INORGANIC CHEMISTRY.1. INTRODUCTION.INORGANIC chemistry is slowly-too slowly perhaps-changing from adescriptive and preparative science to one concerned with valency, structure,and reaction mechanisms-reaction mechanisms which frequently have cometo show many stages and to involve transient atomic groupings. Speakingof these entities in his Presidential Address, C. N. Hinshelwood says " Anew inorganic chemistry of such equilibria is now due, in which the theory ofstructure will become more closely linked than ever to the theory of kinetics.The latter has, so far, often taken the lead in indicating the existence ofunexpected species in small concentration (for example N,O,, S,O,, and soon) but perhaps the compliment may now be returned and a more carefulstudy of the inorganic chemistry of unstable species may help us to predictthe most probable kinetic mechanisms in examples as yet unexplored ".This engaging subject of reaction intermediates from the inorganic angleprovides the theme of the second part of this report : the more generalearlier part deals briefly with other topics of current interest.P.L. R.2. GENERAL.The literature has contained some articles of a general nature whichshould be mentioned. Vapour-pressures for about 300 inorganic compoundshave been tabulated by D. R. Stull,2 E. Rinck has reviewed the equilibriumbetween metals and fused salts, and U. R. Evans oxidation and corrosion.The importance of structure as an essential adjunct to chemistry has beenstre~sed.~5 6,Nuclear physico-chemistry up to 1941 has been dealt with by J.Mattauchand S. Fluegge,s whose comprehensive tables are preceded by a text whichrenders them intelligible to the non-specialist. I n the Liversidge lecture for1946, H. C. Urey treated a special aspect of nuclear chemistry in reviewingthe thermodynamic properties of isotopic substances. These are applicableto the concentration and separation of isotopes and provide an insight intothe temperature conditions in remote geological time.Missing EEements.-F. A. Paneth l o has provided a needed summary ofthe work involved in the discovery and properties of elements 43, 61, 85, 87,93, 94, 95, and 96. It willbe generally accepted that the right to name should rest with the discovererof the first isot,ope (naturd or artificial) but that a name should not surviveRules for naming a new element are advanced.J., 1947, 698.Ann.Chine., 1945, 20, 444.J. 10. Bernal, J., 1946, 643.J., 1947, 562.I n d . Eng. Chein., 194i, 39, 540.J., 1946, 207.A. F. Wells, Nature, 1945, 155, 468.5 J. W. Smith, Science Prog., 1946, 34, 764.* " Nudear Physics Tables ", Interscience, N.Y., 1946.lo Nature, 1947, 159, 8DODD AND ROBINSON : GENERAL. 51repudiation of the discovery. It is not always easy to decide between rivalcla.imants and there should be some court of aesthetic reference. Ifmasurium represents an unfortunate observation, albeit in good faith,nevertheless the name is preferable to technetium, regrettably cacophonous,yet fulfilling all the above criteria.Thenames proposed for them are : 43, technetium, Tc (to replace masurium)from the Greek, 2 ~ ~ ~ 6 5 , artificial, being the first artificially made element ; l185, astatine, At from the Greek diarcr~o~, unstable, being the only halogenwith no stable isotope; l2 87, francium, Fr (to replace actinium-K whichis only suitable for one isotope occurring in a natural radioactive family),proposed by Mlle.Perey , who was responsible for its undisputed discovery.13Examples of the generation and subsequent history of these elements are asfollows :There is little to add to the chemistry of elements 43, 85, and 87.a-decay @-decay2i;Ac ----+ ';,3Fr (Ac-K) 2i:Ra (Ac-X). 13.5 hrs.Isotopes of technetium also occur in the fission products of uranium, a5U.15The most stable occurs thus :&sion B--dec:iy f?--clecay 23aU -----+ ",&to -------+ :9,Tc zo'yrs./ ;:Ru (stable).6.274, 67 hrs.Atomic Weighb.-The twelfth and thirteenth reports of the committeeon Atomic Weights of the International Union of Chemistry are nowavailable ; l6 modifications are restricted to sulphur and copper.That the determination of isotopic abundance ratios can be of highaccuracy l7 is illustrated in the tables of Mattauch and Fliigge.8 Theinternational reports quote mass-spectrographic determinations of theatomic weights of, e.g., potassium, silver, and radium which are in excellentagreement with gravimetric values. M. G. Inghram has determined theabundance ratios for boron, silicon, and tungsten, from which, with dueregard to packing fraction and to the Smythe factor of 1.000275 for conversionl1 C.Perrier and E. Segr6, Nature, 1947, 159, 14.l3 J . Pkp.9. Radium, 1939,10, 435, 439, and ref. (10) where Paneth quotes theproposed name from private communication with Mlle. Perey.la B. Karlik and T. Bernert, Naturwiss., 1946, 33, 23.l5 " The Plutonium Project ", J. M. Siege1 (Ed.), J . Amer. Chern. Soc., 1946,68, 2411.l6 G. P. Baxter, M. Guichard, 0. Honigschmid, and R. Whytlaw-Gray, J . , 1947,1' Ann. Reports, 1946, 43, 315.D. R. Corson, K. R. Mackenzie, and E. Segri., ibid.980 ; G . P. Baxter, M. Guichard, and R. Whytlaw-Gray, ibid., p. 983.Physioal Rev., 1946, 70, 65352 INORGANIC CHE:MIHTR.Y.from the physical to the chemical atomic-weight scale, the chemical atomicweight may be calculated.Boron : IsotopeAbundanco, 0; ...................................Silicon : IsotopeAbundance, 0; (Inghram) ..................(D. Williams, P.Yuster) 30 ......(E. P. Ney, J. H. McQueen) l9 ...Tungsten : Isotope 189. 182.Abundance, 3; ......... 0.122 25-77183.1 i.24LO.18.8329.4.672.694.68182.30.68I I .Sl.1';1 Yci.29.17Atomic weight.10.821 (Calc.)10.82 (1917)2s-086 (Calc.)28.08i (Calc.)28-06 (1947)The new " pykno-X-ray method ", developed after the method employedfor fluorine by C. A. Hutchison and H. L. Johnston,21 is noteworthy. Itinvolves a determination of the ratio of molecular weights of some crystallinecompound of the element, say lithium fluoride, and of some crystallinereference compound such as calcite.This is derived from the densities ofthe two substances and the cell dimensions. T n this way T. Batuecas 2zobtained a value of 28-07.5 for the atomic weight of silicon by coniparison ofsilica with both calcite and rock-salt. Internal accuracies of thc order of1 part in 400,000 parts are obtainable, but the ultimate accuracy dependsupon that of the reference atomic weights. Batuecas's determination hasthe advantage that SiO,: CaCO, and SiO, :NaCl involve oxygen as theonly other element common to both ratios, which, nevertheless, yieldconcordant values. The similar determinations by C. A. Hutchison,D. A. Hutchison, and Johnston23 of the atomic weights of calcium andfluorine, though yielding concordant values, have either carbon, or lithiumand carbon, as common elements in each set of four ratios measured.Electron-dejicient i l l ~ l e c u l e s .