ANALYTICAL CHEMISTRY.1. INTRODUCTION.EAUH year the activities of the analytical chemist become more diversifiedand no fewer than 3600 papers are summarised in Section C of BritishAbstracts for 1949. Though the future considered policy of these Reportswill be to summarise progress of the current year, this volume follows thepattern of its predecessors in presenting accounts of certain discrete topics,viz., organic and gravimetric analysis and the analytical applications ofRaman spectra and of organic reagents, in order to bring up to date thereviews of some topics and to repair other omissions. The report on gasanalysis shows the extent to which physical methods have supplemented orsupplanted classical methods in the past decade, and the section on radio-activation analysis describes a new and powerful technique of great sensitivity.Attention should be drawn to a series of well-documented and illustratedreviews covering work of the past five yea.rs and intended as a basis for itprojected new series of annual reports.1 The 29 articles (totalling 170pages with 3700 references) deal with ultra-violet and visual spectrophoto-metry ; Raman and mass-spectrometry ; the absorption and diffractionof X-rays ; emission and infra-red spectroscopy ; electron and light micro-scopy ; polarogra.phy, amperometry , and electroanalysis ; inorganic (andorganic) microchemical, gravimetric, and volumetric analysis ; distillation,extraction, ion-exchange, indicators, nucleonics, instrumentation, fluori-metry, and statistics. H.M. N. H. I.2. GCRAVIMETRIC ANALYSIS.Balances.-Where rapid volumetric or physical methods of analysis areto be used, the time spent in weighing the sample may well comprise a.nuneconomically large fraction of the whole. Aperiodic balances haverecently been marketed in .which built-in weights operated by control knobsreplace all external and fractional weights, and direct-reading scales displaythe weight of the object which can easily be obtained within one minute.The remarkable balance designed by E. Mettler has only one pan and thesensitivity remains constant a t 0.05 mg. over the range 0-200 g.In adapting a standard analytical balance to work on the micro-scale,the familiar expedient of setting the sensitivity as high as possible and readingthe deflections of the freely swinging beam is provocative of eye-strainwhich can be minimised by a photo-electric attachment described by C.L.Rulf~.~ If the radiation from an a-active source attached to one beam ofa micro-balance is received in an adjacent 3-plate ionisation chamber, theAndyt. Ghem., 1949, $31, 1 to 173. * Cf. R. H. Muller, ibbid., 1948, 80, 29, A , IbM., p. 262IRVINa : GRAVIMETRIU ANALYSIS. 269displacement from a null position cttn be amplified electronically to give asensitivity of 1 pg. per mm. deflection of a spot-gal~anometer.~ By theaddition of twin photo-tubes and a d.c. amplifier J. W. Clarke has de-veloped the principle of the magnetic balance and produced a prototypedirect-reading null instrument which is faster to use and requires less skillthan its mechanical counterpart.Accounts have been published of micro-gram balances employing quartz torsion fibres or helices,? and B. B.Cunningham and L. B. Werner describe measurements of the specificactivity of plutonium which demanded the construction of a Salvioni-typebalance to give a sensitivity of 0.01 pg. with a load of 10 pg.Weights and Weighing.-A. Craig lo discountenances the use of leadin the inner cavity of weights and notes gradual increases in weight even afterlacquering : storage in silk or paper is preferred to a velvet-lined box whichmay promote corrosion.11 C. Herbo l2 points out the series of uselessoperations involved in classical procedures for calibrating sets of weightsby substitution or transposition and describes a simpler method.Althougha new explanation has been advanced l3 to explain the troublesome driftsometimes observed during the course of a weighing, no general agreementhas yet been reached on the desirability of drying non-hygroscopic substancesor on the use of desiccants in the balance chamber.Changes in the weight of a body under isothermal conditions can beused to follow quantitatively the progress of, e.g., corrosion,14 photo-decomposition, weathering, solvation and absorption. Such changes, andthose which result when a body is exposed to a steadily changing tem-perature, can conveniently be studied with recording thermo-ba1an~e.l~Duval and his associates have studied the pyrolysis curves of more than700 substances commonly obtained in gravimetric analysis, and haveestablished the temperature ranges over which the initial precipitates canbe brought to constant weight as a definite hydrate, an anhydrous salt, orsome one or other specific decomposition product.The original papers(Parts I1 to XXVIII) deal in order l6 with compounds of Ca, Sr, Ba, Mg,Be, Li, Na, NH,, K, Rb, Cs, T1, La, Ce, Ne, Sm, Sc, Pr, Eu, Al, Ga, In, U,4 I. Feuer, Analyt. Chem., 1948, 20, 1231. ti Rev. Sci. Instr., 1947, 18, 915.P. L. Kirk, R. Craig, J. E. Gullberg, and R. Q. Boyer, AnaZyt. Chem., 1947,19,P. L. Kirk and F. L. Schaffer, Rev. Sci. Imtr., 1948,19, 785.J . Amer. Chem. SOC., 1949, 71, 1521.427.* E. Salvioni, " Misura di mase comprese fra g. 10-1 a g.lo Analyt.Chem., 1947, 19, 72.l1 F. J. Maffei, Amis Amoc. Quim. B T ~ , 1946,5,53.l2 Analyt. Chim. Acta, 1947, 1, 254.l3 F. E. Beamish, Analyt. Chem., 1949, 21, 144.l4 P. Chevanard, X. Wache, and R. de La Tullaye, BUZZ. SOC. chim., 1944,11, 41.l 6 Y. Jouin, Chim. et Ind., 1947,!58,24; C . Duval, AnalyL Chim. Acta, 1947,1, 341,l6 C. Duval and S . Peltier, ibid., pp. 345, 355, 360; C. Duval and T. DuvaI, &id.,1948, 2, 45, 53, 57, 97, 103, 105, 110, 205; C. D u d and S. Peltier, <bid., pp. 218, 222,226, 228; 1949,3, 183, 186, 189,191; C. Duval and T. Dupuis, ibid., pp. 324,330,335,345, 438, 589, 599.Me8sin8, 1907270 ANALYTICAL CHEMISTBY.Cr, Gd, Th and Mn, and should be consulted for details of the work, which isof fundamental importance for gravimetric practice.The automaticthermo-gravimetric analysis of mixtures (e.g., Ca and Mg precipitated asoxalates, or Cu and Ag in an alloy) can often be carried out once the behaviourof the pure components has been established.17 It has also been shownthat " ashless " filter-papers lose their absorbed water below 75" and keepconstant weight up to 180" : ashing is completed a t 675".lS The weightof Gooch asbestos, dried by 75", remains constant up to 283" only and thendecreases grad~al1y.l~Gravimetric Procedures.-In addition to work with organic reagentsnoted below (p, 271), much attention has been paid, particularly among thealkaline and rare earths and their congeners, to problems of separation andthe techniques of obtaining precipitates in forms suitable for fi1trati0n.l~Tore Holth 2o has studied critically the use of ammonium oxalate in theseparation of calcium from magnesium, though with high Mg : Ca ratios apreliminary separation of most of the magnesium as hydroxide may bedesirable.21 A dense, coarsely crystalline precipitate of magnesium oxalatewhich can be readily filtered and washed is achieved by slowly generatingoxalate ions in situ by the hydrolysis of ethyl oxalate in 85% acetic acidsolution.22 The same process of '' precipitation from a homogeneoussolution " has been applied successfully to the precipitation of ZrO(H,PO,),by means of alkyl phosphates or pyroph0sphates,2~ to the precipitation ofthorium and rare-earth oxalates from monazite by means of methyl oxalateF4and the isolation of thorium from admixture with rare earths by use oftetrachlorophthalic Radioactive pyrophosphate has been used inone method for determining thorium,26 and sodium paraperiodate 27 andm-nitrobenzoic acid have also been examined as precipitants.Radio-active ruthenium being used, an interesting study has been made of itsfire-a.ssay29 which showed that loss of the volatile tetroxide during fusionand cupellation was negligible although the slag and cupel retained significantamounts-bservations which may well have a bearing on the refining ofthe other precious metals. H. M. N. H. I.l7 C. Duval, Analyt. Chim. Acta, 1948,2,432.l8 Idem, ibid., p. 92.*O Andyt. Chem., 1949, 21, 1221; cf. E. R. Wright and R. H.Delaune, Ind. Eng.21 J. A. Greear and E. R. Wright, Analyt. Chem., 1949, 21, 696.za L. Gordon and E. R. Caley, ibid., 1948, 20, 560.23 H. H. Willard and R. B. Hahn, ibid., 1949,21, 293; R. B. Hahn, Microfilm Abstr.,24 H. H. Willard and L. Gordon, Analyt. Chm., 1948, 20, 166.26 L. Gordon, C. H. Vanselow, and H. H. Willard, ibid., 1949, 21, 1323.2e T. Moeller and G. K. Schweitzer, dbid., 1948, 20, 1201.27 M. Venkataramaniah and B. S. V. R. Rao, Cum. S&., 1949,18, 170.9-8 G. H. Osborn, Analyst, 1948, 73, 381.2s R. Thiers, W. Graydon, and F. E. Beamish, AnaZyt. Chenz., 1948,20, 831.Idem, ibid., 1949, 3, 163.Chem. Anal., 1946, 10, 426.1948, 8, No. 1, 25IRVING : ORGANIU REAGENTS IN INORGANIC ANALYSIS. 27 13. ORGANIC REAGENTS IN INORGANIC ANALYSIS.Introduction.-Many extensive studies of groups of related organiccompounds have lately been made in attempts to correlate structure andgroup-reactivity or to procure reagents of greater sensitivity or selectivity,and the recent literature of the subject 3O has been enriched by a notablecontribution from F.Feigl.31The important observation 32 that 2-methyloxine (8-hydroxyquinaldine)differs from oxine in giving no precipitate with aluminium has been confirmedby H. Irving, E. J. Butler, and M. 3’. RingF3 who show that this peculiarity(due in part to steric factors) is shared by 2 : 4-dimethyl- and l-phenyl-8-hydroxyquinoline, l-hydroxyacridine, and 9-hydroxy-1 : 2 : 3 : 4-tetrahydro-acridine, though 5-, 6- and 7-methyl-8-hydroxyquinolines react normallywith aluminium and all these reagents give insoluble complexes with Zn,Cu, Ga, CrIII and FeIII.Mercaptobenzthiazole has been recommended asa precipitant for rhodium,3* and a-furil d i ~ x i m e , ~ ~ n i ~ x i m e , ~ ~ and 1 : 10-phen-anthroline 37 are excellent for palladium. Organic reagents for uranium 38include isatin p - ~ x i m e , ~ ~ and the potentialities of isatin a-oxime 40 andisatin P-semicarbazone and its N-methyl and -benzyl derivatives 41 andisoquinoline 42 have also been explored. Dicyanodiamidine has been used 43to precipitate vanadium as C,H,ON,,HVO, and o-dianisidine forms insolublecompounds with molybdates44 and with copper in the presence ofammonium thiocyanate : 45 in each case there are many interferencesand the precipitates must be ignited to oxide before weighing.Phyticacid (inositol hexaphosphoric acid) precipitates scandium quantitatively 46*O J. F. Flagg, ‘‘ Organic Reagents Used in Gravimetric and Volumetric Analysis,”N. York, Interscience, 1948; F. J. Welcher, “ Organic Analytical Reagents,” Vols. I-IV, N. York, D. van Nostrand Co., 1947-48; J. H. Yoe in “Recent Advances inAnalytical Chemistry,” Interscience, N. York, 1949, pp. 31, 49; P. Wenger and R.Duckert, “ Tables of Reagents for Inorganic Analysis,” Third Report, of the Inter-national Committee on New Analytical Reagents and Reactions, Bade, Wepf & Co.,1948.81 “The Chemistry of Specific, Selective and Sensitive Reactions,” N. York,Academic3 Press, 1949.32 L. L. Merritt and I.K. Walker, In&. Eng. Chem. And., 1944, 16, 387.s3 J., 1949, 1489.34 D. E. Ryan and P. Fainer, Canadian J . Em., 1949, 27, B, 72; cf. Uazzetta, 1948,36 S. A. Reed and C. V. Banks, Proc. Iowa Awd. Sci., 1948, 55, 267.3 6 R. C. Voter, C . V. Banks, and H. Diehl, Analyt. Chem., 1948, 20, 458.38 E. Ware, U.S. Atomic Energy Comrn., Aug. 1946, Rep. MDDC--1432, 20 pp.3* V. Hovorka and Z. Holzbecher, CoU. Trav. chim. Tc%cosl., 1949,14,40.40 V. Hovorka and L. Divis, ibid., p. 116.41 V. Hovorka and Z. Holzbecher, ibid., pp. 186, 248.42 A. E. Spakowski and H. Freiser, Analyyt. Chem., 1949, 21, 986.43 J. Fidler, Coll. Trav. chim. Tche’cosl., 1949, 14, 28.44 F. B. Ubeda and E. L. GonzaIBz, Anal. 2%. QuCm., 1944, 40, 1312.45 F. Buscarons and E.Loriente, ibid., 1948, 44, 215.46 G. Beck, Mikrochern. mikrochim. Acta, 1948, 34, 62.78, 293.D. E. Ryan and P. Fainer, Canadian J . Rm., 1949, 27, B, 67272 ANALYTICAL CHEMZSTRY.as Sc6C6H6P6O2, ,36H20, and after oxidation thallium can be determinedgravimetrically as [ (C,H,),AS]+[TICI,]-.~~Dimethylglyoxime is only sparingly soluble in water at room temperatureand is commonly used in ethyl alcohol or acetone solution. Their solventaction on the red nickel complex, and the danger of contamination by excessof precipitant, can be minimised by using solutions of the sodium or am-monium salt of the reagent-but these do not keep well. Though moresensitive and 17 times more water-soluble than dimethylglyoxime, “ nioxime ”(cyclohexane-1 : 2-dione dioxime) precipitates nickel down to pH 3 butpermits no separation from iron and the complex is not easily filtered.36CycbPentanedione dioxime is still more soluble but it precipitates nickelonly over a restricted pH range.However, ‘‘ heptoxime ” (cycloheptane-1 : 2-dione dioxime) precipitates nickel quantitatively at pH 2.7 and above+*and although it is only 9 times as soluble as dimethylglyoxime it has out-standing advantages. cc-Furil dioxime should not be used in nickel deter-minations, 8s the composition of the precipitate varies with and whilstthe relatively cheap ‘‘ niccolox ” (diaminoglyoxime) gives a stoicheiometricyellow complex which does not creep and is stable to dryingt9 iron andcobalt interfere seriously. The use of wetting agents to reduce the creepingof nickel-glyoxime precipitates has been thoroughly studied by J.N.Ospenson.50 Electron microscopy 51 shows that cobalt and iron separatelyaffect the appearance of crystals of nickel-dimethylglyoxime (withoutaffecting the weight), and confirms the formation of amorphousC O F ~ C ~ , H ~ , ~ ~ ~ , when both are present.Indicator~.5~-The extensive series of papers by G. F. Smith and hiscollaborators dealing with the synthesis of polysubstituted phenanthrolinesand 2 : 2’-dipyridyls, the absorption spectra of their ferrous and ferriccomplexes, and their applications as redox indicators has recently beensummarised.53 I. M. Kolthoff has studied the kinetics of formation anddecomposition of the ferroins (and ferrins), Le., the ferrous (and ferric)trisphenthroline complexes.