n.~~ The alga isimmersed in dilute hydrochloric acid for 24 hours and from theresulting viscous solution a white precipitate is obtained withalcohol. The precipitate yielded crystalline xylose in hydrolysiswith dilute nitric acid.A similar treatment on the case of Dilseaedulis produced a substance which appeared to be similar to themucilages from other marine a l g a It was an ethereal sulphate,contained no xylose, and was oxidisable to mucic acid.Attempts have been made from time to time to isolate poly-saccharides similar to naturally occurring plant gums by the actionin vitro of bacteria, normally associated with a particular plant, onsucrose in artificial media. The work of E. A. Cooper and J. F.21 Nature, 1940, 145, 749.22 Biochem. J., 1921, 15,469; 1922, 16, 578; 1929, 23,425.23 Nature, 1940, 145, 1020.24 Ibid., 1940, 146, 620426 BIOCHEMISTRY.Preston has been recorded in these Reports 25 and the subject isrenewed by R. R.Lyne, S. Peat, and M. Stacey,26 who find thatpolysaccharides of the levan type are formed by B. megatheriurn,Bact. pruni, and Bact. prunicoh under the above conditions. So far,polysaccharides comparable with the gums have not yet beenproduced. Examination of the hydrolysed methylated levansshowed that each could be represented by a chain of 10-12 fructoseunits, and in this respect they resembled the levans produced byB. subtilis. Differences in properties among the levans are thoughtto be due to varying degrees of aggregation of the repeating unit,.A practical point of some importance is the observation that thepolarimetric rotation of the acetates in chloroform solution dependson the water content of the reaction mixture and this probablyexplains discrepancies in results by previous workers.HemiceZZuZoses.-Although much analytical investigation of thecell-wall of plant tissues has been conducted, in most cases it is themttture material which has been examined, and little work by com-parison has been expended on the developing tissue.Such investig-ation would prove of value in the formulation of theories of theorigin of the constituents of mature tissue. A. Allsopp and P.Misra 27 have studied the common ash, the common elm and Scotchpine, whose tissue they divide into three ‘‘ fractions ” : the cambium,together with the differentiating xylem ; newly formed wood ; andmature sapwood. The composition of the first-named fractionshowed similarities with that published for other young tissues. Ahigh pectin content is characteristic of the young cell wall and inthis and the lower lignin : cellulose ratio the cambium and differenti-ating elements contrast chiefly with the mature wood.Even aftervessel differentiation is complete there are small changes in com-position, involving the loss of pectin, increase in encrusting pentosansand in the resistance of the lignin constituents. Theories of originof the constituents of mature cell-walls based on the observed factthat, for instance, a fall in pectin concentration during lignificationindicates that the lignin arose from the pectin, must be treated withreserve, since a change in concentration is no indication of a changein total quantity present.The same remark applies to the suggestionthat hemicelluloses are derived from pectin.Osmotic pressure and viscosity studies on a number of polyosesof wood by E. Husemann 28 indicate a wide range of molecular size,the particles responsible for the development of osmotic pressurebeing molecules and not molecular aggregates. Degree of poly-merisation is of the order of 150-220 units in the case of xylans,25 Ann. Reports, 1938, 35, 378.2 7 Biochern. J., 1940, 34, 1078.26 J., 1940, 237.** Naturwiss., 1939, 27, 595NORRIS : PLANT BIOCHEMISTRY. 427from wheat straw and beechwood, mannan from spruce and arabo-galactan from larch. These values are small compared with t'hatfor beech cellulose, which is given as not less than 1500.The degreeof polymerisation is unaltered on conversion of the products intoacyl derivatives. On the basis of viscosity measurements it isthought that, whereas the mannan and xylans comprise long chains,the arabogalactan molecule consists of multi- branched chains.M. H. O ' D ~ y e r , ~ ~ continuing her investigations on the hemi-celluloses of oak wood, has found that the transition from sap-woodto heart-wood as shown by preparation of hemicellulose A from thetwo types involves definite constitutional changes in this com-ponent. The greater proportion of the carbohydrate residues in themolecule are anhydroxylose and these are combined with uronic acidand methoxyaldobionic acid residues, the former predominating inthe hemicellulose of the sap-wood, the latter in that of the heart-wood.A further difference was also noticed, in that the hemicellulosepreparations from sap-wood all gave the iodine blue colorationtypical of starch, whereas the heart-wood preparations gave noiodine colour. Although a t this stage glucose was not found as ahydrolysis product of the sap-wood preparations, i t was suggestedthat the blue coloration was due to the presence of anhydroglucoseresidues in the hemicellulose.I n a later communication O'Dwyer 30 reports that hemicellulose Amay be split up into two polysaccharide fractions by the action oftaka-diastase and in addition some 10% of the glucose has beenobtained from sap-wood hemicellulose A. The action of water a t100" effects a similar scission, but glucose is not split off.Thepresence of glucose is seen to be the only difference between thehemicellulose A of sap- and h e a r t - ~ o o d , ~ ~ the polysaccharide frac-tions after prolonged hydrolysis with taka-diastase being identical.The complete hydrolysis involves production of two parts of a solublepolysaccharide and three parts of xylose. The molecule of the poly-saccharide appears to consist of six anhydroxylose units combinedwith one methyluronic anhydride unit.I n the latest paper 32 to date similar results are obtained withhemicellulose B. Again the sap-wood hemicellulose only gives theiodine coloration and contains anhydroglucose units. The productsof fission under the action of taka-diastase appear to be the same forall the hemicelluloses examined.Hemicelluloses have been pre-pared by E. Anderson, M. Seeley, W. T. Stewart, J. C. Redd, andD. Westerbreke 33 from various hardwoods before and after chlorin-29 Biochem. J., 1934, 28, 2116.31 Ibid., 1939, 33, 713.33 J . Biol. Chem., 1940, 135, 189.30 Ibid., 1937, 31, 254.32 Ibid., 1940, 34, 149428 BIOCHEMISTRY.ation. Their results confirm much that has been suggested byW. G. Campbell,34 M. H. O’Dwyer, and others with reference to theorigin of hemicelluloses in woods. Two of the woods examined,lemon wood and black locust sap-wood, contained starch and allof the hemicelluloses from these woods gave a blue or pink colorationwith iodine. These hemicelluloses appear to contain anhydroglucosegroups in the xylan chain, and may possibly represent intermediateproducts in the transformation of starch or its degradation productsinto hemicelluloses.The hemicelluloses of birch wood and blacklocust heart-wood did not contain starch, gave no typical colourwith iodine, and were differentiated from the other hemicellulosesin chemical composition, since the xylan groups were combined witha monomethylated uronic acid. The probable number of xylangroups in the chain varies with different hemicelluloses and appearsto approximate to 19 in the largest molecules and 8 in the smallest.The carboxyl group of the uronic acid may be involved in attach-ment of the hemicellulose to some other substance of the cell-wall.This might explain the fact that, although the hemicellulose is notextractable from the wood by hot water, it is nevertheless somewhatsoluble in hot water after extraction with sodium hydroxide.Hemicelluloses prepared by the usually recognised methods arereported from a number of sources, including oat wheat-straw,36 and lucerne hay.37 Those obtained from wheatstraw con-sist mainly of the B-fraction with small amounts of A and C.Thehydrolysis products include xylose, arabinose, and possibly a methylderivative of glucuronic acid ; the xylose predominates. The sameremarks apply to the hemicellulose of lucerne hay.The methods adopted in the pre-treatment of materials employedfor the preparation of hemicelluloses have given rise to some contro-versy, a principal bone of contention being the use of alcoholicsodium hydroxide for the removal of lignin.This method was usedby P. W. Norris and I. A. P r e e ~ e , ~ ~ but was shown later by the latterauthor 39 to involve some loss of furfuraldehyde-yielding material.The procedure was also criticised by A. G. Normanm on similargrounds. Nevertheless, S. Angel1 and P. W. Norris41 found thatin the pre-treatment of the flowers of the hop, no such loss could beobserved after treatment with alcoholic soda, and they suggestedthat the effect of alcoholic soda depended on the material under34 Biochem. J . , 1935, 29, 1068.35 P. W. Krznarich, Cereal Chem., 1940, 17, 457.36 H. D. Weike and M. Phillips, J. Agric. Res., 1940, 60, 781.37 M. Phillips and B. L. Davis, ibid., p. 775.38 Biochem.J., 1930, 24, 59.4 0 Ibid., 1935, 29, 545; 1937, 31, 1579.39 Ibid., 1931, 25, 1304.41 Ibid., 1936, 30, 2159NORRIS : PLANT BIOCHEMISTRY. 429investigation. The subject is again opened by I. A. P r e e ~ e , ~ ~ whoemployed teak sawdust as the raw material, and showed that thereis definite loss of furfuraldehyde-yielding material when the wood issubmitted to either alcoholic or aqueous soda extraction. Theextracted hemicelluloses themselves were not stable under thesereagents. The author concludes that pre-treatment with alcoholicsoda does reduce the lignin content of subsequent preparations ofhemicellulose, but the disadvantage of the treatment outweighs theadvantages. The choice of extractive must depend on the materialand on the purpose in view in the preparation of tjhe hemicellulose.Alcohol extraction may be employed in some cases, or alcoholic sodatreatment in the cold, as employed by H.W. Buston43 and A. G.Norman .40Plant Proteases.-The isolation of papain in crystalline form hasbeen achieved by A. K. Balls and H. Linewea~er,~~ who precipitatedpapaya latex successively with ammonium sulphate and sodiumchloride a t suitable concentrations and pa in the presence of sodiuacyanide. The crystals were usually obtained in fine needles, whichchanged on long standing in water to elongated hexagonal plates.Analysis showed that the enzyme was of a protein nature, containing15.504 of nitrogen, l.Syo of total sulphur, and 174 of cysteine sulphur.A molecular weight of about 30,000 was indicated by osmotic pressuremeasurements and by the evidence of the ultra-centrifuge.Theenzyme was only sparingly soluble in dilute saline solutions, butsoluble in 70% alcohol. The isoelectric point was a t p , 9. Thepresence of an activator such as cyanide or cysteine was essential forthe preparation of crystals of high activity and such activitycorresponded to a maximum value per unit of protein nitrogen forall crystals, although prepared in different ways.The presence of sulphydryl groups as necessary for papain activityis thus again indicated by the preceding work on crystalline papain.Activation-inhibition reactions of a group of similar plant proteasesare investigated by D. If. Greenberg and T. W i n n i ~ k , ~ ~ who employthe bromelin of pineapples ; solanain from the horse-nettle, Xolanumehagnifolium ; a new protease from the latex of milkweed, Asclepiasmexicanca ; and a protease from another milkweed, Asclepiasspeciosar.In accordancc with a recent suggestion i t is proposed touse the suffix " ain " for new plant proteases, and for presentpurposes the last two enzymes are designated asclepain m andascbepain s. All the enzymes with the exception of solanain showreactions which suggest that they are all related to papain, thepresence of a sulphydryl group being necessary for activity. Sola-43 Ibid., 1934, 28, 1028.4 5 Ibid., 1940, 135, 761.4 2 Biochem. J., 1940, 34, 251.4 4 J. Biol. Chem., 1939, 130, 669430 BIOCHEMISTRY.nain is unaffected by oxidising or reducing agents or reagentswhich react with sulphydryl groups. It is not a papainase and itis probable that phenolic groups may be essential for its activity asindicated by inactivations produced by nitrous acid and keten.There is some evidence also that the papainases may require phenolicin addition to sulphydryl groups.pH-Activity curves for the different enzymes, obtained withdenatured hzemoglobin and ovalbumin as substrates, indicate anoptimum pH with hzmoglobin of 6.5-8.5, and with ovalbumin of7-76.The character of the curve has been shown to depend moreon the electrochemical nature of the enzymes than on the degree ofdissociation of the s ~ b s t r a t e s . ~ ~In a third communication 4' the authors discuss the kinetics of theaction of the enzymes, and determine the Michaelis constants ineach case. I n all cases the intermediate compound of enzyme andsubstrate consisted of equimolecular proportions of enzymes andprotein.The heat inactivation of asclepain m and solanain followedthe course of a first-order reaction and these enzymes closelyresembled papain and bromelin in their high critical thermalincrements.Growth Substances.Higher PZants.--l'he mechanism of the pea test for auxin has beenfurther investigated by F. W. Went,48 who finds that the curvatureof split etiolated pea stems under the action of the auxin is due toa loss of sensitivity of the tissues bordering the wound. This lossof sensitivity is essential before auxin eff'ect will take place, and isof the nature of a preparatory action, which is independent of p ,and may be effected by substances which lack growth-promotingactivity.The dual aspect of auxin action is again referred to byF. W. Went,49 who distinguishes two phases in the action of indolyl-acetic acid on etiolated pea stem cuttings. There is an initial effectwhich causes a redistribution of rhizocaline, and such effect isbrought about by phenylacetic acid, which is not in itself a growthpromoter ; the later effect is induced only by indolylacetic acid andits homologues and appears to involve the activation of accumulatedrhizocaline .K. V. Thimann and C. L. Schneider 50 have shown that therelative activities of growth substances as compared with indolyl-acetic acid vary with the species of plant treated and with differentmethods of treatment in the same species.They record results witha number of growth substances, and employ straight growth tests4 6 J . Biol. Chem., 1940, 135, 775. 4 7 Ibid., p. 781.4 8 Bull. Torrey Bot. Club, 1939, 66, 361.4 9 Amer. J. Bot., 1939, 26, 24. Ibid., pp. 328, 792NORRIS : PLANT BIOCHEMISTRY. 431with Pisum, and a new method of curvature test wherein Avenacoleoptiles are slit longitudinally and thus grown in auxin solutions.It is claimed that the test is some 30 times as sensitive as the agarblock method, and that indolylacetic acid may be detected inconcentration as low as 0 . 0 1 pg. per litre.That auxin is present in bound form which is only slowly split upby an action which is probably enzymic is suggested by F.Skoogand K. V. Thimann.51 They find that extraction of auxin withether is complete only after some months and it is assumed that thisis due to the slow hydrolysis of auxin in the bound form. That themechanism may be enzymic is indicated by the fact that addition oftrypsin preparations accelerates the extraction according to theparticular preparation employed.D. M. Bonner 52 has shown that acids of similar molecular struc-ture have the same activity in growth reactions, dependent on pE.Thus, the acid-induced curvature in split sections of pea stemsresults from an increase in active auxin produced after a change inthe internal pH of the cut surface. Dissociation measurements on anumber of acids such as cis-cinnamic, indolyl-acetic, -propionic, and-butyric, and naphthylacetic show that equimolar concentrations ofthe acids have the same activities.The effect of p , in respect ofthe growth reaction of Ave,na coleoptiles has been examined byJ. V. Rakitin and L. M. J a r k ~ v a j a , ~ ~ who find that increasingacidity over the range 6-32-34 enhances the auxin effect. Ob-servations were made with an acetate buffer and with oxalic, citric,malic, and sulphuric acids.N. H. Grace 54 has compared the activity in inducing rooting ofcuttings shown by acids of the w-naphthyl-aliphatic series, and findsactivity for acids up to and including the hexoic acid. Acidshaving an even number of carbon atoms in the side chain have agreater activity than those with an odd number.In experimentson rooting of cuttings of Lonicera tartarica, indolylbutyric acid wasmost effective ; indolylacetic and indolylpropionic acids showed lessactivity, and 5-methylindolylpropionic and indolylvaleric acids noaction at all.A. E. Hitchcock and P. W. Zimmerman55 have examined thecombined effect of mixtures of root-inducing and other substancesand find that in a number of cases the mixtures are more effectivethan the separate components. Such mixtures frequently givegreater numbers of roots, a higher percentage of rooted cuttings and51 Science, 1940, 92, 64.63 Compt. rend. Acad. Sci., U.R.S.S., 1939, 23, 952.5p Canadian J . Res., 1939, 17, C, 247, 373.5 5 Contr. Boyce Thompson Inst., 1940, 11, 143.52 Bot.Gaz., 1938, 100, 200432 BIOCHEMISTRY.concomitant phenomena associated with large amounts of growthsubstances. Mixtures of indolyl-acetic and -butyric acids, and ofnaphthyl- and phenyl-acetic acids with vitamins B, and B, andethylene were used. It appears that the vitamins have no root-inducing function in themselves, but act rather as root formationactivators.It has been recognised for a long time that the concentration ofheteroauxin applied to the plant is somewhat critical, and that ingeneral higher concentrations not merely have little effect, but mayactually inhibit those effects which in lower concentrations arepromoted. This finds support in experiments by Y. Hwang andH. L. P e a r ~ e , ~ ~ and by L. D~hamet.~’ The former found that diluteindolylacetic acid had little effect on oat and bean seedlings, andthat concentrations higher than 2 parts per million retarded growth.Indolylacetic acid is evidently only a growth stimulator whennatural auxin is deficient.The latter worker, employing extremelylow concentrations of indolylacetic acid, finds no effect with thelowest, double the growth rate with a slightly less dilute solution, andinhibition above a certain concentration. Growth of roots ofLupinus albus was observed in these experiments, and the smallamounts necessary to induce the auxin effect may be gauged wheni t is stated that the best response was obtained with solutions ofGrowth substances applied in vapour form were found to give thesame characteristic responses as when applied as aqueous solutions,by P. W.Zimmermann, A. E. Hitchcock, and F. W i l c o ~ o n . ~ ~Similar experiments with vapours and solutions of growth substanceswere later described by P. W. Zimmermann and A. E. H i t c h c ~ c k , ~ ~who examined 54 substances, including 26 reported for the first time.All produced formative effects, although these differed in characterwith the substance used. It was notable that the more active sub-stances employed in vapour form produced emanations from thetreated plants which affected their neighbours. After a period of onehour from exposure to the vapour, when carbon dioxide productionfell below that of controls, the production for the next five hoursexceeded that of the controls. The same authors also compared theefficiency of application of root-inducing substances as dustingpowders and in solution.The two methods appear to have aboutthe same efficiency, but it was found that in the case of applicationin a talc dusting powder, the talc itself had some beneficial influence,in part due to the improvement in water relationships and in partnormality.5 6 Ann. Bot., 1940, 4, 31.6 8 Contr. Boyce Thompson I n s t , 1939, 10, 363.61) Ibid., pp. 481, 461.5 7 Cornpt. rend., 1939, 208, 1838NORRIS : PLANT BIOCHEMISTRY. 433possibly due to the fact that talc appears to contain a physiologicallyactive ingredient, which could be extracted by chloroform.Further investigation 6o of the traumatin (wound hormone)isolated by J. English, J.Bonner, and A. J. Haagen-Smit 61 hasshown it to be A1-decene-1 : 10-dicarboxylic acid, and the nametraumatic acid has been given. The acid promotes wound peridermformation in potato and inhibits germination in seeds of tomato.In this respect decanedicarboxylic acid and sebacic acid were foundto be about half as active as traumatic acid.A new growth substance is reported by S. C. Bausor,62 who hasfound that naphthoxyacetic acid and its sodium and potassium saltsinduce typical responses when applied in lanoline paste. The effectswere recorded in detail in the case of root primordia of tomatoplants.A plant growth inhibitor was discovered by W. S. Stewart,W. Bergren, and C. E. Redemann 63 in cotyledons of radish plants,and a similar substance is reported by R.H. Goodwin 64 in etherealextracts of maize meal and bean shoots ; further confirmation of theexistence of substances which inhibit growth, or a t least mask theaction of auxin, is provided by R. Snow 65 by the isolation of awater-soluble substance from pea leaves soaked in ether.Yecast.-The final elucidation of structure and synthesis of panto-thenic acid has been reported.66 The high structural specificity ofpantothenic acid is indicated as the result of biological examinationof amino-acid analogues of the acid carried out by H. H. Weinstock,E. L. May, A. Arnold, and D. Price.67 Esters of aspartic acid,alanine, lysine, and P-aminobntyric acid were prepared and con-densed with ay-dihydroxy-pp-dimethylbutyric acid-the non-nitrogenous moiety of pantothenic acid.None of the syntheticanalogues was active, although the close relationship of theamino-acids mentioned to 8-alanine is sufficiently obvious.E. E. Snell, R. E. Eakin, and R. J. Williams 68 have utilised thegrowth response of Sacch. cerevisict! in presence of p-alanine andvitamin B6 to determine minute amounts of biotin in naturalmaterials such as autolysed liver, whey solids and cane molasses,which are among the richest sources of biotin. The latter is thoughtto be an a-amino-acid. A similar biological method of assay appliedto pantothenic acid is described by D. Pennington, E. E. Snell, andR. J. Williams,69 who utilise the respoiise of LactobaciZZus casei E.6o Science, 1939, 90, 329; J . Amer. Chern.Soc., 1939, 61, 3434.61 Ann. Reports, 1939, 36, 369.63 Science, 1939, 89, 185.65 Nature, 1939, 144, 906.6 7 J. Biol. Chern., 1940, 135, 343.69 J . Biol. Chern., 1940, 135, 213.62 Amer. J . Bot., 1939, 26, 415, 733.64 Amer. J . Bot., 1939, 26, 130.6 6 See this vol., p. 226.J . Amer. Chem. SOC., 1940, 62, 175434 BIOCHEMISTRY.The requirements of yeast for the recognised bios substances areknown to vary with the strain of yeast. Further confirmation ofthis is forthcoming as the result of recent work by R. J. Williams,R. E. Eakin, and E. E. Snell,'O who also note that the necessity forthese substances appears to diminish after long incubation. Theyalso report the presence in liver extracts of additional unknownstimulants, in support of an earlier paper by E.P. Pratt and R. J.Williams 71 in which it was stated that the respiration and growthof yeast are stimulated to a much greater extent by liver extract thanby any other known stimulant. Pantothenic acid not only exertsa stimulating effect on growth and respiration of living yeast buthas been shown to enhance fermentation by dialysed yeast macer-ation juice. Suggestions that liver extract contains a yeast stirnul-ant find support from the work of B. Alexander and Y. S~bbarow,'~who have obtained an acetone extract from liver which is active asa bios substance but appears on biological test to contain a substancedifferent from any of those a t present recognised as yeast growthstimulants. It is stable to heat and to acid hydrolysis and to someextent to alkaline hydrolysis, and is soluble in organic solvents.Itis precipitated by phosphotungstic acid and unaffected by nitrousacid a t room temperature. F. W. N.9. INDUCED GLYCOSIDE FORMATION IN PLANTS.The full significance of glycosides in higher plants is still far frombeing understood. It has long been considered that the glycosidesrepresent a detoxication mechanism in the plant analogous to theglucuronides excreted by animals when foreign substances areabsorbed. The natural aglycones in plants may be said, on thisview, to be toxic by-products of metabolism held in an immobilisedstate because the plant has no excretory system. It has now beenshown by L. P. Miller that a variety of plants will form glycosidesfrom certain foreign substances which had previously been shownto be physiologically active.This gives experimental support forthe view that in the plant glycoside formation is a detoxicationmechanism, as is the formation of glycuronides in animals. A tpresent the results cannot be considered directly in relation to thenatural glycosides, but the experimental methods used have over-70 J . Amer. Chem. SOC., 1940, 62, 1204.71 J . Gen. Physiol., 1939, 22, 637. 72 J . B w l . Chem., 1940, 135, 341.Contr. Boyce Thompson Inst., 1938, 9, 425; 1940, 11, 271; Science,F. E. Denny, Amer. J . Bot., 1926, 13, 118; Contr. Boyce Thompson Inst.,1940, 92, 42.1937, 8, 473STEPHENSON : GROWTH FACTORS FOR BACTERIA. 435come former difficulties and open a wide field for further investig-ation. Ethylene chlorohydrin, one of the substances found to breakthe dormancy of tubers, is converted by the living tissue intoP-2-chloroethyl-d-glucoside. The same plant was shown to converto-chlorophenol into p-(o-chloropheny1)gentiobioside. Chloral hydr-ate was converted into P-trichloroethylgentiobioside, reduction to thealcohol taking place as in animals.Thus it appears that the foreignsubstance influences the type of glycoside subsequently obtainedfrom the plant. That the foreign aglycones are literally immobilein the plant is shown by the fact that on the subsequent growth oftreated corms no traces of the induced glycoeide can be found eitherin the tops or in the new corm. R. H.10. GROWTH FACTORS FOR BACTERIA.Accessory food factors playing a part in the nutrition of theanimal1 and of yeasts2 have recently been dealt with in theseReports; the subject of bacterial vitamins will now be reviewed.Most of the vitamins known to function in the animal and allthose so far found necessary for any variety of yeast have beenfound necessary for one bacterium or another ; on the other hand,several substances have been shown to be necessary constituents inthe growth of bacteria but have not yet been shown to function inthe animal.It may be recalled, however, that pantothenic acid wasproved to be a necessary factor in the nutrition of certain yeastsbefore it was proved to function in the animal.It is well known that heterotrophic bacteria differ widely in theirfood requirements.Many organisms isolated from soil can becultivated in serial subculture on media consisting of inorganic salts,deriving their nitrogen requirements from ammonia and their carbonfrom carbohydrates or salts of simple organic acids. Others needcomplex media consisting of protein digests with or without carbo-hydrate ; for still more exacting organisms, blood, ascitic fluid, liverextract, yeast autolysate or egg yolk may be necessary ; it is mainlythe last type which provides the material for vitamin studies.According to the view put forward by P. F i l d e ~ , ~ organisms whichrequire a large selection of compounds from which to build up theircell material have lost a number of synthetic powers which moreprimitive organisms possess; this he believes is due t o prolongedcultivation in an environment where tlhey are surrounded by a richassortment of molecules which they are able to use " ready made,"as a result of which they lose synthetic powers.Thus highly para-Ann. Reports, 1939, 36, 340.Proc. Roy. SOC. Ned., 1934, 28, 79.Ibid., p. 369436 BIOCHEMISTRY.sitic organisms resident in the animal body have more exacting foodrequirements when isolated and grown in culture than have most oftheir near relatives living in soil. Even non-parasitic organismssuch as the lactic fermenters have in many cases lost the power tosynthesise essential molecules, such as riboflavin or pantothenicacid, which are present in milk.The inability to synthesise substances essential for the building ofcell material is carried to greater lengths among bacteria than in theanimal world.Not only is the list of accessory food factors alreadyknown for bacteria longer than that for the animal, but many strainsmust be supplied with a formidable selection of amino-acids, somein only very low concentration. For such organisms these amino-acids cannot logically be distinguished in function from vitamins,though for convenience i t is proposed to exclude them from thisReport.I n investigating the nutritional requirements of an exactingorganism, two methods of approach are possible. A medium maybe built up from simple known constituents until i t is adequate forthe growth of the organism. This method is possible only if therequirements of the organism are comparatively simple.The secondmethod, and the one more likely to attain success, is to start with amedium-however complicated-fully adequate for the growththrough serial subculture, and then, by fractionating the variousconstituents, eliminate unnecessary material until the simplestgroup adequate has been attained. If this still contains unknownconstituents, it may be possible to replace them by known vitamins ;otherwise the active substances must be isolated and their constitu-tion determined. When this has been achieved, two concentrationsof the new factor should be given; the minimum concentrationwhich supports visible growth and the minimum required formaximum growth through serial subculture.The first method is exemplified by work of Fildes on Proteus, thesecond by the work of Mueller on C.diphtheriaProteus.-This organism differs from those with the simplestrequirements in only one particular, that is, i t grows aerobically oninorganic salts and ammonium lactate if nicotinic acid only is added(0.1 pg./ml.).4 Nicotinic acid is among the most widespread require-ments of bacteria and i t is possible that the ability to synthesise itis readily lost. Within the Proteus group, however, two degrees ofsynthetic disability are found, all P . morganii strains requiring inaddition pantothenic acid, which produces detectable growth in0.001 pg./ml. and optimal growth in 0.005 pg./mL5* P. Fildes, Brit. J. Exp. Patlb., 1939, 19, 239.M.J. Pelczar and J. R. Porter, P,roc. Xoc. Exp. Biol. Med., 1940, 43, 151STEPHENSON : GROWTH FACTORS FOR BACTERIA. 437The second or analytical method for determining the growthrequirements of a microbe is exemplified in the classical studies ofJ. H. Mueller Thisorganism was cultivated on a medium consisting of (A) a saltmixture, (B) Liebig’s extract, 7.5 mg./ml. ; (C) tryptophan, 0.1mg./ml.; (D) an acid hydrolysate of caseinogen, 10 mg./ml. Thegrowth obtained in 60 hours was equivalent to 0.2 mg. of nitrogen/ml.and i t was the aim of the study to replace this medium by one ofknown composition giving an equal bacterial crop in the same time.(A), (B), and (C) being kept constant, (D) was adequately replacedby a mixture of amino-acids with ethyl alcohol as additional sourceof energy; the acid hydrolysate of casein being now replaced byknown compounds free from contaminating vitamins, (B) wasreplaced by liver extract, from which were isolated two constituentsboth necessary for the growth of the organism, vix., nicotinic acid,’active in 0.1 pg./ml., and p-alanine,S active in 0.1 pg./ml.; a thirdsubstance necessary was present in the liver extract, but wasactually isolated from cow’s urine; this was pimelic acid, active in0.005 ~ g ./ m l . ~ The effect of the last is duplicated by syntheticpimelic acid, but not by any other member of the series tried. Thesethree substances in the concentrations stated completely replaced theliver extract. These findings were confirmed by English workers lofor all strains of C.diphtherim mitis and most of the gravis strainstested. Some still more exacting strains of the latter type failed togrow on this medium, but grew when liver extract was added.From the latter a concentrate was prepared which could be replacedby pantothenic acid l1 (ay-dihydroxy-pp-dimethylbutyrylalanide) .We have then in this group two grades of synthetic disability; themajority of strains can use either p-alanine op pantothenic acid, theexacting gravis strains require pantothenic acid supplied as such.The Lactic Fermentem.