~ ~ ~ 25-Further interest in the hydroboroiw,borohydrides, and allied compounds has been mainly dcvoted to cx tcndiiigH H H'\ / /'(I.) \B BH H\ @ B \B./H(111.)H H H\e/ 8' B ( t \ - . )H/ '\ \H Hthe appliwtjiorr of the hydrogeii-bridge h k , but there rt.iiiains in questioii t IN'nature of the link between the two boron atonis in diborniic. The argumentsl9 E. P. Ney and J. H. McQueen, Phjsicol Rec., 1946, 69, 41.2o D. Williams and P. Yuster, ibid., p. 556.21 J . Amer. Chem. Soc., 1941, 63, 1580.2s Physical Rev., 1943, 62, 3 2 ; 1944, 66, 144; J .Chena. Physics, 1942, 10, 489;24 Cf. Ann. Reports, 1943, 40, 62.22 T. Batuecas, Suture, 1947, 159, 706.1945, 13, 383.2 s Cf. ibid., 1945, 42, 67inay be summarised : It. P. Bell and H. C. Longuet-Higgins 26 proposed aresonance involving canonicals, such as (I)-( IV), and K. S. Pitzer 27 suggesteda " protonated double bond " which may be described as a a-bond betweenthe two borons, plus a x-bond in whose antinodes are imbedded two protons.Recently, A. D. Walsh 28 has suggested a third formulation (V), similar insome respects to the generally accepted structure (VI), of Al,C16, in that theelectron pair involved in the weak covalent link to the inner hydrogen atomsis supposed to be iitilised also for the co-ordinate link. Walsh notes that theconception that bonding electrons are capable under certain conditions offorming co-ordinate links, embraces the idea of " x-bonds " put forwardindependently by M.J. 8. D e ~ a r . ~ ~ A further suggestion,30 whereby each(11 C:l x x x\ / \ /M M I? x-yx '\*( L /' A1H/ \ / ''x c1 X X \ / \ / \H H H C1 c1(V.) WI.) (VII.) (VIIT.)bridge-hydrogen atom is supposed to use an electron-pair and its 1s orbitalin forming two " half bonds ", would appear to be identical with the originalresonance bridge structure (I) and (11). Without implying bias in favourof either formulation the bond may he conveniently written as (VII).Some ground has been cleared by the examination which Bell andLonguet-Higgins 31, 32 have made of the normal vibrations of the bridgedmolecule (VIII), and in which they find quantitative agreement with theRaman spectruni of liquid diborane33 and the infra-red spectrum of thevapour.34 This agreement is only obtained when using their symmetricalbridge model and not when using the ethane-type model, thus reducing theprobability35 of an electrostatic hydrogen bond, B-H .. . B, which wouldnecessarily dispose the hydrogens unsymmetrically between the two borons.Incidentally, Pitzer 36 concludes on theoretical grounds that entropymeasurements can make no useful contribution to the solution of the problem.The extension of the bridged-link (protonated or resonating) to theallied molecules, (a) higher hydroborons, ( b ) metallo-borohydrides, ( c ) covalenthydrides, and (d) metal alkyls, has been made as follows :(a) Pitzer 27 has shown that the higher hydroborons can be formulatedfrom the structural units BH,, BH,*BH,, BH,*BH*BH,, etc.-the so-calledhorines.Each hydrogen in each borine is potentially capable of formingit hydrogen-bridge link between two such units, and does so subject to thefollowing three conditions, (i) such a group is no longer free for the purpose? G *7., 1943, 250.!!& *7., 1947, 89.30 H. E. Rundle, J . Amer. Cllteni. SOC., 1917, 69, 1327.31 Proc. Roy. SOC., 1945, 183, A , 357.33 T. 1'. Anderson and A. B. Burg, J . Client. Physics, 1938, 6, 586.:!4 E'. Stitt, ibid.. 1941, 9, 780.36 . I . Attier. Chen,. SOC., 1947, 69, 194.2 7 , J . - 4 t n p r . C'ttew. S o c . . 1945, 67, 1126.49 J ., 1946, 406.32 Nature, 1948, 155, 328.35 A. Bumwoy, Nature, 1945, 155, 32'854 INORGANIC CHEMISTRY.of linkage if it becomes adjacent to two bridge links [as Bd‘ in the structure(IX) of B,H,], (ii) each boron participates in one bridge only, and (iii) ringscontaining less than five members are unstable. Thus all the hydroborons inStock’s original classification can be formulated and no others are predictedwhich would fall within the adequately studied range. High-molecular-weights polymers are known which may be assigned formulae (BH,*BH,),, or(BH,*BH,), + 2BH3.Longuet-Higgins 37 has pointed out that Stock’s classification has thesignificance in Pitzer’s theory that the more stable B,H, , , members possessonly two borine units each and include all but one of the cyclic structures,whereas, the less stable B,,H,, t 6 types comprise three borine units.Theproperties of the hydroborons, so far as they are known, fit in with these ideas,though information is as yet largely restricted to diborane. The structure( I X) of B,H, does, however, agree well with electron diffraction data.3*HRH2\ 14\e /” H \ A13i--BH2\B/B*\,/H HH /B<T- \H ( X - 1HI Li@ IHIH H /B\H H /(XI.)B,HIH\B/B*\,/H I I H\ A/” H \tH,H’I(XI.) BH,-Be-B-H”/ \H’ H’H’I(b) The preparation and properties of the borohydrides of lithium,beryllium and aluminium were reported in 1941,3B as were the electron-diffraction measurements of J. Y. Beach and s. H. Bauer *O on the aluminiumcompound, AlB,H,,.Where the chemical evidence indicates the covalentcharacter of that compound in contrast to the marked ionic character ofLiBH,, the structures (X) and (XI) might reasonably be e~pected.~’G. Silbiger and Bauer 41 have, however, observed the electron diffractionof beryllium borohydride, and re-examined the corresponding data of Beachand Bauer for the aluminium compound in the light of the suggested bridgestructures (as XI). Although confirming that the Al, Be, and B atoms aredisposed as shown and that the bridge structures predict correct positions ofmaxima, the overall picture is, in their opinion, better interpreted on thebasis of structures (XII) and (XIII). I n each BH, group, three hydrogenatoms (H’) form a symmetrical girdle about the straight line joining themetal, the boron, and the fourth hydrogen atom(H”).I n Be(BH,),, they findthat Be-B-H’ = 85” & 5” and the distance Be-B is. 1.66 & 0-04 A . Here, /\3i J . , 1946, 139.3 8 S. H. Bsuer and L. Pauling, J . Amer. Chena. SOC., 1934, 58, 2403.39 Ann. Reports, 1941, S, 65.40 J . Amer. Ghem. SOC., 1940, 62, 3440. * I Jbid., 1946, 68. 312DODD AND ROBINSON : GENERAT,. 55at least, i t appears that a resonance formulation, involving ionic andone-electron bonds, is appropriate. I n this connection, it would beinteresting to have the structure of the analogous compounds dimethyl-gallium borohydride, GaMe2BH,,24? 42 and dimethyldiborane.( c ) Hydrides of gallium, indium, thallium, and aluminium have beenreported, the first three being volatile, and the last non-volatile, and nodoiibt polymerized, (AlH3),.241 43 Longuet-Higgins and Bell 26 haveadvanced formula (XIV) for the volatile hydrides, and Longuet-Higgins 37has since put forward a two-dimensional giant molecule arrangement (XV)for (AlH3)n, the aluminium atoms being disposed in layers a t the verticesof a hexagonal tessellation and joined together by hydrogen bridges.H H(XIV.) \M-H-M/ / I1 \ (M = Ga, In, Tl.)H' 'H/AI--:-A~H / \ H\ /~i-;-~i€I/ \ n\Al- \H / /=\ H-A1 / H ~i-;-~i"\H/ \I3/ u AI-E-A~ H \H / \ HAl- /= H \ /" "\ AI-E-A~/ \ H H / \ H H //" / H Ir\/ \ i \Al<l-Alu \~i-:-~i A~-~-AI\(XV.)(d) There has been considerable speculation about the constitution ofVarious bridge45 including a " methylated double bond "37Pitzer and H.S. Gutowsky 46 havethe dimers which apparently exist in aluminium alkyls.structures have beenanalogous to Pitzer's protonated bond.recently prepared the following aluminium alkylsAlMe,. ,41Et,. AlPr,. AlPP,.20 M. .p. ........................ 150" -52.5" -107'd:! .......................... 0.752 0.823 0.837 -and some mixed alkyls. By X-ray, infra-red, and cryosopic measurementsthey find dimerization in all the above-formulated alkyls except theisopropyl compound. No higher polymers are found. It is concluded,from these compounds and the mixed alkyls, that one methyl per aluminiumatom will suffice to form stable dimers, though an a-methylene group willallow dimers of less stability.No dimer is possible with only one hydrogenatom on the a-carbon atom. These findings are embodied in the proposedstructure (XVI).4 2 H. I. Schlesinger, H. C. Brown, and G. W. Schliffer, J . Anaer. Chem. SOC., 1943, 65,43 0. Stecher and E. Wiberg, Ber., 1942, 75, 2003.44 L. 0. Brockway and N. R. Davidson, J . Amer. Chem. SOC., 1941, 63, 3287.4 5 A. Burawoy, Nature, 1945, 155, 269.1838.46 J . Amer. Chem. SOC., 1946, 68, 2204Tn this connection it is noteworthy that X-ray determinations showtetramethylplatinum to be tetrameric in the solid Structure(XVII) indicates the position of the platinum atoms and methyl groups :the distances are such that bonding through the four inner methyl groups isalmost certain.Ionic bonds through methyl ions are unlikely, and thepossible bridge link formulation shown (XVII I) occiirs to the R'eporters.Me,II (XVI.)RCH Pt(XVII.) (XVlII.)Lanthnons and Actinons.