% The redox potentials of the ferroins varypredictably and additively with the number and extent of sub~tituents.~~Methyl groups in positions 3 (or 8), 5 (or 6) and 4 (or 7) lower i t by 0.03,0.04 and 0.11 volt, respectively, so the whole range from 1.10 to 0.84 canbe covered smoothly.5 : 6-Dimethyl ferroin (redox potential 0.97 v. inIN-acid), recommended as the best indicator for ferrous-dichromatetitrations,55 may soon be replaced by 4 : 5 : 7-trimethylferroin which has a47 W. T. Smith, AmZyt. Chem., 1948, 20, 937.4s R. G. Voter and C. V. Banks, ibid., 1949, 21, 1320.O9 M. Kuras, Coll. Czech. Chem. Comm., 1947, 12, 198; Mikrochem. mikrochim.rjo Acta Chem. Scand., 1949, 3, 630.61 R.B. Fisher and S. H. Simonsen, Anal@ Chem., 1948, a0, 1107.62 Cf. I. M. Kolthoff, ibid., 1949, 21, 101.63 W. W. Brandt and G. F. Smith, {bid., p. 1313.65 G. F. Smith and W. H. Brandt, Anulgt. Chm., 1949,21, 948.Acta, 1944, 32, 192.fbid., 1948, 20,985; J . Amer. Chem Soc., 1948,70, 2348IRVING : ORGANIU REAGENTS IN INORGAXIC ANALYSIS. 273still lower redox potential of 0.84 v. and the largest molecular extinctioncoeficient of all the ferroins .53 Nickel-dimethylglpxime has been proposedas an external indicator for the same titration and can also be used inacidimetry and in the determination of nickel with cyanide.56 The valueof a-naphthaflavone as a reversible indicator for bromate titrations hasbeen ~onfirmed.~’ M. Taras 58 proposes disodium 4 : 4‘-di-2”-amino-l’’-naphthylazostilbene-2 : 2’-disulphonate and two analogues to replacemethyl-orange or -yellow, €or which artificial colour standards have beendevised by M.L. Nichols and B. L. I r ~ g r a m . ~ ~ 7-Acetamido-2-methyl-quinoline-5-carboxylic a.cid is recommended as a fluorescent indicator,mchanging sharply between pH 7.6 and 8, and Congo-red,61 bromothymol-blue and bromocresol-purple,62 and bromophenol-blue 63 have been usedas adsorption indicators for AgC with CI‘, Agf with CNS‘, and Tl+ with If,respectively. N-Methyldiphenylamine-red can be used for Ag+ with C1’or Br‘ even in strongly acid solutions.64 M. M. Davis, P. J. Schuhmann,and M. E. Lovelace 65 have extended earlier studies of bromophthalein-magenta as an indicator for titrations in benzene to a number of othersulphonphthaleins.Complexing Agents and ‘‘ Complexones.”-Interfering ions can often be“ masked ” by transformation into stable complexes with anions such asF’, Pod3-, CN’, and CNS‘, but though many photometric determinationsand extractive separations depend upon the formation of stable complexesbetween metals and organic reagents, their use as specific masking agentsis still open to development. Tartaric, citric, and other hydroxy-acidshave long been used to ‘‘ hold up ” Cu, Al, Cr, FeIII, etc., sulphosalicylicacid 66 will sequester Be, Ni, and U0,2+, and permits the separation of Mn,TI, or Ti from Fe.Thioglycollic acid masks Fe3+ in the photometric deter-mination of A1 with alumin0n,~7 and in the determination of low concen-trations of aluminium in iron ores 2 : 2’-dipyridyl, by complexing Fe2+,prevents coprecipitation of Fe(OH), with AI(OH),FR = H) and itsderivatives (11 to V), whose properties, first noted in the patent literatureof 1935,69 were thoroughly examined by Schwarzenbach 70-77 who aptlyEspecially noteworthy are iminodiacetic acid (I;56 F. Burriel and F.Pino, Anal. Pis. Qulm., 1949, 45, B, 43.67 R. Belcher, Analyt. Chim. Actu, 1949, 3, 578.5 8 Analyt. Chem., 1948, 20, 680; J . Amer. Water Works A ~ ~ o c . , 1948,40, 468.6* Analyt. Chem., 1948, 20, 1188.6o L. Velluz and M. Pesez, Bull. SOC. chim., 1948,15, 682.61 R. C. Mehrota, J . Indian Chem. Soc., 1948, 25, 541.Idem, Analyt. Chim. Acta, 1949, 3, 69.H. Schlifer, 2.anal. Chem., 1949, 129, 222.65 J . Res. Nat. Bur. Stand., 1947, 39, 221; 1948, 41, 27.66 G. Mannelli, Ann. Chim. appl., 1948, 38, 594.6 7 E. M. Chenery, Analyst, 1948, 73, 501.68 G. F. Smith and F. W. Cagle, Analyt. Chem., 1948, 20, 574.6s Fr. P. 47875, 811938, 822688, 845587.‘O G. Schwarzenbach, E. Kampitsch, and R. Steiner, Helv. Chim. Acta, 1945, 28,63 Idem, ibid., p. 73.828, 1133; 1946,29, 364274 ANALYTICAL CHEMISTRY.termed them " complexones ". Although the stability of most metalcomplexes diminishes rapidly (along a horizontal period) with decrease inthe atomic number of the central atom 78 and falls below that ofthe corresponding aquo-complexes in the case of the alkaline earths andalkali metals, yet chelate-ring formation always enhances stability, and thepolydentate nature of the complexones is such that they can form verystable wa,ter-soluble complexes with magnesium and the alkaline-earthmetals and some even complex significantly with lithium and sodium.Among them the capacity to mask calcium and magnesium (which is ofobvious technical importance in, 'e.g., the removal of lime soaps and dressingfrom textiles, in washing powders, and in photographic developing baths)was found to be especially high in nitrilotriacetic acid 7O (11; the disodiumsalt is marketed as '' Trilon A "), ethylenediaminetetra-acetic acid 72 (I11 ;n = 2 ; the active principle of " Ergalon T " ; the disodium dihydrogensalt dihydrate is " Trilon B "), uramildiacetic acid (IV),73 and exceptionallyso in 1 : 2-diaminocyclohexanetetra-acetic acid (V).', The stability of(IV. 1complexesstill morerare earthNA2*[CH2],*NA,(111.)R*NA2 NA3(1.1 (11.1NH-CO /?p 3 2 p3-NA2 60 \ )CH*NA, CH, CH*NA2NH-CO \ / (V.1(A=CHz*C02H) CH2with the transition metals and with tervalent kations is naturallymarked: a pH of 13 can be reached without precipitation ofhydroxides if (11) or (111) is present.79 That bi- and ter-valentkations can displace hydrogen ions from complexes has been made the basisof a number of volumetric determination^,^^^ 753 76 viz.: ( A ) On addition ofkations to at solution of, e.g., Trilon B, the pH falls from 5 to 3 in consequenceof the reactionMpp+ + H,Y2- --j MYn-* + 2H+and the acid formed can be titrated by using a potentiometric or visual71 G.Schwarzenbach, A. Willi, and R. 0. Bach, HeZv. Chim. Acta, 1947,30, 1303; G.Schwarzenbach, H. Ackermann, and P. Ruckstuhl, ibid., 1949, 32, 1175.72 G. Schwarzenbach and H. Ackermann, ibid., 1947, 30, 1799; 1948, 31, 1029;1949, 32, 1543, 1682.73 G. Schwarzenbach and W. Biedermann, ibid., 1948, 31, 457.74 G. Schwarzenbach, W. Biedermann, and F. Bangerter, ibid., 1946, 29, 811.75 G. Schwarzenbach and W. Biedermann, ibid., 1948,31, 331, 459.7 6 Idem, Chimia, 1948, 2, 56; Helv. Chim. Actu, 1948, 31, 678; G. Schwarzenbach77 G. Schwarzenbach, Chimia, 1949, 3, 1; HeZv. Chim. Actu, 1949, 32, 839; G.78 H. Irving and R. J. P. Williams, Nature, 1948,162, 746.79 H. A. Laitenen and E. Blodgett, J . Amer. Chem.Xoc., 1949, 71,2261.and H. Gysling, ibid., 1949, 32, 1314, 1484.Schwarzenbach and A. Willi, ibid., 1949, 32, 1046IRVING : ORGANIC REAGENTS IN INORGANIC ANALYSIS. 275indicator end-point.K3X or K,Y, the complexing reactions(B) On titration with fully neutralised complexones,M?$+ + X3- -> MXpZp3, or Mn+ + Y4- +cause no change in hydrion concentration, but a jump in pH from 5 to 9marks the appearance of excess of reagent, which is hydrolysed to giveHX2- (or H,Y2-) and hydroxyf ions. The course of a compleximetrictitration can often be followed by employing an indicator sensitive to theconcentration of the metal ions themselves. For instance, murexide(ammonium purpureate) serves as a selective metal-indicator 76 for Cu,for S C , ~ ~ and for Ca, thus permitting the determination of calcium-hardnessin water 74 since the indicator does not respond to Mg : total Mg + Cahardness can be obtained volumetrically by procedure (A) by using thetrialkali salt of Trilon B.Thiocyanate, thiosalicylate, or thioglycollateions serve as indicators for Co and Fe3+, and Eriochromschwarz T permitsthe direct compleximetric titration of Ca, Sr, Mg, Zn, and Cd and by aslight modification Pb, Mn, and Hg.76Complexones can stabilise higher valency states of some metals andmodify normal redox potentials. Since the redox potential of +1*S forCO~+/CO~+ is reduced to 0-6 volt (depending on the pH) by complexing withethylenediaminetetra-acetic acid (as 111), quantitative oxidation by cericions provides a new volumetric method for that element,s1 thoughmanganese and nickel interfere.Bismuthate or lead dioxide oxidisescolourless Mn(I1) to ruby-red Mn(II1) complexonate which can be reducedby standard ferrous or ferrocyanide ; no indicator is needed. SimilarlyCo( 111) complexonates [prepared by oxidation of Co(I1) complexonate a t60°, under which conditions the tervalent manganese complexes are un-stable] can be reduced quantitatively with chromous or titanous solutions.82Essentially the same procedure permits the determination of cobalt polaro-graphically in the presence of large amounts of nickel (or manganese) in abase solution containing Trilon B, for after preliminary oxidation the reduc-tion of the Co(II1) to Co(I1) complexonate gives a good wave with E,-O-1volt : in these circumstances complexonates of A1 or bivalent Co, Ni, andMn are not reduced.s3 Tervalent Cr, Co, Fe, Ti, and Mn give intenselycoloured red, violet, or blue complexes with Trilon B which should servefor their photometric determination.82 Although their efficacy in difficultseparations is demonstrated by their successful application to the classicalproblem of rare-earth fracti~nation,~~ the full potentialities of complexonesas masking agents in microchemical spot-tests and in gravimetric analysishave yet to be developed.s2G.Beck, Analyt. Chim. Acta, 1947, 1, 69.81 R. PZibil and V. MaliEk, Coll. Trav. chim. Tchkcosl., 1949, 14, 413.s2 R. PZbil, ibid., p. 320.83 P. Souchay and T. Faucherre, AnaZyt. Chim. Acta, 1949,3, 252.84 G.Beck and A. Gasser, ibid., p. 41 ; cf. G. Beck, Helv. C h h . Acta, 1946, 29, 357216 ABNALYTICAL UBEMISTRY.The applications of organic reagents to spot-tests, extractive separations,and absorptiometric determinations will be reported on next year..H. M. N. H. I.4. ANALYTICAL APPLICATIONS OF THE RAMAN EFFECT.The general nature of the Raman effect and its chemical applicationswere considered in the Annual Reports for 1934,l and in 1938 a short sectionwas devoted to uses in analytical chemistry. The object of the presentReport is to give a brief account of subsequent developments in the fieldof analysis, illustrated by selected examples. Early in 1949 a more extensivereview appeared which may be consulted for further references.When monochromatic light is passed through a pure transparentsubstance, the spectrum of the small fraction which is scattered contains,in addition to the Rayleigh line, a number of feeble lines of modifiedfrequency-the Raman spectrum.The frequency shifts, relative to theexciting line, are equal to normal vibrational frequencies of the scatteringmolecules and so are characteristic of the substance concerned. In mixtures,the Raman spectra of the components are superposed; but since the spectrausually consist of a relatively small number of more or less sharp lines,they remain distinct (except for fortuitous coincidences). It is upon thisthat the usefulness of the Raman effect for qualitative analysis depends.Since the intensity of Raman scattering is dependent on the concentrationof the molecules concerned, the effect can also be used for quantitativeanalysis.The method is to be regarded as complementary to those based uponultra-violet or infra-red absorption. It is especially useful for componentspresent in relatively large proportion.Although in special cases lowerconcentrations may be detected, the limit is generally about 1%. Forquantitative determinations the minimum may be set at about ‘5% forstrongly scattering species and 10 yo or more for components with intrinsicallyweak Raman spectra.When intermolecular forces between the components of a mixture aresmall, the qualitative and quantitative analysis may be carried out bycomparison with the spectra of the pure components. Such is generallythe case for mixtures of hydrocarbons, t o the analysis of which the Ramaneffect has been mainly applied.When intermolecular forces are stronger,the Raman spectrum of a mixture may differ considerably from a simplesuperposition of the spectra of the components. Such differences havebeen related to intermofecular-compound formation * and ass~ciation.~Ann. Reports, 1934,31, 21. Ibid., 1938, 35, 394.F. J. Taboury and R. Thomassin, C m p t . rend., 1946, 223, 627.* W. G. Braun and M. R. Fenske, AnaZyt. Chem., 1949, 21, 12.li L. Briill, J. Errera, and H. Sack, Rec. T T ~ v . chim., 1940, 59, 284; P. Kohwaram,I d k n J . Physics, 1940,14, 353 ; C. S. Venkateswaran and N. S. Pandya, PTOC. IndianAcad. Xci., 1942, 15, A, 401WOODWARD : ANALYTIOAL APPLICATIONS OF 'PHE m EFFLUT.27'7The exciting light being chosen so as to produce no photochemicaleffects, the determination of Raman spectra leaves the scattering systementirely unaffected and so may with advantage be used for the detectionand estimation of species in equilibria which cannot be " frozen." ThusM.-L. Delwaulle and F. Frangois 6 have detected the ion SnC1,' in solutionsof SnCl, containing excess of chloride ion : and similarly the ion SnBr,'.The same workers have used the Raman effect * to investigate the equilibriumHgX, + HgY, 2HgXY in solution, X and Y being C1, Br, I, or CN;and have demonstrated that mixed halides are formed on mixing stannicbromide with stannic chloride: or titanium tetrabromide with titaniumtetrachloride.1° They find,ll however, that the Raman spectrum of amixture of the tetrabromide and tetrachloride of silicon is a simple super-position of the spectra of the pure components, so that no mixed halideformation occurs in this case.Evidence has been obtainedI2 for thepresence of the species PFClBr in a mixture of PFCI, andPFBr,. Anotherapplication of a similar kind is to the ionization equilibrium of '' strong "acids in aqueous solution. 