-This group of organisms, isolated frommilk and cheese, forms another group of varying vitamin require-ments. Orla-Jensen et aZ.12 first showed that milk which had beenshaken with activated charcoal no longer supported the growth ofmany of these strains, but that when riboflavin together with theon the exacting, H.Y.strain of C. diphtherim.J . Bact., 1935, 29, 515.J. H. Mueller, J . Biol. Chem., 1937, 120, 219.J. H. Mueller and S. Cohen, J . Bact., 1937, 34, 381.J. H. Mueller, J . Biol. Chem., 1937, 119, 121.lo W. C. Evans, VV. R. C. Handley, and F. C. Happold, Brit. J . Exp. Path.,l1 E. T. Stiller, J. C. Keresztesy, and J. Finkelstein, J . Amer. Chem. SOC.,l2 S. Orla-Jenscn, N. C. Otte, and A. Snog-Kjaer, Mem. Acad. Roy. Sci.1930, 20, 41, 396.1940, 62, 1779.Lettres Dartemark, 1936, 6, no. 5 ; Zentr. Bakt. Par., 11, 94, 434, 452438 BIOCHEMISTRY.material eluted from the charcoal was added to the deficient milk,the growth rate attained on untreated milk was almost reached.Later workers have more precisely determined the requirements ofthis group ; l3 the medium consisted of bactopeptone, salts, cystine,and glucose.This was rendered free from riboflavin by exposingthe peptone to light in alkaline. solution ; four strains grew on theriboflavin-free medium (Xtr. luctis, L. arubinosus, L. pentosus, andLeuconostoc mesenteroides) ; L. delbruckii, L. casei, L. gayoni, andB. Zactis midi grew only when riboflavin (0.1 pg./ml.) was added.It is noteworthy that all lactic fermenters tested contain riboflavinwhether it is supplied in their growth niedium orThe activity of various synthetic flavins was tested and comparedwith that of the natural product obtained from milk; only thosecontaining ribose were active and none equalled rib0fla~in.l~I.6 : 7-Dimethyl-9-d-l‘-ribitylisoalloxazine (riboflavin) ...11. 6-Methyl-9-d-l’-ribitylisoalloxazine ...........................111. 7-Methyl-9-d-l‘-ribitylisoalloxazine ...........................IV. 7-Methyl-6-ethyl-9-d-l’-ribitylisoalloxazine ...............6 : 7 -Dime thy1 . 9 -d- 1 ’ -arabi ty lisoalloxazine ..................6 : 7-Dimethyl-9-2- 1’-arabitylisoalloxazine ..................6 : 7-Dimethyl-9- 1’-sorbitylisoalloxazine.. ...................VIII. 9-ZZ-Arabitylisoalloxazine .......................................IX. 9-(dZ-Ribityl)-5 : 6-benzoisoalloxazine ........................X. 6 : 7 : 9-Trimethylisoalloxazine (lumiflavin) ...............XI. 6 : 7-Dimethylalloxazine (lumichrome) .....................XII. Riboflavin tetra-acetate .......................................V.VI.VII.Activity.100507875Inactive9 ,,,It is noteworthy that substitution in the 6 : 7-positions leavessome activity and that the ribityl group is essential and cannot bereplaced by other sugars or by methyl.The effectiveness of the synthetic flavins for bacterial growth isclosely paralleled in animal studies, only I , 11, 111, and IV beingactive in rat growth tests.16The necessity for pantothenic and nicotinic acids for some mem-bers of the group was demonstrated as follows.A medium consistingof an acid hydrolysate of bactopeptone supplemented by salts,sodium acetate, cystine, and riboflavin failed to support growthunless liver extract was added. The latter was fractionated, and aconcentrate obtained active in 0.003 pg. /ml.This was subsequentlyfound to owe its activity to pantothenic acid. Some strains grewpoorly on this medium, and required in addition nicotinic acid(0-05 pg./ml.). Strains requiring pantothenic acid were B. Zcrctisl3 E. E. Snell, F. M. Strong, and W. H. Peterson, Biochem. J., 1937, 31,1 4 F. Schutz and H. Thoorell, Biochem. Z., 1939, 295, 246.l5 E. E. Snell and F. M. Strong, Enqmologia, 1939, 6, 186.1789; J . Bact., 1939, 38, 293; J. Amer. Chem. Soc., 1938, 60, 2825.Kuhn et al., Ber., 1937, 70, 2560STEPHENSON GROWTH FACTORS FOR BACTERIA. 439acidi, L. arabinosus, L. pentosus, L. delbriickii, B. brassictz andXtr.Zactis. Those requiring nicotinic acid in addition were L. cuseiand L. arabinosus.17The strain of Str. lactis isolated from silage and originally knownas Bact. acetyll choline (Keil), in addition to factors already mentioned,requires adermin (vitamin B6), 3-hydroxy-4 : 5-bis(hydroxymethyl)-2-methy1~yridine.l~ This compound is also required by Xtr.hcemolyticus (see p. 440).The Dysentery Group.-& A. Koser et aZ.19 have shown thatnicotinic acid is essential for a number of dysentery bacilli (FEesner,Hissy, Strong, and other unspecified strains). These were sown in abasal medium of fifteen amino-acids, glucose, and salts ; growth wascompletely negative on the basal medium, but full rapid growthoccurred on the addition of nicotinic acid, 0.1 pg./ml.; 0.04 pg./ml.gave slower growth and a visible effect was obtained with 0.01The Pasteurella Group.-A number of organisms of the Pasteurellagroup have been found to require nicotinamide and pantothenic acid(0.1 pg./ml. was used in both cases) ; the former was replaceable bycoenzyme I, but p-alanine did not replace pantothenic acid; thiscase resembles that of the exacting strains of C. diphtheritz gravis.20The Staphylococci.-The growth requirements of this group oforganisms are high. Aerobically they grow on peptone water, butit is clear that this does not function only as a source of amino-acids,for when it is replaced by (say) an acid digest of caseinogen plustryptophan, tyrosin, and cystine, no growth occurs, unless extractof meat or yeast is added.From the latter (in the form of marmite)two active fractions, both necessary, were obtained,21 the onereplaceable by nicotinic acid or amide (0.2 pg./ml.) or by diphospho-pyridinedinucleotide, the other by aneurin (0.02 pg./ml.).The organism can use the two basic components of aneurin ifthese are provided separately, vix., 4-amino-5- aminomethyl-2 -methyl-pyrimidine (0-002 pg./ml.) and 4-methyl-5-~-hydroxyethylthiazole(about 0-01 pg./ml.).The specificity for the thiazole base appears to be complete,closely related compounds tried being quite inactive even at concen-trations 100 to 1000 times that used for the acceptable compound.Thiochrome, for example, will not replace aneurin and 4-methyl-17 E. E. Snell, F. M. Strong, and W.H. Peterson, J . Amer. Chem. Xoc., 1938,60, 2825.1* E. F. Moller, 2. physiol. Chem., 1938, 254, 285.19 S. A. Koser, A. Dorfman, and F. Saunders, Proc. SOC. Exp. Biol. Med.,2o S. Berkman, I?. Saunders, and F. A. Koser, ibid., 1940, 44, 68.21 B. C. J. G. Knight, Biochem. J . , 1937, 31, 731.pg. /ml.1938, 38, 311440 BIOCHEMISTRY.thiazole does not replace 4-methyl-5-~-hydroxyethylthiazole. Theorganism is, however, less specific towards the pyrimidine group, thefollowing substitutes for the aneurin base being active in approxi-mately the same concentration, 'uuz'x., 4-amino-5-thioformamido-methyl-2 -methylpyrimidine and 4 -amino-5-aminomethyl-2 -methyl-pyrimidine ; inactive, however, were 4-hydroxy-5-hydroxymethyl-2 -methylpyrimidine, 4-hydroxy-5-aminomethyl-2-methylpyrimidine,and 4-amino-2-hydroxypyrimidine (cytosine).22The anaerobic metabolism of Staph.aureus is different from itsaerobic metabolism; in the former case i t derives its energy mainlyfrom amino-acids ; anaerobically from glucose and pyruvic acid.I n the latter case an additional factor is required, 'uix., uracil (2 : 6-dihydroxypyrimidine), which is active in 1-5 pg. 1.11. Relatedbases were inactive, 'uiz., 5-methyluracil (thymine), 4-methyluracil,1 : 3-dimethyluracil, 1 : 3 : 4-trimethyluracil, 2-thio-5-methyluraci1,barbituric acid, cytosine, and i~ocytosine.~~Str. haemo1yticus.-Growth of this organism failed on a mediumconsisting of bactopeptone to which were added cystine, glucose, anda formidable list of bacterial vitamins.The addition of meatextract supplied the material lacking, which was subsequentlyidentified as glutamine, which supported full growth a t 30 pg. lml.24A number of related compounds, including glutamic acid, asparticacid, and asparagine, are inactive.25Two other factors for this organism were discovered by treatingthe bactopeptone with alkali, which resulted in an inactive medium.Growth was then obtained on the addition of riboflavin (0.1 pg./ml.)and pantothenic acid (1.0 pg./m1.).26* 27 Finally, vitamin B, wasshown to be necessary by growing the organism on a selection ofamino-acids (replacing protein hydrolysate), riboflavin, and panto-thenic acid plus an aqueous extract of liver. The active part of thelast was adsorbed on lead sulphide and eluted, and the eluateadsorbed on and eluted from fuller's earth. The concentrate thusobtained was active in 1 pg./ml.and could be replaced by vitaminB6.28The C1ostridia.-The requirements of the Clostridia (spore- bearingstrict anmobes) are notoriously high as to both vitamins and amino-This was confirmed by an alternative procedure.2922 B. C. J. G. Knight, Biochern. J., 1937, 37, 966.23 G. M. Richardson, ibid., 1936, 30, 2184.24 H. McIlwain, P. Fildes, G. P. Gladstone, and B. C. J. G. Knight, ibid.,25 H. McIlwain, ibid., p. 1942.26 B. L. Hutchings and D. W. Woolley, J. Bact., 1938, 38, 285.2 7 H. McIlwain, Brit. J. Exp. Path., 1939, 20, 330.B. L. Hutchings and D. W. Woolley, Science, 1939, 90,42.20 H.McIlwain, Brit. J. Exp. Path., 1940, 21, 25.1939, 33, 223STEPHENSON : GROWTH FACTORS FOR BACTERIA. 441acids and few cases have been worked out in detail. CZ. sporogenesfurnished an early example of a bacterial vitamin.30 The activesubstance was found in yeast, moulds, bacteria, and urine ; a highlyconcentrated preparation from yeast was active in 0.02 pg./ml. Itis an acid substance forming a soluble barium salt and a methyl esterdistilling at 80-100"/0~001 mm.31 It awaits further identificationand is probably necessary for other members of the CZostridia.The butyl fermenters display definite vitamin requirements andwork on this group is in a state of active progress. CZ. butylicumgrows on a synthetic medium containing asparagine and glucose asthe only organic compounds and needs in addition only 33* 34It can be used, therefore, as the test organism for the presence ofthis substance, 1.3 x pg./ml.being detectable. The followingtable shows growth as measured by the turbidimeter ; 0 representsthe uninoculated culture, 100 complete opacity.Biotin, pg. /ml.0~0000000.00001 330*00002660.0000530~000100*000200.000660.01330.03330.06660.13320.26640.6660Liv& conc., pg./ml.Growth measured by turbidimeter.2.210.229.054.075.088.094.039.060.07 9.094.094.096.0The closely related CZ. acetobutyzicum requires in addition somefactor or factors obtained from yeast.35, 36 Moreover, the study ofthese fermenters is complicated by the fact that some still unidenti-fied factor (or factors) modifies the course of the fermentation aswell as the 37The Ha3mophilus Group-The vitamin requirements of this grouphave not been determined recently enough to be considered hereand will therefore be only briefly stated. Two factors were foundnecessary in cultivating these organisms, the X factor obtained from30 B.C. J. G. Knight and P. Fildes, Brit. J . Exp. Path, 1933, 14, 112.31 A. M. Pappenheimer, Biochem. J., 1935, 29, 2055.32 F. Kogl and B. Tonnis, 8. physiol. Chem., 1936, 242, 43.33 E. E. Snell and R. J. Williams, J . Amer. Chem. SOC., 1939, 81, 3594.34 W. H. Peterson, L. E. McDaniel, and E. McCoy, J . Biol. Chem., 1940,36 C. Weizmann and B. Rosenfeld, Biochem. J., 1939, 33, 1376.36 A.E. Oxford, J. 0. Lampen, and W. H. Peterson, ibid., 1940, 34, 1588.133, LXXV.C. Weizmann and B. Rosenfeld, Zoc. cit442 BIOCHEMISTRY -blood and the V factor from animal or plant tissues or other bacteria.The former was identified as hzematin, and the latter as diphospho-pyridinenucleotide (coenzyme I) , replaceable by the triphospho-derivative (coenzyme 11). When grown anzrobically, the formercan be dispensed with and it has been suggested that i t is required,in part a t any rate, for the synthesis of catalase, which is unnecessaryin anzerobic life where hydrogen peroxide is not formed. It hasrecently been claimed that cysteine can replace hsmatin in aerobicgrowth; this is regarded as evidence that the hzmatin is requiredfor the synthesis of catalase; in the presence of cysteine hydrogenperoxide would be reduced and catalase rendered unnecessary.Further details of this work are promised.3*Nicotinic acid or amide cannot replace the diphosphopyridine-nucleotide in this group, whereas organisms requiring the former canreplace it by the latter (cf.the case of p-alanine and pantothenicacid in the diphthericc? group).Functions of Bacterial Vitamins.-Organisms which require a givenvitamin can be used as delicate reagents for the detection of thatsubstance in naturally occurring materials and for a rough quantit-ative assay during its isolation. Str. hamolyticus or one of theexacting strains of C. diphtheria gravis can be used to determine thepresence and approximate amount of pantothenic acid in a liver oryeast extract, the only alternative method involving prolongedanimal feeding experiments; six days’ work may thus replace sixweeks’.It is fairly apparent that the vitamins shown to be neces-sary in the special strains which require them supplied in the mediumare of wide-spread importance and exist also in other species whichare able to make them for themselves. This can be shown by usingextracts of non-exacting strains to supply the known requirementsof exacting strains.Some success has been achieved in determining the function ofcertain vitamins by the method first used by Lwoff and Lwoff.This consists in growing an organism requiring the vitamin in amedium containing it in a sub-optimal concentration. The organismso obtained is comparable with a vitamin-deficient animal and by itcomparison of its enzyme systems with those of the normally grownorganism it may be possible to show what chemical mechanism isdeficient.Thus H . parainfluenxm grown in sub-optimal amounts ofcoenzyme I was shown to have decreased powers of oxidisingglucose, etc. ; this power could be augmented by the addition of co-enzyme I to the reaction vessel. With the organism grown normally,the addition of coenzyme I does not affect the oxidation rate.39T. L. Snyder and R. H. Broh-Kahn, Nat~tre, 1938, 143, 133.39 A. Lwoff and M. Lwoff, Proc. Roy. SOC., 1936, B, 122, 360GALE : NITROGEN METABOLISM OF BACTERIA. 443Washed suspensions of Xtaph. aureus grown in sub-optimalamounts of aneurin oxidise and dismute pyruvate a t a lower ratethan suspensions of organisms grown in optimal amounts of aneurin ;the rates in the former case are increased by the addition of aneurinto the reacting vessel, but the rates in the latter case are notaffected.40 The organisms grown in the aneurin-deficient mediumare seen, therefore, to suffer from lack of cocarboxylase (aneurindiphosphate) and to display the same decreased ability to metabolisepyruvate as the tissues of the pigeon sufTering from the same vitamindeficiency.The following table indicates the vitamin requirements amongbacteria, yeasts, and animal tissues :Bacteria.Yeasts................................... Nicotinic acid.. $- +/3- Alanine + +Pantothenic acid + +Riboflavin + +Adermin (vitamin B,) + + Aneurin .......................................... +Uracil .............................................+Asparagine .................................... +Biotin + + Sporogenes vitamin ........................... +Coenzyme I .................................... +Haematin ....................................... +Pimelic acid .................................... +....................................... ..........................................................................................................................................Animal. ++ + + ++M. S.11. NITROGEN METABOLISM OF BACTERIA.Nitrogen Pixation.-The publication by A. I. Virtanen and T.Laine 1 of the experimental details of their work on the symbioticnitrogen-fixing Rhixobium confirms the general scheme set out inthe former’s book and collects the information appearing in variousnotes and abstracts.3 It is claimed that nitrogen is fixed by theorganism with the production of hydroxylamine, which then reactswith oxaloacetic acid produced by the host plant from carbohydrate,as follows :H02CC( :NOH j *CH2*C02H + HO,C*CH (NH,) *CH,*CO,HC6H,,06 --+ HO2dCO*CH,*CO2H40 G.M. Hills, Biochem. J., 1938, 32, 383. Biochem. J., 1938, 32, 412.A. I. Virtanen (1938), “ Cattle Fodder and Human Nutrition,” CambridgeA. I. Virtanen, J. SOC. Chem. Ind., 1935,54,1015; J . Agric. Sci., 1937,27,University Press.332 ; Agric. Col. Sweden Ann., 1938, 5, 429444 BIOCHEMISTRY.The evidence supporting this scheme, which is put forward by(1) The infected plant excretes Z-aspartic acid into the medium(2) The aspartic acid is excreted only from roots infected with(3) Oxaloacetic acid can be detected in the host plant.'(4) The oximinosuccinic acid has been isolated and identified,l* 8but hydroxylainine itself has not been identified.( 5 ) Nitrogen fixation by free-living Rhixobium cultures has beenobserved in the presence of oxaloacetic acid.g(6) The greater part of the nitrogen excretion which is notaccounted for by the aspartic acid consists of p-alanine,1° andsuspensions of Rhixobium will decarboxylate aspartic acid to formp -alanine.l1(7) I n the presence of crushed pea plants, the amino-group ofaspartic acid may be transferred to keto-acids such as pyruvic acidwith the formation of a-alanine, etc.12The conclusions of Virtanen have been criticised by P.W. Wilson,13who has reviewed the work of the Wisconsin school in this field. Hiscriticisms, published before the detailed description of Virtanen'swork, are not completely justified now and are largely disposed ofby the isolation of the 0xime.l Wilson has been unable to obtainsignificant nitrogen-fixation by the free-living Rhixobium culturesin the presence of oxaloacetic acid. The Wisconsin workers haveshown that, the plant and bacteria being treated as one system, therate of nitrogen fixation is dependent upoii the nitrogen pressure,l*independent of the oxygen pressure over a wide range,15 andspecifically inhibited by hydrogen.13Little advance has been made in our knowledge of nitrogen-fixation by the free-living Axotobacter.A. I. Virtanen and T. Laine l6have found aspartic acid as the chief product of excretion, but asthis is much less marked than with Rhizobium, the study has not asyet been elaborated. G. Endres l7 has found oximes formed bythese workers, rests on the following points :around the roots.*.Rhixobium.6A. I. Virtanen and T. Laine, Nature, 1935, 136, 756.5 Idem, Suomen Kem., B, 1937, 10, 32.6 A. I. Virtanen, von Hausen, Synnove, and T. Laine, J . Agric. Sci., 1937,7 A. I. Virtanen and T. Laine, Suomen Kern., B, 1938, 11, 25.8 Idem, Nature, 1938, 142, 165.lo Idem, ibid., p. 2.12 Idem, Nature, 1938, 141, 748.l4 P. W.Wilson, J . Amer. Chem. SOC., 1937, 58, 1256.15 P. W. Wilson and E. B. Fred, Proc. Nut. Acad. Sci., 1937, 23, 503.l6 Suomen Kem., B, 1937,10,2.27, 332.Idem, Suomen Kem., B, 1937, 10, 24.l1 Idem, Enzymologia, 1937, 3, 266.l3 Ergebn. Enzymforsch., 1938, 8, 13.l 7 Annalen, 1935,518,109; 1938,535,lGALE NITROGEN METABOLISM OF BACTERIA. 445Axotoblacter growing in the presence of nitrate. D. D. Woods l8has shown that washed suspensions of CZ. welchii and Bact. co2i(probably many other types) will reduce nitrate first to nitrite andthen to ammonia in the presence of hydrogen and has broughtforward good evidence that the reduction of nitrite proceeds throughhydroxylamine.A. S. Corbet and W. R. Wooldridge l9 have investigated thedistribution of nitrogenous compounds in sewage and the changesthat occur during the activated sludge process.Confirming muchearlier work on soil, their experiments indicate that the followinginteractions and changes can occur in the nitrogen cycle :af'- Organic NThe particular course of the reactions followed varies accordingto the conditions which obtain, especially as regards the proportionsof available carbonaceous matter and nitrogen present and thenature of the compound in which the nitrogen occurs. The reactionsindicated in the diagram are encouraged by the following chemicalconditions :(a) The presence of ammonia and assimilable carbon compounds.( b ) Ammonia present, but little or no assimilable carbon compounds.( c ) Assimilable carbon compounds present together with nitriteand/or nitrate as the most readily available source of nitrogen.( d ) Autolytic changes arising from the absence of nutrientmaterials.( e ) Fixation of nitrogen in the presence of assimilable nitrogen-freeorganic compounds and no readily available source of nitrogen otherthan free gas.Activated sludges appear to contain the enzymes necessary toeffect any of the above changes.Amino-acid Metcsbo2ism.-G.M. Hills 2o has studied the amino-acidmetabolism of certain pathogenic bacteria. Gram-positive coccicontain an enzyme which attacks arginine to form ornithine andammonium carbonate; as neither citrulline nor urea appears to bean intermediate in the reaction, the enzyme has been named arginineBiochem.J., 1938,32,2000.Ibid., 1940, 34, 1015, 1026, 1036. 2o Ibid., p. 1057446 BIOCHEMISTRY.dihydrolase. Gram-positive bacteria produce no ammonia aerobic-ally from other amino-acids with the exception of Shphylococcus,which attacks serine and threonine. Bact. typhosum deaminatesserine, aspartate, threonine, and arginine aerobically, and C .diphtherim appears to attack aspartate only. A. Janke and W.Tayenthal 21 showed that glycine is oxidatively deaminated bywashed suspensions of Bact. coli, Bcsct. uulgare, Ps. Jluorescens, andBcsc. mycoides : glyoxylic acid was isolated as the 2 : 4-dinitrophenyl-hydrazone. has dealt withthe anaerobic breakdown of cysteine and cystine by washed suspen-sions of Bact. coli. Cysteine is broken down by an adaptive cystein-ase, liberating ammonia and hydrogen sulphide in equimolecularquantities. The reaction is partially inhibited by glucose and isspecific for the natural isomer.Cystine is reduced to cysteine beforefurther attack. The enzyme responsible for the breakdown ofcysteine by Propionibact. pentosaceum 23 differs from the cysteinaseof Bact. coli in being accelerated by the presence of glucose, showingno optical specificity, and not requiring the presence of the substrateduring growth for its formation. The products of the breakdownother than ammonia and hydrogen sulphide have not been reported.C. E. Clifton 24 has shown that serine is disrupted anaerobically bywashed suspensions of Cl. botulinum with the formation of ammonia,carbon dioxide, acetic acid, and ethyl alcohol.He suggests thatpyruvic acid is formed as an intermediate.The breakdown of E( +)-glutamic acid and I( -)-aspartic acid bybacteria has been the subject of several investigations. E. Adlerand co-workers 25 have been able to extract the glutamic aciddehydrogeiiase from suspensions of Bact. coli by a modification ofthe method used previously by M. Stephenson 26 for the extractionof lactic dehydrogenase. The extracted enzyme reduces methylene-blue in the presence of glutamic acid and coenzyme I1 and evidenceis put forward to show that the dehydrogenation to iminoglutaric acidis reversible. The reactions involved in the deamination of glutamicacid are :( a ) Glutamic acid + coenzyme =+ Iminoglutaric acid + dihydro-( b ) Iminoglutaric acid + H,O + Ketoglutaric acid + NH,( c ) Dihydrocoenzyme + 402 + Coenzyme + H,O21 Biochem.Z., 1937, 289, 76.22 P. Desnuelle and C. Fromageot, Enzymologia, 1939, 6, 80, 242, 387.23 P. Desnuelle, E. Wookey, and C. Fromageot, ibid., 1940, 8,225.24 PTOC. Soc. Exp. Bwl. Med., 1940, 43, 588.25 E. Adler, V. Hellstrom, G. Gunther, and H. Euler, 2. physiol. Ckem.,26 Bwchem. J., 1938, 22, 605.A series of papers by P. Desnuelle etcoenzyme1938, 255, 14GALE : NITROGEN METABOLISM O F BACTERIA. 447J. R. Klein2' has shown that washed suspensions of HemophiEuspcsrain$uenzce oxidise aspartic and glutamic acids with the liber-ation of ammonia and the formation of acetic acid according to theequations :Aspartic acid + 0, --+ CH3-C02H + NH, + 2C02Glutamic acid + 2+02 + CH,-CO,H + NH, + 3c0,By a study of the metabolism of possible intermediate com-pounds, Klein has established the probable course of the oxidation.For aspartic acid the reaction proceeds : aspartic acid --+ oxalo-acetic acid L_, pyruvic acid acetaldehyde --+ acetic acid.The first and the last step require the presence of a coenzyme.E.P. Gale 28 has shown that there are two mechanisms in Bact.coZi which deaminate aspartic acid anaxobically. One of theenzymes concerned is stable to toluene treatment and is the aspartaseof J. H. Quastel and B. W00lf,~~ which deaminates aspartic acid tofumaric acid. The other enzyme is inhibited by toluene and requiresthe presence of a coenzyme, which can be replaced in vitro by adenos-ine. The two enzymes have been fractionated in a cell-free juiceobtained from Bact.coEi crushed in a bacteria-crushing mill designedby V. H. Booth and D. E. Green.3O Since the fraction containingaspartase I1 also contains fumarase, the immediate product of thedeamination process is not known. A. I. Virtanen and J. Erkama 31claim to have shown the presence of two enzymes attacking asparticacid in B. jluorescens Ziquefaciens, one producing fumaric acid and theother carrying out a hydrolytic deamination to form inalic acid. Ifthe latter statement is proved, it will be the first known case ofbiological hydrolytic deamination.E. F. Gale and M. Stephenson 32 have continued their studies onfactors affecting bacterial doamination.Washed suspensions ofBact. coli deaminate glycine, dl-alanine, and Z( + )-glutamic acidaerobically and dl-serine and I( - )-aspartic acid anmobically.Anzrobic growth conditions increase the formation of the anaerobicdeaminases and decrease that of the aerobic deaminases, whereas thepresence of glucose during growth inhibits the formation of alldeaminases studied to the extent of 85-95%. The activity of theserine deaminase varies greatly with the age of the culture andupon the condition of the organism, the presence of phosphate anda reducing agent being necessary to prevent this activity being loston standing. J. W. Baker and F. C. Happold33 have studied thegroups essential to the tryptophan molecule for the production of27 J . Biol.Chem., 1940, 134, 43.29 Ibid., 1926, 20, 545.31 Nature, 1938, 142, 954.a3 Ibid., 1940, 34, 657.a * Biochem. J., 1938, 32, 1583.32 Biochem. J., 1938, 32, 392, 1583.Ibid., 1938, 32, 855448 BIOCHEMISTRY.indole from that molecule by Bact. coli. Up to the present noattempt to isolate or identify an intermediate substance in thisbreakdown has met with success. By studying the action of the cell-free tryptophanase preparation on a large number of tryptophanderivatives, they have obtained results which suggest that the break-down to indole requires the following structural features : ( a ) a freecarboxyl group, ( b ) an unsubstituted a-amino-group, ( c ) a p-carbonatom capable of oxidative attack. They tentatively suggest a typeof breakdown involving " reductive fission " of the tryptophanmolecule, but there is as yet no experimental evidence for this typeof reaction.The study of the amino-acid metabolism of the Clostridia (strictanzrobes) has been continued by several workers.D. D. Woodsand C. E. Clifton 349 35 have found that CZ. tetanomorphum utilisesmany amino-acids, causing the liberation of ammonia and hydrogen.The course of the breakdown of glutamic acid has been worked outin detail, the products being acetic acid, butyric acid, carbon dioxide,hydrogen, and ammonia. The same products in the same propor-tions are produced by the growth of an unidentified Clostridiumgrowing on a glutamic acid medium and studied by H. A. Barker.36W. Kocholaty and J. C. Hoogerheide 37 have studied the dehydro-genation reactions carried out by CZ. sporogenes and have found thatin certain cases amino-acid molecules can act as both hydrogendonators and acceptors, so that intramolecular reactions occur inwhich one molecule is oxidised and another reduced, thus forminga special case of the intermolecular oxidation-reduction reactionsdiscovered by L. H. Sti~kland.~* C. E. Clifton39 has studied theamino-acid metabolism of CZ. botulinum and finds that i t obtains itsenergy through " Stickland " reactions in a manner similar to thatof CZ. sporogenes.Amine Formtion.-A. I. Virtanen and T. Laine 40 followed uptheir observation that Rhixobium decarboxylates aspartic acid withthe formation of p-alanine by showing that Bact. coZi producescadaverine from lysine. Later 41 they found that Rhixobium willalso decarboxylate glutamic acid to y-aminobutyric acid. A. H.Eggerth et ~ 4 1 . ~ ~ have worked out an improved method for theestimation of histamine in bacterial cultures and A. H. Eggerth 43has investigated the production of histamine by many species of34 Biochem. J., 1937, 31, 1774.36 Enzymologia, 1937, 2, 175.4 0 Enzymologia, 1937, 3, 266.4 1 A. I. Virtanen, P. Rintala, and T. Laine, Nature, 1938, 142, 674.4 4 A. H. Eggerth, R. S. Littwin, and J. V. Deutsch, J . Bact., 1939, 37, 187.43 Ibid., p. 205.35 Ibid., 1938, 32, 345.37 Biochem. J., 1938, 32, 437, 949.39 J . Bact., 1940, 39, 485. Ibid., 1934, 28, 1746GALE NITROGEN METABOLISM OF BACTERIA. 449organisms growing in culture. He has shown that many of thecommon inhabitants of the intestine will produce histamine, especi-ally if glucose is present in the growth medium. By adjusting themedium pH during growth, he showed that histamine is best pro-duced if the pH is low. Growth temperature also plays an importantpart, as some organisms produce more histamine if grown at a lowtemperature than at the normal 37". All these results were obtainedby bacteria growing in various media, so the effects studied may beproduced by action on the growth of the organism rather than onthe histamine-producing mechanism. The conditions under whichbacteria produce certain amines by the simple decarboxylation of thecorresponding amino-acids have been cleared up in a series ofpapers by E. F. Gale.44 In the first case the conditions under whichthe organisms are grown are important ; to obtain organisms possess-ing strongly active amino-acid decarboxylating enzymes, they mustbe grown under acid conditions in the presence of the free amino-acids. The most effective way of doing this is to grow the organismsin a tryptic digest of casein with 2% of glucose, the fermentation ofthe glucose producing the necessary low pH. Organisms grown inthis manner will decarboxylate certain amino-acids quantitativelyto the corresponding amines a t low pH values, the optimum value ineach case depending upon the amino-acid concerned. Thus washedsuspensions of Bact. wZi grown in glucose broth or in broth at p , 5,the physiological limit of growth, will decarboxylate E( + )-arginineto agmatine optimally at pH 4.0 ; E( +)-lysine to cadaverine at pH 4.5 ;I( +)-ornithine to putrescine a t p , 5.0 ; I( -)-histidine to histaminea t pH 4.0 ; and I( +)-glutamic acid to y-aminobutyric acid at pH 4.0.Similarly, Streptococcus fcecalis quantitatively decarboxylates I( - ) -tyrosine to tyramine at pH 5.0. I n all cases the product has beenisolated from a simple mixture of washed suspension of organism,appropriate buffer, and amino-acid, the quantities so arranged thatthe decarboxylation has proceeded to completion. Good yields ofpure product are obtained and i t would seem that this biologicalmethod may be the best method for the large-scale production ofsome of these compounds. Several groups of organisms have beenstudied from this point of view and again the strictly anmobic groupof Clostridia proves interesting : CZ. welchii decarboxylates histidineto histamine a t the exceptionally low optimum pH of 26-3.0, soin this case appreciable amounts of histamine are only producedin uiuo when the organism grows in the presence of fermentablecarbohydrate. Amongst other members of the group, CZ. septiquedecarboxylates ornithine to putrescine at pH 5.5 and CZ. cerofetidumforms tyramine from tyrosine a t pH 5.0. Many organisms, e.g., most4 4 Biochem. J . , 1940, 34, 392, 846, 853, and in the press.REP.-VOL. XXXVII. 450 BIOCHEMISTRY.strains of Bact. coli, Bact. proteus, Cl. welchii, Cl. arofcetidum,Ci'. bz&rmentcans, decarboxylate glutamic acid to y-aminobutyricacid. The distribution of the various decarboxylases shows thateach enzyme is specific for one amino-acid.G. M. Hills 45 has shown that certain Xtreptococci and S~phylococciwill attack I(+)-arginine to produce ornithine by splitting offammonium carbonate. This reaction is carried out by strains ofXtrep. fcecalis and consequently a symbiotic mixture of Xtrep.fcemZis and Bact. coli attacks arginine in an interesting fashion, theproduct of the attack at pH 4-0 being agmatine and at p~ 5-5,putrescine, produced with ornithine as an intermediate s~bstance.~6H. L. A. Tarr 47 has shown that the trimethylaminc in putrid fishis produced by reduction of trimethylamine oxide by bacteria whichpossess an enzyme activating the trimethylamine oxide so that it canbe reduced by any one of a number of dehydrogenase systems.Since only a few of the bacteria infecting putrid fish possess thisenzyme, the trimethylamine production cannot be regarded asa measure of putrefaction.Purine Metabolism.-M. Stephenson and A. R. Trim4* havecontinued the investigations started by C. Lutwak-Mann 49 on thebreakdown of adenylic acid and other adenine compounds by Bact.coli. Muscle adenylic acid is deaminated and dephosphorylated byBact. coli, the dephosphorylation appearing to precede deamination.Adenosine is deaminated to inosine and the ribose is split off andfermented ; the fermentation of ribose in adenosine is about 10 timesas fast as that of free ribose. Adenine is slowly deaminated tohypoxanthine, the rate of deamination being increased some 6-7times by the presence of adenosine-an effect similar to that foundin the deamination of aspartic acid by aspartase II.26 E. F. G.D. J. BELL.J. F. DANIELLI.E. F. GALE.L. J. HARRIS.R. HILL.E. KODICEK.J. R. MARRACK.A. NEUBERGER.F. W. NORRIS.M. STEPHENSON.4 5 Biochem. J., 1940, 34, 1057.4 7 J. Fish. Res. Bd. Can., 1939, 4, 367.48 Biochem. J., 1938, 32, 1740.4 6 E. F. Gale, ibid., p. 853.49 Ibid., 1936, 30, 1405