-For the double misnomer, " rare earths ",more recent practice has been to use '. lanthanides " or '* lanthanates ".J. K. Marsh 48 shows the unsuitability of the terminations and proposes themuch better term lanthanon ; and, now that the trans-uranium elements arecoming to be recognised as a second " rare-earth " transitions series, theyshould properly be called actinons.Marsh 48 has presented the chemistry of the lanthanons from a modernviewpoint, with emphasis on their separation; and B. S. Hopkins 49 hasreviewed the electro-chemical isolation of lanthanons.The properties ofLa, Ce, Pr, Nd, Sm, and Y have been summarised. D. L. Simonenko a hasdescribed a method of producing metals from difficultly reducible oxides,which is applicable to zirconium and the lanthanons. The lanthanonsform soluble complexes 51 with sodium nitrilotriacetate, N(CH,*CO,Na),.With excess of the reagent and of ammonium oxalate, acid precipitateslanthanum at pH 6, then przeseodymium and neodymium; samarium atpH 5 ; gadolinium a t pH 4-5, and erbium at pH 4.0. Four fractionationsQuart. Reviews, 1947, 1, 126.47 R. E. Rundle and J. H. Sturdivnnt, J . Amer. Cheur. ~!!oc., 1917, 69, 1561.4g B. S. Hopkins, Trans. Electrochem. SOC., 1946, 89, preprint S, 113.50 Compt. r e d . Acad. Sci. U.R.S.S., 1946, 51, 303.5 1 0.Beck, Helv. C/tht. -Actti. 1946, 29, 357give pure lanthanum. Hydrated orange ceric oxide is precipitated fromiiitriloacetate solution by hydrogen peroxide, the reaction being specific,sensitive to 8 pg. of Ce per c.c., and useful for purification. T. Moeller 52finds that slight hydrolysis of sulphates of tervalent lanthanons increaseswith decrease in the ionic radius.53 In the separation of uranium minerals, alarge part of the actinium would be concentrated by adsorption on the leadand barium sulphates. Fractional crystallisation of the double nitrates ofthe lanthanons and magnesium is recommended, whereby most of thelanthanum is removed. The actinium may then be separated by fractionalprecipitation of the hydroxides, which gives bett&r results when lanthanumis absent!.Though i t may havebeen under investigation previously,j5 the first positive identification wasannounced in April 1946 by C.D. Coryell lo (who has not yet suggested aname to replace ilkinium). The names of N. E. Ballou, L. E. Glendenin,B. L. Goldschmidt, F. Morgan, and J. A. Marinsky are particularly associatedwith the discovery.15 They have identified several radio-isotopes of whichthe most stable, 14’61, may be produced artificially from the stableneodymium isotope, ‘4,6,Nd, or as the result of p-deca,y of the nucleus l;7,Nd,which occurs as 2.6 yo of the fission products ~f‘ura~nium, 235U.The chemistry of element 61 has been ~onsiderect.~~Much of the work on the trans-uranium elements is known only throughreviews 56-59 which give some of the chemical and nuclear properties ofthese elements 60 and a table of isotopes of elements 90-96.These may bereferred to for further details and for the names of the workers in this field.The more important transmutations involved in generating the new elementsare shown in the accompanying figure, where the sort of nuclear reaction isto be inferred from the direction of the arrows in the key.Neptunium Ti?Np, the first trans-uranium element whose isotope wasdefinitely established,61 derives its name from the first trans-Uranus planet.Of its two isotopes subsequently found, 2iiNp is an a-emitter 57 (2.25 >: 106yrs.) and 1;:Np is a p-emitter (2.0 days).62Plutonium was discovered in 1940 as the very active a-emitter ?i?Puj2 J .Physical Chem., 1946, 50, 242.53 ill. Bachelet, J . Ci~inz. physique, 1946, 42, 98.64 S. Takvorian, Ann. Chitic., 1945, 20, 113.5 5 M. L. Pool, L. L. Quill, D. C. Macdonald, and J. D. Kurbatov, PJj,+~icd Rev.,1!142, 61, 106; C. S. Wu and E. Segrb;, ibid., p. 203.-s6: 1,. S. Foster, J . Chent. Edtic., 1945, 22, 619.j7 C. T. Seaborg, Chem. Eng. News, 1945, 23, 2190.j3 Idem, ibid., 1946, 24, 1192.ba Idem, Science, 1946, 104, 379.6 1 E. 31. McMillan and P. H. Abelson, Physical Rev., 1940, 57, 1186.5p Ident, ibid., 1947, 25, 358.1. Yerlman, Chenz. Eng. S e w s , 1946, 24, 3032; G . T. Seaborg, E. M. McMillan,A. C. Wahl, and J. W. Kennedy, PI/y+al RPV., 1946, 69, 36658 INORGANIC CHEMISTRY.(50 yrs.) and is named after the second trans-Uranus planet.A less activeisotope ";Pu (30,000 yrs.), identified 63 in 1941, decays into ",","U, the parentAc-U of the natural radioactive actinium family. Both 239P~~ and 235U are" fissionable " isotopes, the former being more important because it is moreeasily separated from uranium than is 235U from non-fissionable 23eU.Atomic No. 92 93 94 95 96Seaborg and M. L. Perlman 57 have found plutonium in small quantity( 1 part in 1014) in pitchblende and carnotite; 64 this might arise by way of238U by capture of neutrons whose emission from uranium products isattributed to U-X, (2,3tA~). The spontaneous fission of 238U (half-life,1 0 1 6 yrs.) may be an alternative source of neutrons.59 The atomic-bombproject prevented publication of these findings before 1946.6570, 555.63 J.\V. Kennedy, G. T. Seaborg, E. Segr6, and A. C. Wahl, Physical Rev., 1946,dl M. I. Corvalen, ibid., 1947, 71, 132.Oir See, however, H. D. Smyth, " Atomic Energy ", H.M.S.O., 1945; e..q., pp. 60, 77DODD AND ROBTNSON : GENERAT,. 59The trans-uranium elements form a group resembling the lanthanons andbeconling known as the actinon series. Elements 95 and 96 are the latestto be d i s c ~ v e r e d . ~ ~ High-energy bombardments with 4 0 4 MeV. He+ +ions of 238U and 239Pu using the 60-inch Berkeley cyclotron enabled Seaborg,R. A. James, L. 0. Morgan, and A. Ghiorso t o identify the isotopes of 95 and96. The nomenclature, americium and curium, follows that employed forthe corresponding members of the lanthanon series, vix., europium andgadolinium.Cerium recalls the Curies just as gadolinium honours Gadolin.Several hundred milligrams of neptunium have been worked up as purecompounds, and sufficient plutonium has been recovered for the physical andchemical properties to be accurately known.60, 61, G6 Neptunium exhibitsvalency states 3,4,5, and 6, with the shift in stability, in contrast to uranium,towards the tervalent state. NpT1 resembles uranium and is precipitatedwith sodium uranyl acetate. NpLV shows great similarity to thequadrivalent lanthanons and NplI1 is quantitatively carried down asfluoride with cerium. The chemistry of plutonium has been comprehensivelysurveyed by B. G. Harvey, H. G. Heal, A. G. Maddock, and (Miss)E.L. R ~ w l e y , ~ ~ and valency states 3, 4, and 6 have been established. Theirrelative stabilities are illustrated by the following scheme :(a) ~e+++ I , K2crso, ‘ purr 0, \ puIv ( b ) hot acid KMnO,, KeS,O,so,, U+++, I’ ‘ SO,, [Fe(CN),]‘-, H,O,Pllr’I ~where (a) and (71) refer to different concentrations of plutonium solution,(a) at tracer concentrations and (b) at about 1 mg. per C.C. PuTTr formsbright blue solutions, the sulphate, chloride, and perchlorate being readilysoluble in dilute acids, and the fluoride insoluble. PiiT17 forms pale pinksolutions (except the nitrate, which is green) and PuTv ions are morereadily hydrolysed than PuIrl ions. In the sexivalent state plutonium,like neptunium, resembles uranium and forms a brownish-yellow ammoniumplutonate, orange plutonyl nitrate, pale mauve sodium plutonyl acetate, andforms a complex compound with “ oxine ” in which the plutonium contentshows i t to be the analogue of uranyl “ oxinate ”.PurL- forms complexeswith the commoner organic reagents more readily than PnI1 I, which wasnot well precipitated by any of the reagents used.The a-radioactivity of the available isotopes of americium and curiumprecludes any attempt to obtain and handle amounts of more than 1 mg.Hence much of this work has involved ultramicrochemical techniques, ofwhich the acknowledged pioneers were P. L. Kirk and A. A. Benedetti-Pichler. The order of magnitude is illustrated by R. B. Cunningham a.ndL. B. Werner’s preparation of a pure plutonium compound from a 2microgram sample of the element, and by Cunningham’s isolation of pureamericium hydroxide, Am(OH),, from even smaller samples of startingn ~ a t e r i a l .