0. Redlich and J. Bigeleisen l3 have deter-mined the nitrate-ion concentration in solutions of nitric acid by comparisonof the intensity of its Raman spectrum with that in solutions of sodiumnitrate. Similar measurements have also been made l4 for perchloricacid.Excitation of Spectra.-The Raman effect is relatively feeble andintense irradiation of the sample is required.Recently A. C. Menzies andJ. Skinner l5 have described an efficient arrangement in which the sampletube and mercury arc lamps are surrounded by a water-cooled enclosurecoated internally with magnesium oxide, which has a very high reflectioncoefficient in the visible region. The 4358 A. line of mercury is generallyused as exciting line, but when obtained from ordinary high-pressuremercury arc lamps it is accompanied by a, continuous background whichtends to obscure weak Raman lines and renders quantitative photometrymore difficult. The intensity of this background may be diminished byunder-running normal lamps l5 or by the use of low-pressure arcs withcooled mercury electrodes.16 Filters may also be used, both t o reduce thebackground and also to isolate appropriate mercury lines.Useful trans-mission data for filter solutions are given by R. F. Stamm.17 Solid filtershave also been used.18 Objectionable fluorescence of the sample may beCompt. rend., 1940, 211, 65.BuEE. Xoc. chim., 1940, 7 , 359.Ibid., 1941, 212, 761.Compt. rend., 1944, 219, 64.lIL Ibid., 1944, 219, 336. lo Bid., 1945, 220, 173.la M.-L. Delwaulle and F. Franqois, ibid., 1946, 223, 796,l8 J . Amer. Chem. Soc., 1943, 65, 1883.l4 0. Redlich, E. K. Holt, and J. Bigeleisen, ibid., 1944, 66, 13.l5 J . Sci. Instr., 1949, 26, 299.l6 D. H. Rank and J. S. McCartney, J . Opt. Soc. Amer., 1948, 38, 279.l7 Id. Emg. Chem. Anal., 1945,17, 318.B.L. Crawford and W. Horwitz, J . Chem. PhysiCe, 1947, 15, 268; Q. Glocklerand J. F. Haskin, ibid., p. 759278 ANALYTICAL CHEMISTRY.removed by adsorbents 19v20 or by the addition of quenchers.21 The lastwork referred to gives a review of analytical applications of the Ramaneffect up to 1939. Owing to the extreme feebleness of the effect for gases,all the applications have been to the liquid phase.Photo~aphic Method.-Until recently (see below) photography wasthe only method of obtaining Raman spectra, and all the work so farreferred to in this Report was done photographically.Apparatzm-Owing to the relatively low intensity of the effect, a luminousspectrograph with a fast camera is desirable. Rank, Scott, and Fenske l9describe a 3-prism instrument with an f 4-5 camera, and an account of agrating instrument with an f 3.6 camera is given by Stamm.17 The volumeof sample is usually in the neighbourhood of 10 ml., and with efficientexcitation exposure times of the order of minutes (or even less 15) may beused.Qualitative Analysis.---Applications to mixtures (predominantly organic)have been numerous, and some typical examples must suffice.Rank,Scott, and Fenske,lg Stamrn,17 and A. V. Grosse, E. J. Rosenbaum, andH. F. Jacobsonz2 have investigated the applicability of the method tohydrocarbon mixtures ; and motor spirits, both natural z3 and synthetic,2*have been analysed by means of the Raman effect. The method has beenfound useful for the analysis of the products of various organic reactionsand in the field of natural products.25Quantitative Analysis.-Whereas €or qualitative work it suffices toobserve the positions of the Raman lines, for quantitative analysis it isnecessary to undertake the more difficult measurement of intensities.Forthis purpose the photographic plate is notoriously a somewhat inconvenientand inexact agent. In addition, the dependence of the intensity of aselected Raman line of a component upon its concentration must be known.For mixtures in which intermolecular forces are not large the dependence isa linear one, and comparison with the intensity for the pure componentis all that is required. Such simple linear dependence has been verifiedfor hydrocarbon mixtures by various workers.26 Where there is reason todoubt linearityZ7 it becomes necessary to use as standards a number ofIs D.H. Rank, R. W. Scott, and M. R. Fenske, Ind. Eng. Chem. Anal., 1942,14,816.2O M. R. Fenske, W. G. Braun, R. V. Wiegand, D. Quiggle, R. H. McCormick, andD. H. Rank, Analyt. Chenz., 1947, 19, 700.2 1 See J. Goubeau in " Physikalische Methoden der malytischen Chemie " by W.Bottger, Leipzig, Akdemische Verlagsgesellschaft, 1939.22 Id. Eng. Chem. A d . , 1940,12, 191.23 See, e.g., J. Goubeau and V. von Schneider, Angew. Chem., 1940, 53, 531 ; S.Midzushima and T. Tobiyama, J. Chem. Soc. Japan, 1944, 65, 374; S. Midzushima,T. Tobiyama, and H. Shirakawa, $bid., p. 549.24 M.-L. Delwaulle, F. Franqois, and J. Weimann, Chim. et Id., 1946, 56, 292.25 For literature references see ref.(3).26 See refs. (17) and (19) ; also H. Gerding and A. P. van der Vet, Reo. Tmv. chim.,87 P. Traynard, Bull. Xoc. chim., 1945, 12, 981.1945, 64, 257WOODWARD : ANALYTICAL APPLICATIONS OF "FIE RAMAN EFFECT. 279mixtures made up with known proportions. Various schemes for deter-mining concentrations from measured intensities by the use of standardshave been considered by Stamm 17 and by Goubeau.21 Intensity com-parisons between unknown and standards may be facilitated by the additionof a known amount of a reference substance such as carbon tetrachloride l9or carbon disulphide.l7 Compositions of major components of hydro-carbon mixtures containing up to four compounds have been determined 28within about &2%, and the method has been successfully applied l7 to thequantitative analysis of solutions containing sodium nitrate and nitrite.For rough purposes, mere visual estimate by a practised observer 22 gavepercentages within about &lo.A method based upon line widths insteadof intensities has also beenPhotoelectric Method.-Owing to the characteristics of the photo-graphic plate, intensity determinations from microphotometer traces requirefor each plate and wave-length the use of a calibration curve obtained fromstandard intensity marks. Due correction has also to be made for con-tinuous background. The procedure is lengthy and the accuracy attainableis not high. It was therefore a notable advance when in 1942 D. H. Rank,R. J. Pfister, and P.D. Coleman first showed30 that Raman spectra couldbe recorded by using, in place of a photographic plate, an exit slit and aphotomultiplier cell as detector. In 1946 D. H. Rank and R. V. Wiegand 31gave a. full description of a grating spectrograph and photoelectric recordingapparatus suitable for use in quantitative analysis by means of the Ramaneffect. The spectrum is scanned by rotation of the grating and correspond-ing motion of the exit slit and photocell assembly; the photo-current isamplified by a d.c. unit and operates a galvanometer, whose deflectionsare recorded photographically. The intensity scale of the record is linear.A practical difficulty arises from the random " noise " of the photocell,which tends to give a fluctuating background to the record and must beminimised by cooling the cell with solid carbon dioxide.Similar arrange-ments, but using prism spectrographs, have subsequently been describedby J. Chien and P. Bender 32 and by P.-0. Kinell and P. T r a ~ n a r d . ~ ~ Morerecently, C. H. Miller, D. A. Long, L. A. Woodward, and H. W. Thompson34have given a description of a photoelectric instrument in which advantageis taken of the fact that, with mercury lamps run off the 50 c./sec. ax.mains, the exciting light (and hence also the Raman scattering) pulsatesa t 100 c./sec. The photo-current is amplified by a band-pass a.c. unit andrectified by a homodyne system. This has the advantage that the Ramansignal is preferentially amplified as compared with the random " noise,"28 See refs.(17) and (19); also J. Goubeau and L. Thaler, Angew. Chem., 1941, 54,26; 2. Elektrochem., 1941, 47, 150; E. J. Rosenbaum, C. C. Martin, and J. L. Lauer,Ind. Eng. Chem. Anal., 1946, 18, 731.2s G. Duyckaerts and G. Michel, Analyt. Chim. Acta, 1948, 2, 750.30 J . Opt. SOC. Amer., 1942, 32, 390.31 Ibid., 1946, 36, 325.a8 Acta. Chem. Smnd., 1948, 2, 193.34 Proc. Physical SOC., 1949, 82, A , 401.32 J . Chem. Physics, 1947,15,376280 ANaLYTICAL CEEMISTRY.which is thus largely eliminated from the record without the necessity forcooling the photocell. A commercial recorder of the pen type is used.The advent of photoelectric recording in place of photography enablesreliable intensity measurements to be made with greater speed, and opensup new possibilities in the application of the Raman effect to quantitativeanalysis. It is understood that complete photoelectric instruments willsoon be available commercially in this country as well as in America!.In their 1946 paper 31 Rank and Wiegand gave results for 18 syntheticmixtures (2-5 components) of aromatic hydrocarbons containing up to10 carbon atoms.These were all successfully analysed qualitatively andthe percentage compositions determined to within approximately &2. Theinstrument was used by M. R. Fenske et aE.,2* who give reproductions of theRaman spectra records of 172 pure hydrocarbons and also the so-called'' scattering coefficients '' of the lines, i.e., the intensities (as measured byrecorded deflections) relative to that of the Av = 459 cm.-l line of carbontetrachloride determined under the same conditions. These scatteringcoefficients are used in the quantitative analysis of synthetic hydrocarbonmixtures containing up to 6 components, linear dependence of intensityupon concentration being assumed.The determined percentages areusually within &2. Unfortunately, owing to the fact that, as in otherspectrographic methods of analysis, relative intensity measurements in theRaman effect depend on the nature of the instrument used, the scatteringcoefficients of Fenske et aE. (so useful in connection with the particularinstrument with which they were determined) cannot be used for quantitativeanalytical purposes with other spectrographs. Each worker should employstandards measured with his own instrument.The reasons for the variationof scattering coefficients from instrument to instrument have been con-sidered by Rank,35 who gives a method of correcting for one of them (dueto the different polarisation of Raman lines). Application of this correctionwould allow the use of the published scattering coefficients for roughquantitative analysis with any spectrograph. L. A. W.6. ANALYSIS OF ORGANIC COMPOUNDS.Determination of Carbon and Hydrogen.--During the past five yearsvery few drastic modifications of standard methods of combustion analysishave been proposed, and though the trend towards micro-analysis continues,American analysts 1t have stressed the greater reliability of the macro-method, particularly for discriminating between substances which differin their carbon and hydrogen contents by only a fraction of a per cent.Automatic combustion units, several of which have been described: are,55 Analyt.Chem., 1947,19, 766.D. D. Wagman and F. D. Rossini, J. Res. Nat. Bur. Stand., 1944,32,95.D. D. TunniclB, E. D. Peters, L. Lykken, and F. D. Tuemmler, Id. Eng. Chem.* R. 0. Clark and C. H. Stillson, AnuZyt. Chem., 1947, 19, 423; A. Steyemark,Anal., 1946, 18, 710.Id. E w . Chem. Anal., 1945,17,523INGRAM AND WATERS : ANALYSIS OF ORGANIC COMPOUNDS. 281on the whole, now regarded as safe for routine work with compounds of knowncharacteristics. Amongst these, attention may be directed to the semi-micro apparatus of F. 0. Fischer * which is of the Pregl type, and is claimedto give an absolute accuracy of &0.02% provided that compounds of widelydiffering hydrogen content are not burned consecutively.Several modifications of the standard Pregl technique of micro-combustionhave aimed a t hastening the whole operation by employing a faster gas flowthrough the tube, and the extreme limit in this direction would appear tohave been reached by V.L. Les~her,~ who allows hydrocarbon vapours toinflame in an oxygen stream of velocity 300 ml./minute. R. Belcher andC. E. Spooner's method 6-first developed for coal analysis ' f o r thecombustion of organic substances in oxygen in an empty silica tube main-tained at 800" has also been advocated by Russian workers s and has beentested critically for the analysis of compounds containing only carbon,hydrogen, and oxygen.G. Ingram9 found it necessary to place a plugof copper oxide, or preferably silica wool, in the hot tube to prevent thepassage of particulate carbon, and A. F. Colson lo has used a silica spiral toensure the complete oxidation of the combustible gases. However, E. C.Horning and N. G. Horning 11 and P. Gouverneur l2 have shown that inanalysis on a centigram scale good results can be obtained in packed tubeswith flow rates of oxygen, or air, of as much as 25-50 ml./minute. Timeonly will show whether these rapid-flow methods will ultimately replace thepresent standard procedures, for they may not prove to be applicable for theanalysis of very volatile, or thermally unstable, substances.Much attention has been given to the removal of the oxides of nitrogenwhich are formed in the combustion of all nitrogenous substances, especiallyif they are burnt in oxygen rather than in air,l3 and, more particularlywhenever catalytic tube fillings, such as platinum, are used.Since leaddioxide is often considered to be a source of error in the determination ofcarbon and hydrogen, in that some preparations yield high blanks and behavecapriciously with regard to their equilibrium between water and carbondioxide contents of the flowing gases, the use of an external absorbent foroxides of nitrogen has often been advocated. A. E. Heron,14 who hasdealt with the combustion of aliphatic nitro-compounds, increases thelength of the lead peroxide layer inside the tube and, in certain cases, supple-ments it by a spiral bubbler containing chromic acid-sulphuric acid mixture,which he places after the water absorption tube.Liquid absorbents forInd. Eng. Chem. Anal., 1949, 21, 827.J., 1943, 313.M. 0. Korshun and V. A. Klimova, Zhur. Anal. Khim., 1947, 2, 274 (abst. inAnalyst, 1948, 73, 351); M. 0. Korshun and N. S . Sheveleva, C m p t . r e d . (DokZady)Acad. Sci. U.S.S.R., 1948,60,63 (Chem. Abs., l948,42,6270d).lo Ibid., p. 541.Ibid., p. 1247.Fuel, 1941, 20, 130.