ISSN:0365-6217    

DOI:10.1039/AR9403700382

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

ANALYTICAL CHEMISTRY.INTRODUCTION.SINCE only a limited number of topics can be dealt with in detaileach year, it is inevitable that much of the steady and solid growthof analytical chemistry, particularly along its well-developedbranches, escapes attention. In this Report the analytical usesof distillation, colorimetry, fluorescence, and electrolytic methodshave been chosen for detailed description. Though all the methodsare basically physical, their analytical applications cover the wholefield of chemistry.In the analysis of organic substances, distillation, includingazeotropic distillation, is frequently employed, possessing as itdoes the evident advantage that separation is effected withoutchemical change.The technique of colorimetric analysis was summarised in theReport for 1936 and apphations in Inorganic Chemistry receiveddetailed treatment.The importance of traces of organic substancesin biological fields and the recognition that colour reactions areamong the most sensitive tests continue to stimulate the applicationof colorimetric methods. Examples of the utility of fluorescencemethods in analysis have been given in the Annual Reports of thepast decade, and it is clear that they must now be accepted asestablished analytical techniques.The analytical uses of electrochemical phenomena are manifold,and the Annual Reports for 1933, 1934, and 1938 contained surveysof three branches, vix., electrometric, oxidation-reduction potential,and polarographic methods, respectively. Electrodeposition, whichis wholly inorganic in its scope, is now added to this group.FRACTIONAL AND MOLECULAR DISTILLATION.1.General.-Distillation has been defined as “ the art of separatingsubstances by the condensation of their vapours into liquid fractionsof constant boiling point ” (Thorpe’s Dictionary, 4th Edn., Vol. 4,p. 34). Where only one component of a mixture is volatile, itcan be obtained in a pure state by distillation if not decomposedin the process. Where all constituents of a mixture are volatile,generally, the more volatile constituents of the mixture increasein concentration in the vapour phase when the mixture is partiallyvapourised and equilibrium is established between the liquid an452 ANALYTICAL CHEMISTRY.the vapour phase; in such cases, by repeated distillations, or byfractionation, the mixture may be separated into its constituents.In the separation of pure substances from mixtures, the difficultiesmost likely to occur are due either to the presence of several com-ponents with nearly equal boiling points, or to the presence of oneor more components in relatively small quantity.Improvedfractionating columns have overcome these difficulties in somemeasure. A third difficulty, due to the formation of constant-boiling mixtures or azeotropes, cannot, however, be overcome inthis manner.To the chemist, distillation methods of analysis, in which thecomponents are recovered unchanged, have advantages overchemical methods, in which one or more of the components arechemically altered. In many cases, too, a mixture which cannotbe analysed readily by chemical methods may be separated intoits components by fractionation through an efficient column, Theideal fractionation yields a series of sharply defined fractions,each distilling at a definite temperature; after each fraction hasdistilled, the temperature rises rapidly, no liquid being distilled asan intermediate fraction.I n such an ideal fractionation, if tem-perature is plotted against volume of distillate, the graph obtainedis a series of alternate horizontal and vertical lines resembling astaircase. A more or less sloping break indicates the presence ofan intermediate fraction, and the amount of such fraction can beused as a criterion of the performance of different columns.Asan example of the order of separation which can be obtained in thelaboratory, a mixture of benzene, toluene, and xylene has beenseparated, practically quantitatively, into the three individualliquids.It is desirable to define some of the terms which are used in discuss-ing fractionation columns. The capacity of a column is the measureof the amount of vapour and liquid which can be passed counter-current to each other in a column without causing it to choke orprime. The eflciency of a column is defined as the separating powerof a definite length of column, and is measured by comparing theperformance of the column with that calculated for a theoreticallyperfect plate column under similar conditions. For this com-parison, i t is necessary to distil a mixture in the column, and recordcertain liquid-vapour equilibrium data.The theoretical columnwhich will give the same values is then calculated, and the numberof theoretically perfect plates so determined is divided by thelength of column employed to make the separation. The reciprocalof this efficiency is called the height of equivalent theoretical pkzte1 J . SOC. Chem. Ind., 1935, 54, 2 9 7 ~ GRIFFITHS, M.ACLENNAX , AND WHALLEY. 453(H.E .T.P.) .2 I n comparing relative efficiencies of fractionatingcolumns, the operating procedure requires to be standardised. Abinary mixture whose liquid-vapour equilibrium data are knownshould be used, and the mixture should be one which is easilyanalysed by a simple determination.For instance, refractiveindex serves to evaluate mixtures of n-heptane and cyclohexane,benzene and carbon tetrachloride, or benzene and dichloroethane.2. Types of Column.-Very long columns have been used inattempts to increase the separating power of a rectifying still. A36.1 ft. packed laboratory glass column was used by J. H. Bruunand S. T. Schicktanz; and subsequently M. R. Fenske, C. 0.Tongberg, D. Quiggle, and D. S. Cryder * extended the length of alaboratory column to 52 ft. A very high degree of separation isobtainable in such columns, but their erection is rarely practicable.The purpose of packing in a fractionating column is to effect asintimate a contact as possible between the ascending vapour anddescending liquid without too great a reduction in capacity. Apacking will not give a good separation unless it has a low H.E.T.P.,nor will i t prove satisfactory unless the capacity is adequate.Afurther requirement for packings which will give sharp separationsis a low hold-up, i.e., the packing must not retain or hold up a largequantity of the liquid being distilled.A complete and extensive study of the relative fractionatingefficiencies of analytical distillation columns has been made byW. J. P~dbielniak,~ using different types of columns and differentcolumn packings. Various factors influencing the separation ofmixtures into sharp fractions have been considered, amongst thembeing the following :-(i) Time of distilktion. There is always an optimum time ofdistillation-in the case of the Podbielniak type of column about3 hours-below which accuracy is sacrificed, and above which theslightly improved separation does not justify the extra time taken.This is defined as the ratio between the numberof mols.of vapour returned as refluxed liquid to the rectifyingcolumn and the number of mols. of final product, both per unittime. This should be varied according to the difficulty of fraction-ation, rather than be maintained constant. A high efficiency ofseparation requires a high reflux ratio.(iii) HoZd-up of column. The hold-up should be reduced to aminimum compatible with scrubbing effectiveness and an adequate(ii) ReJclux ratio.2 W. A. Peters, junr., I n d . Eng. Chern., 1922, 14, 476.3 Bur.Stand. J . Res., 1931, 7, 851.I n d . Eng. Chem., 1936, 28, 644.I n d . Eng. Chern. (Anal.), 1933, 5, 119, 135, 172454 ANALYTICAL CHEIKISTRY.column capacity. Uniform, continuous, and geometrically sym-metrical packings were found to be most suitable.Slight heat losses completely disturbthe delicate equilibrium of the column, and almost perfect thermalinsulation is required to separate components boiling only a fewdegrees apart.The fractionating columns described have a metal reflector-type vacuum jacket for thermal insulation at all temperaturesfrom - 190" to 300", within which jacket any of a number ofdistilling tubes of different diameter may be inserted for the fraction-ation of gaseous or liquid samples. These distilling tubes containspiral, continuous, uniform wire-coil packings.Using columnsapproximately 4 ft. long, Podbielniak has obtained in a singledistillation lasting a few hours separations equal to those givenby very tall columns operated a t slow rates. The columns describedare particularly suitable for the fractionation of hydrocarbonmixtures.The construction and performance of three vacuum fractionatingcolumns of the Podbielniak type with capacities of 5 - 5 0 g.suitable for the distillation of hydrocarbons, alcohols, andparticularly esters have been described by A. Klem.6A modified Dufton fractionating column has been used by W. J.Gooderham for the analysis of hydrocarbon mixtures. A Duftonspiral is surrounded with a vacuum jacket which prevents thetendency of the spiral to choke a t temperatures above lOO", andenables extremely sharp fractionations to be obtained at hightemperatures. The fractionating surface consists of a spiral of10-12 mm.pitch-made by winding a soft copper wire arounda stiff core (e.g., nichrome)-which makes a tight, sliding fit in astrong tube of internal diameter 4.5 inm. The whole is enclosedin a silvered evacuated jacket.Fenske and his co-workers 7 y *, have studied various types ofpacking in fractionating columns. In investigations on columnsof diameters from 0-6 to 2.0 in., the apparatus was worked withsuitable liquid pairs, e.g., benzene and carbon tetrachloride, undertots1 reflux until equilibrium was established, the height of atheoretical plate equivalent to the packing being then determined.Efficiency was found to decrease with increase in diameter of thecolumn; and corrosion of the packing had, in general, an adverse(iv) ThermaE insukstion.ti Nature, 1938, 142, 616.7 M.R. Fenske, C. 0. Tongberg, and D. Q . Quiggle, Id. Eng. Chem., 1934.26, 1169.C. 0. Tongberg, S. Lawroski, and M. R. Fenske, ibid., 1937, 29, 957.* I&m, ibid., 1938, 30, 297GRIFFITHS, MACLENNAN, AND WHALLEY. 455effect on performance. The most efficient packings, which had alarge surface area, and a high percentage free space, were one- andtwo-turn helices of wire or glass, and carding teeth 3’’ wide. Owingto surface tension, which causes the condensate to coalesce on adry packing but to spread uniformly on a wet one, the highestefficiency in the column was obtained when the packing was wettedby flooding just before the start of fractionation.Fractionatingcolumns with helix-type packings showed higher throughput andefficiency than columns containing the many other types of packingused.The effect of packing in laboratory columns has been studied byother investigators. A. R. Glasgow, junr., and S. T. Schicktanz l ohave studied the relationship between the efficiency (q), liquid hold-up(v), ball diameter (d), and total surface ( a ) in fractionating columnspacked with glass, lead, and copper balls with diameters between2 and 4 mm. For a column of diameter 2.6 cm. packed with suchballs, the thermal conductivity of the balls was without effect,and approximately -q o c a p and v cca.Silicon carbide has been used as packing by H. J.Hall and G . B.Bachmann.ll Test data for mixtures of benzene and dichloro-ethane proved the efficiency of this type of packing, which has theadvantage that it is not attacked during the distillation of organicbases, corrosive sulphur compounds, and unstable halides. Fine-mesh wire-gauze packings, which gave an efficiency up to 20 platesper foot of column length, have been described by D. F. Stedman.12L. B. Bragg 13 has described a laboratory column using a conicaltype of Stedman packing, and has recorded efficiency test data fora mixture of benzene and dichloroethane. Columns equivalent to200 theoretical plates can be built in a laboratory of average height.M.L. Selker, R. E. Burk, and H. P. Lankelma14 have describeda &foot, efficiently insulated column, packed with close-fitting,multiple, concentric glass tubes, which has an efficiency of about85 theoretical plates, and a small hold-up. A fractionation columnemploying absorbents has been used in the separation of benzeneand methyl alcohol.15Enumeration of all the fractionating columns described in recentyears is not possible in a limited space, but a few may be mentioned.Particulars of a jacketed, heat-insulated column about 2 metreslong containing 100 bubble-cap plates (H.E.T.P. = 2 cm.) of thelo J . Res. Nut. Bur. Stand., 1937, 19, 593.l1 I n d . Eng. Chem. (AnaZ.), 1938, 10, 648.la Canadian J . Res., 1937, 15, B, 383.l3 I n d .Eng. Chem. (Anal.), 1939, 11, 283.l5 R. J. Hartman and D. H. Jung, PTOC. Indiana Acad. Sci., 1437, 46, 118.l4 Ibid., 1940, 12, 352456 ANALYTICAL CHEMISTRYtype described by J. H. Bruun l6 have been given by J. H. Bruunand W. B. M. Fau1coner.l' The column has a separating powerequivalent to about 70 theoretical plates, and very good separationsof n-heptane (b. p. 98.4") and toluene (llo"), and of benzene (80.2")and ethylene chloride (83.7") were obtained. A similar and shortercolumn has been described by J. H. Bruun and S. D. West 18 forthe distillation of low-boiling compounds, and in this column,n-propane of unusually high purity was prepared. l9 Variousdesigns of fractionating columns have been given by J. H. Simons.20A small low-temperature rectifying column with capacity of ca.5 C.C.of liquefied gas, which operates at constant pressure, and issuitable alike for purification and fractionation of small amountsof gases generated in reactions, and for liquids boiling a t - 130"to - 50", has been described by the same investigator.21H. Vigreux 22 has described a fractionating column for distillationat atmospheric or lower pressures.3. AppZimtions.-Some interesting separations by fractionationhave been made in recent years. J. R. Huffman and H. C. Urey 23used a fractionating column made up with alternate stationaryand rotating cones which provided a large reaction surface, to effectan approximately five-fold increase in concentration of the oxygenisotope of atomic weight 18.A few hundred C.C. of water containing0.85% of H,lsO were obtained in this way. Rectification has beenused in the separation of neon into its isotopes, and neon fractionswith atomic weights 20.043 and 20.785 were obtained. A 38-foldenrichment of deuterium has been effected, and positive resultsobtained in the separation of the isotopes of oxygen.24cisButene-2 (b. p. 3.73") and transbutene-2 (0.96") have beenseparated by G. B. Kistiakowsky, J. R. Ruhoff, H. A. Smith, andW. E. Vaughan 25 in an all-glass column 5 m. long and 18 mm. indiameter packed with glass spirals of the type described by C. D.Wilson, G. T. Parker, and K. C. Laughlin.26Distillation of decalin through a fractionating column a t a pressureof 10 mm. of mercury, followed by several fractional crystallisations,l6 Ind.Eng. Chem. (Anal.), 1936, 8, 224.l7 Ibid., 1937, 9, 192.l9 M. M. Hicks-Bruun and J. H. Bruun, J . Amer. Chem. SOC., 1936, 58,810.2o Ind. Eng. Chem. (Anal.), 1938, 10, 29.21 Ibid., p. 648.s3 Ind. Eng. Chem., 1937, 29, 530.24 W. H. Keesom and H. Van Dijk, 7th Congr. intern. Froid, 1st Comm.intern. Rapports et Commun., June 1936, 103; W. H. Keesom, H. VanDijk, and J. Haantjes, Physica, 1934, 1, 1109.l* Ibid., p. 247.22 Ann. Palsif., 1938, 31, 26.z 5 J . Amer. Chem. SOC., 1935, 67, 876. Ibid., 1933, 55, 2795GRIFFITHS, MACLENNAN, AND WHALLEY. 457has enabled W. F. Seyer and R. D. Walker 27 to isolate cis- andtrans-decahydronaphthalenes, and a column of the Fenske type35 cm. long and 1.5 cm.in diameter, packed with glass helices, hasbeen used by M. L. Sherrill and E. H. Launspach 28 to purify cis-pentene-2.4. MicrodistiZkbtion.-The development of micromanipulationhas resulted in attempts to achieve in microfractionation the degreeof precision obtained in macro-work with carefully designed columns.L. C. Craig 29 has described a column 100 mm. high for the dis-tillation of 04%-2.0 g. of material. The hold-up of the column issmall, approximately 0.1 g., and in using a test mixture of benzeneand carbon tetrachloride, a separation corresponding to approxim-ately 8 theoretical plates was obtained. S. D. Lesesne and H. L.Lochte 30 have designed a semi-microfractionating column whichuses a metal band 37.5 cm. long rotating at 1000 r.p.m.in place ofpacking. The apparatus can be used for quantities of 1-10 c.c.,at atmospheric pressure, and the separation obtained was equivalentto 13 plates a t total reflux. G. von Elbe and B. B. Scott 31 havedescribed an apparatus for high-vacuum fractional distillationwithout gravitational reflux. The mixture to be fractionated(0.01-0.10 c.c.) is sealed in a long glass tube evacuated to mm.of mercury around which a temperature gradient is maintained fora short distance by a thermostat system. The mixture tends toaccumulate at the low-temperature end of the gradient, and bypulling the tube slowly and uniformly through the gradient towardsthe warm end, the mixture is made to distil continuously withinthe gradient. The method is illustrated by the separation of p -and m-xylene.Other methods of microdistillation described include an apparatuswith indented fractionation column suitable for quantities of 0.2-0.5 c.c.,32 an apparatus for vacuum distillation of ~ 0 .5 g. of high-boiling fatty acids and their a distillation capillary forfractional distillation of low-boiling liquids in quantities as smallas 0.024.10 C.C. ,34 and a microdistilling apparatus for fractionatingunder reduced pressure quantities of material of the order5. Molecuhr Distillation.-In molecular distillation, the permanentgas pressure is so low (10-6 atm.) that i t has very little influence on0.5-2.0 g.3527 J . Amer. Chem. SOC., 1938, 60, 2125.2s Ind. Eng. Chem. (Anal.), 1937, 9, 441.30 Ibid., 1938, 10, 450.32 J.W. Young, Mikrochem., 1936, 21, 133.33 E. Klenk, 2. physiol. Chem., 1936, 242, 260.34 A. 0. Gottler and J. Fine, Ind. Eng. Chem. (Anal.), 1939, 11, 469.35 S. A. Shrader and J. E. Ritzer, ibid., p. 54,28 Ibid., p. 2562.s1 Ibid., p. 284458 ANALYTICAL CHEMISTRY.the speed of distillation, or even on whether distillation takes placeor not. The distillation velocity is then determined by the speedat which the vapour from the liquid being distilled can flow throughthe pipe connecting the still and condenser under the driving forceof its own saturation pressure. If the distance from the evaporatingliquid to the condenser is made commensurate with or less than themean free path of a molecule of distillate vapour in the residualgas a t the same density and pressure, the molecules which leave thesurface will for the most part not return.A form of apparatus inwhich a cooled condensing surface is supported a few em. above ashallow, heated pool of liquid, the whole enclosed in an evacuatedchamber, offers the least hindrance to the flow of vapour from theevaporating to the condensing surface. The rate of distillation isthen determined by the rate a t which the liquid surface is capableof producing vapour. Where the evaporating liquid is a chemicalindividual, the rate of evaporation will be pc/s g. per sq. em. persee., p being the density of the saturated vapour a t the given tem-perature, c the mean molecular velocity, and s the mean free pathof a distillate molecule.If the liquid is a mixture, the rate ofevaporation of the rth component will be p,c,/s g. per sq. em. persec. The separation obtained in molecular distillation, therefore,depends on p,cT, unlike the separation obtained in ordinary dis-tillation (where the vapour is in equilibrium with the liquid) whichdepends on p,. Since c, is inversely proportional to the squareroot of the molecular weight and the value of p, is in general greatestfor constituents of least molecular weight, p,c, is greatest for con-stituents of least molecular weight. By molecular distillation, animproved separation can be obtained, e.g., of isotopes. Indeed,the first application of molecular distillation was the separationof the isotopes of mercury.36 Subsequently, the method wasapplied to organic mixtures, e.g., to petroleum fractions by C.R.B ~ r c h , ~ ' and to paraffin wax by E. Washburn, J. H. Bruun, andM. M. Hi~ks.3~ Molecular distillation is, indeed, the only methodby which substances of high molecular weight can be distilled withoutdecomposition, and it is a method by which i t may be possible toavoid formation of constant-boiling mixtures.An interesting account of the apparatus and methods used inmolecular distillation has been given by K. C. D. H i ~ k r n a n . ~ ~ Theapparatus described is suitable for distillation of cotton-seed orcod-liver oils, natural waxes, lubricating oils, synthetic polymerisedmaterial, etc. Since under molecular distillation conditions, there36 J. N. Bronsted and G.von Hevesy, Phil. Mag., 1922, [vi], 43, 31.37 PTOC. Roy. SOC., 1929, 123, A, 271.38 BUT. Stand. J . Res., 1929, 2, 461. 39 I n d . Eng. Chem., 1937, 29, 968GRIFFITHS, MACLENNAN, AND WIIALLEY. 459is virtually no foreign gas, evaporation takes place whenever thereis a difference in temperature between distilland and condenser.Increase in the absolute temperature and in the temperaturedifference increases the rate of distillation, but there is no abrupttransition available for record as a boiling point. Hickman measuresa property of the distilland analogous to the boiling point, viz., the" elimination curve," secured by plotting rate of elimination, i.e.,rate at which material appears in the receiver, against the increase inabsolute distillation rate with temperature.The temperature ofmaximum elimination of the unknown is compared with that of" pilot " substances which are added in small amounts, e.g., methyl-indigo, celanthrene-red 3B, etc. The " elimination curve " has beendiscussed mathematically by N. D. Embree.4* Hickman 41 hasdistilled fish-oils in molecular stills, and has separated a vitamin-rich fraction evaporating below 190" and a mixture of glyceridesevaporating at 190-250". Two elimination curves for vitamin-Awere found with maxima a t 125" and 220°, probably correspondingto the free vitamin and to a mixture of its esters. Vitamin-D,likewise, showed maxima for the free vitamin (162') and its esters(230-250'). Further work on the molecular distillation of vitamin-D has been published by K.C. D. Hickman and E. LeB. Gray,42and work on vitamins-A and -D by J. G. Baxter, E. LeB. Gray, andA. 0. T i ~ c h e r . ~ ~ Molecular distillation has been applied to theconcentration of the oil-soluble vitamins A , D, and E,44 and thepossibility of applying molecular distillation to large-scale industrialunits has been considered by C. R. Burch and W. J. D. Van Dijck.45Molecular stills have been described for the concentration, fraction-ation, and purification of free fatty acids, sterols, vitamins, etc.466. Axeotropic Mixtures.