~ ~ So far it appears that in aqueous solution only tervalentamericium is stable. Curium presents similar difficulties in even greaterF. Xtrrtssmm and 0. Hnhn, Naturwiss., 1942, 30, 256. G7 J . , 1947, 1010degree. Its intense a-activity induces decomposition of the water in whicahit is dissolved, and recoil effects are marked. Pure preparations have notbeen reported, but the only stable state of curium in aqueous solutionappears to be the tervalent one.The strong evidence that these elements belong to an actinon transitionseries has been reviewed by V. ST. Goldschmidt,68 by G.E. Villar,69 andby Seaborg.GO This would involve utilising the 5f orbitals as under A in theaccompanying table, rather than the ‘‘ standard ” configuration givenunder R.-4. €3.7 -____- L--- r-------_hl----- 7811ell. 0. 13. 0. 0. T’. 0.2 ILrt 88 = 78 + - 2 Ci -- 2 _- 2 (j __- 1 ~ 89 = 78 + - 2 t i 1 2 - - - 0 1 2 -)Th 9 0 = 78 f 1 d 6 1 2 -. - ’ L ( j d 2Pa 9 1 = 7 8 + 2 s G 1 2 _ _ , 6 3 2 .)U 9 2 = 7 8 + 3 2 G 1 2 __ 2 6 4 2Np 93 = 7 8 + 4 2 6 1 2PI1 9 4 = 78 +- 5 2 6 I 2Am 95 = 78 -i- A 2 0 1 2Cm 96 = 7 8 -I- 7 2 (i I 2Orbitals. 5j’. (is. Gp. &I. 7s. 5f. t k . 611. Btl. i s .Since the difference in energy between 5f and 6d shells is no doubt less thanbetween 4f and 5d and may well be of the order of chemical binding energies,the starting pointt of the actinon series may be less marked than that of thelanthanons, as appears from the variety of points chosen by different authorson chemical grounds.In its sexivalent state a t least uranium supportsthe “ standard ” configuration rather than the transition configuration ofits less stable tervalent state. Proceeding from uranium to higher atomicnumbers the tervalency clearly asserts itself at elements 95 and 96. A moreimportant criterion for accepting the actinon series, as such, than theoccurrence of lower valencies in earlier members, or even the possiblepresence of a single 5f electron in thorium, is the predominance of tervalencyin americium and cnrium.R. E. D.P. L. R .3 . SOME INTERMEDIATE COMPOUNDS IN IK‘ORG ANIC REA4CTIONS.This section deals with a number of unstable intermediate compoundsappearing in many inorganic reactions which have been investigated par-ticularly in recent years and have not been reviewed in detail.W. A.Waters’s book deals mainly with organic free radicals and discussesbriefly SL few inorganic radicals. Although relatively few intermediates havebeen ‘. isolated ” in a “ pure state ” their existence can be deduced fromkinetic measurements and is often made probable by some circumstantialNorsk Fysisk Tidsskr., 1941-1042, 3, 179.69 Bol. facultad in,q. Montevideo, 1946, 3, No. 2.“ The Chemistry of Free Rrtdirds ”, Clarendon Press, Oxford, 1940evidence, such as absorption spectrum, electrical conductivity, or magneticsusceptibility. These compounds are discussed here primarily from theirinorganic aspect without necessarily entering into full discussion of thekinetics of the reactions concerned.The unstable intermediates which play ail important r6le in the mechanismof many chemical processes are normally present in low concentration sinceinstability is the essence of chemical reactivity.They fall roughly into twogroups : (1) Free radicals or radical ions (" odd ions "), which, on accountof a free (unpaired) electron, enter into reactions with low heat of activationand thus appear in low stationary concentrations in a pseudo-equilibrium orstationary state. (2) Unstable addition or complex compounds which,though their ordinary valency requirements are generally satisfied, appearonly under suitable conditions in certain chemical equilibria ; examples areNO, and SnCli -.(1) Free Radicals.Free radicals are generally ephemeral and their presence is soiiietiiiiesdifficult to establish and usually rests on evidence derived from the kineticstudy of a number of chemical reactions.Cases in point are HO, and OH.Their importance springs from the fact that many chemical processes proceedin a sequence of simple elementary reactions which consist individually ofno more than univalent changes, such as the breaking of a single bond,the transfer of a hydrogen atom, or the transfer of an electroi~.~ These" univalent changes " being admitted, it is clear that the reaction must gothrough unstable intermediates each possessing an odd number of electrons.The Radical HO,.-Among the simplest, and probably most importantradicals are those compounded of hydrogen and oxygen which appear forinstance in the reactions between these elements.Here, following the ideasset out above, HO, would be expected to appear as a precursor in theformation of H,O, from niolecular oxygen. The existence of this radicalwas suggested first by H. S. Taylor,* and F. Haber 5 inferred that i t playedaii essential part in the reaction of hydrogen with oxygen. Its formationseems to take place according to the equation H + O,(+ M) = HO,(+ M)(M is a third body). Thus the formation of HO, is to be expected in thereaction between hydrogen atoms (e.g., from a discharge tube) with molecularoxygen.P. Harteck and K. H. Geib,6 who studied this reaction from apreparative point of view, found that i t leads eventually to the formation ofa very unstable compound of the composition of hydrogen peroxide whichdcconiposes into water and oxygen at teniperatures above - 115". Thereis evidcncc to suggest t h t HO, radicals are formed from oxygen ailti hydrogeniitoiiis produced photochemically by means of excited niercury atoins :Hg* -l- H, = HgH + H. Hydrogen peroxide, which is a product in thesereactions, is generally attributed to the interaction of the two radicals :C. N. Hinshelwood', J . , 1947, 691.3 J. Weiss, Naturwiss., 1935, 23, 64.Natunuias., 1931, 19, 450.2. phyaikal. Chem., 1926, 120, 183.Ber., 1932, 05, 155162 INORGANIC CHEMISTRY.2H0, = H,O, + 0,.7 The formation of HO, has been held responsible forthe inhibiting action of molecular oxygen in the photochemical interactionbetween hydrogen and chlorine,* in which hydrogen atoms appear a8 one ofthe chain carriers; and similarly, in the inhibition of the photochemicaldecomposition of hydrogen iodide by molecular ~ x y g e n .