* Analyst, 1948, '43, 548.l1 Analyt. Chem., 1947, 19, 688.l8 Anal. Chirn. Acta, 1948, 2, 510.l8 A. E. Heron, A&y8t, 1947, 72, 142.l4 Ibid., 1948 73, 314282 ANALYTICAL CHEMISTRY.nitrogen oxides have been used by many others 99 11, l5 in conjunction withthe rapid-flow methods, though Colson lo places a second lead peroxide tubefor this purpose between the water and the carbon dioxide absorptiontubes. External absorbents can be used only when water produced duringthe combustion is prevented from condensing in both the beak end of thecombustion tube itself and in the inlet of the weighed anhydrone tube.Any liquid water in either of these positions would retain nitrogen oxides,and so vitiate the hydrogen figure.In the rapid-flow methods the fast gasstream drives the water vapour well into the desiccant layer before there isany condensation; in the conventional combustion method this is difficultto ensure, and consequently lead peroxide with all its potential failingsis often preferred still.16 A. Bennett,17 however, has eliminated the use oflead peroxide by using nitrogen containing only a little oxygen as the carriergas, and packs his tubes with a layer of copper oxide followed by one of re-duced copper.This departure from the current method of burning substancesin pure oxygen in the presence of oxidation catalysts is a return to theconceptions of Liebig and Dumas which, in view of Heron's findings,l3*14may prove to be particularly valuable for the analysis of nitro-compoundsand the like.It now seems to be agreed that a heated silver packing, if long enough,is adequate for the removal of both halogens and sulphur. The fouling ofcombustion tubes through the volatilization of silver halides is still, however,one of the major drawbacks of micro-combustion for which, as yet, no remedyhas been proposed.Determination of Nitrogen.--In connection with the Dumas method afew modifications in the design of apparatus can be noted.W. K. Noyce l8back-flushes the combustion tube with carbon dioxide when refilling betweenanalyses, and so saves time in routine work. Other workers l9 haveimproved the designs of micro-nitrometers so as to rninimise the foulingaction of the concentrated potassium hydroxide.The Kjeldahl method continues to receive much study, especially inconnection with protein analysis. Mercuric sulphate,2O selenium, andselenium dioxide 21 appear to be the favourite oxidation catalysts, andboric acid solution continues to gain favour as the absorbing agent.In astatistical study of the Kjeldahl method, P. E. Machemer and W. M. &Nab 22find that steam-distillation of the ammonia is much safer than direct boiling.l5 A. Etienne and R. Mileur, Ann. Chim. analyt., 1946, 28, 215; I. Irimescu and B.l6 R. 0. Clark and G. H. Stillson, Ind. Eng. Chem. Anal., 1945,17,520.l7 Analyst, 1949, 74, 188.Popescu, 2. anal. Chem., 1948,128,185.l8 Arurlyt. Chem., 1949, 21, 877.E. Stehr, Ind. Eng. Chem. Anal., 1946, 18, 513; A. Muller, Mikrochem., 1947, 33,2o A. Hiller, J. Plctzin, and D. D. Van Slyke, J. Biol. Chem., 1948, 176, 1401; R. L.21 G. Frey, Helv. Chim. Acta, 1948, 31, 709; R. Jonnard, Ind. Eng. Chem. Anal.,2a Aruzlyt. Chim. Acta, 1949, 3, 428.192.Shirley and W.W. Becker, I n d , Eng. Chem. Anal., 1946, 17, 437.1945, 17, 246INGRAM AND WATERS : ANALYSIS OF ORGANIC COMPOUNDS. 283R. Chand,23 however, advocated distillation in a closed system, and I(.Marcali and W. Rieman 24 have eliminated the distillation altogether byneutralising the digestion mixture and then titrating the ammonium salt toa second end-point with phenolphthalein after adding formaldehyde.Potassium bi-iodate, KH(IO,),, has also been proposed as a receiver for theammonia, which can then be titrated with thiosulphate, after addition ofpotassium iodide and starch indicator.25 P. McG. Shuey 26 has again directedattention to the fact that chlorides may cause loss of nitrate nitrogen, byvolatilization of nitrosyl chloride, when the usual Kjeldahl method is usedfor determining total nitrogen in organic matter.Determinations of Halogens and of Sulphur,-Micro-chemical modifi-cations of almost every current method of halogen or sulphur analysis havenow been proposed, but only a few of these seem to merit special note.M.A. M. Fleuret 2' has developed an interesting centigram-scale methodof fusion analysis. He decomposes his compounds in molten silver nitrate,and thus obtains silver halides which can be weighed directly, whilst silversulphate, or arsenate, can be extracted with ease. Granulated magnesia,has been advocated for the fusion analysis of chlorides,28 and calcinedmanganite for use with iodides.29 Good accuracy is claimed for modificationof the Stepanov method,30 and catalytic reduction using Raney nickelhas been advocated for use with chlorides or bromides other than volatilearomatic ~ubstances.~~Much attention has been paid to the determination of fluorine in organiccompounds.Several workers 32~33 combust fluorides in silica tubes contain-ing quartz chippings. If it is desired to estimate carbon a t the same time,the silicon fluoride thus formed can be absorbed on aluminium oxide kepta t 175" and weighed, whilst simultaneously chlorine can be retained on silverwool and weighed as silver chloride.33 Potassium fluoride, which formsK,SiF,, is another absorbent for fluorine in elementary analysis for carbon.32For fluorine estimation, however, the silicon tetrafluoride is usually collectedin a bubbler containing water, or alkali, after which i t may be estimatedgravimetrically as PbClF or titrated with thorium nitrate.34 An improvedprocedure for this uses Solochrome-blue as indicator.3524 Ind.Eng. Chem. Anal., 1946,18, 709.27 Bull. SOC. china., 1945, 12, 133.z8 J . Indian Chem. SOC., 1947, 24, 167.25 R. Ballentine and J. R. Gregg, Analyt. Chem., 1947,19, 282.26 Ibid., p. 882.28 J. Anelli, Rev. farm. (Buenos Aires), 1945, 87, 61.2s A. Horeau, Compt. rend., 1946, 220, 89.30 A. K. Ruzhentseva and V. S. Letina, Zhur. Anal. Khirn., 1948, 3, 139 (Chem.Abs., 1948, 42, 7656f); K. Shishido and H. Sagi, Analyt. Chem., 1948, 20, 677; J.Decombe, Bull. SOC. chim., 1948, 38, 353.*l M. Pesez and P. Poirier, ibid., p. 379; A. Schwenck, I d Eng.Chem. Anal., 1943,15, 576.s2 N. S. Nikolaco, Chern. Age, 1946, 54, 309.3s R. D. Teston and F . E. McKenna, Analyt. Chem., 1947,19, 193.sa R. H. KimbalI and L. F. Tufts, ibid., p. 150; MT. C. Schumb and K. J. Radimer,$bid., 1948, 20, 871.R. F. Milton, H. F. Liddell, and J. E. Chivers, Analyst, 1947, 78, 43284 ANALYTICAL CHEMISTRY.On the micro-chemical scale J. F. Alicino, A. Crickenburger, and B.Reynolds 36 have reintroduced Van der Meulen's iodimetric method 37 fordetermining bromine, after the catalytic combustion of an organic bromide.The bromine is collected in sodium hydroxide solution, and then oxidised tobromate with sodium hypochlorite. Excess of the reagent is removed bybuffering with sodium dihydrogen phosphate and then boiling with sodiumformate, and the bromate is finally decomposed with potassium iodide and atrace of ammonium molybdate.G. L.Stragand and H. W. Safford38 determine sulphur, after catalyticcombustion, by absorption on weighed silver gauze, kept a t 650". In thepresence of halogens, other than fluorine, the weighed silver is extractedwith boiling water, and the amount of leached silver sulphate is then foundby difference. In sharp contrast to this, M. 0. Korshun and N. E. Helman 39decompose sulphur compounds in hydrogen, using a platinum catalyst,absorb the hydrogen sulphide in an acetic acid solution of zinc sulphate,and finally titrate iodimetrically. The Carius and Burgess-Parr bombmethods for sulphur analysis have been completed by the use of knownvolumetric methods of sulphate determinati~n.~~Determination of Oxygen.4.Unterzaucher's direct method of con-version of oxygen into carbon monoxide 4 1 has now been modified to analysison the centigram scale and an accuracy of 0.2% is now claimed for it.42Helium has been used in place of nitrogen as the carrier gas.43 The method,however, still seems to be outside the scope of most laboratories.Group Analysis.-Modifications of the Zeisel procedure have beendescribed by D. 0. Hoffman and M. L. Wolfrom44 for the determinationof alkyloxy-groups in acetals and in easily volatile alcohols in which thesamples are carefully introduced below the surface of the reaction mixture.By using an electrically heated bath and a water condenser maintained at40°, H.E. Fierz, D. E. Pfanner, and F. Oppliger 45 claim to be able to estimatealkyloxy- and alkylimino-groups to an accuracy of .+0.2%, and also todetermine methoxyl and ethoxyl separately. They use nitrogen rather thancarbon dioxide as the carrier gas.R. GI.. Stuart 46 has made a useful study of the trans-esterification reactioninvolved in acetyl group determination, and other workers have described86 Analyt. Chem., 1949, 21, 755.37 Chem. Veekblad, 1931, 28, 238; 1934, 31, 558.3* Analyt. Chem., 1949, 21, 625.39 Zavod. Lab., 1946,12, 754 (Chem. Abs., 1947,4l, 3326e).40 E. C. Wagner and S. H. Miles, Amlyt. Chem., 1947, 19, 274; A. Sfeyerm~rk,41 Ber,, 1940, '73, 391.42 V. A. Aluise, R. T. Hall, F. C. Staats, and W. W. Becker, Analyt.Chem., 1947,43 W. W. Walton, F. W. McCulloeh, and W. H. Smith, J . Res. Nut. Bur. Stand.,44 Analyt. Chem., 1947,19, 225.46 Helv. Chim. Acta, 1946, 20, 1463.48 AnaZyst, 1947, 72, 235.E. Bass, and B. Littman, ibid., 1948, 20, 587.19, 347; R. A. Dinerstein and R. W. Klipp, ibid., 1949, 21, 545.1948, 40, 443SMALES : RADIOAU!I!IVA~ON ANALYSIS. 285modifications of the normal procedure.4'7 4* R. B. Bradbury,48 for instance,is a strong advocate of the use of toIuene-psulphonic acid, and titrates theacetic acid iodimetrically, making a correction for any sulphur dioxide by ablank determination.For determination of hydroxyl groups the method of acetylation withacetic anhydride in pyridine, and subsequent titration of the unused reagentafter decomposition with water, now appears to have won general favour.Potentiometric titration greatly increases the accuracy of this analysis 49except for colourless solutions.Primary and secondary amines thiols,and some aldehydes interfere seriously with this procedure.A few modifica.tions of the Zerewitinoff method of determining activehydrogen have been describedJ50 and lithium aluminium hydride has beenproposed as an alternative reagent.51G. I.W. A. W.6. RADIOACTIVATION ANALYSIS.Although radioactivation analysis was first used in 1936, it is still in theearly stages of development and indeed this subject has not been previouslyreported on as a special technique of analysis. The basis of the method,the advantages offered, and the possibilities of applying it will therefore bediscussed rather more fully than is customary in these Reports.andsince reviewed by B.Goldschmidt? G. T. Seab~rg,~ G. E. Boyd: and L.T ~ r d a i , ~ is simple; an element is detected and determined by the formationof a radioactive isotope, which can then be subjected to radioactive assay, aprocedure with the inherent possibilities of extreme sensitivity and specificity.Theory.-If an element is placed in a homogeneous flux of constantenergy of positively charged particles or neutrons, then the rate of growthof the number of radioactive atoms N* with time is given byThe essential basis of the method, suggested by G. von HevesydN*/dt = fC,,+N - AN*which on integration for the period of irradiation becomesN* = ~ c ~ ~ .N ( 1 - e-At) /AI f E. Weisenberger, Mikrochem. Mikrochim. Acta, 1947, 33, 51.48 Analyt. Chem., 1949, 21, 1139.C. L. Ogg, W. L. Porter, and C. 0. Willits, Ind. Eng. Chem. Anal., 1945,17, 394;A. Robertson and W. A. Waters, J., 1948, 1585.so A. P. Terentev and K. D. Shcherbakova, J . Gen. Chem. U.S.S.R., 1946,16, 855;R. H. Lehman and H. Basch, Ind. Erq. Chem. Anal., 1945, 17, 428; P. M. Maginnityand J. B. Cloke, Analyt. Chem., 1948, 20, 978.61 H. E. Zaugg and B. W. Horron, ibid., p. 1026.Kgl. Danske Videnskab. Selskab., Math.-fys. Medd., 1936, 14, [ 5 ] ; 1938, 15, [ll].Bull. Soc. chim., 1939, 6, 718.Amlyt. Chem., 1949, 21, 335.Chem. Rev., 1940, 27, 266.Atomics, 1949, 1, 101.* E. Pollard and W. L. Davidson, '' Applied Nuclear Physics," New York, J.Wiley& Sons, 1942286 ANALYTICAL CHEMISTRY.where f is the flux of bombarding particles in units of particles/sq. cm./sec. ;is the isotopic cross section for the nuclear reaction in units of sq. em. pertarget atom; N is the number of target atoms; and A is the radioactivedecay constant which is connected with the half life T,,, by the relationA = 0*693/T,,,.The amount of activity At, in disintegrations per second, exhibitedby the atoms N* produced up to a time t , is given by the expressionAt = AN* = foac.j'jT(l - e-ht) = fo,,J7 (1 - e-0693tiT11~ )So far it has been assumed that the element is monoisotopic; however,considering a weight W g. of an element of atomic weight M , if 6 is theabundance of the particular isotope giving rise to the activity, then the abovebecomesAt =foa,.6W x 6 x 1023 (1 - e-0693t/T~f~ ) /J!fThe factor (1 - e-0693t1T1rt) has been called the saturation factor, 8,which may vary between zero and unity, having a value of Q when the irradi-ation time t is the half-life T,,,.After the irradiation is stopped the activityformed will decay with its characteristic half-life. A quantitative descrip-tion of this is given by E. Rutherford, J. Chadwick, and C. D. Ellis.' Thusfor high activity for a given mass, there should be high values for the fluxand activation cross-section ; while if other things are equal then sensitivityis greater for lower atomic weight elements, and those with high relativeabundance of the particular isotope concerned.The half-life of the isotopeformed does not control the inherent sensitivity of the method. It can,however, be a practical limitation where long irradiation times are necessaryto obtain sufficient activity.The character of radiation emitted by the active isotope formed mustalso be considered in evaluating the sensitivity for a particular isotope.Boyd discusses this in more detail. In general, unless specialised detect-ing equipment is employed, the greatest sensitivity in activation analysiswill be attained when @-particle radiation is measured.General Practical Technique.