-When a mixture of two or more com-ponents is distilled, the components normally separate in order oftheir boiling points. In many cases, however, the vapour pressure-composition curves show at particular concentrations maximumor minimum values corresponding respectively with minimum ormaximum boiling points. The term " azeotropes " was appliedby J.Wade and R. W. Merriman47 to such mixtures, It is usualin such mixtures to find a minimum boiling point, and the mixtureis said to exhibit positive azeotropism or azeotropism of the first4 0 I n d , Eng. Chem., 1937, 39, 975.4 3 Ibid., 1938, 30, 796.4 4 W. Jewell, T. H. Mead, and J. W. Phipps, J . SOC. Chern. Ind., 1939, 58,4 5 Ibid., p. 3 9 ~ .4 6 S . B. Detwiler, junr., and K. 5. Markley, I n d . Eng. Chm. (Anal.), 1940,4 7 J., 1911, 99, 1104.41 Ibid., p. 1107.48 Ibid., 1937, 29, 1112.6 6 ~ .12, 348460 ANALYTICAL CHEMISTRY,kind; negative azeotropism or azeotropism of the second kind isassociated with a maximum boiling point.About 200 negativeazeotropes are known.48 The present discussion is restricted topositive azeotropism, but similar considerations hold for negativeazeotropism. On distilling a mixture exhibiting positive azeotrop-ism, the azeotrope distils over first, at constant boiling point, andany excess of either constituent remains in the still. It is possibleby careful fractionation of such a mixture to separate ( a ) theazeotrope, and ( b ) one of the components, if present in excess,Lists of azeotropic mixtures are given in Sydney Young’s “Dis-tillation Principles and Processes ” (London, 1922), in MauriceLecat’s book “ La Tension de Vapeur de MBlanges de Liquides :L’AzBofropisme ’ ’ (Brussels, 1938) , and in other papers .49Azeotropes may be separated into their constituents by chemicalor physical methods.If one of the components is acidic, basic,or contains halogen, etc., chemical methods of analysis may betried. Physical methods, however, are to be preferred, e . g . ,solubility of one of the components in a third liquid. If thequalitative composition of a binary (i.e., two-component) azeotropeis already known, the quantitative composition may be evaluatedby comparing density, refractive index or other convenient propertywith the appropriate values for known synthetic mixtures. Forexample, azeotropic concentrations have been determined bydensity measurements accurate to a few parts per million 5o byusing the twin pyknometer method of measuring density.51Athird substance may be added which forms either another binaryazeotrope with one of the components or a tertiary azeotrope withboth components, the new azeotrope having a lower boiling pointthan the original.On distillation, the new azeotrope distils first,leaving behind the other constituent or part of it. Such “ thirdsubstances’’ may be selected from the tables given in Lecat’sbook.R. Wright 52 has considered the case where there is addition ofa third substance with selective action on one of the constituents.Addition of a relatively non-volatile material yields a distillate a tfirst richer in the constituent in which the added substance isinsoluble. The method was applied to the azeotrope n-propylalcohol-water, which contains 28% of water.By adding 25%Distillation methods may also be applied in several ways.4 a M. Lecat, Ann. Soc. sci. Bruxelles, 1934, 54, B, 283.49 M. Lecat, ibid., 1936, 55, B, 253; 1936, 56, B, 41, 221.50 M. Wojciechowski, Nature, 1938, 141, 691.51 E. R. Smith and W. Wojciechowski, Rocz. Chem., 1936, 18, 104.62 J., 1938, 1720GRIFFITHS , MACLENNAN , AND WHALLEY. 461by weight of potassium hydroxide to the azeotrope, in one dis-tillation, material was obtained containing only 8% of water,which, on fractionation, yielded the azeotrope and pure n-propylalcohol.The reverse process, separation of liquids by first making themform azeotropes with a third substance, is illustrated by the separ-ation of butanes and butenes by distillation of their azeotropeswith sulphur dioxide, the butane azeotropes having lower boilingpoints than any of the butene azeotr~pes.~~ In a similar manner,(i) fuse1 oil may be analysed by fractionating the mixture of binaryazeotropes formed by the individual alcohols and carbon tetra-chloride; and (ii) mixtures of aliphatic acids may be separated byfractional distillation of the binary azeotropes which are formedwith benzene and toluene.54 S.Young and (Miss) E. C. Fortey 55 de-duced the empirical relationship for a positive azeotrope that theratio of the weight of constituent not in excess in the originalmixture to the weight of distillate below the mid-point, i.e., thetemperature mid-way between the boiling points of the two con-stituents, is equal to the proportion of that constituent in the binarymixture. Thus, if the proportion of that constituent in the azeotropeis known, the composition of the original mixture can be deduced.Azeotropes may be separated into their components by distillationunder reduced pressure.R. W. Merriman 56 showed that a t760 mm. pressure, the ethyl alcohol-water azeotrope contained4.4% of water, whereas a t 70 mm. pressure there was completeseparation of the two constituents. Other investigators haveconsidered the effect of pressure on aze~tropes.~~ In the systemethyl alcohol-benzene, E. Le. Borgne and M. Schmitt 68 havestudied the azeotropic concentration at pressures ranging from 89to 1160 mm. of mercury.Azeotropes may serve to identify unknown organic liquids ifthey boil without decomposition a t atmospheric pressure. Themethod, which is described in Lecat’s book, is to measure thedifference in temperature between the boiling point of one of a53 M.P. Matuszak and F. E. Frey, Ind. Eng. Chem. (And.), 1937, 9, 111.54 (i) S. T. Schicktanz, A. D. Etienne, and W. I. Steele, ibid., 1939, 11, 421 ;6 6 J., 1902, 81, 717, 739; 1903, 83, 45.5 6 J., 1913, 103, 628.5 7 A. A. Sunier and C. Rosenblum, I d . Eng. Chm. (Anal.), 1930, 2, 109;W. Swientoslawski and B. Karpinski, Compt. rend., 1934, 198, 2166;A. Bouzat and M. Schmitt, ibid., 1934, 198, 1923 ; W. D. Bonner and M. B.Williams, J . Physical Chem., 1940, 44, 404; P. W. Schutz and B. E. Mallonee,J . Amer. Chem. SOC., 1940, 62, 1491.5 8 Bull. SOC.sci. Bretagne, 1936, 13, 72.(ii) S . T. Schicktanz, W. I. Steele, and A. C. Blaisdell, ibid., 1940, 12, 320462 ANALYTICAL CHEMISTRY.series of standard liquids and that of the binary azeotrope formedby the standard and the unknown.COLORIMETRIU ANALYSIS.To the physicist, colorimetry relates to the process of completespectrophotometric analysis or of determining a colour in terms oftrichromatic coefficients, whereas to the chemist it generally refersto the process of comparing the colour of a medium with a standardcolour, with the object of establishing the identity or the con-centration of the substance responsible for the colour in the medium.Occasionally, the chemist finds it advantageous to employ thetrichromatic system, as, e.g., in measuring the colour of sugarproducts and in defining a minimum blue colour which antimonytrichloride shall produce with cod-liver oil in order that the lattershall conform with British Pharmacopoeia requirements.2 Theprinciples underlying the selection of a correct terminology inrelation to instruments and processes are discussed by M.G. Mellon.3I n colorimetric analysis, a photoelectric photometer or colori-meter has advantages of speed and accuracy over direct visualcomparison methods, particularly when a number of determinationsof the same substance by a particular colour reaction are beingmade, and the last few years have been marked by the productionof a large variety of such instruments. It is of interest thatthermopiles may be used in place of photocells, thereby permittingthe construction of a thermoelectric abs~rptiometer.~The investigation of a proposed colorimetric method now involvesthe determination of the spectral region in which light absorptionis a maximum, and the provision of a light source correspondingwith this region.The colour matching is then a matter of photo-metry. With the application of colour filters which transmitrelatively narrow spectral bands, the gap between photoelectriccolorimetry and photoelectric spectrophotometry is not large.5The fraction of incident light transmitted by the coloured mediumis observed a t time intervals to determine whether any changesoccur as a result of any photochemical reaction or any incom-pleteness of the reaction which produces the coloured substance.The effective extinction coefficient under the optimum conditionsand then the applicability of Beer’s law are determined.6 The1 E.Landt and H. Hirschmuller, 2. Wirts. Zuckerind., 1938, 88, 247.British Pha.rrnacopu&, 1932, p. 596.Ind. Eng. Chern. (Anal.), 1939, 11, 80.H. H. Willard and G. H. Ayres, ibicl., 1940, 12, 287.ti Cf. Ann. Reports, 1938, 35, 394.6 R. H. Muller, Ind. Eng. Chem. (Anal.), 1939, 11, 1GRIFFITHS, MACLENNAN, AND WHBLLEY. 463errors involved and the sensitivity of photometric methods havebeen discussed.'From the point of view of colorimetry, organic compounds maybe divided into two classes. One of these consists of substanceswhich can be determined directly in virtue of their own colour;the other comprises substances which afford a suitable colour onlywhen treated with a reagent, If the substance to be determinedis the only coloured substance in the medium, or if any absorptiondue to other substances can be made ineffective by the use ofsuitable colour filters, the concentration of the substance in questioncan be determined in terms of the fraction of incident light absorbed.For example, as little as 0.1 pg.of chlorophyll in acetone can bedetermined by observations with a vacuum thermocouple in theregion of the red absorption band and carotenoid pigments do notinterfere, * or a photoelectric colorimeter with suitable filters maybe used.gAlthough the compound to be determined is sufficiently coloured,it may be associated with other coloured compounds which mustbe removed, otherwise they would contribute to the magnitudeof the colorimetric observations.This applies to the determinationin foodstuffs of carotene, which is of considerable importance asthe precursor of vitamin-A. The pigments, after extraction bymeans of solvents and saponification, are separated by partitionprocesses employing various solvents.lO The resulting '' carotene "solutions appear, in several cases, t o contain a considerable pro-portion of colouring matter which is not carotene. The yellowpigment extracted from flour is an example of this.ll In the caseof extracts of leaf material, the xanthophyll may be removed fromthe carotene by shaking the petroleum solution with speciallyprepared magnesium hydroxide.12 Some foodstuffs, e.g., tomatoes,contain lycopene or other red pigments which are not completelyremoved by the xanthophyll adsorbent, but a specially preparedmagnesium carbonate was found to adsorb both xanthophyll andlycopene but not car0tene.1~ The percentages of the differentcoloured components of paprika vary, and since the total colourS.E. Q. Ashley, Ind. Eng. Chem. (Anal.), 1939, 11, 72.E. S. Johnston and R. L. Weintraub, Smithsonian Misc. Coll., 1939,H. G. Petering, W. Wolman, and R. P. Hibbart, Ind. Eng. Chern. (Anal.),98, 1.1940, 12, 148.lo W. M. Seaber, Analyst, 1940, 65, 266.l1 V. E. Munsey, J . Assoc. Off. Agric. Chem., 1938, 21, 331.l2 G. S. Fraps and A.R. Kernmerer, <bid., 1939, 22, 190.lS G. S. Fraps, A. R. Kemmerer, and S. M. Greenberg, &id., 1940, 28, 422,659464 ANALYTICAL CHEMISTRY.does not adequately evaluate the article, these pigments are separatedchromatographically and determined colorimetrically.14A means of eliminating the effect of interfering coloured sub-stances is illustrated in the determination of riboflavin (or lacto-flavin, a constituent of the vitamin-& complex) in yeast and yeastproducts. Coloured substances in the filtered hydrochloric acidhydrolysate are reduced with sodium dithionite (hyposulphite orhydrosulphite) a t pH 3 - 5 4 , the reduced riboflavin is reoxidisedby shaking with air, and the light transmission is measured photo-electrically. Then the vitamin is 90% reduced with sodiumdithionite and the residual colour is determined.The differencebetween the two readings is a measure of the riboflavin.15 I n thecase of riboflavin extracts from other materials, some interferingsubstances are removed by preliminary oxidation with potassiumpermanganate before the light absorption measurements are made.16I n the majority of cases, the substance to be determined iscolourless or insufficiently coloured for that property to be useddirectly as a means of determination a t the dilution in question.Therefore, a reagent is sought which gives a sufficiently intenseand stable colour. As an example, as little as 0.5 mg. of nitro-ethane in 25 ml. of aqueous solution, when treated with sodiumhydroxide, acidified, and then immediately treated with ferricchloride, affords a red colour which is compared with similarlyprepared standards.Nitropropane and nitrobutane can be deter-mined similarly, but the colour fades too rapidly in the case of theisopropane and isobutane compounds, and nitromethane does notgive a colour. l7 Small quantities of 2-methyl-1 : 4-naphthaquinone,which has some of the characters of vitamin-K-the blood-coagul-ation vitamin-are determined by means of the yellow colourproduced when the alcoholic solution is treated with alcoholic:ammonia and ethyl cyanoacetate, followed by potassium hydroxide.18It is recognised that many colour reactions depend upon the-presence oPcertain groupings in the molecule of the compound tobe determined. Diazotised p-aminoacetophenone is said to be a:specific reagent for the thiazole grouping in aneurin (vitamin-B,,thiamin), giving a purple-red substance l9 which is soluble in certain(organic solvents, thereby affording a colorimetric determinationA.E. Schumacher and G. F. Heuser, Ind. Eng. Uhem. (Anal.), 1940, 12,16 G. Lunde, H. Kringstad, and A. Olsen, 2. physiok. Chem., 1939, 260,l7 E. W. Scott and J. F. Treon, Ind. Eng. Chem. (And.), 1940, 12, 189.l8 J. L. Pinder and J. H. Singer, Analyet, 1940, 65, 7 .l4 L. Cholnoky, 2. Unters. Lebensm., 1939, 78, 157.203.141.H. J. Prelubda and E. V. McCollum, J . BioE. Chem., 1939, 127, 495GRIFFITHS, MACLENNAN, AND WHALLEY. 465of the vitamin-B, content of various natural products,20 but aneurinphosphates escape determination unless hydrolysed.21The recognition of nicotinic acid as an antipellagra vitaminhas stimulated the study of its colour reactions,22 and the yellowcolour produced when the pyridine ring reacts with cyanogenbromide and an aromatic amine is also given by nicotinamide,nicotinuric acid, and nicotine, the relative colour intensitiesproduced by equimolar concentrations being, respectively,100 : 142 : 42 : 8 when metol is the aromatic amine.The totalnicotinic acid (free and combined) in a mixture is determined assuch if a preliminary hydrolysis with an alkali is carriedp-Aminoacetophenone gives a more intense colour than anilineor metol, and in order to obtain reproducible results the colourreaction is conducted a t pn 6 and the solutions are protected as faras possible from light.24 The last method determines as little as1 pg.of nicotinic acid per g. of material after this has been hydrolysedwith alkali and certain substances precipitated by alcohol. Beforehydrolysis, nicotinamide gives only 20% of the colour producedby nicotinic acid, whereas nicotindiethylamide gives a more intensecolour. Pyridine, a-aminopyridine, and nipecotic acid give muchless intense colours than nicotinic acid, and many amino-acids andthe pyridine derivatives adermin, trigonelline, and quinolinic acidgive negative results. The difficulties to be faced in adjusting theconditions in order to make a colorimetric reaction specific or evenhighly selective are illustrated by the observation that certaincereals appear to contain more nicotinic acid (determined colori-metrically) than corresponds with the biological activity.26 Theabove results also illustrate that the intensity of the colour producedby a particular reaction is affected by changes in the attachmentsto the reactive grouping of the molecule under investigation as wellas by those of the reagent.The nature of other substances present along with that to bedetermined may make a preliminary separation desirable, but thismay not always be practicable.I n such cases, the selectivity ofthe reagent producing the colour is particularly important. Thecreatine and creatinine contents of meat and other extracts and ofbiological fluids are diagnostically significant.After preliminarytreatment, the creatinine, into which creatine is converted, is2o M. E. Auerbach, J . Amer. Pharm. ASSOC., 1940, 29, 313.21 D. Melnick and H. Field, junr., J . Biol. Chem., 1939, 127, 531.22 W. R. Ashford and R. H. Clark, Trans. Roy. SOC. Canada, 1939, [iii], 33,23 E. Bandier, Biochem. J., 1939, 33, 1787.24 L. J. Harris and W. D. Raymond, ibid., p. 2037.2 s E. Kodicek, ibid., 1940, 34, 712, 724.111, 29466 ANALYTICAL CHEMISTRY.usually determined by means of the yellow colour produced withpicric acid and sodium hydroxide, but 3 : 5-dinitrobenzoic acid isfound to be a more selective reagent, albeit the colour producedis less stable.26 Micro-determination in plasma and serum maybe effected by conversion into methylguanidine with yellowmercuric oxide and measurement of the red colour produced withcc-naphthol and hyp~bromite.~~FLUORESCENCE METHODS.If a substance emits light while it is absorbing radiation, it issaid to fluoresce.Fluorescence differs from the Raman effect 28in that the latter is due to the scattering of light by molecules,whereas fluorescence results from the absorption of radiation.Fluorescent light is generally of longer wave-length than that ofthe incident radiation, and when spectrally resolved is usuallyfound to consist of bands, the nature and wave-lengths of whichare characteristic of the substance concerned. For example, thecharacteristic fluorescence bands shown by europium, ytterbium,samarium, and thulium can be used for the detection of smallamounts of these elements,29 and fluorescence spectrum measure-ments allow the detection of 0.01, 0.1, and 100 p.p.m.of terbium,cerium, and europium, respectively, in s~lution.~O m-Diamines aredetected by the bright yellowish-green fluorescence of the diamino-acridines formed by heating with zinc chloride, glycerol, and oxalicacid. A phenolic group interferes with the test, but if the zincchloride is half replaced with stannous chloride, m-orientation inaminonitro- or dinitro-compounds can be detected by thefluorescent diaminoacridines prod~ced.~l The fluorescence of theporphyrins has been investigated extensively, and the determinationof the concentration of these substances by the fluorescence spectraof their solutions is des~ribed.~~ Increased excretion of porphyrinis associated with lead poisoning, and the characteristic redfluorescence appears to be the most sensitive and accurate methodof determining coproporphyrin-III.33Fluorescence is, however, often markedly influenced by numerous26 E.Komm and H. Pinder, 2. Unters. Lebensm., 1939, 78, 113.28 Ann. Reports, 1938, 35, 394.2o K. Przibram, Mikrochirn. Acta, 1938, 3, 68.30 A. Zaidel, J. Larionov, and A. N. Filippov, J. Gen. Chem. Russia, 1938,31 A. Albert, J., 1939, 920.32 A. Keys and J. Brugsch, J. Arner. Chem. SOC., 1938, 60, 2135.33 M. Meyer, Siiddeut. Apoth.-Ztg., 1939, 78, 138.A. Riegert, Compt. rend. Soc. Biol., 1939, 132, 535.5, 943GRIFFITHS, MACLENNAN, AND WHALLEY. 467factors operating during observation, and it is the recognition andcontrol of these that must precede the acceptance of a fluorescenceprocedure a8 a trustworthy analytical method.Some of thesepoints will be illustrated in the sequel.Light 8ources.-Ultra-violet radiation of wave-length between3000 A. and 4000 A. is generally employed for exciting the fluoresc-ence, which is usually of low intensity relative to the source.Consequently, the exciting radiation must be of high intensity and,for quantitative work, should be constant for long periods. Inaddition to modern types of mercury arc, an iron vacuum arc andan iron arc cored with iron-carbon mixture satisfy the conditions.MVisible light is removed from the exciting radiation by means of alight filter, such as Wood’s glass.Instruments.-There is a fairly close parallel between the typesof instrument used in fluorescence studies and in colorimetry.The colour and intensity of the fluorescence may be matchedsubjectively in terms of trichromatic coefficients, e.g., in the Donald-son ~olorimeter.~~ The fluorescence may be compared visuallyor photoelectrically with standards in a photometer or may beobserved spectrometrically and recorded spectrographically. Thelast process may involve exposing a sensitive photographic platefor many hours, but in cases where the spectrum is specific for asubstance, it is of value in identification and it also facilitates theselection of colour filters and standards having the appropriatespectral characteristics.Methods.-It is not unusual for solutions of fluorescing organicsubstances to undergo changes-photochemical, oxidative, etc.-which alter the characteristics of the fluorescence and make suchsolutions unsatisfactory as standards, particularly in routine work.For instance in the determination of aneurin (vitamin-B,) byoxidation to thiochrome, which is determined fluorometrically,quinine sulphate lacks stability as a standard but a piece of fluorsparhas been found suitable.36 The riboflavin content of milk is rapidlydetermined with a mean error of 5 2 % by comparing photoelectricallythe fluorescence of the filtered acetone-milk mixture with that ofa piece of uranium glass which is calibrated against solutions ofpure riboflavin under the same conditions. The fluorescence is sointense that amplification of the photocell current is unnecessary,but fluorescein and riboflavin are not used as standards in routinework owing to photodecomposition.However, this decompositionis not so rapid as to introduce a significant error if, as in the deter-34 C. E. White, Ind. Eng. Chem. (Anal.), 1939, 11, 63.36 R. Donaldson, Proc. Physical Soc., 1935, 47, 1068.36 R. G. Booth, J . SOC. Chem. Ind., 1940, 59, 181468 ANALYTICAL CHEMISTRY.mination of the vitamin in the sample, a single observation lastingonly 3 seconds is s~fficient.~’The fluorometric determination of an ingredient of a mixturedoes not always depend on its own fluorescence or that of the mainproduct of a reaction. For example, the determination of 0.05-1.0% of o-nitrophenol in p-nitrophenol depends on the observationthat o-aminophenol (e.g., produced by reduction of o-nitrophenol)reacts with benzoic acid to form phenylbenzoxazole and associatedfluorescent by-products, whereas rn- and paminophenols do not.The fluorescence of the washed benzene extract is compared withthat obtained similarly from mixtures of known cornp~sition.~~Fluorescence methods are also valuable in inorganic analysis,as, e.g., in the detection and determination of small quantities ofaluminium. The first direct test to differentiate this metal fromberyllium depends upon as little as 0.2 pg.of the former producingan orange-red fluorescence with Pantochrome Blue Black R (thezinc salt of 4-sulpho-2 : 2’-dihydroxyazonaphthalene) in presenceof acetic acid, but certain elements interfere.39 It has not beenpossible to adapt this reaction for quantitative work, but morin,the dyeing principle of fustic wood, gives an intense green fluoresc-ence with aluminium, and by adding enough reagent to give maxi-mum fluorescence and a t the same time keeping the concentrationof alcohol constant, as little as 0.0005 pg.of aluminium a t a con-centration of 0.1-1 a 2 mg./l. can be determined fluorometrically.