~ G. C. Eltenton lohas attempted to show the presence of HO, radicals in low-pressure flames ofoxygen and simple aliphatic hydrocarbons, by coupling the reaction chamberto a Dempstw type of mass spectrometer. Under these conditions thedetection of HO, is rendered difficult by presence in great quantity of normalmolecules of oxygen (mass 32) and by the contribution of 160170 of massidentical with the radical.This author actually found some indication ofHO, in that, after ignition of the gaseous mixture which reduces the con-centration of the two interfering species there was an increase in the ioncurrent for the mass corresponding to HO,. Recently G. J. Minkoff l1 hassummarised some of the chemical evidence for the existence of the HO, in t,hcgas phase, particularly with regard to some recent work of (Sir) A. Egertonand G . J. Mii&off,l2 and he has also discussed certain possible energy surfacesconnected with the formation of HO,.Additional inforination comes from reactions in the solid state and insolution. HO, is presumably formed as an intermediate when molecularoxygen reacts with hydrogen dissolved in palladium metal,13 and also inelectrolysis, by the depolarising action of molecular oxygen a t a metalcathode, since in both cases the hydrogen atoms present on the surface canreact with oxygen molecules.When molecular oxygen is passed over molten potassium a t 300°, a deeporange-coloured peroxide is formed corresponding to the general formula(KO,), which was previously given the formula K,O,.I4 Recently, E.W.Neuman l5 investigated the compound, a t the suggestion of Pauling, whopointed out that the only reasonable electronic structure for the 0;- ionwould be : 0 : 0 : 0 0 : , but an ion of this constitution should be dia-magnetic and presumably colourless. However, the peroxide proved to beparamagnetic with a Bohr magneton number of 2.04 at room temperature.This is close to the value of 1-73 B.m.which corresponds to the spin momentof an unpaired electron, and to the value of 1.83 B.m. as calculated byvan Vleck for a molecule with doublet separation. The discrepancy has beenattributed to experimental error or, possibly, to a mutual interaction of themagnetic dipoles in the condensed phase. The investigation leads to theformula KO, and hencc to the assumption of 0,- ions in the crystal lattice, a. . . . . . . .. . . . . . . .7 J. R. Bates and D. I. Salley, J . Arner. Chena. SOC., 1834, 56, 110.8 31. Bodenstein and P. W. Schenclr, 2. physikal. Chern., 1933, B, 20, 420.0 G. A. Cook and J. R. Bates, J . Amer. Chetn. SOC., 1935, 57, 1775.10 J . Chem. Physics, 1947, 15, 456.11 Faraday SOC. Discussion on " The Labile Molecule ", Sept.1947 (in the press).12 Proc. Roy. SOC. (in the press).13 J. Weies, Tram. Farachy SOC., 1935, 31, 668.14 W. Traube, Ber., 1912, 45, 2201, 3319. l5 J . Ghem Physics, 1934, 2, 31WEISS : SOME INTERMEDIATE COMPOUNDS IN INORGANIC REACTIONS. 63fact confirmed by W. Kassatochkin l6 who, from X-ray diffraction, found thati t was an ionic crystal coinposed of K+ and 0,- ions, somewhat similar instructure to the lattice of potassium chloride, but elongated in one directionon account of the shape of the 0,- anion. A. Baeyer and V. Villiger l7 foundthat the oraiige-red peroxide is also formed from ozone and solid potassiuinhydroxide and that i t is decomposed rapidly by water, giving off oxygen.F. Haber and H. Sachsse l8 have found that the reaction between sodiumvapour and oxygen at low pressures, which they studied by the method ofdiffusion flames, involves the intermediate formation of NaO, according to(i) Na + O,(+ M) = NaO,(+ M) (If = a third body)(ii) NaO, + Na = 2Na0 -- Na,O,.A study of the absorption spectrum of ozone in solution in the presence ofacids and bases, and particularly in concentrated solutions of potassiuinhydroxide (down to temperatures of - 40°), has furnished evidence l3 thatthe ozone presumably interacts in these solutions according to the equations(i) 0, + KOH = KO, + HO,, (ii) 0, + OH- = 0,- + HO,.The secondreaction is probably of importance for all the reactions of ozone in aqueoussolution, particularly for that between ozone and hydrogen peroxide.Thereis some evidence l39 19 to show that the following chain reactions play asignificant part : (i) 0, + HO, = 20, + OH, (ii) 0, + OH = 0, + HO,.I?. Haber and R. Willstatter 2o and F. Haber and J. Weiss 21 have assumedthe presence of HO, in the photochemical and thermal reactions of hydrogenperoxide which are discussed on p. 66.The study of a number of the reactions of molecular oxygen in solution(e.g. , processes of autoxidation and the quenching of fluorescence by molecularoxygen) has made it very likely that the first step in these reactions consistsin the univalent reduction of 0, by a single-electron transfer, to give the0,- ion. In autoxidation in solution the electrons are transferred successivelyfrom the reducing agent (electron donor) according to 0, + electron ---+ 0,-.22In aqueous solution the latter ions are in equilibrium with HO,.These nowtake up a second electron from the donor, HO, + electron .+ HO,-, thusleading to the formation of H,O, which has been isolated in many autoxid-ation processes. Whether or not H,O, is actually found depends entirely onthe rate of any subsequent reactions in which it is involved. This isillustrated by the autoxidation of various acceptors (e.g., leuco-dyes, 39 22 23s ,4C'oinpt. rend. U.R.S.S., 1946, 47, 193; W. Kassittochkirl and W. Kotov, J . (:lteijt.Phgsics, 1936, 4, 458.l7 Ber., 1902, 35, 3038.l8 2. physikal Chern., Bodenstein Vestband, 1931, 831.l9 H. Taube and W. C. Bray, J . Anher. ChnL. SOC., 1940, 62, 3357.2o Ber., 1931, 64, 2844.21 (a) Naturwiss., 1932, 20, 948; ( b ) Proc. Roy.SOC., 1934, 147, A, 338.2 3 J. Weiss, Nature, 1934, 135, 648.23 J. E. Lu Valle and A. Weissberger, J . Anzer. Clwn. SOC., 1947, 60, 1567, 1576,24 J. H. Baxendale and S. Lewin, Tram. Paraday SOC., 1946, 43, 126.182164 INORGANIC CHEMISTRY.metal ions) which have been studied by various authors.ferrous ions the reactions are as follows : 3,22I n thc cas0 ofFe2’ -1- 0, + Fe3 + 0,-0 - H-- A 2 + -HO,Fe2 + HO, =+ Fe”- + H0,-H02- + H- =+= H,O,Here the 11)-drogen peroxide is decomposed further by the ferrous ions (seeApart from the interaction of two HO, radicals, giving H,O, and O,,there is also the possibility of an interaction according to 2H0, = 0, + H20.It is possible that the traces of ozone which are sometimes observed inautoxidation processes are due t o this cause.I n the quenching of fluores-cence of, e.g., polycyclic hydrocarbons (anthracene, benzpyrene) by molecularoxygen 25-28 the electron is donated by the excited molecule (A*), on theabsorption of a light quantum, and this may either return to the ground statcwith the emission of light, thus emitting fluorescence, or else may reactwith molecular oxygen : A*[+ $1 + O,[+ +]+A [.f] + O,-[.fJ-][.f]. Thearrows in the brackets indicate that this and similar processes proceed withouta change in the multiplicity and therefore should take place with an enhancedp r ~ b a b i l i t y . ~ ~ I n accordance with this scheme, products of the formula AO,have frequently been isolated.In view of the importance of 0,- and HO,, attempts have been made toobtain values for the electron afinity of the 0, inolecule and for the bindingenergy of the H atom.The first is not yet known with much certainty;calculations from a cycle process 30 have yielded ca. 2-7 e.-v. ; D. T. Vierand J. E. Mayer 31 found 3-2 e.-v. by some direct measurement’s. The bondenergy 0,-H is also not known accurately.estimated it at 44 kcals. According to thc gencral theory given byW. Heitler,32 it is possible to estimate the energy values of the reactions0, + H = HO, + Q1 kcals; HO, + H = H,O, + Q2 kcals.on the assumption that the oxygen atom is in a 3S ground state and that theoxygen-oxygen bond is a pure spin-valence.Although the values of Q1 and&, depend on the amount of Coulombic energy assumed, t,hey are not greatlyinfluenced by this factor.1). 