-From the foregoing, it is seen that, pro-vided the magnitude of the flux, reaction cross-section, and half-life are known,a determination of the absolute disintegration rate should enable the calcula-tion of the absolute mass of the constituent to be determined.However,accurate knowledge of the flux and accurate determination of the absolutedisintegration rate are not always possible, but in practice these difficultiesmay be avoided by making use of a comparative procedure, i.e., the simul-taneous irradiation of samples with standards of the same general com-position, a device quite common in analytical chemistry. After irradiation,if chemical separations are necessary, the samples and standards are dissolved,inactive carrier for the constituent desired is added (and also usually hold-back carriers for other active elements to prevent difficulties from adsorption,etc., as in conventional radiochemical practice), the necessary chemical' " Radiations from Radioactive Substances," Cambridge Univ.Press, 1930SMALES : RADIOACTIVATION ANALYSIS. 287separations performed, and then aliquots mounted for radiochemical assay.Where the chemical yield is not quantitative, a correction may usually beapplied from a knowledge of the mass of inactive carrier added and thatfinally isolated, the latter being measured by any usual analytical techniquebut most often gravimetrically. Corrections for self absorption of theradiation may be necessary if the weights of samples and standards assayedare different, particularly if weak @-emitters are being measured.8 Thenthe mass of X, the constituent originally to be determined, is obtained bycomparing corrected counting rates of samples and standard thus :Total activity from element X in unknown __ mass of X in unknownTotal activity from element X in standard - mass of X in standardNormally the radiochemical purity of the samples and standards wouldbe checked by absorption and decay measurements.Requirements.-The requirements other than the usual analyticalfacilities are obviously (a) an activation source and (b) a counting mechanism.Dealing with ( b ) first, there is now commercially available in this countrystandard p-counting equipment comprising Geiger-Muller tube, usuallyof the end-window type, and lead castle for shielding i t ; a power pack forsupplying the high voltage necessary for the tube, and a scaling unit forchoosing a proportion of the counts to be fed into a counting meter usuallyof the Post Office type.So far as (a) is concerned, the various types and their advantages and dis-advantages will be dealt with individually, special emphasis being laid on thetwo generally available possibilities, i.e., the laboratory radium-berylliumor other similar neutron sources, and the chain reacting pile.Laborcztory Neutron Sources.-These depend upon the bombardment ofberyllium either by a-psrticles from radium (or radon) or polonium, or byy-radiation of maximum energy greater than 1.63 Mev., such as that fromartificially prepared 60-day lz4Sb ::Be + :He ---+ 'ZC + in + 1-6 MeV.:Be + hv _t :Be + inPolonium and antimony both have the disadvantage that they are relativelyshort-lived, decaying with half-lives of 139 and 60 days, respectively, althoughthe former has the great advantage of emitting very little radiation other thana-particles, and hence shielding is much simplified.Because of general availability, however, the radium-beryllium sourceas a permanent unit is useful, and a description of its use for demonstratingactivation has been given by W.H. Hamill, R. R. Williams, and R. H.S ~ h u l e r . ~The slow neutron flux obtainable with a 500-mg. radium-berylliumsource in ~ l . paraffin-wax moderator is of the order of lo4 neutronslsq. cm./sec.and an example of the usefulness of this can be given. The most sensitiveW. F. Libby, Analyt. Chem., 1947,19, 2. J . Chem. Educ., 1949,26, 210, 310288 ANALYTICAL CHEMISTRY.example is lMDy which has a natural abundance of 27.6% and an activationcross-section of 2620 barns (Le., 2620 x sq.cm.) for slow neutrons,lOand the product of its activation, ‘;:Dy, has a half-life of 2.5 hours.The activity to be obtained per g. of natural dysprosium on irradiationto saturation, i.e., for a few days, will belo4 x 2620 x 10-% x 0-276 x 1 x 6 x x 60164 AE == 1-6 x log dis./min. per g.or if irradiated only for one half-life, i.e., 2.5 hours, the activity obtainedwould be 8 x lo5 dis./min./g. The reasonable assumptions being madethat a normal Geiger counting apparatus will register 10% of the p-particlesemitted from the 165Dy and that 8 counts/min, registered above background(-8 c./m.) is a reasonable figure for positive determination, then i t may beseen that 0-1 mg.of dysprosium will be determinable immediately afterirradiation.There are several elements with activation cross-sections and isotopeabundances which can give useful results, though with somewhat lowersensitivity than dysprosium, with the radium-beryllium source, e.g., Ag,Au, Eu, Ho, In, Ir, Lu, Mn, Pr, Re, Rh, Sc, Sm, Ta, Tb, Tm, Yb, W, all havevalues for the product 8 x aae. of >10 (cf. 720 for Dy). On the other hand,there are a number of elements from which practically no activity can bedetected after irradiation for a short period followed by a short time for decay.This is the case with, e.g., Al, B, Be, C, Cb(Nb), F, Fe, Li, Mg, N, Ne, 0, S,Sn, Ta, Ti, T1, V, where the element has either a low cross-section or anexceptionally long or short half-life.Thus it becomes feasible to determinein one or other of these elements small amounts of those listed above.A summary of the slow neutron atomic activation cross-sections (i.e.,product of fractional natural isotope abundance and isotopic activationcross-section) is given by Boyd: and fuller information is available fromSeren et aZ.10 or from I(. Way and G. Haines.llIt can be seen from these examples that, even without chemical separationin some wses, the Ra-Be source makes possible certain otherwise difiEicultanalytical determinations, e.g., rare earths, rare metals, etc., but in generalthe 500-mg. Ra-Be source is useful only for quantities of the order of milli-grams of the favourable elements.Some applications have been reported,e.g., von Hevesy and H. Levi1 determined dysprosium in yttrium, andeuropium in gadolinium, and B. Goldschmidt and L. Meitner l2 have alsoused this method for rare earths, and R. Dope1 l3 has discussed the determin-ation of iridium in platinum.The Atomic Pile (Nuclear Reactor) as a Neutron Source.-This may be10 L. Seren, H. N. Friedlander, and S. H. Turkel, Physical Rev., 1947,72, 888.11 “Thermal Neutron Cross Sections for Elements and Isotopes H-Bi,” 17.5.1st Ark. Mat. Astr. Fys., 1941, 27, A , Pt. 3, No. 17, 1-18.Atomic Energy Commission, AECD-2138.Phydkal. Z., 1945, 44, 261S U E S : RADXOAOTIVATION ANALYSIS. 289considered as an available source since irradiation facilities are availablein this country on request.14Considering first thermal neutrons only, the general discussion givenunder “ Laboratory Neutron Sources ” applies, except that in the Harwellpile, a flux of more than 1011 neutronslsq.cm./sec. is available. If this iscompared with the figure of lo* neutronslsq. em. Isec. previously discussed,which gave a sensitivity of milligram quantities for certain elements, it canreadily be seen that amounts of 10-10 g. or less become determinable. It isthis enormous potential sensitivity coupled with the specific identification,by half-life and energy, of the particular isotope formed, which really givesthis method its attraction. A further attraction is the possibility of over-coming one of the troublesome analytical difficulties in the normal handlingof such small quantities of materials as 10-6 g.or less, i.e., loss by absorption.In the case of radioactivation analysis, once the irradiation is completedit is quite permissible to add relatively large quantities of the inactiveelement concerned, and provided exchange between the active and theinactive isotope is established, the problem of handling sub-microgramquantities disappears.With these higher fluxes the scope of analytical determinations indicatedon p. 288 must be modified, and reference should be made to the full lists l o p ll.for the evaluation of particular problems, although it still remains generallytrue that slow neutron activation analysis will not be possible for the lighterelements because of their very low cross-sections and the short-lived isotopesformed.Thus,Boyd lists the detection of potassium and cmium in sodium salts by resolu-tion of the decay curve, the analysis of mixtures of sodium and potassiumby differential absorption measurement (using the difference in maximumenergy between %Na, max.@-energy 1.39 Mev.; *2K, max. @-energy3 6 8 MeV.), the analysis of a manganese-aluminium alloy for manganese(cf. H. M. Clarke and R. T. Overman 15), and the interesting determinationof stable isotope abundance ratios for copper and chlorine.16 Other examplesare the determination of traces of thulium in erbium,l7 the gallium andpalladium contents of iron meteorites,18 the relative abundance of rheniumin Nature,19 and the analysis of the micro-composition of biological tissue.20A brief mention is made of the analysis of zirconium-hafnium contents ofmixtures of their oxides.21 The use of activation for qualitative analysisExamples of the use of the pile in this way are now appearing.l4 “ Radioactive and Stable Isotopes,” available from Isotope Information Office,16 U.S.Atomic Energy Commission, MDDC-1329.J. W. Kennedy and C. T. Seaborg, PhysicaZ Rev., 1940, 57, 843.1 7 B. H. Ketelle and G. E. Boyd, J . Amer. Chem. SOC., 1947, 69, 2800.H. Brown and E. Goldberg, U.S. Atomic Energy Commission, AECD-2296.l@ Idem, PhysicaZ Rev., 1949, 76, 1260.*O C. A. Tobias and R. W. Dunn, U.S. Atomic Energy Commission, AECD-11 S . A. Reynolds, G. C. Bell, and C. 0. Muelhouse, AnuZyt.Chem., 1949,21, 1214.A.E.R.E., Harwell, nr. Didcot, Berks.2099-B.BEP.-VOL. XLVT. 290 ANALYTICAL CHEMISTRY.is illustrated by R. Lindner,22 who irradiated yttrium rare earths beforepassing them through an ion-exchange column. The special case of detectionof fissionable elements by slow neutron activation (followed by detectionof fission products) must also be mentioned.The discussion so far has been concerned only with slow neutrons, forwhich in general the reaction is the straightforward capture of a neutron(n-y reaction) t o give an isotope of the same element with an increase inmass of one unit. However, other nuclear reactions can occur in the pile,more particularly if there is any appreciable fast neutron flux at the siteof irradiation.These processes can be either advantageous or otherwise.The disadvantage is that active products other than those expected mayarise due to n,p, n,a, n,Zn, etc., reactions (see, e.g., Pollard and Davidson 23for fuller discussion), thus necessitating possible modification of the chemicalmethods for isolation of the element desired and making the comparativerather than the absolute technique necessary. A discussion of the con-taminants arising from these and other factors both for pile and cyclotronirradiation is given by W. E. Cohn.24 The advantage is in the possibleextension of the method to those elements which do not give suitable isotopesby the n,y reaction. A useful example of this is the detection of oxygenor lithium. When these two elements are irradiated with slow neutronstogether, e.g., as lithium carbonate, the lithium undergoes a unique reactionOLi + ln --+ *He + 3H; the tritons so produced react with the oxygenproduced, whereas normal slow-neutron irradiation of either lithium oroxygen separately gives no suitable active isotope.Other SOurces.-Other sources, e.g., the cyclotron, with its flux possi-bilities at least as high as those in the pile, and the electrostatic generator(both with their possible variation of bombarding particle), are attractiveand have been used for activation analysis.Thus G. T. Seaborg and J. J.Living~od:~ using cyclotron deuterons, detected 6 p.p.m. of gallium in iron,and also demonstrated the presence of small amounts of copper in nickel,iron in cobalt, and phosphorus and sulphur in various substances. D.T. P.King and W. J. Henderson 26 used a-particles to activate traces of copper insilver by the a-n reaction (;$XI + :He I_, :;Ca + in and ;~CU + :He --+",Ca + in), and R. Sagane, M. Eguchi, and J. Shigata 27 used deuterons todetect 10 p.p.m. of sodium in aluminium (23Na + 2H --+ 24Na + lH).M. von Ardenne and F. Bernhard,28 using an electrostatic accelerator toproduce deuterons, determined carbon in steel down to 0.05%.Some disadvantages must be mentioned, e.g., the extreme care requiredowing to the introduction of small amounts of impurities from recoil atoms16 ,O + iH + ':F + in. Active fluorine (half-life, 112 mins.) is thus22 2. Naturforsch., I, 1946, 67.24 " The Origin, Detection, Identification, and Removal of Radioactive Con-25 J .Amer. Chem. Soc., 1938, 60, 1784.27 J . Phys. Math. Xoc. Japan, 1942, 16, 383.28 2. Physik, 1944,122, 740.25 Ref. (6), pp. 75 et seq.taminants in Tracers," U.S. Atomic Energy Commission, MDDC- 1643.28 Physical Rev., 1939, 56, 1169SPENCE : GAS ANALYSIS : CHEMICAL METHODS. 