*0However, various ions interfere, and beryllium, gallium, indiumand the rare earths give a, fluorescence similar to that of aluminium.There are numerous examples of fluorescent substances beingused as indicators. Aluminium may be determined by titratingalkali fluorides in presence of sodium chloride with an alcoholicsolution of aluminium chloride containing morin, the end pointbeing characterised by fluorescence occurring when Al”’ ceases tobe removed as the non-fluorescent AlF,’”.41 The end-points ofthe iodine-sodium thiosulphate and the arsenious oxide-bromatetitrations in the presence of coloured ions are indicated bya-naphthaflavone, of which the ultra-violet-excited blue fluorescenceappears after removal of free iodine or bromine.42Limitations of space permit reference to only a small proportionof recent work, but a full account of the diverse applications ofD. B.Hand, Ind. Eng. Chem. (Anal.), 1939, 11, 306.W. Seaman, A. R. Norton, and 0. E. Sundberg, ibid., 1940, 12, 403.30 C.E. White and C. S. Lowe, ibid., 1937, 9, 430.4 O Idem, ibid., 1940, 12, 229.4 1 A. Okac, Coll. Czech. Chem. Comm., 1938, 10, 177.4 2 H. GotB, Sci. Rep. TGhoku, 1940, 29, 1GRIFFITHS, MACLENNAN, AND WHALLEY. 460fluorescence analysis as late as 1939 has been published by J. A.Radley and J. Grant.43ELECTROLYTIC ANALYSIS.Electrodeposition has previously received only occasional andbrief mention, and a survey of recent advances in this field ofanalysis, in which standard methods 1 already exist for the deter-mination in certain circumstances of a large number of metals, isdeserved. That the material to be determined is usually obtainedin elemental form and weighed as such, instead of as some com-pound to the weight of which a conversion factor has to be applied,appeals to the analyst : but the appeal may be mainly mthetic,and the justification of higher accuracy should be proved for eachmethod and not presumed.General Considerations.-The potential difference, x , between ametal and a solution of its own ions is given by the Nernst equation0.0002T [ion] log -, or x = xo + os log [ion] k Vx = -2)where v is the valency of the ion, xo is the standard potential whenthe ions are a t concentration of 1 g-ion/l., and the other symbolshave their usual significance.At 17" this equation becomesx = x,, + (0-058/v) log [ion] . . . . - (1)The standard potential, as defined above, of the hydrogen electrodeis taken as zero on the hydrogen scale o€ potential, to which thepotential of any other electrode is usually referred.If the hydrogenelectrode is combined with some other electrode, and the e.m.f.of the resulting cell measured, this e.m.f. is equal to the singlepotential of the other electrode, and the sign of this potential ispositive or negative according as the other electrode is the positiveor negative pole of the cell, e.g., Na, - 2-7 v. ; Fe, - 0.44 v. ; Ni,- 0.23 v. ; Sb, 0.1 v. ; Cu, 0.34 v. ; Ag, 0.8 v. From equation ( 1 )i t will be seen that the lower the concentration of a metal ion insolution, the less positive (or more negative) is the potential of themetal electrode, e.g., the cathode of an electrolytic cell, in contactwith it. To deposit a metal, the potential of the cathode must beslightly more negative than the equilibrium value, and since xchanges by O.O58/u volts for a ten-fold change in concentration, toeffect a reduction to one ten-thousandth (i.e., 1 g.to 0.0001 g.)43 " Fluorescence Analysis in Ultra-Violet Light," 3rd Edition, Chapman& Hall, 1939.H. J. S. Sand, " Electrochemistry and Electrochemical Analysis," Vols.I and 11, Blackie & Son, Ltd., 1940470 ANALYTICAL CHEMISTRY.for a bivalent ion, the cathode potential has to change by- (0.058/2) x 4 = - 0.116 v., and double this valuefor a univalention. Hence, provided the standard potentials of two metals differby more than 0.2 volt, practically complete electrolytic separationshould be possible, the metal with the more positive potentialbeing deposited first.H. J. S. Sandy2 realising the fundamentalimportance of the cathode potential in electrolytic depositions andseparations, initiated for its measurement and control the use of anauxiliary electrode (usually a saturated potassium chloride calomelelectrode) in the system. The construction of a suitable electrodevessel has been de~cribed,~ and recent developments based upon itsuse are reported later.According to the simple theory outlined above, in the electrolysisof a N-acid solution, hydrogen should be evolved a t a potential of0 v., and with a N-alkali solution a t a potential of - 0.8 v. In fact,a potential in excess of that calculated, i.e., an overpotential, usuallyhas to be applied before hydrogen is evolved, and the value of theoverpotential varies with the nature of the metal and its physicalstate.The existence of this phenomenon makes the depositionof certain metals with a more negative potential than that ofhydrogen practicable. Overvoltages are most marked in the caseof gases, and i t is to reduce the anodic overvoltage, or anodicpolarisation, of oxygen that reducing agents are frequently addedto solutions for electrolysis. Lindsey and Sand have investigatedthe efficiency of the sulphates and hydrochlorides of hydrazineand hydroxylamine as anodic depolarisers under varying con-ditions of temperature and current density, and conclude thathydrazine salts appear to be efficient under all conditions, whilstthe efficiency of hydroxylamine, though smaller, is improved bythe presence of chlorine and copper ions.The overpotential duringthe deposition of most metals is usually very small ((( 0.1 v.), andhence the deposition of metals at the cathode is an almost trulyreversible process to which the Nernst equation can be applied.Iron, cobalt, and nickel, with overvoltages of about 0.25 v., wereobserved by S. Glasstone to be exceptional, though at 90" theybehaved practically normally.The deposition of a metal from a solution of its complex ionsfrequently requires a more negative cathode potential than froma simple salt solution, since the effective concentration of simplemetal ions is reduced by complex formation. This change indeposition potential, similar in its effect to an overvoltage, is in3 A. J. Lindsey and H. J.S . Sand, AnaEyst, 1934, S9, 329.J., 1907, 91, 373.Idem, ibirE., 1935, 60, 739. J., 1926, 2887GRIBFITHS, MACLENNAN, AND WHALLEY. 47 1fact a concentration polarisation, and Glasstone so ascribes thehigher cathode potentials necessary for the deposition of the metalsfrom argenti- and cupro-cyanide solutions. Its effect is clearlyof analytical use, since the shift in the deposition potentials may incertain cases permit the electrolytic separation of two metalswhich when they are present wholly as their simple ions is notpracticable.Since it is the concentration of the ions in the immediate vicinityof the cathode that determines the potential difference, n, thedesirability of stirring the electrolyte becomes evident. At lowcurrent densities, the ions migrate and diffuse into the cathode layera t a rate equal to that of deposition, but on increasing the currentdensity a limiting value is reached beyond which the cathodepotential rises until the deposition of other ions occurs.Thislimiting value of the current density is raised considerably bystirring and by increasing the temperature, and both devices arecommonly employed to facilitate speedy deposition. The mixingof the products of the anodic and cathodic reactions by vigorousstirring may, however, in certain cases be undesirable. Hightemperatures, though decreasing the resistance, increasing diffusion,and often assisting in the formation of a good adherent deposit,reduce the overvoltage of hydrogen and may cause oxidation ofthe deposit. 0.K. Kudra has examined the influence of tem-perature on cathode processes with particular reference to theformation of black deposits of copper and cadmium. The decreaseof the hydrogen overvoltage and the consequent evolution ofhydrogen is generally unwanted : it often causes spongy and non-adherent deposits, possibly due to the formation and subsequentdecomposition of hydrides. When mercury or Wood’s metalcathodes are used,8 however, the liberation of hydrogen providesa simple means of agitating the solution.General Determinations.-The influence of iron salts on the de-position of copper has been investigated by J. G. Fife and S.T~rrance.~ During electrolysis, oxidation of ferrous to ferric ionsoccurs a t the anode, and since ferric ions cause redissolution of theelectrolytically deposited copper, they must be reduced at thecathode before deposition of copper can take place.Smallquantities of iron can be kept reduced by the addition of hydrazineor hydroxylamine salts, but when large amounts are present thistreatment is inadequate. The difficulty may be overcome byJ., 1929, 690.Mem. Inst. Chem. Ukrain. Acad. Sci., 1938, 5, 239.H. Paweck, 2. anal. Cham., 1927, 72, 225; 1929, 79, 115.Analyst, 1937, 62, 30472 ANSLYTICAL CHEMISTRY.separating the revolving anode, which is immersed in an indifferentelectrolyte, by a parchment diaphragm from the cathode and thesolution to be analysed, the potential of the cathode being controlledthroughout the analysis.Alternatively, the internal electrolyticmethod referred to below may be employed. M. Geloso and P.Deschamps l o investigated the iron-copper system and observedthat the critical iron concentration is lowered by increase in tem-perature and in the rate of agitation of the solution. M. Geloso 11reports that re-dissolution of the copper can be prevented by theaddition of a fluoride, which effectively removes the iron from thealternating oxidation-reduction processes by the formation of thecomplex ion (FeF,)”’. The completion of the deposition of copperfrom dilute nitric acid solution is obtained, according to J. A.Schemer, R. K. Bell, and W. D. Mogerman,12 if a trace of chlorideis added either during the dissolution of the alloy or before elec-trolysis.The influence of glycerol, acetone, methyl and ethylalcohols on the cathode efficiency of copper deposition from acidifiedcopper sulphate solution has been quantitatively investigated,l3and the increase in efficiency with increasing concentration of non-electrolyte up to a limiting value is attributed to the reduced abilityof the solution to redissolve deposited copper, whilst the possibilityof a complex cation of bivalent copper and the non-electrolyte isrecognised.I n an examination of the influence of acetone and methyl andethyl alcohols over a wide range of concentrations on the depositionof siZver,l* pronounced discontinuities in the interrelationship ofthe concentration of non-electrolyte and cathode efficiency arethought to indicate the formation of complexes.R. Taft andL. H. Horsley l5 have made an extensive examination of the effectof the addition of 140 organic compounds and 30 inorganic saltson the deposition of silver from silver nitrate solution under standardconditions. Colloids of molecular weight greater than 250 werefound to produce abnormal or striated deposits, whilst certainhigher aliphatic and cyclic acids and inorganic salts gave finecrystalline deposits. The mechanism of the deposition of silverfrom silver nitrate solution has been elucidated by microscopicexamination, and the deduction made that the number of silvernuclei forming was inversely proportional to the concentration10 Bull. SOC. chim., 1939, [v], 6, 1100.11 Ibid., p.1238.12 J. Res. Nat. Bur. Stand., 1939, 22, 697.l3 S. S. Joshi, D. H. Solanki, and T. V. S. Rao, J. Indian Chem. SOC., 1938,14 S. S. Joshi and S. Padmanabhan, ibid., p. 176.1 5 Trans. Electrochem. SOC., 1938, 74, 77.15, 167GRIFFITHS, MACLENNAN, AND WHALLEY. 473of silver nitrate between 0.1 and Details for an improvedprocedure for the deposition of silver from potassium cyanidesolution have been given.17Electrolytic methods have frequently been employed in micro-analysis, and a recent example is the observation that 0.4-30 pg.of gold may be electrolysed quantitatively by using gold-free leadcathodes,l* the lead being subsequently eliminated by the usualmethods, and the gold bead measured micrometrically. Theelectrolytic microdetection of traces of copper on shears which hadbeen used for cutting copper wire has been described by G .W.Baker.19 A good positive result was obtained with only 0.0002 mg.of copper per sq. mm. The precision of the microelectrolyticdetermination of copper has been examined.20The separation of lead from bismuth was effected in earliermethods 2s 21 by first depositing the bismuth a t a controlled potential,reducing agent being added, and the lead being subsequentlydeposited as metal. This method was recognised by E. M. Collin z2to have certain disadvantages owing to the great tendency of thelead to oxidise on drying, and to the deleterious effect on the platinumcathode, and a modified method was devised in which, as the reduc-ing agent during the deposition of bismuth a t 80--85", hydrazinehydrate was used, this material being advantageous in that it wassubsequently easily destroyed by sodium peroxide, thus enablingthe lead to be deposited as dioxide.Collin deduced an empiricalconversion factor for the lead dioxide deposit to lead, and con-cluded that its value was dependent on the conditions of depositionand drying. The determination of lead as the electrolyticallydeposited dioxide has recently been the subject of investigationby W. T. Schrenk and his c o - ~ o r k e r s , ~ ~ who found that for smallquantities the composition of the anodic deposit was SPbO,,PbO,H,O,i.e., substantially pure PbO,.The electrolytic deposition of antimony 24 from dilute sulphuricacid solution has been shown to be rapid and accurate (averageerror of 0.1% of the amount determined), even though basicl6 A.T. Wahrarnian and S . A. Alemian, Acta Physicochim. U.R.S.S., 1937,7, 95.l7 D. Tschavdarov, 8. anal. Chem., 1938, 112, 258.l* M. G. Raeder and 0. S. Kyllinstad, Mikrochem., 1939, 27, 112.lQ Analyst, 1936, 61, 603.ao W. M. MacNiven and R. A. Bournique, Ind. Eng. Chem. (Anal.), 1940,21 A. Lassieur, " Electroanalyse Rapide," Paris, 1927, p. 108.22 Analyst, 1929, 54, 654.23 P. H. Delano and W. T. Schrenk; T. G. Day and W . T. Schrenk, School24 S. L. Jovanovitch, 2. anal. Chern., 1938, 114, 415.12, 431.Mines Met. Univ. Missouri Bull., 1935, Tech. Ser. No. 2, 7, 31474 ANALYTICAL CHEMISTRY.antimony sulphate is precipitated in the solution.The methodmay be applied to the analysis of antimony ores.The frequently encountered difficulty of removing the residualquantities, which may amount to several mg., of tin in the elec-trolytic deposition of the metal from chloride solution has variouslybeen attributed to the formation of the gaseous hydride, stannane,a t the anode, to the volatilisation of stannic chloride during thepreparation of the solution, to the redissolution of some of thedeposits during washing, and to mechanical loss due to the poorquality of the deposit. F. G. Kny-Jones, A. J. Lindsey, and A. C.Penney 25 have examined this problem, and from earlier work 26g 27they deduced that loss of tin as stannane does not occur, whilsttheir own experiments show that the addition of ammoniumchloride to prevent the volatilisation of stannic chloride is dis-advantageous, as it promotes redissolution of the deposit.Thecoating of the platinum cathode with copper diminished dissolutionof the tin on washing, and the use of a low current density ensureda good adherent deposit. The controlled cathode potentialtechnique was used; hydroxylamine was added t o prevent theanodic evolution of chlorine.Arsenic may be determined by the electrolytic Reinsch test,but direct deposition of the metal is not easy, since the concen-tration of arsenic ions in solutions of its compounds is low, andany liberation of hydrogen results in the formation of arsine.S. Torrance 28 has examined this problem and has found that thequantitative deposition requires the presence of chloride ions, thesimultaneous deposition of copper, and that the arsenic should bein the arsenious form.I n practice, these conditions were satisfiedby preliminary reduction of the hydrochloric acid solution bysulphurous acid, the addition of copper (about five times the amountof arsenic present), and electrolysis, an auxiliary electrode beingused to enable the cathode potential to be controlled. Copperarsenide was deposited quantitatively.The deposition of bismuth from chloride solutions has long pre-sented difficulties, in that spongy deposits were liable to be formed,accompanied by a sharp rise in the auxiliary potential at an earlystage of the electrolysis, possibly due to the formation at the cathodeof bismuth oxychloride. P.G. Kny-Jones 29 has found that theaddition of oxalic acid prevents the formation of basic salts, and25 AnaZyst, 1940, 65, 498.36 F. Paneth and E. Rabinovitch, Bey., 1924, 57, B, 1877.a7 A. Schleicher and L. Toussaint, “ Electroanalytische Schnellmethoden.”2 * Analyst, 1938, 63, 104.Stuttgart, 1926, p. 191.*@ Ibid., 1939, 64, 172GRIFFITHS, MACLENNAN, AND WHBLLEY. 475electrolysis a t 80--85", using either a saturated calomel electrodeor D. J. Brown's auxiliary electrodem (i.e., a platinum wire onwhich a small amount of bismuth is first deposited) to control thecathode potential, gave satisfactory results. The method may beapplied to the determination of bismuth in the presence of leadand tin, and hence to the analysis of ternary alloys of these threemetals.The same worker has also found31 that good depositionmay be made from sulphuric and nitric acid solutions by conductingthe electrolysis at a high temperature from strongly acid solution,the controlled potential technique being used with hydrazinesulphate as depolariser.The determination of cobalt in ores has been described.32 Thecobalt is deposited with the nickel (since their deposition potentialsare very close), and subsequent separation of the two metals ismade chemically. Similarly, in .the analysis of nickel bronzes,S. Torrance 33 obtained the cobalt codeposited with the nickel.Recently, however, Torrance 34 has developed a method for separat-ing cobalt from nickel by deposition as cobaltic oxide at 90-95"on the anode, using a diaphragm electrode, as previously described,to separate the cathode from the cobalt solution.In the earliermethod of A. Coehn and D. Gl%~er,3~ deposition of cobalt on thecathode was prevented by the addition of potassium dichromatea8 a depolariser, and careful adjustment of the electrical conditions,but the deposition of cobaltic oxide was slow.The ability of any technique to play a substantial part in theanalysis of a complex material is a fair indication of its value, andconvincing evidence of the utility of electrolytic methods is to befound in their application to the analysis of white brass,36nickel br0nzes,3~ and aluminium alloys.33Electrolytic Marsh 17ests.-The electrolytic reduction of arsenicand the subsequent decomposition by heat of the arsine has formedthe basis of a number of methods for the determination of thiselement, and that of P.S. A~monier,~' in which a mercury cathodeis used, has long been accepted as a standard procedure. A rapidcontinuous method has recently been described by H. C. Lock-The arsenical solution flows slowly over a cadmiumcathode, and details of drying the gas and the deposition ofarsenic are given. The cathodic reduction of arsenic has also been30 J . Arner. Chem. SOC., 1926, 48, 582.31 Analyst, 1939, 84, 575.sa M. M. Fine, U.S. Bur Mines, 1938, Rept. Invest. 3370, 69.33 Analyst, 1938, 63, 488.35 2. anorg. Chem., 1903, 33, 9.3 e S. Torrance, Analyst, 1937, 62, 719.37 J. SOC. Chem.Ind., 1927, 46, 341.84 Ibid., 1939, 64, 109.38 Analyst, 1939, 64, 657476 ANALYTICAL CHEMISTRY.investigated by L. Cambi and G. G. Mon~elise,~~ who report that inzinc sulphate solution reduction is inhibited if pyridine is presenttogether with magnesium or aluminium salts (separately they havenegligible effect) owing to the formation of a cathode film of pyridineand basic salts.It is known that germanium is more readily reduced to mono-germane in alkaline than in acid solution, but it was not until 1934that the electrolytic reduction in alkaline solution was attempted.As arsenates are not easily reduced in this medium, S. A. Croase’smethod 40 (which is based on J. Grant’s determination of antimonyas stibine 41) enables germanium to be determined in the presenceof large amounts of arsenic.It was found that nickel was the mostsatisfactory cathode metal of a number (nickel, cobalt, stainlesssteel, lead, magnesium, iron, and copper) examined. Optimumvalues of current density and alkali concentration were determined.Internal Electrolysis.-Though C. Ullgren 4 2 in 1868 plated smallquantities of copper from sulphate solutions which formed theelectrolytes of Zn-Pt, Cd-Pt, and A1-Pt cells, no further develop-ment of this attractive method, which requires no external sourceof current, was made for over 60 years. H. J. S. Sand43 and hisco-workers evolved and used a technique suited to the rapid deter-mination of small quantities of metals. A platinum-gauze cylinderas the cathode, and a pair of anodes of baser metal (e.g., zinc orlead), enclosed in parchment bags and straddling the cathode andexternally connected to it, make a typical set up. The solutionto be analysed is the catholyte, and its composition is so adjusted,e.g., by the addition of reducing agents, that redissolution of thedeposited metal due to aerial oxidation is prevented. The anolyteis a solution of a salt of the anode metal of a higher concentrationthan that in the catholyte so that deposition of the anode metalmay not occur.E. A.Collin 44 has determined bismuth and copper in lead bullion, bismuthin lead ores, and cadmium and copper in sulphate solutions ofspelter and zinc ores. J. G. Fife 45 preferred the use of amminechlorides to sulphates for the separation of small amounts of cadmiumfrom zinc, and the same author has determined nickel in zinc,46 silverin galena and in pyrites,47 mercury (0.7-7 mg.) in andRend. Ist. Lomb. Sci. Lett., 1936, [ii], 69, 392; Chem. Zentr., 1936, 11,3643.The method has been found useful in a number of cases.40 Analyst, 1934, 59, 462, 747.4 2 8. anal. Chern., 1868, 7 , 442.4 4 Ibid., pp. 312, 495, 680.4 6 Ibid., p. 683.4B Ibid., 1938, 63, 650.41 Ibid., 1928, 53, 626.4 3 Analyst, 1930, 55, 309.4 5 Ibid., 1936, 61, 681.4 7 Ibid., 1937, 62, 723GRIFFITHS, MACLENNAN, AND WHALLEY. 477copper in the presence of large quantities of cadmium 53 by theinternal electrolytic technique. Reference has already been madeto the difficulties in the electrolytic determination of small amountsof copper in the presence of large amounts of iron : the separationmay be satisfactorily made by internal electrolysi~.~~J. L. Lurie and his co-workers49 have found that internal elec-trolytic processes may be conducted without the use of a diaphragmor stirrer. They emphasise the risk of deposition of the metal onthe anode (i,e., " cementation "), resulting in a fall of the currentand subsequent cessation of deposition on the cathode. Cement-ation on the anode is favoured by a large potential difference :when a nickel anode was used for the deposition of copper, cement-ation occurred, but deposition on the cathode was complete when alead anode was employed. Technical modifications and pro-cedures for the determination of copper in ores,51 cadmium andcopper in zinc alloys,52 and tin in aluminium alloys 52 have alsobeen described.J. G. A. GRIFFITHS.G. W. G. MACLENNAN.H. K. WHALLEY.49 J. L. Lurie and L. B. Ginsberg, Ind. Eng. Chem. (Anal.), 1937, 9, 424;J. L. Lurie and M. I. Troitzkaia, Ann. Chim. anal., 1938, 20, 61 ; Zavod. Lab.,1937, 6, 33.B. L. Clarke, L. A. Wooten, C. L. Luke, Ind. Eng. Chem. (Anal.), 1936,8, 411.51 N. N. Eberg, Zavod. Lab., 1938, 7, 239.52 B. L. Clarke and L. A. Wooten, Trans. Electrochem. SOC., 1939, 76, 339.53 Analyst, 1940, 65, 562