65).Bodenstein and SchenckThis is illustrated by the following figures : 33Coulombicenergy, 7;. Q1, kcals. Q 2 , kcals.33 53 10120 48 042 5 14;. J . Jj~owur~ a i d A. Nortoil, Tra)hs. 2f’uruday Soc., 1‘339, 35, 44.2 6 15. J . Bowen and A. H. JVilliams, ibid., p. 765.2 7 H. W’eil-Malherbe and J. Weiss, Nature, 1942, 149, 471.$ 8 J. Weiss, Trans. Faraday SOC., 1946, 42, 133.29 A. Terenin, Acta Physicochim. U.S.S.R., 1943, 18, 211.30 J. Weiss, Traits. Faraduy SOG., 1935, 31, 966.32 “ Handbuch der Kadiologie,” Yol. 11, Akadem. Verlagsgosellsch., Leipzig, 1934.35 J. Weiss, Faradey SOC. Discussion (see ref.1 1 ) .31 J . Cheni. Physics, 1944,12, 28WEISS : SOME INTERMEDIATE COMPOUNDS IN INORGANIC REACTIONS. 65The Hydroxgl Radical (OH).-This radical was postulated in themechanism of the reaction between hydrogen and oxygen by Haber et uL3'Subsequently, K. F. Bonhoeffer and H. Reichhardt 34 established its presenceby the absorption spectrum of heated water vapour in excess of oxygen,where the following equilibrium holds : 40H 2H,O + 0,. The absorp-tion spectrum of OH radicals in the gas phase was then investigated in greaterdetail by Oldenberg et ~ 1 . 3 ~ with spectroscopes of high resolving power. Inthis way OH radicals were shown to be present in such systems as the H,-0,flame and the electric discharge through water vapour. These authorsdetermined also the transition probabilities of the lines of the OH bandspectrum (f-values) and were thus able to measure quantitatively theconcentration of the OH radicals.By this means they obtained thedissociation constant, and the heat of dissociation, of water vapour withconsiderable accuracy : 36H + OH ---+ H,O 1- 118.2 0.7 kcals. (at 0" K.)0 + H ---+ OH + 110.1 -j= 0.9 kcals. (at 0" K.)"he absorptioii spectrum is at present undoubtedly the best method for thedetection of OH radicals in the gaseous phase, but as Eltenton lo has pointedout, OH is very difficult to detect by the mass spectrometer.Hydroxyl radicals presumably appear in a great number of reactions inthe gas phase and in solutions, some of which have been cited already in thediscussion of the HO, radical.Among the most important and frequentlyinvestigated of these is that between hydrogen and oxygen. In his recentBakerian Lecture, C. N. Hinshelwood 37 has represented the principalelementary processes by the following scheme :H = 2HOH + H, = H20 + HH + 0, = OH + 00 + H, = OH + HH + 0, + M = HO, + MHO, + H, = H20 + OHHydroxyl radicals play an important part in the reactions of hydrogenperoxide in solution, and consequently also in autoxidation processes whereinhydrogen peroxide is formed as an intermediate. One of the better knownis that between hydrogen peroxide and ferrous iron salts which, after thevery early work of Sohonbein, was investigated later by W. Manchot andG . Lehmann,38 with the following results : (a) I n the presence of excess ofj'errous s d t .A given amount of dilute hydrogen peroxide was added slowlywith stirring to a solution containing an excess of ferrous sulphate. Underthese conditions, when the ferrous concentration in the reaction space was31 2. physikal. Chem., 1928, 139, 75; K. F. Bonhoeffer and F. Haber, ibid., 1928,137, 263; K. F. Bonhoeffer and T. G. Pearson, did., 1931, B, 14, 1.35 0. Oldenberg and F. F. Rieke, ibid., 1938, 6, 169, 779; 1939, 7, 487;L. Avramenko, Acta Physicochim. U.S.S.R., 1943, IS, 58; G. Damkohler and R. Edse,Naturwiss., 1943, 31, 310.36 R. J. D y e r and 0. Oldenberg, J . Chem Physics, 1944, 12, 351.37 Proc. Roy. SOC., 1946, A , 188, 1. 38 Annalen, 1928, 460, 179.REP.-VOL. XLIV.66 INORGANIC CHEMISTRY.always high compared with the concentration of the H,O,, the reaction wasrepresented by 2FeS0, + H,O, + 4H20 = 2Fe(OH), + 2H2S0,. ( b ) In thepresence of excess of hydrogen peroxide. A given amount of a dilute solutionof ferrous sulphate was added to a dilute solution of hydrogen proxide sothat the latter always remained in excess. For these conditions the reactionchanged to 2FeS0, + 3H202 + 2H,O = = 2Fe(OH), + 2H,SO, + 0,.I n order t o explain their results these authors assumed the intermediateformation of a quinquevalent iron oxide (Fe,O,) which was supposed t o reacteither (a) with excess of ferrous salt yielding ferric salt, or ( b ) with excess ofperoxide to give ferric salt and oxygen. It has been shown, however, thatthis explanation cannot be correct, for under suitable conditions (i.e,, highratio of [H202]/[Fe2+] in the reaction space), much more than the one moleof oxygen demanded by Manchot's equation is produced in the oxidationof one mole of ferrous t o ferric salt.This led to the formulation of thereaction as a chain mechanism. It is found that, as a first approximation,the reaction can be described in its main course by the following fourelementary processes :Fe2+ + H,O, = Fe3+ + OH- + OH . . ( 1 , 1 )Fe2+ + OH = Fe37 + OH- . . . . . (1, 2)I n the case when the ferrous salt is in excess, no oxygen is formed. I n thepresence of excess of peroxide the latter competes with the Fe2 for tlhe OHradicals and the following two chain reactions come into operation :H202 + OH = HO, + H20 .. . . ( 1 , 3 )H 2 O 2 + H O , = 0 , + H 2 0 + OH . . . (1, 4)The possibility of a chain in the hydrogen peroxide decomposition is clearlydemonstrated by the photochemical decomposition where quantum yieldsup to 100 have been observed.39 The photochemical primary process due toabsorption (in the near ultra-violet) consists in the formation of OH radicals :H,O, + hv --+ 20H.4" Chain-breaking processes are represented here bythe interaction of two radicals and their consequent disappearance from thesystem :The interaction of radicals is also of importance in the ferrous salt reaction,particularly at sufficiently low ferrous-ion concentrations, since the abovescheme is evidently inadequate as it is known that the chain length decreasessomewhat with increasing hydrogen-ion concentration. This could point toa Participation of H0,- or 0,- in the chain reactions, but it is also possible thatsome of the chain-breaking processes are favoured by hydrogen ions in thatthese might lead t o the formation and subsequent participation of H20,T(from the equilibrium HO, -t Hi + H,O,- ) or of H,O (from the equili-brium HO + H' + H,O* ) in these processes.HO, + OH = 0, + H20 .. . . (1, 5)39 G . Kornfeld, 2. physikal. Chein., 1936, B, 29, 208.40 H. C. Urey, L. H. Dawsey, and F. 0. Rice, J . Amer. Chem. Soc., 1929, 51, 1371WEISS : SOME INTERMEDIATE COMPOUNDS IN INORGANIC REACTIOXS. 67If one starts with ordinary '' neutral " solutions of ferrous salts andhydrogen peroxide, hydrogen ions are used up in the course of the reactionand basic ferric salt is precipitated and eliminated from the reaction.I nsufficiently acid solutions, however, where ferric ions are present, the rapidoxidation of the ferrous salt is followed by the slow decomposition of hydrogenperoxide which, under these conditions, assumes the character of a catalyticprocess, in the course of which ferric ions are periodically reduced. Accordingt o Haber and Weiss,21 the ferric-ion catalysis is not a chain reaction underordinary conditions. It can be indicated by the schemeFe3I +H0,---+Fe2++H0, . . . . ( 1 , 6)which is followed by the chain reaction (1,4) and reoxidation of the ferrous ionsaccording t o ( 1 , 2). Originally the reaction Fe3+ + HO, = Fe2 + + H + + 0,was thought to be responsible for the formation of oxygen under theseconditions, but recent investigations 41 have shown that this reaction isprobably unnecessary.The system hydrogen peroxide-ferrous salt, known as Fenton's reagent,is of considerable interest in that according to the above discussion itgenerates hydroxyl radicals in solution.This has been demonstratedrecently by M. G . Evans et u Z . , ~ ~ who showed that the system is capable ofinitiating the chain polymerisation of certain monomers (methyl meth-acrylate, acrylonitrile) in aqueous systems and also that hydroxyl groups areactually taken up into the growing polymer chains.I n the presence of water, cobaltic ions give rise to the formation of oxygenat room temperature. It has been suggested3 that this is due t o theintermediate formation of OH radicals according toCo3' + OH- = Go2' + OH; Co3' + H20 = CO'.