291and v~latilization.~ The heat to be dissipated, and the rather small areaavailable in using a cyclotron beam impose limitations on the type and sizeof material which can be irradiated. In general, activation cross-sectionsusing fast neutrons or charged particles are lower than those for slow neutrons ;and, for high-energy particles, the multiplicity of nuclear reactions which mayoccur simultaneously are unfavourable from the analytical viewpoint.It is not intended to deal in detail with these sources here, however, sincethey cannot as yet be regarded as being generally available.Their possi-bilities in giving more favourable nuclear reactions, in some cases, than thepile must not be overlooked, and in considering whether the activationmethod is applicable for individual cases it is worth remembering all thepossible reactions with both neutrons and charged particles. A full list ofthe isotopes with the nuclear reactions involved and the original references,is given by G. T. Seaborg and I. Perlrnax~.~~Finally, the extension of the use of this method may depend to a largeextent not only on the wider provision of suitable sources, but also on theavailability of rapid and specific isolation procedures, Le., the new toolpresented to the analytical chemist at the same time brings its own challenge.A. A.S.7. GAS ANALYSIS.(i) Chemical Methods.Considerable effort has been made of recent years towards the improve-ment of existing methods of gas analysis, and the period has also beennotable for the appearance of a number of essentially new techniques.The subject has been reviewed by V. J. Altieri 1 and the standard pro-cedures have been described in a series of articles by W. D. Vint.2 A detailedaccount of a representative selection of micro-methods is also given byK. M. Wilson in “ Methods of Quantitative Micro Analysis.”The methods of most general applicability still largely depend on pro-cesses of combustion or chemical absorption.The well-known Orsatapparatus, for instance, which was brought to a high degree of refinementby M. Shepherd,* has recently been modified by the introduction of anabsorption train consisting of a series of glass loops containing smallquantities of liquid or solid absorbents, the liquid absorbents being held ona length of wick or similar material wound round a glass rod. The gassample, which can be considerably smaller in the modified apparatus, ispassed back and forth round the loop until reaction is complete.According to H. W. Deinum and J. W. DamY6 there is it loss of accuracyin the determination of methane and ethane with the standard Orsatz0 Rev. Mod. Physics, 1948, 20, 585.New York : American Gas ASSOC., 1945.Metallurgia, 1947,35, 153, 255, 294; 36,47, 157, 276, 333; 37, 317.Collected and edited by R. F. Milton and W. A. Waters, Edward Arnold & Co.,London, 1949.4 J . Res. Nat. Bur. Stand., 1931, 6, 121.6 M. Shepherd, ibid., 1941, 26, 351. Anal. Chim. Acta, 1948,2, 50292 ANALYTIOAL CHEMXSTBY.apparatus owing to absorption of carbon dioxide by copper oxide, etc.They have shown, however, that copper oxide does not absorb carbondioxide a t 800°, nor does it give up appreciable quantities of oxygen a t thistemperature; only 0.01 ml. was picked up by a stream of nitrogen in 10minutes. The cooler ends of the copper oxide tube should therefore bepacked with quartz to avoid uptake of carbon dioxide and correctionsshould also be applied for deviations from the perfect-gas law.Othermodifications of the general design and new forms of absorption equipmenthave also been pr~posed.~The Haldane apparatus, t o which the Orsat apparatus is closely related,still continues to be widely used.8 Details of a special form of stopcockdesigned to reduce dead space are given by F. S. Cotton? together withoperational instructions leading to greater accuracy. In the case wherethe sample is completely absorbable, use of a subsidiary burette containingnitrogen, which can be introduced as a diluent in the later stages, is alsorecommended. loThe Bone and Wheeler constant-volume gas-analysis apparatus hasbeen in use a t the Fuel Research Station for many years, and L. J.Edge-combe l1 has given an account of the equipment, reagents, and analyticalprocedure which have been adopted. Other constant-volume proceduresare described by Z. Szabo and I. Soos,12 by G. Wagner,13 and by P. T.Sprague, C. A. Sprague, and A. Soller,14 who use an absorption systemcontaining steel-wool wetted with a suitable reagent. C. H. Bamford andR. R. Baldwin 15 have developed a constant-volume type of apparatus inwhich the gas is removed from the measuring bulb by means of a Toplerpump and passed round a circuit containing liquid-air traps and a copperoxide tube or a platinum spiral. After removal of the uncondensed gas,carbon dioxide can be separated from the condensate by raising the tem-perature to -78", and its pressure determined after transference to themeasuring bulb.An accuracy of -&0.03% is claimed. This apparatuscan, if necessary, be used for semi-micro-work. Other types of semi-micro-apparatus have been described by R. K. Goltz and by S. G. Demidenkoand B. A. Geller.16 Details of a modified form of the manometric VanSlyke apparatus are given by M. Shepherd and E. 0. Sper1ing.l' Thereagent is forced into the absorption vessel by mercury, producing a fountainL. L. Vaydaand J. A. Stein, B.P. 601,018/5.9.44; C. A. Sprague, Assr. to HaysCorpn., U.S.P. 2,312,285/23.2.43; P. T. Sprague and C. A. Sprague, V.S.P. 2,180,322114.11.39; C. M. Blair and J. H. Purse, Id. Eng. Chem. Anal., 1939,11, 166; A. R.Anderson, ibid., 1946,18, 70.See, e.g., H. Enghoff, Acta Medica Skand., 1946, Suppl.1'90, 307.J. Lab. Clin. Med., 1939,24, 1178.lo H. C. Bazett, J . Biol. Chem., 1941,139, 81.Fuel, 1946, 25;, 163, 171.l2 2. anal. Chem., 1943, 126, 219, 22.'* U.S.P. 2,179,867/14.11.39.l7 J , Res. Nat. Bur. Stctnd., 1941, 26,341.1s Oesterr. Chem.-Ztg, 1940, 43, 71.l5 J., 1942, 26.Zavod. Lab., 1939, 8, 1078; 1948,14, 601SPENCE : GAS ANALYSIS : CHEMIUAL METHODS. 293which gives very rapid absorption. The addition of reagents to the VanSlyke apparatus can be simplified by substituting a syringe pipette for theusual Hempel type of pipette.18An entirely new form of gas-analysis apparatus, consisting of a trainof absorption or reaction vessels each having a soap film flow-meter im-mediately following it in the train, has been developed by W.J. Gooderham.l9The gas entering the flow-meters can be either by-passed or switched to thecalibrated soap-film tube. All the soap-film tubes can be switched on oroff together, thus permitting an instantaneous reading of the volume changesoccurring in the gas due to passage through the various reagent tubes.Conventional absorbing solutions or oxidation systems are used, andthe absorption tubes consist of a. glass spiral down which the solutionflows.The determination of individual constituents of gas mixtures has beenthe subject of numerous papers. The use of the copper sulphate-p-naphtholreagent for carbon monoxide has been discussed by J. I. Tscherniaeva20and by L. B. Berger,al and a gravimetric determination utilising the reactionbetween carbon monoxide and red mercuric oxide at 175-200°22 is ofsome interest on account of the highly advantageous gravimetric factor(7.14) arising from the distillation of the mercury produced in the reaction.Determination of carbon dioxide in fuel gases by preliminary condensationin a liquid-air trap and subsequent transfer to an apparatus similar to thatof C.H. Bamford and R. R. Baldwin l5 is claimed by R. R. Baldwin 23 tobe accurate to 0.001 yo for small carbon dioxide concentrations and to0.25 % for higher concentrations. Combustion of hydrogen, carbonmonoxide, and methane by means of palladised asbestos 24 and platinisedsilica gel 25 has been recommended, but various authors 26 have reporteda reaction between platinised asbestos or a platinum wire and oxygen whenheated above 700".An error due to this reaction can be avoided if com-bustion mixtures deficient in oxygen are used. The determination ofacetylene colorirnetrically by means of the Ilosvay reagent 27 and titri-metrically after reaction with potassium mercuric iodide 28 has also beendiscussed. Fluorine has, of recent years, become increasingly importantboth industrially and in the laboratory. The determination of Auorine ingas mixtures can be conveniently carried out by displacement of brominela R. Wennesland, Skund. Arch. Physiol., 1940, 83, 201.l9 J . Soc. Chem. Id., 1940, 59, 1; Analyst, 1947, 72, 520.2o Zuvod. Lab., 1939, 8, 1092.21 U.S. Bureau of Mines, 1947, Rept. Invest. 4187; see also P. R.Thomas, L. Down,a2 J. D. XlcCullough, R. A. Crane, and A. 0. Beckman, {bid., 1947, 19, 999.24 R. Vandoni, Mdm. Sew. chim. de l'ktut, 1943,30,18, 272.25 K. A. Kobe and R. A. McDonald, Ind. Eng. Chern. Anal., 1941,13,457.L. K. Nash, ibid., 1946, 18, 505; C. E. Ransley, Anaiyst, 1947, 7$2, 504.27 H. A. J. Pieters, Chem. VeekbEad, 1947, 72, 504.and H. Levin, Arutlyt. Chem., 1949,21, 1476.J., 1949, 720.F. R. Brooks, Anulyt. Chem., 1949, 21, 1433; J. G. Eanna and 8. Siggia, ib.td.,p. 1469294 ANALYTICAL CHEMISTRY.from sodium bromide and determining the bromine absorptiometrically,29or by displacement of chlorine from sodium chloride and determining thechlorine by absorption in alkaline arsenite solution with subsequenttitrati~n.~* Other components can be determined if the second method isused since the residual gas after absorption of the chlorine can be analysedseparately.When a high degree of accuracy is required, departures from the ideal-gas laws must be considered, and corrections have been calculated byJ.J. Leendertse and F. E. C. Scheffer 31 for a number of the more importantbinary mixtures. The correction does not exceed 0.2% in the case of mostcommon gases at 50 molecular per cent. at N.T.P. but rises to 0.6% forn- butane-nitrogen mixtures.Therehave been numerous variations of the original Krogh screw-controlledburette for analyses under constant pressure. Absorption tubes for usewith aqueous absorbents are described by J. A. Christiansen and I. Wulff,32who have also introduced a quartz hair-pin capillary immediately abovethe burette.Methane is quantitatively combusted when the hair-pin israised to a bright yellow heat. In the case of the well-known apparatusof Blacet and Leight0n,3~ in which the gaseous components are absorbedon solid reagents, the chief improvements reported 34 relate to the micro-burette. That described by s. s. Burke comprises a capillary microburetteand compensation tube immersed in a thermostated container with micro-meter-screw control of the mercury. An improved combustion coil hasbeen devised by R. N. Smith and P. A. L e i g h t ~ n , ~ ~ who also give detailsfor the analysis of nitric oxide-nitrogen, nitric oxide-hydrogen, and nitrousoxide-ammonia mixtures.P. F. Scholander36 has described a simple apparatus for the micro-analysis of respiratory gases, similar in principle to that due to T.C. S ~ t t o n , ~ 'in which the absorption chamber is an integral part of the burette. Thecapillary burette is vertical and has a small horizontal absorption chamber,open to the atmosphere, attached to the upper end. Reagents and gassamples are introduced and removed by means of small glass syringes andmercury displacement is read from a micrometer screw gauge. It is claimedthat 10 cu. mm. samples can be analysed with an accuracy of 0.1%. In amore elaborate apparatus,38 for samples of the order of 0.3 cu. mm., themenisci are observed by means of a dissecting microscope. Citrate solutions29 L. Nash, U.S. Atomic Energy Commission, MDDC-2158; E.Staple, J. G.30 R. H. Kimball and L. E. Tufts, U.S. Atomic Energy Cornmission, MDDC-195.31 Rw. Truv. chim., 1940, 59, 3.3% K g l . Dumke Videlzskab. Selskab., Mat.-fys. Medd., 1945, 22, No. 4, 23.33 Ind. Eng. Chem. Anal., 1931, 3, 266.34 D. C. Grahame, ibid., 1939, 11, 351; S. S. Burke, Analyt. Chem., 1949, 21, 633.35 Ind. Eng. Chem. Anal., 1942, 14, 758.36 Rev. Xci. Imtr., 1942, 13, 264.88 P. F. Scholander and H. J. Evans, J . Biol. Chem., 1947,169,651.Micro-methods of gas analysis have attracted much attention.Schaffner, and E. Wiggin, MDDC-1610.37 J . Sci. Instr., 1938, 15, 133SPENCE : GAS ANALYSIS : CHEMICAL METHODS. 295were found to be the best confining liquids. Another apparatus of this type 39with a piece of thermometer tubing as burette has been used for the analysisof between 0.3 and 1 cu.mm. of gas. In this case the confining liquid is asaturated solution of lithium chloride.A somewhat different constant-pressure apparatus, also due to P. F.S~holander,~~ comprises two absorption bulbs, a compensating bulb, anda microburette connected t o one another and to a multi-way stopcock, allmounted in a water-bath. The mercury in the burette is controlled bymeans of a micrometer screw, whilst that in the other vessels can be con-trolled by a levelling bulb. The pressure in the burette is equated withthat in the balancing tube by means of a drop of liquid in the connectingcapillary. 0*5~-Sulphuric acid is used as washing agent, 0*25~-sodiumhydroxide for the absorption of carbon dioxide, and a mixture of sodiumhydrosulphite (dithionite) (nine parts) and sodium anthraquinone-p-sulphonate (one part), dissolved in 0*25~-sodium hydroxide, for oxygen.The use of indigo-carmine has, however, been recommended instead ofsodium anthraquinone-P-s~lphonate.