ISSN:0365-6217    

DOI:10.1039/AR9403700451

出版商:RSC    

年代:1940

数据来源: RSC

摘要:

INDEX O F AUTHORS' NAMES.ABADIE, P., 56.Abbasy, M. A., 384.Abbott, 0. D., 384.Abeles, A., 369.Abelson, P. H., 11.Abramowitz, A. A., 401.Achmatowicz, O., 381.Achtermann, T., 354.Ackermann, P. G., 28.Adair, G. S., 410.Adam, N. K., 103.Adams, F. H., 230, 251.Adams, J. R., 333.Adams, R., 201, 222, 225, 230, 369,374, 375, 376.Addison, C. C., 103.Adler, E., 446.Aeschbacher, R., 342, 346.Agallidis, E., 250.Agliardi, N., 250.Ahmann, C. F., 384.Albert, A., 321, 466.Albert, C., 151.Alemian, S. A., 170, 473.Alexander, A. E., 414.Alexander, B., 434.Alexander, W. A., 94, 95, 272.AlexBeff, W., 107.Allen, A. O., 274.Allen, F. L., 267, 268.Allmand, A. J., 94.Allsopp, A., 426.Almy, G. M., 84, 97.Alrich, (Miss) F. R., 136.Altar, W., 59.Alvarez, L.W., 17, 21.Amdur, I., 36.Anderaon, E., 427.Anderson, H. H., 126, 161.Anderson, J. S., 315.Anderson, L. C., 254.Anderson, P., 170.Anderson, R. J., 223, 224.Anderson, S., 84, 97.Anderson, T. F., 148.Ando, T, 359.Andrews, J. S., 331.Andrieux, J. L., 131.Andronikasckvili, E., 170.Angelescu, E., 107.Angel], s., 428.Anisimov, K. N., 130.Ant-Wuorinen, J., 128.Appleyard, (Miss) M. E. S., 88.Archibald, W. J., 31.Arcus, C. L., 391.Ardenne, M. von, 413.Arnold, A., 389, 390, 433.Arrhenius, S., 51.Asahina, Y., 373, 374.Ashburn, L. L., 390.Ashford, W. R., 465.Ashley, S. E. Q., 463.Ashton, R., 218.Askew, F. A., 347, 414.Astbury, W. T., 422.Aston, J. G., 177.Atanasiu, I. A., 128.Atkins, B.E., 34, 154.Audrieth, L. F., 66, 137.Auerbach, M. E., 465.Aumonier, F. S., 475.Aumiiller, W., 340, 342, 343.Avery, W. H., 177.Axelrod, A. E., 388.Aykroyd, W. R., 388.Ayre, C. A., 220.Ayres, G. H., 462.Babcock, S. H., 389, 390.Baber, T. D. H., 28.Babor, M., 128.Bacharach, A. L., 394.Bachman, C., 402.Bachmann, G. B., 455.Bachmann, W. E., 259, 261, 263,283, 285, 287, 332.Bacon, F., 258.Baddeley, G., 77.Badger, R. M., 46, 177.Bagchi, P. N., 221.Bahl, R. K., 128.Bailey, C. R., 177.Bailey, W. A., 331.Bak, B., 65.Baker, B. R., 374, 375.Baker, J. W., 247, 248, 447.Baker, W., 370, 371.Baker, W. O., 57, 172.47480 INDEX OF AUTHORS’ NAMES.Bakken, R., 135.Balakhovski, S. D., 308.Ball, C. D., 361.Ballard, K.H., 136.Ballou, G. A., 423.Balls, A. K., 419, 429.Baltes, J., 214.Bamford, C. H., 85, 86, 273, 274,276, 277.Bamman, F., 401.Bandier, E., 465.Bandow, F. 325.Banerji, G. G., 387.Banga, I., 385.Banks, T. E., 10.Barbot, A., 213.Bardeen, J., 35, 156.Bargellini, G., 370.Barger, G., 377.Barker, G. W., 473.Barker, H. A., 448.Barrett, P. A., 188, 313, 314, 316,317, 367, 368.Barrow, R. F., 275.Barry, V. C., 425.Barschall, H. H., 8.Bartell, F. E., 102.Bartlett, P. D., 237, 238, 242.Bartz, Q. R., 403.Bassett, J., 171.Bastick, R. E., 34, 154.Bateman, L. C., 237, 238, 239, 240,Bates, R. W., 403.Batty, J. W., 298.Bauer, K., 321.Bauer, S. H., 46, 62, 144, 149, 150.Baumann, C. A., 385.Baumgarten, P., 127.Bausor, S.C., 433.Bawden, F. C., 412.Bawn, C. E. H., 271, 274.Baxter, J. G., 459.Beach, J. Y., 38, 44, 46, 59, 64, 66,70, 148, 176, 177.Beckman, A. O., 90.Beer, R., 220.Bell, F. O., 422.Bell, R. K., 472.Bell, R. P., 25, 49, 52.Beller, A., 369.Bellinghen, R. van, 211.Bembry, T. H., 375.Benedict, W. S., 91.Bennett, G. M., 77, 211.Bent, H. E., 252.Berm, C. A., 168.Berenblum, L., 397.BBres, T., 308.Berg, M. A., 227.Berg, W. F., 171.Bergmann, A. J., 403.Bergmann, E., 82.246.Bergmann, W., 363.Bergren, W., 433.Bergstrom, F. W., 137.Berkman, S., 439.Bernal, J. D., 105, 412.Bernstein, H. J., 65.Bernstein, S., 362.Besson, J. A., 145.Best, C. H., 404.Beste, G. W., 242.Beth, E., 30, 31.Beyer, H. G., 22.Beynon, J.H., 363.Biggs, B. S., 57.Bijvoet, J. M., 186, 187.Billings, 0. B., 130, 186.Billingsley, W. W., 404.Biltz, W., 131, 135, 184.Biniecki, S., 376.Binkley, S. B., 392.Binzer, I., 219.Birch, A. J., 225.Birch, S. F., 204.Birch, T. W., 388, 392.Bircumshaw, L. L., 125.Birkhaug, K. M., 391.Birse, E. A., 89, 90.Bischoff, F., 401.Biscoe, J., 398.Bishop, W. S., 57, 177.Bistrzycki, A., 369.Bixler, M. E., 48.Blacet, F. E., 84, 273.Black, S., 390.Blaisdell, A. C., 461.Blanchard, A. A., 129.Bleick, W. E., 26.Blicke, F. F., 257, 369.Bloch, F., 21.Bloom, G. I. M., 48.Blum-Bergmann, O., 287.Boas, M. A., 392.Bock, H., 323.Bodenstein, M., 93, 94.Bodnar, J., 420.Boeder, P., 115, 412.Baggild, J. I<., 8.Boehm, G., 412.Borner, E., 337.Boese, A.B., jun., 208.Bottcher, C. J. F., 50.Bogert, M. T., 290.Bohr, N., 8, 10.Boivin, A., 394.Bolland, J. L., 91.B o m e r , H., 128, 184.Bonner, D. M., 431.Bonner, J., 222, 433.Bonner, W. D., 461.Bonsnes, R. W., 402.Booher, J. E., 96.Booher, L. E., 383.Booth, E. T., 9, 11INDEX OF AUTHORS’ NAMES. 481Booth, H. S., 150, 151, 152, 153.Booth, R. G., 467.Booth, V. H., 447.Borchert, L. H., 169.Borders, A. M., 233.Borelius, G., 170.Borgne, E. L., 461.Borisova, L. M., 130.Born, M., 28, 33, 174.Bose, P. K., 370.Bosworth, A. W., 215.Bottomley, A. C., 404.Bourdillon, J., 407.Bourne, G., 394.Bournique, R. A., 473.Bowat, A., 461.Bowden, S. T., 253, 254, 256, 257,Bowen, E.L., 37.Bowman, K. L., 400.Bowman, K. M., 389.Boyd, R. N., 32, 33.Bozarth, A. R., 152.Bradbrook, E. F., 313, 315.Bradley, A. J., 170.Brady, S. E., 213.Bragg, L. B., 455.Bragg, (Sir) W., 168.Bramley, A., 155, 162.Brandenburger, H., 364.Braschos, A., 153.Brauer, G., 133.Braun, J., 104.Brauns, M., 266.Braunwarth, V., 96.Eraybrooke, F. H., 200.Brazer, J. G., 384.Brenken, B., 369.Brenschede, W., 93, 94, 95.Bresler, S., 175.Breslow, D. S., 195.Brewer, A. K., 154, 155.Breywisch, W., 360.Bridjpnan, P. W., 173.Bridgman, W. B., 172.Briggs, L. H., 378.Briggs, T. R., 136.Brillouin, L., 169, 174.Brockmann, H., 296, 297, 300, 309.Brockaay, L. O., 36, 37, 39, 40, 46,Brody, E., 174.Broeker, C. E., 157.Brsnsted, J.N., 123, 458.Broh-Kahn, R. H., 442.Brooker, L. G. S., 365.Brookman, E. F., 289.Brooks, B. T., 205.Brooks, D. B., 201.Brostr0m, K. J., 8.Brown, C. J., 185.Brown, D. J., 475.Brown, H., 35, 156, 162.259, 269.68, 70, 74, 76, 77, 185.REP.-VOL. XXXVIT.Brown, H. C., 68, 140, 147, 148, 210,Brown, J. B., 211, 212, 213, 215, 392.Brown. J. H., 258.288, 289.Brown; N. C.; 370.Brown, W. B., 215.Brdck, H., 26.Briingger, H., 349.Brugsch, J., 325, 466.Bruns, W., 127.Bruun, J. H., 453, 456, 458.Buch-Anderson, E., 15.Buchanan, J., 425.Buchholz, K., 362, 366.Buckingham, R. A., 26,27, 29, 3Buckley, H. E., 171.Bud6, A., 56.Buchner, E., 280.Bull, H. B., 411.Bunge, W., 263, 264, 267.Bunker, J. M., 361.Bunning, E., 302.Burawov.A.. 73. 186. 298.36.35,Burch, e.’ R.; 458, 459.Burg, A. B., 139, 140, 141, 142, 143,144, 146, 147, 148.Burger, M., 130.Burgers, J. M., 123, 410, 411.Burk, D., 393.Burk, R. E., 455.Burnett, D., 33.Burr, G. O., 212, 213, 392.Burrows, G. J., 132.Burton, M., 80, 82, 84, 85, 87, 91, 271,Burwell, J. T., 46.Buston, H. W., 429.Butenandt, A., 333, 339, 340, 341,Butler, G. C., 397, 398.Butler, J. A. V., 168.Butler, R. E., 389.Buttenvorth, E. C., 281, 282.272, 273.344, 345, 352, 355, 358, 363.Cain, C. K., 267, 374.Caldwell, C. C., 66.Calingaert, G., 201.Callison, E. C., 383.Callow, A. E., 96.Cambi, L., 476.Campbell, D. A., 395.Campbell, (Miss) I. G. M., 71.Campbell, J., 404.Campbell, W. G., 428.Cannan, R.K., 405.Carius, G., 417.Carson, S. C., 177.Carter, G. P., 291.Carter, M. G. R., 211.Carter, P. W., 304.Cartwright, C. H., 53.482 @IDEX OF AUTHORS’ NAMES.Cassidy, M. C., 152.Catchpole, H. R., 403.Caucbois, Y., 16.Cavell, H. J., 73.Chabovski, G. L., 126.Challinor, S. W., 394.Chang, K. J., 378.Chao, T. H., 288.Chaplin, H. O., 233.Chapman, S., 33, 34, 154.Chargaff, E., 223, 224.Cheesman, G. H., 136.Chen, G., 401.Chen, H., 378.men, M. C., 214.C h h e , M., 131.Chibnall, A. C., 217, 218, 219, 220.Chick, F., 208.Chick, H., 389, 394.Chiong, Y. S., 169.Cholnoky, L. von, 290, 292, 299, 301,Chow, B. F., 401, 402.Christian, W., 419.Chuang, C. K., 378.Chuit, P., 222.Church, M. G., 238, 239.Clack, K.D. G., 136.Claes, A,, 34.Clarenz, L., 219.Clark, D., 185.Clark, G., 175.Clark, J. H., 374.Clark, R. H., 465.Clark, W. M., 314, 413.Clarke, B. L., 477.Clarke, D. L., 256, 269.Clarke, F. H., 398.Clarke, H. T., 287.Clarke, L., 201.Clayton, W., 102.Cleave, A. B., 34.Cleclrley, H. M., 389.Clemo, G. R., 379.Clews, C. J. B., 178.Clifton, C. E., 446, 448.Cline, J. E., 92, 252.Clusius, K., 35, 155, 157, 163.Clutterbuck, P. W., 215.Cobler, H., 341.Cockbain, E. G., 109.Coehn, A., 475.Coghill, R. D., 398.Cohen, K., 26.Cohen, S., 437.Cohn, E. J., 402.Cole, R. H., 50.Cole, W., 332.Coleman, G. H., 130.Collin, E. A., 476.Collin, E. M., 473.Collins, G. B., 17.Collins, S. C., 29.464.Collins, V. J., 342.Collip, J.B., 402, 404.Colowick, S. P., 418.Compagnon, P., 369.Condon, E. U., 24.Conn, J. B., 177.Connerade, E., 258, 266.Conrad, M., 320.Cook, A. H., 315.Cook, C. A., 403.Cook, W. R., 34.Cooley, R. A., 20.Coomber, D. I., 69.Coombs, (Miss) E., 195.Coop, I. E., 48, 49, 52, W, 69, 70, 72,Cooper, E. A., 425.Cooper, G. R., 411.Cooper, K. A., 239.Coops, J., jun., 211, 219, 220.Copenhaver, J. W., 254.Corbet, A. S., 445.Corey, R. A., 77.Cori, C. F., 418, 420.Cori, G. T., 420.Cork, J. M., 19.Corner, J., 29, 30, 31, 33.Corruccini, R. J., 178.Corson, D. R., 14.Cortis-Jones, B., 325.Corwin, A. H., 331, 332.Coryell, C. D., 314.Cossel, H. von, 376.Cosslett, V. E., 42.Cottini, G. B., 389.Coulson, C. A., 33, 49.Cowley, E.G., 50.Cowling, T. G., 33.Cox, E. G., 182, 185.Crafton, H. C., 201.Craggs, H. C., 94.Craig, L. C., 379, 457.Cramer, P. L., 202, 277.Crawford, B. L., jun., 177.Crawford, M. F., 24.Criegee, R., 336, 357.Zroase, S. A., 476.Zrooks, H. M., 352.Zross, P. C., 45.2rowfoot, (Miss) D. M., 218.=ryder, D. S., 543.hllinane, N. M., 367.2unninaham. J. P.. 86.75.>unniniharn; R. L:, 92.h d . F. H., 373.2urtis, A. C., 384.hthbertson, G. R., 252.Deft, F. S., 390.Dalma, G., 380.Dalton, H. R., 419.Daly, E. F., 177INDEX OB AUTHORS’ N ~ S . 483Dammerau, I., 267.Dane, (FrE.) E., 375.Daniel, D., 369.Danielli, J. P., 400, 414.Danielli, J. M., 400.Daniels, F., 92.Danilov, V., 170.Dannenberg, H., 333.Darke, W.F., 111.Dart, F., 175.Datta, S. C., 230.Davidson, N. It., 61.Davies, A. W., 385.Davies, G., 233.Davies, T. H., 314, 325, 413.Davis, B. L., 428.Davis, E., 392.Davis, T. W., 85, 272, 273.Dawson, E. R., 394.Day, H. G., 387.Day, J. N. E., 230.Day, T. G., 473.Dean, H. K., 102.De Beer, E. J., 403.De Boer, J., 29, 31.De Bruyn, P., 30.De Bruyne, J. M. A., 52.Debye, P., 38, 42, 44, 48, 50, 56, 163,Degard, C., 40, 44.Degering, E. F., 206, 207, 275.De Lange, J. J., 178.Delano, P. H., 473.Delbpine, M., 135.Deming, L. S., 72.Dennison, D. M., 172, 177.Denny, F. E., 434.Denstedt, 0. F., 402.Dent, C. E., 313, 316.Deppe, M., 359.Derjaguin, B., 101.Derksen, J. C., 422.Dervichian, D. G., 105, 171.Desai, D. M., 118.Deschamps, P., 472.Desnuelle, P., 446.Desreux, V., 419.Detrick, L.E., 392.Detwiler, S. B., jun., 459.Deutsch, J. V., 448.De Vault, D. C., 19.Devaux, H., 414.Devonshire, A. F., 31, 173.De Vries, T. A., 185.De Waal, H. L., 376.Dewar, D. J., 272.Dickel, G., 35, 155, 157, 163.Dickinson, R., 366.Dickinson, R. G., 86.Diebold, W., 218.Diederichsen, J., 284.Dieke, G. H., 82.Dierichs, H., 370.169.Dijck, W. J. D. van, 459.Dijk, H. van, 456.Dillon, T., 425.Dilthey, W., 370.Dimroth, K., 293, 334, 336, 360.Dingemanse, E., 403.Dingle, J. H., 400.Dippy, J. F. J., 234.Discherl, W., 362.Dixon, M., 388, 415.Dod6, M., 20.Dodson, R. W., 9.Doescher, R. N., 169, 177, 268.Doisy, E. A., 332, 392.Dolby, D. E., 215, 422.Dolidge, N., 119.Donaldson, R., 466.Dootson, F.W., 154.Dorfman, A., 439.Dorfman, M., 252.Dorp, D. A. van, 387.Dothie, H. J., 181.Doughty, J. F., 389.Downing, J. R., 312.Draper, H. D., 109.Drem, H. P., 201.Drigalski, W. von, 383.Drikos, G., 251.D m m , P. J., 312.Drummond, J. C., 393, 394.Dube, G. P., 101.Dubourg, J., 171.Du Bridge, L. A., 20.Dufraisse, C., 369.Duhamet, L., 432.Duncan, D. R., 136.Dunn, J. L., 356.Dunn, M. S., 392.Dunn, W. J., 388.Dunning, J. R., 11.Dunstan, A. E., 200,204.Dutcher, J. D., 362.Dutt, P., 370.Dutton, F. B., 152.Du Vigneaud, V., 394.Dybowski, C., 381.Dye, W. B., 207.Eakin, R. K., 433, 434.Eberg, N. N., 477.Eberhardt, R., 225.Eddy, C. R., 50.Edgar, G., 201.Edlund, K.R., 205.Edsall, J. T., 116, 412.Eggerth, A. H., 448.Eggleston, L. V., 385.Egle, K., 304.Egloff, G., 201.Ehrenstein, M., 342, 358.Ehrhart, G., 340, 342, 343.Ehrmann, K., 152484 INDEX OF A1Eichenberger, E., 349.Einstein, A., 11 1.Eirich, F., 124.Eisenstein, A., 33.Elbe, G. von, 457.Elkes, J. J., 414.Elliott, G. A., 247.Elliott, N., 74.Ellis, B., 357.Ellis, C., 200.Ells, V. R., 85.El Ridi, M. S., 291.El-Sadr, M. M., 389.Elsom, K. O., 387.Elvehjem, C. A., 226, 385, 387, 388,Embree, N. D., 459.Emelbus, H. J., 38, 151.Emerson, R., 308.Emster, K. van, 280.Ende, M. van den, 408.Endermann, F., 317.Endres, G., 444.Engel, R. W., 390.English, J., jun., 222, 433.Enskog, D., 33, 153.Epifanski, P.F., 355.Erdey-Grhz, T., 171.Erikson-Quensel, I. B., 405.Erkama, J., 447.Erlenmeyer, H., 284, 364.Ernst, P., 331.Errera, J., 53.Erschow, A., 26.Estermann, I., 48.Etienne, A. D., 461.Etzler, D. H., 85.Eucken, A., 57, 171, 172, 177.Euler, H. von, 294, 295, 297, 298,302, 386, 387, 446.Euw, J. von, 336, 343.Evans, D. P., 195, 233, 234, 235.Evans, E. A., 385.Evans, G. R., 20.Evans, H. M., 401, 402, 403, 404.Evans, R. C., 131, 186.Evans, T. W., 205.Evans, W. C., 437.Evering, B. L., 270.Evers, W. L., 201.Ewell, R. B., 30.Eyring, H., 177.Eyster, E. H., 46.Ezoe, H., 10, 12.390.Fairbrother, F., 51.Fairclough, R. A., 234.Fajans, K., 20.Paltis, F., 380.Fankuchen, I., 412.Farinacci, N. T., 239.Farkas, A., 99.HORS’ NAMES.Farkas, L., 89, 99.Farmer, E.H., 214, 215, 216.Farmer, S. N., 347.Faulconer, W. B. M., 456.Faxen, H., 30.Feeney, R. E., 389.Feher, F., 125.Fein, H. D., 389.Feit, W., 128.Feldman, J. B., 384.Fell, N., 398.Fenske, M. R., 453, 454.Ferguson, A. L., 68.Fermi, E., 10, 24.Fernholz, E., 361, 362.Ferns, J., 229.Ferry, J. D., 56.Fevold, H. L., 401, 402.Fichter, F., 283.Fidler, F. A., 204.Field, H., 387.Field, H., jun., 465.Fieser, L. F., 348.Fife, J. G., 471, 476.Fildes, P., 435, 436, 440, 441.Filippov, A. N., 466.Fine, J., 457.Fine, M. M., 475.Finkelstein, J., 228, 389, 437.Finkle, P., 109.Finn, A. E., 60.Firmenich, G., 339.Fischer, E., 56, 169, 179, 217.Fischer, H., 312, 317, 318, 319, 320,321, 322, 323, 324, 325, 326, 327,328, 329, 330, 331, 332.Fischer, W.H., 332, 354, 363.Fleischer, G., 341, 344.Fleischer, M., 137.Fleischmann, R., 160.Flerov, 10.Flodin, A. W., 141.Flodin, N. W., 146.Fock, V., 25.Folkers, K., 228, 389.Folkins, H. O., 270.Folley, S. J., 403, 404.Foote, H. W., 137.Forbes, G. S., 92, 96, 126.Fordyce, C. R., 223.Fortey, (Miss) E. C., 461.Foster, R. T., 373.Fourt, L., 395.Fowler, R. D., 9.Fowler, R. H., 29.Fox, D. L., 308, 311.Fraenkel-Conrat, H., 401, 402, 403.Fraenkel-Conrat, J., 402.Frahm, H., 50.France, H., 265, 282.Francis, F., 211, 217.Franck, J., 275.Frank, F., 169INDEX OF AUTHORS’ NAMES. 485Frank, F. C., 56, 57, 62, 170,Franke, W., 94, 95.Frankel, J., 212.Frankel, S.P., 153.Franz, H., 42.Franz, K., 82.Fraps, G. S., 463.Frary, S. G., 152.Fraser, H. F., 388.Fraser, R., 48.Fraser, R. G. J., 48.Frazer, A. C., 414.Fred, E. B., 444.Freeman, G. G., 394.Freiberg, J. K., 404.Frenkel, B. E., 332.Frenkel, J., 174, 175.Fresenius, N. F., 66.Freud, J., 403.Freundlich, H., 117, 119.Frevel, L. K., 185.Frey, F. E., 461.Fricke, H., 56.Fricke, H. H., 391.Fridenson, A., 332.Friedel, M. G., 105.Friedman, B. S., 201.Frohlich, P. K., 204.Fromageot, C., 446.Frost, D. V., 390.Fry, E. G., 337.Fry, E. M., 348.Frye, D. A., 314.Fuchs, K., 28.Fuchs, O., 48, 59.Furth, R., 169.Fujii, K., 332, 354.Fukushima, T., 373.Fuller, H. Q., 84.Furry, W. H., 35, 154, 156.Furth, R., 33, 174, 175.Gabler, R., 175.Gabriel, C.L., 207.Gatzi, K., 334, 336, 337, 345.Galachow, G., 26.Gale, E. F., 447, 449, 450.Gamble, E. L., 145, 151.Gardner, J. A., 359.Gardner, J. H., 258.Gardner, W. U., 342.Garner, W. E., 99, 172.Gassmann, A. G., 233.Gatzi-Fichter, M., 229.Gaunt, W. E., 397.Gee, G., 123.Geiger, A., 297.Geiger, H., 329.Geissman, T. A., 369.Gelissen, H., 282.Geloso, M., 472.Gettler, A. O., 457.175. Greyer, E., 379.Ghosh, R., 375.Gibson, C. S., 73, 186.Gilbert, E. C., 178.Gillam, A. E., 265, 291, 292, 298.Gillespie, L. J., 153.Gilliland, E. R., 203.Gilman, H., 285.Gilmont, P., 129, 145.Gingrich, N. S., 32, 33, 46.Ginsberg, E., 251, 252, 254.Ginsberg, L. B., 477.Ginsburg, S. R., 70, 182.Girard, A., 332.Girard, P., 56.Gladstone, 0.P., 440.Gkser, D., 475.Glasgow, A. R., jun., 455.Glasoe, G. N., 9.Glasstone, S., 36, 37, 51, 77, 470, 471.Glazebrook, H. H., 273.Glickman, S. A., 121.Glockler, G., 35, 160.Godden, W., 359.Goebel, H. L., 69.Goebel, W. F., 396.Goeppert-Mayer, M., 33.Gold, M. H., 369.Goldberg, L., 25.Goldberg, M. W., 332, 341, 346, 349.Goldberg, W., 285.Goldhaber, M., 13.Goldschmid, O., 124.Gombas, P., 25.Gomberg, M., 253, 254, 257, 259,Gommers, S. C., 397.Gooderham, W. J., 454.Goodeve, C. F., 98, 116, 124.Goodhart, R., 387, 388.Goodson, L. H., 350, 353, 354.Goodwin, R. H., 433.Goodyear, G. H., 226.Gordon, H., 417.Gordon, J. J., 207, 234.Gordy, W., 206.Gorin, E., 85, 87, 177.Gorrie, D.R., 392.Goss, F. R., 50.Gossner, B., 186.Got& H., 468.Grabar, P., 399.Grace, N. H., 431.Grace, N. S., 136.Graef, E., 370.Graham-Little, (Sir) E., 394.Grahame, D. C., 14, 83, 84, 273.Granick, S., 268.Grant, J., 469, 476.Grattan, J. F., 402, 404.Gray, E. LeB., 459.Gray, T. J., 270.Greaves, J. D., 404.283486 INDEX OF AUTRORS’ NAMES.Green, D. E., 415, 417, 447.Green, (Miss) N. D., 213.Greenberg, D. M., 429.Greenberg, S. M., 463.Greep, R. O., 401, 402.Gregg, A. H., 77.Grieve, W. S. M., 279, 281.Grfith, R. O., 96.Grinten, W. van der, 40, 156.Groll, H. P. A., 205.Gropper, L., 30, 31.Grosse, A. V., 9, 11, 203.Grossman, A. J., 66, 70, 182.Groth, W., 96, 161.Groves, L. G., 48, 53, 54, 55, 248.Grube, G., 133, 136.Griin, A., 220.Griissner, A., 229.Grundmann, C., 172, 295, 297, 299,Grunow, H., 380.Guareschi, I., 226.Gtinther, E., 363.Gimther, G., 297, 298, 446.Gtintzel, B., 359.Gugelmamn, W., 294, 309.Guggenheim, E.A., 29.Guggisberg, H., 412.Guilbert, H. R., 383.Guinier, A., 167.Gundermann, J., 115.Gurin, S., 402.Gurney, R. W., 168.Guth, E., 115, 175.Guy, J. B., 211.Gyorgy, P., 392, 393.307.Haagen-Smit, A. J., 222, 433.Haantjes, J., 456.Haas, P., 425.Haberland, H. W., 328.Hackenberg, E. G., 126.Haefele, W. R., 209.Hagedorn, A., 347, 348.Hagiwara, T., 11.Haig, C., 384.Haist, R. E., 404.Halbach, H., 329.Halford, R. S., 19, 169, 237.Hall, )I. J., 455.Hamaker, H. C., 101, 117, 170.Hamburger, L., 170.Hamer, (Miss) F.M., 366.Hamilton, J. G., 15.Hammer, C., 170.Hammett, L. P., 239, 242.Hammick, D. L., 51, 60, 96.Hampson, G. C., 54, 60, 66, 71, 72,Hanby, W. E., 280.Hand, D. B., 468.Handley, W. R. C., 437.73, 77, 186.Hanes, C. S., 418, 419, 421, 422.Hannum, C., 209.Hansen, F., 399.Happold, F. C., 437, 447.Haraldsen, H., 135.Harington, C. R., 397, 398.Harker, D., 191.Harkins, W. D., 157, 206.Harman, R. A., 235.Harms, H., 50, 68.Harmsen, H., 187.Harper, S. H., 298.Harris, I., 125.Harris, I. W. H., 136.Harris, L. J‘., 384, 387, 388, 390, 393,Harris, R. S., 361.Harris, S. A., 228, 389.Harris, W. E., 254, 259, 269.Harrison, G. E., 34.Harrison, S. F., 28.Hart, G. H., 383.Hart, M. C., 361.Harteck, P., 89, 161.Hartel, H.von, 274.Hartley, G. S., 71, 103, 104, 106, 109.Hartman, R. J., 233, 455.Hartree, D. R., 25.Hartree, E. F., 416.Hartree, W., 25.Hartsuch, P. J., 212.Harvey, E. N., 414.Hasegawa, M., 371.Hass, H. B., 205, 207, 208, 275.Hass6, H. R., 28, 34.Hattori, J., 354.Hattori, S., 371.Haurowitz, F., 313.Hauschild, K., 367.Hausen, S. von, 444.Hauser, C. R., 195, 249.Hauser, E. A., 118, 121.Hausmann, E., 333, 355.Havens, R., 49.Havorka, I?., 206.Hawkins, E. G. E., 423.Haworth, E., 298.Kaworth, J. W., 282.Haworth, R. D., 212.Haxby, R. O., 10.Haxel, O., 8.Healy, T. V., 373.Hearne, G., 205.Heath, H. R., 34.Heath-Brown, B., 361.Kechenbleikner, I., 286.Heidelberger, H., 398, 400.Keilbron, I.M., 279, 282, 297, 298,Heilmeyer, L., 329.Heimbrecht, M., 131.Hein, F., 271.Keitler, W., 25.465.304, 306, 356, 361, 363, 366INDEX OF AUTHORS’ NAMES. 