+ H20.followed by H20 + OH + H ', 20H = H,O + 0, 20( + M) = O,( + M)Ceric ions which do not react under ordinary conditions, are capable ofundergoing a similar reaction 43 on irradiation in the near ultra-violet, whenpresumably an electron is transferred to the excited ceric ion (Ce4' *) fromone of the water molecules in its hydration shell, resuIting again, as above, inthe formation of H20i and eventually molecular oxygen. It is also verylikely that OH or H,O' is a precursor in the anodic formation of molecularoxygen in electrolysis.@Recently it has been suggested45 that, if water is irradiated by X-rays,y-rays, cc-particles, or neutrons (recoil protons), these ionising radiations cause;t splitting of the water molecules into radicals : H,O ---+ H + OH.? This4 1 J.Weiss, Paraday SOC. Discussion (see ref. 11).42 J. H. Baxendale, M. G. Evans, and G. S. Park, Tram. Faruduy SOC., 1946, 42,43 D. Porret and J. Weiss, Nature, 1937, 139, 1019.IQ 0. J. Walker and J. Weiss, Truns. Furaday SOC., 1935, 31, 1011.t The sign +has been adopted in radiation chemistry following recent American155; J. H. Baxendale, M. G. Evans, and J. K. Kilham, ibid. p. 668.J. Weiss, Nature, 1944, 153, 748; Trans. Faraduy SOC., 1947, 45, 314.practice68 INORGANIC CHEMISTRY.presumably is brought about by the intermediate formation of H,O+ andH,O-, the latter decomposing according to the equilibriumH20- s H + OH-.In agreement with this view are some experiments of F.S. Daintjon,*6 whowaa able to initiate the chain polymerisation of acrylonitrile in aqueous systemsby irradiation with y-rays.More recently it has been shown 47 that simple organic substances(benzene, nitrobenzene, benzoic acid) undergo hydroxylation if treated withhydrogen peroxide and a, ferrous salt and also by irradiation with y-rays,X-rays, or neutrons in aqueous systems, thus giving further support to theformation of hydroxyl radicals under these conditions.The SH Radical.-According to Forbes et u1.,** absorption of ultra-violetlight by hydrogen sulphide in the gas phase or in chloroform solution corres-ponds to the photochemical primary process :H2S + h v + HS + H followed by, e.g., H -t HS --+ H, -1 S ant12HS --+ H,S2 or 2HS -+ H2S + S.Very similar results had been obtained 49 in the investigation of tlwphotochemical primary process corresponding to the light absorptlion ofSH- ions and S2- in aqueous solutions, which can be represented by ( i )SH- + H20 + h v --+ SH + H,O-, (ii) S2- + H,O + h v I_, 8- + H20-(S -+H+ +Z SH)The Radical NOH.-The existence of this radical was assumed byA.Angeli 50 and later by F. Raschig 51 in the reaction between nitrous acidand sulphur dioxide. M. L. Nichols and C. W. Morse 52 assumed its form-ation in certain reduction processes involving oxides of nitrogen, and similarviews were put forward by L. Cainbi,53 who also assumed the intermediateformation of NOH in the decomposition of the so-called " blue acid ".54According to L.Andrussow 55 and M. BodensteinYb6 NOH is one of tlicprimary products in the catalytic oxidation of ammonia on platinum catalysts.Bodenstein found nitrous oxide and nitric acid among the reaction productsand suggested that these were formed according to (i) 2HN0 = N20 + H20,(ii) HNO + 0, = HNO,.The formation of NOH has been assumed if nitrogen oxide is added to astream of hydrogen atoms from a discharge This was confirmed by4 6 Nalure, 1947, 160, 268.4 7 G. Stein and J. Weiss, in course of puhlicRt,ion.' 8 W. H. Avery and G. S. Forbes, J . Anier. Ckeire. A'oc., 1!)38, 80, 1003; CX. S.Forbes, J. E. Cline, and B. C. Bradshaw, ibid., p.1431.4 * H. Fishgold and J. Weiss, Nature, 1936, 137, 71.51 " Schwefel und Stickstoff Studien ", Verlag Chemie, 1924.52 J. Physical Chem., 1931, 35, 1239.53 Atti R. Accad. Lincei, 1933, 17, 204.s p E. Bed, K. Winnacker, and H. H. Saenger, 2. aiaorg. Chem., 1933,211, 379.5 s Ber., 1926, 59, 458; 1927, 60, 536. 5 6 2. EZektrochem., 1935, 41, 466.57 H. M. Smallwood, J . Amer. Chenz. SOC., 1929, 51, 1985.Ber., 1904, 37, 2396WETS : SOME INTERMEDIATE COMPOUNDS IN INORGANIC REACTIONS. 69P. Harteck,58 who showed that if hydrogen atoms are allowed to react withnitrogen oxide at liquid-air temperatures a compound is formed of thegeneral formula (HNO),, which decomposes with the formation of nitrousoxide and water.It is very suggestive that the first product of reduction of NO in solutionis the NO- ion formed by an electron transfer from a suitable reducingagent (donor) according to NO + electron+ NO-, followed by NO- +H i +NOH, and leading thus to the formation of nitrous oxide accordingtjo the above equation.This is in agreement with the formation of nitrousoxide if nitrogen oxide is acted on by reducing agents such as Ieu~o-indigo,~~alkaline pyrogallol,60 sulphur dioxide,6 or stannous chloride.62Similarly NO- is presumably formed in the quenching of fluorescence ofvarious substances by nitrogen oxide in solution 63 (e.g., polycyclic hydro-carbons). This is analogous to the quenching action of molecular oxygenwhich is thereby converted into 02-. Just as in the latter case, where more orless stable peroxides are formed, the quenching of fluorescence by nitrogenoxide can lead to the formation of " nitroxides " of the general formula(A, NO); these are generally less stable than the peroxides, but they can beisolated under suitable conditions.64The N, Radical.-Alkali azides in aqueous solution show an absorptionspectrum in the ultra-violet which is simiIar to that of the alkali bromides oriodides.It has been suggested 65 that the photochemical primary processis connected with the electron-affinity spectrum as proposed earlier byJ . Franck and F. Haber 66 for the halogen ions, vix., N,- + H20 + h v N, -+ H,O-. This leads to the intermediate formation of the azide radical, whichis probably also formed by chemical oxidation, for example, in the transferof an electron to the ceric ion, N3- + Ce4' j N, + Ce3'.The N, thenbreaks up, ultimately giving molecular nitrogen.The Monothionic Acid Radical (SO,H).-Franck and Haber 66 suggestedthat this radical plays an essential part in the thermal and photochemicaloxidation of sulphites in solution. The autoxidation of sulphite can beinitiated by metal ions, e.g., cupric ions, or by irradiation 67 in the ultra-violet, where the SO;- and HS0,- ions show strong absorption.68 Thephotochemical primary process corresponding to this light absorption can berepresented by the equations S@- + H,O + h v + SO,- + H20-, followedby H,O- -+ OH- + H and 2H = H and the equilibrium SO3- +HBer., 1933, 66, 423. 59 W. Manchot, Ber., 1906, 39, 3510.61 G.Lunge, Ber., 1881, 14, 2196. 6o C. Oppenheimer, Bet-., 1903, 36, 1744.62 G . Chesneau, Compt. rend., 1899, 129, 100.61 Idem, J . , 1944, 541.65 J. Weiss, Tram. Faraday SOC., 1947, 43, 119.6 6 J. Franck and F. Haber, Sitzungsbe?.. Preuss. Akad. Wiss., 1931, 250.6 7 H. L. J. Backstrom, Me&. Vente,nsk. Akad. Nobelimt., 1927, NO. 16, 6, 22;J . Anher. Chem. SOC., 1927, 49, 1460; H Alyea and H. L. J. Biickstrom, ibid., 1929, 51,90.H. Weil-Malherbe and J. Weiss, Nature, 1943, 151, 449.