~~ W.A. Nierenberg and C. Williams 42have constructed an apparatus based on that of Scholander capable ofhandling samples less than 0-6 ml. in volume with an accuracy in the caseof simple binary mixtures of &0-04 %.W. B. Price and L. Woods43 have adapted Krogh’s bubble method44for the analysis of micro-bubbles of gas occurring in glass. The bubbleis collected under glycerol and its diameter measured when i t is held undera microscope slide. It is then exposed to various absorbing solutions suchas cadmium acetate in glycerol for hydrogen sulphide, potassium hydroxidein glycerol for carbon dioxide, aqueous alkaline sodium hydrosulphite(dithionite) for oxygen, ammoniacal cuprous chloride for carbon monoxide,and colloidal palladium in saturated sodium picrate solution for hydrogen.In the case of the glycerol absorbents, the bubble is transferred by means ofa micropipette to a small horizontal glass cylinder immersed in the absorbent,and the cylinder rotated a few times to facilitate contact with the gas.Theaqueous absorbents are contained in a capillary tube along which the gasbubble is allowed to travel. E’or accurate work, however, glycerol and otheralcohols should be used with caution as confining liquids for samples con-taining soluble gases such as carbon dioxide.45Various special methods have also been developed for the analysis ofgas from biological systems.A rapid determination of one constituentof a sample of respiratory gas can be carried out in a very simple apparatusdescribed by P. F. S~holander.~~ A small bulb with microburette attachedis filled with the absorbing solution and connected to a levelling bulb bys9 W. E. Berg, Science, 1946, 104, 575.41 C. D. Stevens, P. van Fossen, J. K. Friedlander, B. J. Rattermann, and M.Inatome, Ind. Eng. Chem. Anal., 1945, 17, 598.42 U.S. Atomic Energy Commission, MDDC-529.43 Analyst, 1944, 69, 117.45 K. A. Kobe and G. E. Mason, Ind. Eng. Chern. Anal., 1946, 18, 78.46 J .Biol. Chem., 1942, 146, 169.40 Rev. Sci. Instr., 1942, 13, 27.44 Xkand. Arch. Physiol., 1908, 20, 279296 AIPN&YTTUAL CHEMISTRY.means of rubber tubing. The rubber tubing is perforated by the end of thesyringe containing the gas sample which can then be injected into theabsorption bulb. Analyses for carbon dioxide and oxygen can be carriedout simultaneously by using duplicate equipment. Techniques for theanalysis of blood gas developed by A. H. Whitely 47 and by P. F. Scholanderand L. Irving48 involve, respectively, a micro-form of the Van Slykeapparatus and a method employing centrifugation for the co&rol of thesample. The Cartesian diver device of J. Needham, V. Rogers, and S. C.Shen49 and the mica-pla.te method of N. G. Heatley, I.Berenblum, andE. Chain 5o for tissue gases do not appear to have received a more generalapplication.Constant-volume methods such as that of Bone and Wheeler have longbeen favoured for macro-scale analyses, and a micro-apparatus based onthis principle has been devised by R. Spen~e.~1 The gas sample, whichmay be either at atmospheric pressure or in a system under reduced pressure,is drawn through a three-way stopcock into a capillary burette attached toa 200-ml. bulb, by displacement of mercury. After closure of the stopcock,the sample is compressed to the 50 cu. mm. mark on the microburette andits pressure observed on a manometer connected to the bulb. It is thendriven through the other arm of the three-way stopcock into a capillaryglass loop previously evacuated by means of a small Topler pump.Theloop consists of a number of segments of capillary tubing connected togetherby waxed joints and contains two simple micro-non-return valves and oneor more small cavities for solid reagents. Oscillation of the mercury abovethe three-way stopcock causes the gas to circulate round the loop. Anyconvenient solid adsorbent can be introduced, and loops for low-temperaturecondensation or for high-temperature combustion may be used. Afterabsorption, which can be followed on the manometer, the mercury is loweredto the bottom of the 500 ml. bulb, the stopcock closed, and the mercuryonce more raised to the burette mark. Since the volume of the capillaryloop is less than 0.1% of that of the bulb, a correction need only be appliedto the second pressure reading when the highest accuracy is desired.W. L.Haden and E. S. Luttrop 52 have described an apparatus of a similar kindfor permanent incorporation in a vacuum system. The mercury bulb andmicroburette are connected, in this case, to a two-way stopcock, the otherarm of which leads to a small bulb below a cone and socket joint. A capillaryside tube with a stopcock leads off from the small bulb. Solid reagents canbe attached to the sealed off tip of the cone which projects into the upperpart of the small bulb. K. W. Saunders and H. A. Taylor 53 modified thisapparatus by introducing a special four-way stopcock and a platinum coilfor combustions, and i t has recently been further modified by C.S. Stover,W. S. Partridge, and W. M. Garrison,54 who have replaced the special stop-47 J . Biol. Chem., 1948,174, 947.48 PTOC. Roy. Soc., 1939, B, 127, 336.s1 J., 1940, 1300.68 J . Chem. Phggica, 1941,9, 686.48 Ibid., 1947, 169, 661.6a I d . Eng. Chem. A d . , 1941., 13,571.64 Analyt. Chem., 1949, 21, 1013.Biochem. J., 1939, 33, 53SPENCE : GAS ANALYSIS CHEMICAL METHODS. 297cock by a standard three-way stopcock and added a second three-way stop-cock leading into the bottom of the large mercury bulb, for admission of thesample or for evacuation. Another variation was introduced by L. K.N a ~ h . ~ ~ I n this case, the 500-ml. bulb is surmounted by a three-waystopcock one of the two upper arms of which leads to a 10-ml. bulb and a10-ml.graduated burette whilst the other is connected to the vacuum lineand to a capillary complex consisting of three absorption lines in parallel.One line contains a platinum wire catalyst for oxidation of hydrogen andcarbon monoxide at 450" and for the oxidation of methane at 950°, thesecond line contains ascarite far the absorption of carbon dioxide, and thethird line contains a trap for low-temperature condensations. Circulationof the gas is controlled by stopcocks and by movement of the mercury inthe bulb. The apparatus is intended for the analysis of samples of theorder of 1 ml. and therefore belongs to the semi-micro-class. L. E. J.Roberts and P. C. Davidge 56 have described a somewhat similar apparatusfor smaller volumes (up to 0.5 ml.) with the microburette below the three-way stopcocks leading to the absorption loops.A very simple arrange-ment is possible, with the elimination of all stopcocks, if carbon monoxideor oxygen is to be determined in binary mixtures with inert gases. Samplesmay be introduced or removed through capillary tubes of greater thanbarometric height leading into the base of the usual mercury bulb. Com-bustions can be carried out over a platinum coil situated between the bulband the microburette, the whole apparatus being arranged similarly to thestandard McLeod gauge.57Variations of the classical high-vacuum technique for the analysis ofsmall amounts of gas obtained, for instance from electric light bulbs, haverecently been described.5* C. E.Ransley separates hydrogen by diffusionthrough a palladium tube at 700", and carbon monoxide and methane aredetermined by combustion over a platinum wire at 500" and 1150°, respec-tively. Kenty and Reuter remove hydrogen and carbon monoxide byignition with oxygen, the excess of oxygen subsequently being determinedby reaction with a heated tungsten filament.Numerous chemical methods for the determination of minor or traceconstituents of gas mixtures have also been reported, as, e.g., the determi-nation of small amounts of hydrogen s ~ l p h i d e , ~ ~ the oxides of sulphur,w theoxides of nitrogen,61 and hydrogen cyanide.62 A particularly valuables5 I d . Eng. Chem. Anal., 1946, 18, 505.66 Atomic Energy Research Establishment Report No. C/R.470.67 F. C . Tompkins and D. M. Young, private communication.8s C. E. Ransley, Analyst, 1947, 72, 504 ; C. Kenty and F. W. Reuter, Rev. Sci. Instr.,1947,18,918 ; see also C. H. Prescott and J. Morrison, Ind. Eng. Chem. Awl., 1939,11,230.6s H. A. J. Pieters, Chem. Weekblud, 1947, 43, 455; E. Field and C. S . Oldbach,Id. Eng. Chem. Awl., 1946, 18, 665.6o E. W. F. Gilham, J. SOC. Ohem. I d . , 1946, 65, 370.61 R. Kieselbach, I d . Eng. Chem. Anal., 1944,16,766; J. F. Flagg and R. Lobene,US. Atomio Energy Commission, MDDC-971.H. F. Taylor, Gr, J., 1947, 252,293298 ANALYTICAL CHEMISTRY.method for the determination of small quantities of carbon monoxide in air,which was originally developed at the Royal Aircraft Establishment,Farnboro~gh,~~ depends on the colour change occurring in a tube of silicagel impregnated with ammonium molybdate and a palladium salt.Themethod, as subsequently modified by the U.S. National Bureau of Standards,wis capable of the detection of less than one part of carbon monoxide in5 x lo8 parts of air in twenty minutes or of physiologically significantquantities in 1-5 minutes.The use of impregnated filter discs for the qualitative detection of gasessuch as arsine and stibine in presence of hydrogen sulphide has been describedby C. L. Wilson,65 and the Feigl-Rossler apparatus for qualitative micro-gas analysis has been improved by R. Belcher.66 Detection and deter-mination of gases evolved in the analysis of carbonates, oxalates, sulphides,sulphites, etc., can be carried out in a simple apparatus due to J.G. Reynolds,67in which a slow stream of air is aspirated over the test solution in a V-shapedtube through a small volume of reagent in a tube inserted in one arm ofthe V. R. S.(ii) Physical Methods.There have been a number of interesting developments in the applicationof physical methods, and several general reviews dealing with the subjecthave appeared. Since the applications of infra-red absorption and of massspectra to gas analysis were last mentioned in the 1946 Report, this seemsto be a good occasion for bringing these subjects up to date and for mention-ing several other useful techniques.give a completereview of this subject with bibliography up to the end of 1948. Since thenthere has been a further general article by M.Boivin,' a description of aninstrument for continuous gas analysis with high-speed response,8 andfurther work on the combination of low-temperature fractional distillationwith mass ~pectrometry.~ C. W. Key10 describes the rapid analysis ofstack gases for totaI sulphur and sulphur dioxide, and the analysis of hydro-carbon gases by a combination of infra-red and mass spectrometry has beendiscussed by D. Milsom, W. R. Jacobi, and A. R. Rescorla.lfThermal Conductivity.-Recent work has consisted of modifications of64 Bnalyt. Chem., 1947, 19, 77.6 6 Metallurgia, 1947, 35, 310.MWS Spectrometry.4, A. Hipple and 141. ShepherdJ. D. Main Smith, R.A.E. Report CH. 324, Aug. 1941.6 5 Analyst, 1940, 65, 407.67 Ibid., 1948, 37, 160.R.H. Miiller, I d . Eng. Chem. Anal., 1941, 13, 667.G. Wagner, Oesterr. Chem.-Ztg., 1941, 44, 176.A. L. G. Rees, Austral. Chem. Imt. J . and Proc., 1947, 14, 23.W. A. Cook, Amer. Id. Hyg. Assoc. Quart., 1947,8,42.H. A. J. Pieters and T. W. van Dam, Het Gas, 1948,68,199.Analyt. Chem., 1949, 21, 32.J. A. Hunter, R. W. Stacy, and F. A. Hitchcock, Rev. Sci. Instr., 149,20,33,331.C . E. Starr, J. S. Anderson, and V. M. Davidson, Analyt. Chem., 1949, 21, 1197.Calif. Oil World, 1949, 42, No. 4, 3, 6 , 7, 25.Analyt. Chem., 1949, 21, 547.Chim. Anal., 1949, 31, 80SMALES : GAS ANALYSIS : PHYSICAL METHODS. 299the previous methods or of adaptations to new problems; thus the thermalconductivity of a gas before and after combustion has been used, e.g., forcombustibles in natural gas after addition of oxygen,12 for oxygen in itsmixture with one or more gases such as CO, CO,, or CH, after addition ofhydrogen,13 and for hydrogen with addition of oxygen. R.H. Cherry l4has discussed general features involved in the determination of water vapourby this method, which was a t one time considered to be rather unpromisingowing to the occurrence of a thermal conductivity maximum in watervapour-air mixtures. As Cherry points out, this only precludes operationin the range 1 2 4 7 % of water vapour by volume and there are manyoccasions involving determination outside this range. Similar maxima arefound for water vapour in nitrogen and oxygen and might be expected incarbon monoxide and possibly in acetylene, ethylene, and ethane, but arenot to be expected in mixtures other than these.R. Edse and P. Harteck l5have pointed out the advantages of thermal conductivity over gas densityafter using the desorption technique; smaller quantities of gas and activecharcoal are necessary and the gas is recoverable; they illustrate this inthe analysis of hydrogen-deuterium and other isotope mixtures. C . A.Hansen,lS H. A. J. Pieters,l7 F. Lieneweg,ls and C. C. Minter have alldiscussed the industrial applications, and the last author 2o has also dis-cussed the effects of pressure changes on thermal conductivity. W. J.Clark 21 suggests a device for automatic compensation for such pressurechanges. The use of thermal conductivity for analysis of gases associatedwith internal-combustion engines has been the theme of several patents:,and the application to fluorine-nitrogen mixtures 23 is another example ofthe wartime emergence of fluorine as an industrial gas.Heat of Combustion or Reaction.-The heat of reaction when a gasburns a t a filament or in the presence of a catalyst has frequently been usedfor the determination of combustible gases, e.g., exhaust gases from internal-combustion engines,,, oxygen in flue 25 and other gases 26 after addition ofhydrogen, and low concentrations of carbon monoxide in air.27Electrolytic Conductance after Absorption.-The main application ofthis method is 'in the determination of carbon dioxide after absorption inl2 R.Weber, U.S.P. 2,399,96517.5.46.l3 G.A. Perley and J. B. Godshalk, B.P. 567,974112.3.45.l4 Analyt. Chem., 1948, 20, 1033.l5 Angew. Chem., 1939, 52, 32; 1940, 53,210.l6 @en. Elect. Rev., 1940, 43, 166.lE Arch. tech. Messen., 1942,138, T 125; 1943, 140, T 17.l9 J . Chem. Educ., 1946, 23, 237.2o Analyt. Chem., 1947,19, 464.22 D. E. Olshevsky, U.S.P. 2,154,862118.4.39; W. J. Willenborg, U.S.P. 2,255,551123 E. Staple and E. R. Grilly, U.S. Atomic Energy Commission, MDDC-1565,24 B. Miller, U.S.P. 2,152,439/28.3.39; 2,219,540/29.10.40.25 A. P. Sullivan, U.S.P. 2,310,472/9.2.43.26 G. Cohn, Analyt. Chem., 1947, 19, 832.27 M. Katz and J. Katzman, Canad. J . Res., 1948,26, F, 318.l7 Chem. Weekblad, 1940, 37, 316.2 1 U.S.P. 2,472,64517.6.49.9.9.41 ; H. Laub, U.S.P.2,256,395/16.9.41300 ANBLYTICAL CHEMISTRY.barium hydroxide solution, either directly or indirectly as in the deter-mination of hydrocarbon gases after combustion with oxygen 28 or for watervapour and oxygen in gases containing no other oxygen compounds, bypassing through charcoal a t lOOO", giving carbon monoxide which is thenpassed over iodine pent~xide.