487Hektoen, L., 396.Helberger, H., 319.Helberger, J. H., 313, 315, 316.Helferich, B., 363.Hellmann, H., 24, 25, 27, 28.Hellstrom, H., 302, 315, 446.Hellund, E. J., 35.Helmholz, A. C., 17.Helmholz, L., 181, 185.Helms, A., 184.Hemmendinger, A., 20.Hendricks, S. B., 46, 72.Hennaut-Roland, (Mme.), 201.Hennig, H., 127.Hennion, G. F., 344.Henriques, F. C., jun., 81, 84.Henry, L., 208.Herbert, D., 417.Hercik, F., 398.Hermann, K., 105.Hermans, P. H., 382.Herr, D. S., 273.Herriot, R. M., 419.Herrle, K., 326, 330.Herrmann, C. V., 152.Herzberg, G., 46, 80, 82.Herzfeld, K., 29, 33.Herzfeld, K.F., 53, 84, 273.Hess, K., 104, 115.Heuser, G. F., 389, 464.Heusner, A., 345.Heuvel, F. A. van den, 214, 216.Hevbr, D. B., 315, 316.Hevesy, G. von, 20, 458.Hevesy, G. C., 392.Hewitt, L. F., 407, 408.Hey, D. H., 265, 277, 279, 281, 282,283, 284, 285, 309.Heyl, F. W., 361.Heymann, E., 121.Heyn, F. A., 20.Hibbart, R. P., 463.Hibshman, H. J., 207.Hiby, J. W., 163.Hickinbottom, W. J., 284.Hickman, K. C. D., 458, 459.Hicks-Bruun, M. M., 456, 458.Higasi, K., 50, 72, 211.Hildebrand, J. H., 32, 109.Hildebrand, W., 336, 343.Hillel, R., 372.Hills, G. M., 443, 445, 450.Himel, C. M., 255, 258.Hinshelwood, C. N., 232, 233, 234,Hinton, C. L., 424.Hirschbold-Wittner, (Frau) F., 156,Hirschfelder, J.O., 30.Hirschmann, F., 363.Hirschmann, H., 332.Hirschmiiller, IT., 462.Hirst, A. A., 34.235, 272.157.Hirst, E. L., 423.Hisaw, F. L., 401, 402.Hitchcock, A. E., 431, 432.Hoard, J. L., 46, 184, 186.Hobbs, M. E., 29, 50.Hodge, E. B., 207, 275.Hofelmann, H., 332.Hogberg, B., 386, 387.Hoehn, W. M., 337, 340.Holtje, R., 131.Honigschmid, O., 156, 157.Hoffman, W. A., 210.Hoffmann, U., 171.Hofmann, K., 334, 338, 342.Hohlweg, W., 338.Holaday, D., 326.Holley, C. E., 35.Holley, C. X., jun., 161.Holmberg, B., 230.Holmes, H. L., 380.Holst, G., 134.Holt, J., 14.Holtsmark, J., 30.Holzinger, L., 380.Hoog, H., 201, 203.Hoogerheide, J. C., 448.Hopkins, R. H., 422.Hopkins, S. J., 219, 397.Horn, E., 274.Horn, L., 170.Homer, L., 381.Hornhardt, H., 284.Horsley, L.H., 472.Horvitz, L., 141, 142.Howard, F. L., 201.Howard, J. B., 177.Howell, C. E., 383.Hoygaard, A., 391.Hsing, C. Y., 378.Hubbard, N. E., 392.Hubner, H., 300, 310.Hiickel, W., 286.Huffman, J. R., 456.Huggins, M. L., 115.Hughes, E. D., 230, 236, 237, 238,Hughes, E. L., 389.Hughes, E. W., 192.Hughes, J. V., 48.Hugill, J. A. C., 69.Hultgren, R., 183.Hulubei, H., 16.Hume, E. M., 212, 392.Hume-Rothery, W., 171.Hunt, M., 374.Hunter, L., 233.Hunter, R. F., 274.Hunziger, F., 345.Kurd, C. D., 210.Huse, G., 178.Husemann, E., 426.Hutchings, B. L., 440.Hutchinson, C. A., 160.239, 240, 246, 248, 249488 INDEX OF AUTHORS’ NAMES.Hutson, J. M., 96.Hwang, Y., 432.Hylleraas, E.A., 24.Ibbs, T. L., 34, 154.Ikawa, M., 10, 12.hndest, H., 283.Ingle, D. J., 337.Ingold, C. K., 209, 230, 236, 237, 238,Ingold, E. H., 230.Ingraham, M. A., 307.Inhoffen, H. H., 338.Ipatiev, V. N., 201, 202, 204.Iredale, T., 86.Itterbeek, A. van, 30, 34.239, 240, 246, 248, 249, 251.Jackson, J., 138.Jackson, W., 57.Jacob, (Miss) A., 374.Jacobs, L. J., 397.Jacobs, W. A., 379.Jacobsen, R. P., 333, 348, 363.Jacobson, L. E., 56.Jaff6, W., 295.Jahn, F. P., 272, 273.James, R. W., 43.Janke, A., 446.Janke, W., 268.Jenkelevitsch, Z. A., 126.Janneck, W. H., 96.Jantzen, E., 211.Jarkovaja, L. M., 431.Jenkins, F. A., 157.Jenkins, G. I., 60, 77, 96.Jenkins, H. O., 62, 232, 234, 235.Jennings, F.B., 66.Jensen, A. T., 170.Jemen, H., 25.Jensen, H. (Baltimore), 401, 402, 404.Jensen, K. A., 73.Jentschke, W., 8.Jerchel, D., 302.Jersild, M., 406, 409.Jewell, W., 459.Johns, H. E., 404.Johnson, A. W., 298.Johnson, J. R., 146, 223.Johnson, K., 207.Johnston, E. S., 463.Johnston, 8. A., 207.Johnstone, W. R., 85.Jolliffe, N., 389.Jonas, L., 387.Jones, B., 77, 211.Jones, E. M., 333, 350, 351, 352, 354.Jones, E. R. H., 356, 361.Jones, J. K. N., 423.Jones, P. L. F., 77.Jones, R. C., 35, 154, 156.Jones, R. N., 348.Jones, T. T., 97.Jones, W. E., 298.Joshi, S. S., 472.Josten, W., 370.Jovanovitch, S. L., 473.Jukes, T. H., 226, 389, 390.Juliusburger, F., 119.Jung, D. H., 455.Jungers, J. C., 86.Junusov, S., 377.Kabat, E.A., 398, 412.Kagi, H., 344, 345.Kahn, B., 28.Kahovec, L., 178.Kahr, K., 320, 321.Kainrath, P., 372.Kaischew, R., 170.Kalbfell, D. C., 20.Kalckar, H., 418.Kamen, M. D., 13, 20.Kamerer, E., 417.Kamuscher, H., 203.Kandelaki, B., 119.Kane, G., 33.Kanner, M. H., 8.Kanngiesser, W., 319.Kao, Y. S., 378.Kaplan, J. F., 255.Karagunis, G., 251.Kardos, R., 171.Karpinski, B., 461.Karrer, P., 294, 295, 296, 297, 298,300, 301, 302, 307, 308, 309, 310.Karweil, J., 66.Kass, J. P., 212, 213.Kast, W., 169, 175.Kathol, J., 334, 338.Katz, J. R., 175, 423.Katzin, B., 337.Kaufmann, H. P., 214.Kausche, G. A., 412, 413.Keeble, (Sir) F., 394.Keenan, H. C., 404.Keesom, P. H., 34.Keesom, W. H., 31, 34, 456.Keilin, D., 416, 419.Kekwick, R.A., 405, 407, 409.Kemmerer, A. R., 463.Kemp, J. D., 176.Kemper, W. A., 171.Kendall, E. C., 337.Kennedy, A. M., 177.Kennedy, J. W., 14, 19, 158, 159.Kennedy, T., 359.Keresztesy, J. C., 228, 389, 437.Kern, D. A., 203.Ketelaar, J. A. A., 38, 67, 185.Keutner, E., 56.Keyes, F. G., 29.Keys, A., 325, 466.Kharasch, M. S., 209,210,285,288,289INDEX OF AUTHORS’ NAMES. 489Kidd, F., 394.Kienitz, H., 136.Kiessig, H., 104, 115.Kiessling, W., 321.Kikodze, G., 119.Killian, D. B., 344.Kimball, G. E., 70, 165,Kimbo, M., 211.Kimura, K., 10, 12.Kimura, T., 340.King, A., 111.King, (Miss) A. M., 172.King, C. G., 391.King, H., 377.King, J. D., 389.King, L. C., 361.Kingdon, K.H., 11.Kinzer. G. D.. 84.79.Kiprianov, G.’ I., 332.Kirkwood, J. G., 27, 29, 31, 50, 177,Kirsch, W., 175.Kistiakowsky, G. B., 38, 82, 176,Kitchener, J. A., 98.Klages, G., 56, 169.Klarrer, W., 333.Klatt, W., 137.Klebsattel, C. A., 204.Kleczkowski, A. B., 397, 398.Klein, J. R., 388, 447.Klem, A., 454.Klemm, W., 129, 135, 184, 314.Klenk, E., 218, 219, 457.Kloetzel, M. C., 259.Knaggs, (Miss) I. E., 167.Knight, B. C. J. G., 439, 440, 441.Knipp, J. K., 26, 27.Knox, L. H., 237.Kny-Jones, F. G., 474.Kocholaty, W., 448.Kodicek, E., 388, 465.Kogl, F., 441.Kohler, F., 221.Koehler, J. S., 172, 177.Kohler, L., 221.Koenig, H., 296, 307.Koster, H., 354.Kohlhaas, R., 366.Kohlrausch, K. W. F., 178.Kohn, H. I., 388.KO-, E., 466.Kon, G.A. R., 347, 348, 349.Kon, S. K., 291.Konovalova, R. A., 377.Iiorsching, H., 163, 164.Kosack, H., 261.Koschara, W., 375.Koser, S. A., 439.Kotake, M., 381.Koton, M. M., 285.Kotscheschkov, K. A., 130.Koyanagi, H., 214.232.177, 456.Kraak, H. H., 31.Kraemer, A., 286.Krasny-Ergen, W., 35, 161.Kraus, O., 186.Krebs, H. A., 385.Kremers, H. C., 128.Krevelen, D. W. van, 202.Kringstad, H., 464.Krishnan, R. S., 10.Krohnke, F., 337.Kroon, D. B., 403.Kruck, W., 259, 260.Krueger, J., 333, 340, 351, 353, 358.Kruse, H. D., 389.Kruyt, H. R., 101.Krznarich, P. W., 428.Kubaschewski, O., 133.Kubo, M., 66.Kuck, J. A., 146, 268.Kudra, 0. K., 471.Kiichler, L., 92.Kiilkens, H., 153.Kiindig, W., 364.Kuhn, R., 172, 221, 293, 295, 296,297, 299, 300, 301, 302, 303, 306,310, 311, 392, 438.Kuhn, W., 115, 175, 410, 412.Kuhr, E., 358.Kung, H., 383.Kunitz, M., 419.Kurtz, S.S., jun., 201.KUSS, E., 142.Kusumoto, S., 381.Kuwada, S., 332, 339.Kylin, H., 299, 304, 306.Kyllinstad, 0. S., 473.Lacher, J. R., 176.Laidler, D., 167.Laidler, K. J., 235.Laine, T., 443, 444, 448.Lamb, W. E., 8.Lambert, A., 282.Lamble, C. G., 402.Lambrecht, R., 320, 322.Lampen, J. O., 441.Lampitt, L. H., 394.Landsteiner, K., 400, 406.Landt, E., 462.Langedijk, S. L., 201.Langenbeck, H., 266.Langenbeck, W., 266.Langmuir, I., 101, 117, 118, 400, 414.Langsdorf, A., 19.Langseth, A., 65.Lankelma, H. P., 206, 455.Lapworth, A., 216, 221, 229.Larionov, J., 466.Lark-Horovitz, K., 33.Larsen, C.D., 361.Lassieur, A., 473.Latner, A. L., 220490 INDEX OF AUTHORS’ NAMES.Laubengayer, A. W., 46, 128, 130,Laubereau, O., 319.LauEer, M. A., 412.Laughlin, K. C., 201, 456.Lauaspach, E. H., 457.Lauritsen, T., 8.Lautsch, W., 273, 322.Lawrence, A. S. C., 104, 106, 107,108, 114, 412.Lawroski, S., 454.Lawson, J. L., 19.Lawton, S. E., 282.Leary, R. E., 252.Leavitt, J. J., 368.Le Beau, D. S., 118, 121.Lebeau, P., 152.Lecat, M., 460.Lederer, E., 297, 298, 299, 301, 307,Lee, J. van der, 221.Lee, M., 402.Leenderste, J. J., 202.Leermakers, J. A., 84.LeFBvre, (Mrs.) C. G., 54, 71.LeFBvre, R. J. W., 48, 54, 66, 71.Lehrman, A., 126.Leighton, P. A., 84, 92, 93, 99, 272.Leighton, W.G., 92, 93.Lemberg, R., 302, 325, 329.Lennard-Jones, J. E., 29, 31, 34, 59,Lennette, E. H., 407.Lenz, W., 25.Leo, M., 266.Leong, P. C., 387.Le ROUX, L. J., 247.Lesesne, S. D., 457.Lett&, H., 356.Leuchs, H., 380.Levene, P. A., 218.Levenson, H. S., 234, 235.Levi, G. R., 180.Levi, H., 20.Levie, L. H., 403.Levin, G. I., 403.Levin, R. H., 363.Levine, S., 101, 117.Levine, V. E., 391.LBvy, H. A., 76.Lewis, G. L., 54, 66, 179, 365.Lewis, G. N., 73, 148.Lexer, E. W., 391.Li, C. H., 401, 402, 403, 404.Liang, P., 353.Libby, W. F., 19.Libowitzky, H., 325, 329.Lichtenwald, H., 331.Light, A. E., 403.Lightbown, I. E., 204.Lin, K. H., 382.Linde, W., 375.Lindsey, A. J., 470, 474.186.308, 309, 311, 323, 325, 392.173.Lineweaver, H., 419, 429.Linnett, J.W., 177.Linsert, O., 361.Linstead, R. P., 188, 312, 313, 314,Linton, E. P., 75.Lipmann, F., 386, 417.Lippincott, S. B., 208.Lipton, M. A., 385, 387.Lister, M. W., 43, 74.Littwin, R. S., 448.Liutenberg, A. I., 357.Livingood, J . J., 19, 20.Llewellyn, F. J., 181.Lochte, H. L., 467.Lockwood, H. C., 475.Lockwood, W. H., 314, 329.Lonnberg, E., 312.Logemann, W., 334, 336, 338, 343,London, F., 25, 26, 27.Long, C. N. H., 337, 403.Long, F. A., 206.Long, H., 384.Long, R. S., 376.Longenecker, H. E., 391.Longsworth, L. G., 405, 406, 409, 414.Lonsdale, (Mrs.) K., 167, 178, 314.Loon, J. van, 215.Loring, H. S., 412.Lowe, C. S., 468.Lowry, T. M., 232.Lubbock, D., 393.Lucas, R., 174.Luck, J.M., 404, 423.Luckett, S., 424.Ludwig, C., 161.Luetscher, I. A., 407, 408.Luttringhaus, A., 366, 367.Luke, C. L., 477.Lukens, F. D. W., 387.Lundberg, W. O., 392.Lunde, G., 464.Lundgren, H. I?., 398, 399, 400.Lurie, J. L., 477.Lutwak-Mann, C., 450.Lu Valle, J. E., 67.Lvov, A., 442.Lvov, M., 442.Lydh, R., 135.Lyman, C. M., 225.Lyne, R. R., 426.Lyons, W. R., 403, 401.Lythgoe, B., 297, 304, 390.315, 316, 317, 363, 367, 368.354.Maass, O., 34.McBain, J. W., 104, 106, 111, 120.McBain, (Mrs.) M. E. L., 120, 207.McBee, E. T., 205.Macbeth, A. K., 379.McBurney, C. H., 226.McCarrison, (Sir) R., 394INDEX OF AUTHORS’ NAMES. 491McCleary, R. F., 207, 275.McCollum, E.V., 387, 464.McCoy, E., 441.McCulloch, W. C., 394.McCullough, J. D., 135, 178.McCutcheon, 5. W., 212.McDaniel, L. E., 441,MacDonald, D. G. H., 388.MacDonald, S. F., 322.McParlane, A. S., 406, 409.MacGillavry, C. H., 186.McGovkin, A., 371.McGowan, J. P., 394.McHenry, E. W., 388.McIlwain, H., 440.MacInnes, D. A., 405, 409, 414.McKee, R. W., 392.Mackenzie, G. M., 394.MacKenzie, K. R., 14.McKeown, A., 96.Mackinney, G., 292, 308.Maclean, I. S., 212, 215, 216, 217, 221,McMeckin, T. L., 408.Macmillan, D., 169.McMillan, E., 11, 12, 20.McNamara, W. L., 392.MacNiven, W. M., 473.McNulty, B. J., 249.Macrae, T. F., 389, 390.McShan, W. H., 401.Macwood, G. E., 34.Madden, R. J., 388.Maddock, A. G., 151.Madigan, S., 170.Maws, R., 271.Magat, A., 169.Maggs, J., 99.Mair, B.J., 201.Maitra, M. K., 385.Majewski, K. W., 28.Major, R. T., 227.Makin, F. B., 279.Malkin, T., 211, 217.Mallonee, B. E., 461.Malmberg, M., 297, 298.Malone, J . G., 68.Malone, M. G., 68.Malsch, J., 56, 169.Maman, A., 201.Mamoli, L., 333, 340.Mamotenko, M. F., 26.Mann, F. G., 73, 131, 132, 133, 186.Mann, P. J. G., 386.Mann, T., 419.Manning, M. F., 25.Manske, R. H. F., 376, 377.Manunta, C., 214, 309.Marcus, R., 346.Mardles, E. W. J., 120.Margenau, H., 27, 28, 31.Margolis, E., 288.Mariens, P., 30.361, 392.Marion, L., 377.Mark, H., 42, 175.Marker, R. E., 201, 333, 340, 348,349, 350, 351, 352, 353, 354, 355,357, 358, 362.Markley, K. S., 459.Marrack, J.R., 400.Marschner, R. F., 271.Marsden, R. J. B., 66, 69.Marshall, J., 20.Martin, G., 56.Martini, H., 139.Martschevski, A. T., 333.Marvel, C. S., 230, 251, 252, 254, 255,Masch, L. W., 333.Masing, G., 170.Mason, H. L., 337, 340.Massengale, J. T., 287.Massey, H. S. W., 26, 30, 31, 35, 36.Masterman, S., 230, 249.Matsukawa, T., 332, 354.Mattox, W. J., 203.Matuszak, M. P., 461.Mauthner, H., 354.Mavity, J. V., 203.Mavroska, D., 171.Maxwell, L. R., 46, 72.May, E. L., 389, 390, 433.Mayer, J. E., 26, 28.Mayo, F. R., 209, 285, 288.Mead, T. H., 459.Meamber, D. L., 403.Mecke, R., 49, 82.Meerwein, H., 202, 280.Mehl, J. W., 411, 412.Meier, E., 327.Meites, J., 403.Meldahl, H. F., 338, 342, 344, 345,Meldolesi, G., 330.Mellanby, H., 384.Mellon, M.G., 462.Mellor, D. P., 132, 314.Melnick, D., 387, 465.Melville, D. B., 393.Melville, H. W., 89, 90, 91, 97, 98.Mendel, H., 219.Menke, W., 306.Menschikov, G. P., 376.Mericola, C. F., 153.Merriman, R. W., 459, 461.Mesbe, H. J., 271.Mesquita, B., 309.Mesrobcanu, L., 394, 395.Metzger, W., 331.Meyer, C. M., 226.Meyer, E., 256.Meyer, H., 220, 325.Meyer, Jules, 349, 355.Meyer, Julius, 136.Meyer, K. H., 117.Meyer, M., 466.258, 261.346492 INDEX OF AUTHORS’ NAMES.Meyer, R. K., 355, 401.Meystre, C., 334, 336, 337.Miall, S., 38.Michael, A,, 209.Michaelis, L., 268.Michel, J., 171.Michels, A., 29, 31.Michnevitsch, G. L., 170.Mielke, K. H., 358, 359.Miescher, K., 332, 333, 336, 338, 339,344, 345, 354, 363.Milatz, J.M. W., 169.Miles, A. A., 395, 397.Miller, E., 169.Miller, E. P., 33.Miller, L. P., 434.Milsted, J., 274.Minder, W., 15.Mime, J. L. van der, 110.Misch, F., 175.Misra, P., 426.Misra, R. D., 33.Mitchell, H.K., 226,227, 229,389,390.Mittenzwei, H., 320, 321.Mitter, P. C., 221.Miyasaka, M., 332.Mizushima, S., 66, 178.Modern, F., 399.Moller, E. F., 439.Moller, H., 330.Moewus, F., 302, 303.Mogerman, W. D., 472.Mohr, C. B. O., 35, 36.Moissan, H., 152.Moldavski, B., 203.Moldenhauer, W., 130.Monnier, R., 346.Monroe, L. A., 203.Monselise, G. G., 476.Montgomery, E. H., 387.Mooney, R. C. L., 181.Mooney, R. L., 25.Moore, M. L., 361.Moore, T., 212, 297, 385.Moore, W.J., 272, 273.Moore, W. T., jun., 86.Moran, T., 394.Morey, G. H., 208..Morf, R., 301.Morgan, S. O., 57.Morgan, W. M., 379.Morgan, W. T. J., 394.Morikawa, K., 91.Morino, Y., 178.Morrell, J. C., 203.Morrell, R. S., 214.Morris, B. S., 132.Morris, W. C., 151.Morrison, A. L., 298.Morton, A. A., 286, 287.Morton, R. A., 290, 294, 385.Mosher, L. M., 361.Mosley, V. M., 46.Mott, N. F., 30, 168.Mottram, E. N., 216.Mottram, V. H., 394.Moullin, E. B., 57.Muller, Adolf, 219, 331.Muller, Alex, 57, 167, 172, 177.Muller, E., 259, 260, 263, 264, 265,267, 268.Muller, F. H., 48, 50.Muller, H., 56.Mueller, John Howard, 437.Muller, John Hughes, 186.Muller, K. L., 94, 95.Muller, M., 356.Mueller, M. B., 251, 252, 254, 255,Muller, P., 338.Muller, R.H., 462.Muller-Rodloff, I., 263, 264, 267.Mukerjee, L. N., 111.Mulder, D., 260, 261.Muller, G. J., 174.Mulligan, M. J., 202.Mulliken, R. S., 179, 293.Mumm, O., 233, 284.Mund, W., 89.Mundie, L. G., 80.Munsey, V. E., 463.Muralt, A. von, 116, 412.Murphree, E. V., 204.Murray, L. A., 48.Myers, G. N., 414.261.Nadj, M. M., 130.Naegeli, C., 364.Naggatz, J., 360.Nakamiya, J., 308.Nakao, M., 373.Nann, H., 136.Naruse, N., 118.Natanson, G. L., 93.Nathan, W. S., 247.Nauta, W. T., 260, 261.Nawrocki, C. Z., 363.Nazmi, F., 176.Nebel, R. W., 242.Neckers, J. W., 128.Needham, J., 412.Neisser, K., 341.Nelson, E., 17.Nelson, J. M., 419.Nesmejenov, A. N., 130.Neufeld, H. H., 402.Neugebauer, T., 25.Neuhoff, H., 264.Neumark, W., 170.Neunhoeffer, O., 280.Neurath, H., 410, 411.Newkirk, A.E., 130, 186.Newling, W. B. S., 232.Niederl, J., 202.Niel, C. B. van. 307INDEX OF AUTHORS’ NAMES. 