~~ An automatic apparatus for sulphurdioxide has also been described.30 An interesting method for the deter-mination of water vapour in gases has been described by E. R. Weaver andR. Riley31 depending on the difference in conductivity of a thin film ofphosphoric acid with alteration in the water vapour content of gas flowingover it. The extension to the determination of oxygen in combustiblegases after combustion is suggested.Ionization Potential.-Unlike many of the other physical methods, theionization potential of a gas is an almost specific property and a number ofpatents using this principle have been listed.32Polarography ebnd Amperometric Titration.-This method is an obviousone for the determination of oxygen in industrial gases, and P.Beckmann 33has described continuously indicating polarographic method for oxygencontained in the gas obtained during the carbonization of oil shale.D. W. E. Axford and T. M. Sugden 34 make use of amperometric titrationfor the determination of sulphur trioxide in its mixtures with dioxide ; afterabsorption of the gases in sodium hydroxide solution, the sulphur dioxideis removed, after acidification, by a stream of nitrogen, the remainingsulphate being titrated amperometrically with lead nitrate solution.Magnetic Susceptibility.--H.Rein 35 has described a continuous methodfor the determination of oxygen depending on the decrease of thermalconductivity of oxygen in a magnetic field. The arrangement is similarto the usual thermal conductivity method in that the gas stream is dividedinto two channels each with a heated platinum wire connected in a Wheat-stone bridge arrangement. The extent of disequilibrium on application ofa strong magnetic field to one of the channels is a memure of the oxygencontent. Applications have also been discussed by L. Pauling36 andN. S ~ h w a r z . ~ ~Emission Spectr~sco~y.-As W. F. Meggers says in a recent review,38(' recent attempts to detect and determine gases spectrographically aresolely represented by experiments with halogens excited either by ultra28 G.I;. Hassler, U.S.P. 2,230,59314.2.41; Csn. P. 395,346118.3.41; B.P. 537,486138 N. Shurmovskaya and L. Kupriyanov&, Zhur. Anal. K h h . , 1948, 3, 41.80 M. D. Thomas, J. 0. Ivie, and T. C. Fitt, Ind. Eng. Chem. Anal., 1946,18,383.81 J . Res. Nat. Bur. Stand., 1948, 40, 169.92 W. Jaeger, D.R.-P. 696,05418.8.40;38 Chem. and Id., 1948, 791.86 U.S.P. 2,416,344125.2.67.87 Applied Sci. Research, 1947, A, I, 47.88 Analyt. Chm., lM9, 21, 29.24.6.41 ; Dutch P. 52,836115.7.42,A. P. Solovov, Russ. P. 67,531131.7.40;L. T. Winkler, U.S.P. 2,387,550/23.10.45.J., 1946, 901.D.R.-P.742,690121.10.43; Zen&., 1940, I, 2204SMALES: GAS ABNALYWS: PHYSICAL METHODS. 301high frequency electric fields 39 or in hollow cathode discharges.” Thedetermination of as little as 0.01 fig. of fluorine and 0.2 pg. of chlorine iscited in the latter paper. Another method for fluorine, vix., evolution assilicon tetrafluoride followed by spectrographic determination of the silicon,has been described by R. W. Spence:41 1 pg. of fluorine was the lower limitfor determination although 0.01 pg. was detected ; the greatest difficultyin attaining such sensitivity was the reagent blank.Raman spectroscopy has been used but little in gas analysis.Absorption Spectroscopy.-(a) X-Ray. Only within recent years hasit become possible to measure X-ray absorption precisely and convenientlyand it is likely that it will assume increasing importance.Preliminarystudies using a photomultiplier tube as detector with a polychromaticX-ray beam have been reported on hydrogen, methane, air, oxygen, andmethyl and a general review by H. A. Liebhafsky43 should benoted.Further applications of this technique,usually with simple filter instruments, fall into two general types : absorptionof light (i) by the gas itself 44 (e.g., nitrogen peroxide) or after reaction toform another gas (e.g., oxygen by reaction with nitric fluorine byreaction with sodium bromide 46) or (ii) after reaction of the gas in solutionto give a light-absorbing product (e.g., sulphur dioxide by reduction ofchromate!’ oxygen by reaction with reduced sodium anthraquinone-p-sulphonate) .48The appearance of the General Electric Company’s mercury-vapourdetector, and the extension of this principle to a general purpose unit forlight-absorbing vapours 49 should be noted.(c) Infra-red.The outstanding advance in this connection has beenthe development of industrial gas analyzers using no dispersing system.These are very sensitive and capable of detecting selectively a few partsper million of many gases or vapours which have strong absorption in theinfra-red.H. W. Thompson 50 has summarized the basic features of these instru-(b) ‘Visible and uttra-uiokt.9D A. Gatterer and V. Frodl, Richerche Spettroscop., 1946, 1, 201; Spectrochim,4O J. R. McNally, G. R. Harrison, and E.Rowe, J . Opt. SOC. Amer., 1947, 3’9, 93.4l U.S. Atomic Energy Commission, MDDC-310.42 E. H. Winslow, H. M. Smith, H. E. Tanis, and H. A. Liebhafsky, Analyt. Chem.,4s Ibid., 1949, 21, 17.44 I. N. Kuzminyka, E. Ya-Turkan, and E. I. Savinkova, Zauod. Lab., 1941, 10,46 D. G. C. Eare, U.S.P. 2,389,046/13.11,45.46 L. K. Nash, Analyt. Chem., 1949, 21, 980.47 P. V. Moskalev, Lab. Prakt. (U.S.S.R.), 1940,15, 26.48 L. J. Brady, Analyt. Chem., 1948,20,1033.4e M. B. Jacobs, “Analytical Chemistry of Industrial Poisons, Hazards and6o Ann. Repwts, 1945, 42, 16,Acta, 1948, 3, 214.1947, 19, 866.139; R. H. Parker and J. K. Dixon, U.S.P. 2,417,321/11.3.47.Solvents,” Interscience Publishers Inc., New York, 1941, p. 375302 ANALYTICAL CBEMISTRY.ments, which in general may be classed either as the positive 51 or thenegative filter type, and instruments of both types are now availablecommercially.R. D. Miller and M. B. Russell 53 have suggested the name" autodetector " for the positive filter type and describe a simple laboratory-constructed analyzer for minimising drift, one of the major difficulties ofthese gas analyzers. They use a drop of fluid in a capillary for observingpressure-volume changes in the detector cell and measure its position byuse of a photoelectric position indicator effectively amplifying its motionabout 200 times, enabling changes of less than 10 p.p.m. in the range0-300 p.p.m. of carbon dioxide in air to be detected.H. W. Deinum 54 has described the use of the Baird Associates infra-red gas analyzer for CO,, CH,, CO, H,O, c6H6, and C,H,*CH,, and a com-parison has been made of the infra-red and the iodine pentoxide methodsfor the determination of carbon monoxide in 1lline-damp.5~ A review givingnumerous references to infra-red analysis for specific gases has been givenby R.B. Barnes and R. C. Gore,56 and more recently the adaptation of thePerkin Elmer instrument for continuous determination of six differentcomponents has been de~cribed.~'(d) Micro-wave. Largely as a result of wartime developments, work inthe micro-wave region, Le., 0.3-20 cm.-l, has become a practical possibility.Just as the electronic spectra of molecules occur largely in the visible andultra-violet, and vibrational spectra in the infra-red, so the rotational-energy spectra are observed most conveniently in this micro-wave region.As yet little analytical work has been carried out, but a considerable numberof laboratories, particularly in the United States, are equipped with micro-wave spectrographs for the study of molecular structure and the spins andmoments of nuclei.However, the technique is particularly suitable for,and in fact at present is limited to, gas analysis, although of course moleculeswith no dipole moment, such as carbon dioxide and methane, cannot bedetected and those with dipole moments of less than 0.1 Debye unit offerserious difficulties. The particular attraction of this region lies in the factthat the resolution available is so great that interferences from overlappingspectra are almost completely eliminated.Equipment for micro-wave spectro-scopy is still in an early stage of development ; nevertheless, it seems to theReporter that the industrial analyst should certainly consider its possibilitiesand create the demand for commercial instruments. One useful referencemay be given as a starting point for further reading.58Fractional Distillation.-This is one of the most useful physical methodsavailable to the gas analyst, particularly where complex mixtures of hydro-carbons are concerned. There have been a number of improvements in5 1 K. F. Luft, 2. tech. Physik, 1943, 24, 97; F. I. Callisin, Nature, 1947, 159, 167.62 W. 8. Baird, J. Opt. SOC. Amer., 1945, 35, 7998.53 A d y t . Chem., 1949, 21, 773.55 A.Jager and W. Grebe, Gliickauf, 1949,85, 294.56 Analyt. Chem., 1949,21, 7.5 7 J. U. White, MI. D. Liston, and R. G. Shard, ibid., p. 1156.68 B. P. Dailey, ibid., p. 540.54 Rec. Trav. chim., 1948, 67, 725SMALES : GAS ANALYSIS : PHYSICAL METHODS. 303apparatus, particularly in new types of packing for the fractionation columnand there has been a recognition of the limitations of the distillation methodof analysis with a tendency towards combination of this process with variousphysical, particularly infra-red and mass spectroscopy: and chemicalmethods. A. Rose 59 has fully reviewed this subject recently.Acoustical Methods.-The acoustical gas analyzer is based on theprinciple that the velocity of sound in a gas is a function of the averagemolecular weight of the gas.The apparatus described by C. E. Crouthamel and H. Diehl 6o is typicaland uses an audio-frequency oscillator to generate a signal operating a smallspeaker placed at one end of a brass tube through which the gas sampleflows. A sensitive microphone at the opposite end of the tube gives amaximum or minimum signal depending on the resonance conditions of thetube, which in turn are related to the average molecular weight of the gasby the relationship f = kyT/2M, where f is the natural frequency of theresonator, k. is a constant depending on the dimensions of the resonator,y = c3-/cv, i.e., ratio of specific heats of the gas mixture at constant pressureand volume, T is the absolute temperature, and 2M the molecular weight.Thus if T is maintained constant, the natural frequency is a function primarilyof the average molecular weight. The method is useful for effectively binarymixtures and has been found to work well for hydrogen in a mixture withair or the common gases, and for carbon dioxide in air, but is less sensitivefor methane, oxygen, and ethylene, and insensitive for carbon monoxidein air. The application to mixtures of helium, oxygen, and nitrogen hasalso been described.61The optical acoustical method described by M. Vengerov,62 dependingon the absorption of light in gases, is similar in principle to the infra-redgas analysers previously mentioned.Gas Density.-A critical study of eleven commercial instruments fordetermining specific gravities of gases has been made by the National Bureauof standard^.^^Interferometry.-A few further applications of this well-known methodcan be noted; 1%. A. Patty 64 has determined organic vapours in gases,A. J. Anthony 65 gives a description of the technique of gas analysis usingthe Zeiss laboratory interferometer, and H. Dierkesmann 66 describes theexact analysis of oxygen-nitrogen mixtures.Diff usion.-This method is particularly applicable to hydrogen and onepaper in particular 67 should be noted. This describes a universal gaso-61 W. B. Dublin, W. M. Boothby, and M. D, Marvin, Science, 1939, 90, 399; Proc.Stag Meetings Mayo Clinic, 1940, 15, 412.62 Compt. rend. Acad. Sci. U.R.S.S., 1938, 19, 687; 1946, 51, 195; Nature, 1946,158, 28; Zavod. Lab., 1947, 13, 426.6s Smith, Eiseman, and Creitz, U.S. Bur. Stand., Misc. pub. M. 177 (1947).64 J . Ind. Hyg. Toxicol., 1939, 21, 469.65 2. ges. exptl. Med., 1939,106, 561.6 7 L. P. Pepkowitz and E. R. Proud, Analyt. Chem., 1949, 21, 1000.Analyt. Chem., 1949, 21, 81. 6o Ibid., 1948, 20, 515.66 Ibid., 1940, 107, 736304 ANALYTICAL CHEMISTRY.metric micro-method for the determination of hydrogen in organic, inorganic,and metallo-organic compounds or low-melting metals. Reduction bysodium or magnesium metal is carried out in a sea.led iron capsule throughthe walls of which the evolved hydrogen diffuses into a vacuum system.Miscellaneous.-Miscellaneous work on several other techniques hasbeen described, e.g., analysis by electron scattering,68 by vapour pressure,69€or hydrocarbons by adsorption on active charcoal, 70 followed by fractionaldesorption; for organic sulphur compounds by adsorption on silica gel 71followed by hydrogenation to hydrogen sulphide and colorimetric deter-mination of this; and there is an important section concerning the analysisof gases evolved from metals by vacuum extraction and fusion.72A. A. S.G. INGRAM.H. M. N. H. IRVING.A. A. SMALES.R. SPENCE.W. A. WATERS.L. A. WOODWARD.S. S. West, Ceophy&ics, 1943, 8, 404; J. Hillier, U.S.P. 2,468,261/26.4.41.@@ J. Smittenberg, Rec. Trau. chim., 1948, 67, 703.?* K. Bratzler, Oel u. Kohle, 1943, 39, 953; N. C. Turner, U.S.P. 2,398,817-8/7L J. K. Fog0 and M. Popowsky, Analyt. Chem., 1949, 21, 773.?* D. Lipkin and M. L. Perlman, U.S. Atomic Energy Commission, MDDC-294;C. N. Rice, MDDC-356; L. Brewer, MDDC-366; R. L. Seifert, L. 0. Gilpatrick,T. E. Phipps, and 0. C. Simpson, AECD-2331; Holm, J. Res. Nut. Bur. Stand.,1941, 28, 245; T. Somsiya, J . SOC. Chem. In&. Japan, 1942, 45, 183; G. W. Keilholtzand M. J. Bergin, Instruments, 1949, 22, 320; P. Klinger, Arch. Eisenhiittenw., 1949,20, 151 ; I€. F. Beeghly, Analyt. Chem., 1949,21, 241.23.4.46