493Nier, A. O., 11, 35, 154, 160.Nieuwland, J. A., 344.Nilakantan, P., 167.Nishina, Y., 10, 12.Noack, K., 321.Noble, E. G., 367.Nodzu, R., 371.Noller, C. R., 207, 222, 225, 350, 353,Nordsieck, H. H., 184.Norman, A. G., 428, 429.Ndrrie, M., 325.Norris, A., 51.Norris, F. W., 428.Norrish, R. G. W., 81, 82, 85, 87, 88,276, 277, 289.North, H. B., 342.Norton, A. R., 468.Norton, H. M., 231.Norton, J. A., 209.Norton-Wilson, J., 50.Noyes, W. A., jun., 81, 82, 84, 85,354.273.Nunn, L. C. A., 212, 215, 216, 392.Nutting, G.C., 206.Oberlin, M., 355.Obermann, B., 261.Obst, H., 296.Ochoa, S., 385, 386, 387, 417.O’Colla, P., 425.O’Connor, W. F., 312.O’Donovan, D. K., 402, 404.O’Dwyer, M. H., 427, 428.Oerskov, J., 414.Oestreicher, A., 321.Ogg, R. A., 237.Ohta, T., 297, 307.Okac, A,, 468.Oldenberg, O., 20.Oldham, J. W. H., 170, 172.Oleson, J. J., 390.Olliver, M., 391.Ollsen, L. O., 92.Olsen, A., 464.Olson, A. R., 237.O’Neal, R. D., 13.Oncley, L., 56, 411.Ono, S., 424.Onsager, L., 35, 50, 154.Oppenauer, R. V., 362.Orekhov, A. P., 377.Orla-Jensen, S., 437.Ornstein, L. S., 169, 175.Orr, (Sir) J. B., 393, 394.Orth, H., 312.Orth, P., 268.Ortiz-Velez, J. M., 322.Osaha, H., 76.Osborn, G., 263.Osborne, D., 169, 177.Osterberg, H., 30.Oswald, A., 308.Ota, T., 353.Othmer, D.F., 203.Otte, N. C., 437.Ouchakov, M. I., 355.Overhoff, J., 201, 283,Overstreet, M. R., 384.Oxford, A. E., 442.OZegOWBki, w., 265.Paden, J. H., 331.Padgett, A. R., 206.Padmanabhan, S., 472.Paemel, 0. van, 30, 34.Page, G. H., 28.Pahl, M., 171.Paice, E. S., 214.Paland, J., 333, 334.Paley, R. E. A. C., 29.Palmer, K. J., 46, 64, 66, 67, 74, 176,Palmer, R. C., 103.Paneth, F. A., 273, 474.Pannekoek-Westenburg, S. J. E.,Pappenheimer, A. M., 441.Pappenheimer, A. M., jun., 398, 399,Paquot, C., 315.Parfentjew, I. A., 400.Parker, G. T., 456.Parry, E. G., 304.Partington, J. R., 50, 69.Partridge, S. J., 395.Patat, F., 46, 82.Patek, A. J., 384.Patelski, R.A., 369.Patrick, W. A,, 171.Patterson, J. A., 207, 275.Paul, H., 341, 345.Pauling, L., 39, 40, 46, 67, 74, 148,149, 177, 179, 180, 185, 251, 314,364.177.388.400.Paweck, H., 471.Pearlman, H., 36.Pearse, H. L., 432.Pearson, T. G., 273, 275.Pease, D. C., 374.Peat, S., 423, 426.Pedersen, K. O., 398, 406, 409.Peiser, H. S., 131, 186.Peiskar, H., 66.Pelczar, M. J., 393, 436.Pelmore, D. R., 57, 177.Pemberton, J., 390.Pencharz, R. I., 403, 404.Penney, A. C., 474.Penney, W. L., 46.Pennington, I)., 390, 433.Peppel, W. J., 261.Percival, E. G. V., 425.Perkins, (Miss) M. E., 314, 413494 ENDEX OF AUTHORS’ NAMES.Petering, H. G., 463.Peterlin, A., 120.Petem, D., 339.Pekrs, R. A., 385, 387, 417.Peters, W.A., jun., 453.Peterson, W. H., 438, 439, 441.Petrjak, 10.Petrow, V. A., 355, 356,Peyronel, G., 180.Pfankuch, E., 413.PfeSer, P., 315.Pfiffner, J. J., 337, 342.Philippoff, W., 104, 115.Phillips, M., 428.Phillips, N. W. F., 92, 272.Phillips, P. H., 390.Philpot, F. J., 406.Philpot, J. St. L., 395, 406.Phipers, R. F., 304, 306, 356.Phipps, J. W., 459.Pick, H., 92.Pickard, R. H., 357.Pickering, S. U., 106.Pickett, L. W., 265.Piekara, A., 50.Pikrard, J., 40.Pkrce, W., 169.Pierson, E. H., 207.Pike, H. H. M., 59.Pike, R. H., 394.Pillemer, L., 396.Pim, F. B., 204.Pinder, H., 466.Pinder, J. L., 464.Pines, H., 202, 204.Pink, R. C., 111.Piper, S. H., 211, 217, 218, 219.Pirenne, M. H., 40, 44.Pirie, N. W., 395, 405, 412.Pitkethly, K.C., 203.Pittman, M., 388.Pitzer, K. S., 176, 177.Plambeck, L., 340, 358.Platt, B. S., 388.Platz, B. R., 392.Ploetz, T., 283.Plotze, E., 56.Ylotnikov, V. A., 126.Plummer, C. A. J., 367.Podbielniak, W. J., 453.Polanyi, M., 231, 274.Polghr, A., 291, 299.Poljakova, A. M., 381.Pollard, A., 219.Pollock, H. C., 11.Polly, 0. L., 270.Polson, A., 411.Polya, J. B., 298.Pontecorvo, B., 20.Pool, M. L., 20.Pool, W. O., 208.Pope, 6. G., 399.361.357,Popescu, D. M., 107.Popper, E., 287.Poppick, I., 126.Porter, C. R., 314.Porter, E. F., 400.Porter, J. R., 393, 436.Potapenko, G. M., 56.Potavian, E., 126.Potts, R. H., 208.Potvin, R., 92.Powell, G., 375.Powell, H. M., 69, 73, 166, 178, 180,185, 186.Powney, J., 103.Prankl, F., 8.Prasad, M., 11 8.Pratt, E.F., 227, 434.Preece, I. A., 428, 429.Prelubda, H. J., 464.Preobrashenski, N. A., 381.Preobrashenski, V. A., 381.Preston, G. D., 167.Preston, J. F., 426.Prevost, C., 230.Price, D., 389, 390, 433.Price, J. R., 372.Priou, R., 369.Pruckner, (Frl.) F., 323, 329.Pryce-Jones, J., 124.Przibram, K., 466.Purcell, R. H., 275.Purdie, D., 73, 131, 132, 133, 186.Pyrah, L. N., 384.Quastel, J. H., 386, 389, 447.Quayle, 0. R., 231.Quiggle, D., 453, 451.Rabi, I. I., 36.Rebinerson, A. I., 117.Rabinovitsch, E., 235, 275, 474.Rachevski, F. A., 308.Racinski, B., 381.Radley, J. A., 469.Radschenko, J., 170.Raeder, M. G., 473.Raistrick, R., 394.Rakitin, J.V., 431.Ralston, A. W., 208.Raman, (Sir) C. V., 167.Ramaswamy, R. L., 62.Ramm, W., 169.Ramsden, (Miss) E., 209.Randall, J. T., 66.Rao, B. S., 134.Rao, G. G., 98.Rao, M. R. A., 134, 135.Rao, P. S., 371.Rao, T. V. S., 472.Raper, K. S., 221.Raper, K., 215, 379INDEX OF AUTHORS’ NAMES.Rasetti, F., 21.Rau, B., 325.Raymond, W. D., 388, 465.Razuvaiev, G. A., 285.Read, (Mrs.) A. T., 288.Rebay, A. von, 313, 316.Reber, R. K., 268.Redd, J. C., 427.Redemann, C. E., 433.Reed, C. E., 118.Reed, C. F., 206.Rehaag, H., 169.Reich, H., 337, 339.Reichstein, T., 225, 229, 334, 336,337, 338, 343, 345.Reid, C., 151.Reid, E. E., 207, 229.Reinecke, R., 330, 331.Renning, J., 259.Renz, J., 375.Ricci, J. E., 272.Rice, F.O., 84, 85, 270, 271, 272, 273.Rice, K. K., 270.Rice, W. W., 177.Richards, R. B., 286.Richards, W. T., 171.Richardson, G. M., 286, 287, 440.Richou, R., 400.Riddle, O., 403.Rideal, E. K., 101, 103.Rieger, W. H., 255.Riegert, A., 466.Riemann, T., 359.Riemenschneider, R. W., 212.Rimington, C., 408.Rinderknecht, H., 401.Rintala, P., 448.Ritchie, M., 94, 163.Ritter, D. M., 143, 144.Ritzer, J. E., 457.Rix, W., 170.Roberts, D. I., 254.Roberts, I., 231, 233.Robertson, A., 371, 373.Robertson, J. M., 46, 71, 166, 178,Robinson, (Lady), 218.Robinson, E. S., 399.Robinson, J., 114.Robinson, J. R., 411.Robinson, (Sir) R., 218, 225, 248,Robinson, W. D., 387.Robles, H. de V., 364.Rochelmeyer, H., 378, 379.Rodahl, K., 391.Rodebush, W.H., 48, 50.Roebuck, J. R., 30.Rogers, C. F., 376.Rogers, M. T., 181, 185.Rohrmann, E., 236, 348, 349, 350,351, 352, 353, 354, 355, 357, 358.Roll, A., 170.188, 191, 192, 312, 317.371, 372, 379, 380, 351.495Rolla, L., 128.Rollefson, A. H., 49.Rollefson, G. K., 80, 83, 84, $5, 95,96, 99, 273.Rollefson, R., 49.Rollett, A., 212.Roosen-Runge, C., 359.Rose, C. S., 393.Rose, F. W., jun., 201.Rosen, N., 25.Rosen, R., 204.Rosenberg, H. R., 225, 332.Rosenblum. C.. 461.Rosenblum; L.’A., 389.Rosenfeld, B., 442.Rosenheim, O., 355, 356.Roseveare, W. E., 30.Rosin, S., 36.Ross, F., 373.Rossinskaja, I. M., 130.Rothan, A., 400, 402.Rothemund, P., 312.Rowe, G. A., 368.Roy, M. F., 254.Ruben, S., 13, 20.Ruegger, A., 294, 297, 301.Ruff, G., 399.Ruff, O., 150.Ruhoff, J.R., 456.Ruigh, W. L., 361, 362.Rundall, F. G., 188, 316, 3Runnicles, D. F., 104.Ruoff. P. M.. 271.Ruschig, H.,. 340, 342, 843.Rushbrooke, G. S., 33.Ruska, H., 413.Russell, H., 75.Russell, (Sir) J., 394.Rust, F. F., 205.Rutgers, A., 174.Rutgers, I. J. J., 332.Ruzicka, L., 332, 334, 338, 339, 342,344, 345, 346, 349, 354, 355, 380.Rydbom, M., 297, 302.Rysselberghe, P. van, 207.Sadron, C., 123.Sahai, P. N., 219.Saigh, G. S., 275.Sakan, T., 381.Salah, M., 384.Salaman, R. N., 393.Salmon, C. S., 111.Samant, K. M., 356.Samec, M., 422.Samuel, R., 82.Samuels, H., 214.Sand, H. J. S., 469, 470, 476.Sanderson, R. T., 148.Sando, C.E., 212.Sandor, G., 399, 400.Sandulesco, G., 888496 INDEX OF AUTHORS’ NAMES.Sankaran, G., 384.Sassaman, W. A., 136.Sata, N., 118.Sauer, M. E., 385.Saunders, F., 439.Saunier, R., 171.Sauter, E., 175.Saylor, C., 175.Scanlan, J. T., 213.Scarborough, H., 392.Schafer, K., 30, 176, 177.Schaefer, V. J., 400.Scheer, J. van der, 398.Scheffers, H., 48.Scheka, I. A., 126.Schenck, F., 362.Schemer, J. A., 472.Scheunert, A., 297.Schewe, J., 154.Schicktanz, S. T., 453, 455, 461.Schiessler, R. W., 207.Schirmer, F. B., 128.Schlager, J., 371.Schlegel, H., 131.Schleicher, A., 474.Schlenk, W., 255, 259, 266.Schlesinger, H. I., 139, 140, 141, 142,143, 144, 146, 147, 148.Schlessinger, L., 86.Schmahl, N. G., 154.Schmall, K., 56.Schmeisser, M., 130.Schmerling, L., 201.Schmid, G., 170.Schmidlin, 2 53.Schmidt, J., 341, 352.Schmidt, O., 170.Schmidt, S., 414.Schmidt-Thome, J., 333, 340, 341,Schmitt, M., 461.Schmitz-Dumont, O., 153.Schneider, C.L., 430.Schneider, H., 392.Schneider, I., 206.Schneider, O., 135, 184.Schon, K., 309.Schoenauer, W., 284.Schoenheimer, R., 222.Schopff, C., 373.Schoepfle, C. S., 259.Schopp, K., 301.Scholz, C., 338, 339.Schomaker, V., 42, 65, 67, 70, 75, 79,177, 181, 364.Schoon, T., 38.Schrenk, W. T., 473.Schroder, E., 172.Schroder, C. G., 323.Schubert, M. P., 268.Schuchowitzky, A. A., 26.Schutz, F., 438.Schiilman, J. H., 109.345.Schultan, H., 129.Schultze, R., 329.Schulz, G. V., 289.Schumacher, A.E., 389, 464.Schumacher, H. J., 93, 94, 95,Schumb, W. C., 151.Schuster, K., 138.Schutz, P. W., 461.Schwab, G., 301.Schwab, G. M., 250.Schwarze, W., 375.Schwarzenbach, G., 268.Scott, A. D., 240.Scott, B. B., 457.Scott, E. W., 464.Scott, J. A., 370.Scott Blair, G. W., 112.Seaber, W. M., 463.Seaborg, G. T., 12, 14, 19, 20, 158,Seaman, W., 468.Sears, F. W., 42.Sebrell, W. H., 389, 390.Seefried, H., 287.Seegmiller, C. G., 152.Seeley, M., 427.SegrB, E., 9, 14, 19, 20.Seigle, L. W., 207, 275.Selby, W. M., 208.Selker, M. L., 455.SenfY, H., 314.Senti, F., 191.Serini, A., 334, 336, 338, 343.Seshadri, T. R., 371.Seybold, A., 304.Seyer, W. F., 457.Seyle, H., 404.Shaffer, M. E., 400.Shah, C. S., 379.Sharma, V., 370.Shedlovsky, T., 402, 409.Shemill, M.L., 457.Shih-Chang Shen, 412.Shils, M., 387.Shinowara, G. Y., 211, 212, 215.Shookhoff, M. W., 232.Shorter, A. J., 182.Shortley, G. H., 24.Shoupp, W. E., 10.Shrader, S. A., 457.Shriner, R. L., 230.Shute, H. L., 103.Siddiqui, S., 370.Sidgwick, N. V., 37, 68, 69, 73, 166,Siedel, W., 326, 327, 330, 332.Siegel, S., 168.Signer, R., 121, 123, 412.Sillars, R. W., 177.Simamura, O., 209.Simha, R., 411.Simmonds, W. H. C., 371.96.159.180INDEX OF AUTHORS’ NAMES. 497Simola, P. E., 385.Simons, E. L., 57.Simons, J. H., 456.Simpson, J. C. E., 361.Simpson, M. E., 401, 402, 403, 404.Sinclair, H. M., 387, 388.Singer, J. H., 464.Singh, S., 128.Sirk, H., 169.Sisler, H.H., 137.Sisson, E. W., 215.Sitte, K., 175.Skau, E. L., 363.Skinner, H. A., 182.Skoog, F., 431.Slater, J. C., 24, 25, 26, 27, 28, 30.Slotta, K., 341.Smallwood, H. M., 53.Smelser, Q. K., 402.Smith, E. A., 177.Smith, E. L., 107.Smith, E. R., 460.Smith, F., 424.Smith, G. B. L., 138.Smith, H., 167.Smith, H. A., 212, 234, 235, 456.Smith, H. H., 392.Smith, J. A. B., 219.Smith, J. C., 211, 218.Smith, J. H. C., 302, 307.Smith, J. O., 271.Smith, L. I., 210.Smith, R. A., 202.Smith, R. E., 272.Smith, R. L., 94.Smith, R. N., 92, 93.Smith, W., 175.Smits, A., 174.Smittenberg, J., 201.Smyth, C. P., 48, 50, 52, 54, 67, 66,67, 68, 70, 172, 179, 182, 365.Snell, 2. E., 227, 229, 389, 390, 433,434, 438, 439, 441.Snog-Kjaer, A., 437.Snow, R., 433.Snyder, H.R., 146.Snyder, T. L., 442.Sober, H. A,, 385.Soffge, K. H., 131.Sorensen, N. A., 306, 309, 310, 311.Sok, G., 264.Sokolov, I., 170.Solanki, D. H., 472.Soley, M. H., 15.Solmssen, U., 294, 295, 296, 297, 301,307, 309.Soper, H. R., 349.Soret, C., 161.Soroos, H., 130, 201.Southgate, H. A., 201.Spiith, E., 371, 372, 376.Sparks, W. J., 204.Spence, J. C., 394.Spielman, M. A., 223, 224, 355, 363.Spies, T. D., 390.Sprague, R. H., 365.Spring, F. S., 355, 356, 359, 363.Springall, H. D., 46.Sprules, F. J., 376.Squire, 0. V. V., 94.Stacey, M., 426.Stachel, A., 327.Stanberry, S. R., 227, 390.Stanford, S. C., 206.Stanley, R. H., 390.Stanley, W. M., 412.Starling, W.W., 355, 356.Staudinger, H., 123, 290.Stauff, J., 104.Stavely, H. E., 345, 346, 361,Steacie, E. W. R., 92, 270, 272.Stedman, D. F., 455.Steele, W. I., 461.Steenbock, H., 307, 392.Steger, A., 215.Steiger, M., 337, 338, 343.Steigman, J., 9.Steiner, H., 203.Steinhofer, A., 290.Stene, J., 306, 311.Stenhagen, E., 225, 407.Stenzl, H., 283.Stephens, W. E., 10.Stephenson, M., 446, 447, 450.Stephenson, O., 281.Stern, A., 313, 323, 329.Stern, W. G., 398.Stetten, D. W., 222.Stevens, P. G., 207.Stevens, T. O., 342.Stevenson, A. F., 24, 25.Stevenson, D. P., 42, 46, 59, 65, 66,Stewart, D. W., 20, 160.Stewart, G. W., 168.Stewart, H. C., 414.Stewart, W. S., 433.Stewart, W. T., 427.Stickland, L. H., 448.Stier, E., 319.Stiller, E.T., 228, 389, 437.Stillwell, W. D., 150.Stitt, F., 176.Stock, A., 138, 139, 142, 144, 148.Stocken, L. A., 417.Stoner, G. G., 211.Stoner, H. C., 404.Stopher, E. G., 422.Storms, L. B., 233.Stosick, A. J., 71, 72.Strain, H. H., 292, 293, 299, 302, 306,Stranstthan, J. D., 48.Straub, F. B., 419.Straus, W., 302, 308.362.70, 75, 77, 79, 176, 177, 181.309498 INDEX OF ATJTEQRS' NAMES.Streeter, 8. F., 28.Strell, M., 321, 322.Strobel, E., 331.Strong, F. M., 389, 438, 439.Strotzer, E. F., 135, 184.Stuart, H. A., 48, 169.Style, D. W. G., 274.Subbarow, Y., 434.Subrahmanyan, V., 417.Sue, P., 133.Siitterlin, W., 139.Sugden, S., 48, 53, 54, 73, 247, 267,Sugiura, Y., 25.Sugiyama, G., 340.Suida, W., 354.Surnner, F.B., 311.Sundberg, 0. E., 468.Sunier, A. A., 461.Sutherland, E., 214.Sutherland, G. B. B. M., 46, 47, 174.Sutter, M., 337.Sutton, L. E., 37, 48, 51, 53, 60, 61,62, 66, 68, 69, 71, 72, 74, 75, 77, 182.Svedberg, T., 406.Svensson, H. I., 407.Swain, G., 355.Swaminathan, M., 388.Swern, D., 213.Swientoslawski, W., 461.Swinehart, C. F., 150, 151.Swinney, R. E., 394.Swirles, B., 25.Sydenstricker, V. P., 389.Synerholm, M., 353.Synnove, 444.Szabo, A. L., 231.268.Taft, R., 472.Taher, N. A., 238, 240, 248.Tait, T., 204.Takebayashi, M., 209.Takeda, Y., 297, 307.Tamamushi, B., 315.Tamblyn, J. W., 96.Tank6, B., 420.Tam, H. L. A., 450.Taurog, A., 273.Tavastsherna, N. I., 356.Tavel, P. von, 123.Tayenthal, W., 446.Taylor, A., 167.Taylor, D., 94, 163.Taylor, F.A., 218.Taylor, H. A., 85, 87, 271, 272, 273.Taylor, H. S., 86, 88, 99, 160, 271,Taylor, J. F., 314, 413.Taylor, M. D., 205.Taylor, N. W., 263.Taylor, T. I., 35, 160.T'lylor. T. W. J.. 36.272, 273.Taylor, W., 239.Tchoubar, D., 346.Teichert, W., 135, 1%.Teller, E., 270.Teoreil, T., 414.Teschner, F., 267, 268.Tessmar, K., 380.Thayer, S. A., 392,Theilacker, W., 265.Theorell, H., 324, 438.Thiele, F., 286.Thielert, H., 315.Thiessen, P., 175.Thimann, K. V., 430, 43i.Thode, H. G., 160.Thoma, O., 375.Thomas, C. D., 46.Thomas, L. B., 92.Thomas, L. H., 24.Thomas, R. M., 204.Thomas, T. L., 257.Thompson, F. C., 14.Thomson, D. L., 404.Thomson, G., 50.Thys, L., 30.Tiedeke, C., 211.Tietz, E., 265.Tiffeneau, M., 346.Tiggelen, A.van, 89.Tilk, W., 129.Tilman, G., 283.T i m , E. W., 233, 235.Timmermans, J., 172.Timmennans, M. J., 201.Tischer, A. O., 459.Tischer, J., 305, 306.Tiselius, A., 398, 405, @7, 414.Todd, A. R., 374, 375, 390.Tonnis, B., 441.Tohmatsu, S., 315.Tolksdorf, S., 401, 402.Tongberg, C. O., 453, 454.Topley, W. W. C., 394.Torrance, S., 471, 474, 475.Toussaint, L., 474.Toyama, Y., 214.Trautmann, G., 359.Trautz, M., 152.Travers, M. W., 270.Trendelenburg, F., 42.Treon, J. F., 464.TrifLescu, T., 126.Trikojus, V. M., 402.Trim, A. R., 450.Trimm, B., 49.Troitzkaia, M. I., 477.Troitzki, G. V., 308.Truesdail, J. H., 226.Trunel, P., 67.Tschacdarov, D., 473.Tschesche, R., 347, 348, 353.Tschinneva, A.D., 332, 355.Tschitschibabin, A. E., 203INDEX OF AUTHORS’ NAMES. 400Tschopp, E., 332.Tsuchiya, T., 214.Tsujimoto, M., 214, 217.Tsukamoto, T., 353.Tuey, G. A. P., 188, 316, 317, 367.Tullar, B. F., 333.Turkevich, A., 64.Turner, C. W., 403.Turner, D. L., 352, 353, 354.Turner, E. E., 71.Turner, L. A., 11, L6.Turner, L. B., 204.Tuzson, P., 291, 292, 294, 300, 309.Tyray, (Frl.) E., 372.Ubbelohde, A. R., 167, 169, 170, 172,Uehling, E. A., 35.Ueno, Y., 353.Uhlenbeck, G. E., 28, 30, 31, 35.Ujhelyi, E., 308.Ulich, H., 66, 126.Ullgren, C., 476.Ulman, M., 122.Ulshafer, P. R., 353, 358.Ungley, C. C., 387.Unsold, A., 26.Urey, H. C., 26, 160, 231, 233, 456.Uschakov, M. I., 332, 357.Uyei, N., 403.Uyeo, S., 72.Uyldert, I.E., 403.270.Valensi, G., 171.Vanderbilt, B. M., 207, 275.Van Dyke, H. B., 401, 402.Van Heuverswyn, J., 342.Van Vleck, J. H., 49.Vaughan, W. E., 205, 456.Vaughn, T. H., 344.Verheus, J., 203.Veberwasser, H., 364.Veen, A. G., 388.Verkade, P. E., 211, 220, 221.Verleger, H., 46.Verweel, H. J., 187.Verwey, E. J. W., 102.Vesper, H. G., 95.Vestling, C. S., 312, 314, 413.Vigreux, H., 456.Vincent, W. B., 186.Vine, H., 71.Vinti, J. P., 27.Virtanen, A. I., 443, 444, 447, 448.Vivian, D. L., 207.Vloed, A. van de, 172.Volker, O., 309.Vold, M. J., 104.Vold, R. D., 104, 105.Volman, D., 84.Volmer, M., 170.Volqvartz, K., 123.Vorlander, D., 175.Vozaki, K., 214.Wade, J., 459.Wadlund, A.P., 167, 168.Wagner, K. H., 297, 383, 384.Wahl, A. C., 159.Wahramian, A. T., 170, 473.Waisman, H. A., 226.Wakeham, H. R. R., 32, 33.Wald, G., 302.Waldman, B., 17.Waldmann, L., 156.Walke, H., 14, 20.Walker, A. O., 140.Walker, E. W., 283.Walker, J., 394.Walker, O., 297.Walker, 0. J., 153, 284.Walker, R. D., 457.Wall, C. N., 33.Wall, F. T., 68, 161.Wall, T., 35.Walling, C., 209.Wallis, E. S., 230, 251, 362.Walsh, V. G., 414.Walstra, W. K., 31.Walter, G. F., 265.Walter, J., 44, 177.Walters, W. D., 271, 272.Walther, H., 321.Wan, S. W., 214.Wang, S., 373.Wang, Y. L., 387.Warburg, O., 419.Ward, A. L., 201.Wardlaw, W., 181, 182.Wark, E. E., 103.Wark, I.W., 103.Warren, B. E., 32, 46.Warsi, S., 370.Wasastjerna, J. A., 29.Washburn, E., 458.Waterman, H. I., 202.Waters, W. A., 277, 278, 279, 280,Watkins, T. F., 257.Watson, H. B., 202.Watson, H. E., 62.Watson, W. W., 159, 160.Waugh, D. F., 414.Wearn, R. B., 369, 374, 375.Weber, H. M., 364.Weber, K., 268.Weber, P., 205.Weber, S., 34.Webley, D. M., 386.Webster, W. L., 170.Wed, A. J., 400.Weil-Malhorbe, H., 387.Weill, P., 346.281, 285500 INDEX OF AUTHORS’ NAMES.Weimer, N., 209.Weinberger, E., 330.Weinstock, H. H., 226, 227, 229, 389,390, 433.Weintraub, R. L., 463.Weise, J., 280.Weiss, J., 268.Weiss, R., 369.Weiss, T., 333.Weissberger, A., 54, 66.Weitnauer, G., 377.Weizmann, C., 441.Welch, A. J. E., 164, 181.Welch, M.S., 418.Welge, H. J., 90.Welkor, W. H., 396.Wells, A. F., 133, 185.Wells, W. H., 10.Wen-Po, W., 30, 33, 34.Wenderoth, H., 318, 321.Wendt, G., 302.Wenning, H., 366.Went, F. W., 430.Wenzke, H. H., 69.Werder, F. von, 360.Werkman, C. H., 385.Wessely, F. von, 373.West, S. D., 456.West, W., 86.Westenbrink, H. G. K., 387, 388.Westerbreke, D., 427.Westhaver, J. W., 154.Westheimer, F. H., 232.Westphal, U., 341.Westphalen, T., 356, 357.Wettstein, A., 333, 336, 339, 342, 363.Weygand, C., 175.Whalley, W. B., 371.Wheeler, D. V., 212.Wheeler, J. A., 10.Wheland, G. W., 251, 268.Whetstone, J., 423.Whitaker, M. D., 22.White, A., 402, 403.White, A. H., 50, 57, 177.White, C. E., 467, 468.White, J. D., 201.White, P. C., 288.Whitmore, F. C., 201, 204, 231.Wibaut, J. P., 201.Wiberg, E., 138, 144.Wieland, H., 253, 283, 287, 375, 381.Wiener, N., 29.Wierl, R., 144, 148.Wiese, O., 362.Wiesemann, W., 264, 268.Wiessmer, P., 364.Wigner, E., 30.Wiig, E. O., 90, 96.Wijk, A. van der, 118.Wikswo, J. P., 136.Wilcoxon, F., 432.Wild, G. L. E., 284.Wilds, A. L., 332.Wilkinson, R., 279, 282.Wilkinson, S., 375.Willard, H. H., 462.Williams, A. O., 25.Williams, E. C., 205.Williams, E. F., 217, 218, 219, 220.Williams, E. J., 20.Williams, J. W., 398, 399, 400.Williams, M. B., 461.Williams, P. C., 401.Williams, R. J., 226, 227, 229, 389,390, 433, 434, 441.Williams, W. L., 342.Willstaedt, H., 290, 308.Wilsmore, N. T. M., 208.Wilson, A. L., 208.Wilson, C. D., 456.Wilson, D. W., 402.Wilson, E. B., jun., 176.Wilson, J., 394.Wilson, P. W., 444.Windaus, A., 354, 357, 358, 359, 360.Winkler, H. G. F., 121.Winnick, T., 429.Winstein, S., 238.Winterfeld, K., 376.Winterstein, A., 293, 297, 303.Wintersteiner, O., 332, 337, 362.Wirth, T., 220.Wirtz, K., 163, 164, 175.Wirz, H., 355.Wiselogle, F. Y., 261, 263, 268.Wissler, A., 412.With, T. K., 290.Wittig, G., 261, 269, 289.Wittle, E. L., 333, 340, 352, 358, 362.Wizinger, R., 366.Wohl, M. G., 384.Wohlisch, E., 175.Wojciechowski, M., 460.Wolf, K. L., 48, 50, 59, 68.Wolff, A., 358.Wolff, H., 374.Wolman, W., 463.Woo, T. M., 106.Wood, H. G., 385.Wood, L. J., 103.Wood, W. C., 235.Woods, D. D., 448.Woods, D. W., 445.Woodwrtrd, (Miss) I., 178, 188, 312,Wookeg, E., 446.Wooldridge, W. R., 445.Woolf, B., 447.Woolley, D. W., 226, 389, 440.Woolman, A. M., 348, 349.Wooster, C. B., 252.Wooten, L. A., 477.Worden, A. N., 389.Work, C. E., 390.Wormall, A., 397.317INDEX OF AUTHORS' NAMES. 501Wouthuysen, S., 174.Wright, R., 460.Wu, T. Y., 179.Wuis, P. J., 260, 261.Wunderer, A., 323.Wyckoff, R. W. G., 181, 186, 398.Wyk, A. van der, 175.Wyman, J., 51.wu, c. s., 9.Yager, C. B., 130.Yago, M., 339.Yanagita, M., 373.Yeng, E. F., 388.Yasaki, T., 10, 12.Yates, J., 357.Yeddanapalli, L. M., 86.Yost, D. M., 169, 177.Yost, W. J., 25.Young, F. G., 404.Young, G. T., 423.Young, J. W., 457.Young, S., 460, 461.Yudkin, J., 387.Yuill, M. E., 397, 398.Zachariasen, W. H., 168, 180, 185.Zaidel, A., 466.Zdanow, V., 26.Zechmeister, L., 290, 291, 292, 294,Zeile, K., 325.Zerfas, L. G., 388, 415.Ziegler, K., 268.Zilva, S. S., 391.Zimmerman, P. W., 431, 432.Zubrys, A., 297.Zuffanti, S., 207.ZugrBvescu, I., 259. ' Zugriivescu, s., 259.~ Zuhlsdofl, G., 360.299, 300, 301, 308, 309.Zuiderweg, F. J., 203.Zumbusch, M., 135, 184.Zumwalt, L. R., 177.Zwiauer, K., 133

ISSN:0365-6217    

DOI:10.1039/AR9403700479

出版商:RSC    

年代:1940

数据来源: RSC