** H. W. Albu and P. Goldfinger, 2. physikal. Chent., 1932, B, 16, 33870 INOWANIC CHEMISTRY.A -HSO,. .This has been confirmed by the work of F. Haber and 0. H.Wansbrough- Jones,69 who found that irradiation of sulphite solutions gavedithionic acid, resulting from recombination of the radicals, ZHSO, = H2S,06,and molecular hydrogen.Some of the radicals may also undergo dismutation :2HSO, = SO, + H,SO,, leading to the production of some sulphate. Theaction of cupric ions in initiating the autoxidation is connected with theformation of the same radical : Cu2. + SO:- + Cu $- SO,-. Previousexperiments 70 had shown that the reaction between copper sulphate and analkali sulphite, in the absence of oxygen, leads to formation of dithionate. Inthe presence of molecular oxygen, SO,H can initiate a chain reaction. Beliefin the chain character of this autoxidation reaction is supported by the fact thatvery small amounts of cupric ions are capable of starting the reaction, and alsoby the characteristic action of inhibitors.Furthermore, this is confirmed bythe photochemical autoxidation initia$ed by light absorption in the ultra-violetwhere, according to the foregoing discussion, SO,H radicals are produced. Aswas found by H. L. J. BackstrOmys7 several tens of thousands of oxygenmolecules are used up per one light quantum absorbed by the sulphite. Thesehigh quantum yields (10,000 or more) clearly establish the chain character ofthis reaction. Other reactions in the autoxidation chain reaction can berepresented by the following elementary processes : 71(i) HSO, + 0, ---+ SO, + HO, (iii) SO,- + H20, + SO, + OH- + OH(ia) (SO,- + 02+ SO, + 02-) (iv) SO:- + OH + SO,- + OH-(ii) HO, + SO:- + SO,- + HO,-- (v) SO,H + OH --+ SO, + H20The UO,.Radical Ion.-The existence of this radical ion has been sug-gested by A. H. Carter and J. Weiss 72 in the quenching of the fluorescenceof uranyl salts in solution and in the photosensitised decomposition andoxidation of various substances by uranyl salts.Absorption of light by uranyl salts leads to formation of an excited ionaccording to UO: + + hv -+ UO$+ * ; the latter can either return to theground state with the emission of fluorescence, UOpt * --+ UOg + hv’, or,in the presence of a suitable electron donor (e.g., iodide or oxalate ions), thefluorescenee is quenched. This process consists in the transfer of an electronfrom the donor t o the excited uranyl ion, leading to the formation of UO,’ ,UO; * + I- -+ U02+ + I, and represents the primary process of thequenching of fluorescence which is identical with the photosensitised formationof iodine.More recently W.E. Harris and I. M. Kolthoff 73 have suggested that theformation of this radical ion occurs in the polarographic reduction of uranyl6s 2. physikal. Chem., 1932, B, 18, 103.70 H. W. Albu and H. D. Graf von Schweinitz, Ber., 1932, 65, 729 ; H. Baubigny,Compt. rend., 1912, 154, 701; Ann. Chim. Phys., 1914, 1, 1.P. Goldfinger and H, D. Graf von Schweinitz, 2. piqsikal. Chem., 1933, B, 22,211; H. L. J. Backstrom, ibid., 1934, B, 25, 122.72 Proc. Roy. SOC., 1940, A , 174, 351.73 J . Anter. Ghem. SOC., 1945, 87, 1484WETSS SOME INTERMEDIATE COMPOTJNDS IN INORGANIC REACTIONS. 7 1salts, where it is supposed to represent the first stage : UO; ~ + electron +UO, ' .This has also been confirmed by H. G. Heal.'*(2) Ions derived from Cation-Anion Complexes.According to the classification given above, these complexes belong tothe second group of unstable intermediates. They appear in ionic equilibriaand their stability varies greatly. It has been pointed out recently byC. N. Hinshelwood2 that complexes of this nature are probably of greatimportance as intermediates in the mechanism of many reactions.(i) Complexes between Cations and Halogen Ions.-The identification of thesecomplexes in solution is based on some evidence from kinetic data, but a tpresent rests mainly on the study of absorption spectra in solution. Theinterpretation of the absorption spectra is sometimes rather ambiguous, andi t is necessary to study the spectra systematically under different conditions.H.Fromherz and Kun-Hou Lih 75 analysed the spectra of dilute solutionsof lead halides in terms of superimposed band due to Pb2" and PbX(X = halogen ion). The corresponding association constants K,, defined byK , = [PbXr]/[Pb2i][X-], are given in the following table. They wereB ssociation constants (K,) and absorption maxima of lead h.alidecomplexes, PbX .K , * An,,,. (mp).PbCl i- ........................... 12.9 226PbBrt ........................ 14.1 236.8Pb1'- ........................... 29.0 364.5determined by H. Fromherz 76 from the change in the relative intensity ofthe short-wave band due to Pb2 ions and the long-wave bands which could beassigned to the PbXt ions.Higher association complexes are not presentunder these conditions. This was confirmed by experiments on solutions oflead bromide and iodide in the presence of a large excess of Pb2] added aslead perchlorate. Under these conditions the equilibrium of any higherassociates should be shifted towards the formation of PbX , but no appre-ciable change was observed.However, in the investigation of absorption spectra of the lead halides inconcentrated solutions of alkali halides, more highly associated species werefound which were attributed to the formation of PbX:-. On dilution, theseparticular bands diminished in their intensity, but no other bands wereclearly distinguishable until the spectrum was gradually transformed intothat of a mixture of Pb2+ and PbX .The authors have discussed thesechanges in terms of a gradual dissociation of the PbXi- complex, passingthrough PbX,- and PbX, to PbX+ and Pb2+. A similar explanation wasgiven by H. Fromherz et al. for the bands of the halides of Cu2+, Hg2+, Zn2 b,7p Nature, 1946, 157, 228.75 2. physikal. Chem., 1931, A , 153, 321; 1933, A , 167, 103; see also E. Doehlemann7R Ibid., 1931, A , 153, 376.and H. Fromherz, ibid., 1934, A , 171, 35372 INORGANIC CHEMISTRY.Cd2 , and Sn2 which were investigated in solutions containing alkali halides,where equilibria of the type MeX, + 2X-In many spectra of these complexes the extinction curves in the far ultra-violet rise to very high values not encountered in the spectra of the freeanions or of the binary associates. Analysis of the extinction curves of thestannous halides in solution in the presence of alkali halides shows clearly theabsorption bands due to the SnXS- ions.There is a regular shift towardslower frequencies in the series chloride-bromide-iodide in the spectra of thesecomplexes. 'l'he following t!al)le i R has4 on figiires given by H. Fromherzand H. J .MeX2,- are established.Wave-length (mp) of bund maxima of the complexes of the form Mexi-.Co-ordinating ion. Zn2+. Cd2+. fin2+. HgZi. C U ~ ~ . Pbzt'.c1- ........................... - 187-5 218.5 228.5 250 272........................... 215.5 245 250 98 1 304 Br -I- .............................. 238.5 2557 "0 ..d '3 9 3 I 363.3-In the crystalline state the co-ordinately " saturated " complexes arefavoured a t the expense of the " non-saturated " forms, a condition arisingprobably from reasons of symmetry, but in solution this would not generallybe expected to be the case.In solutions of FeX,, E. Rabinovitch andW. Stockmayer 78 found that the intermediate forms FeX2 +, FeX,+, and FeX,predominate over the Fe3 + and FeXi- species throughout a wide range ofconcentrations. This view was reached from a study of the spectra ofsolutions of ferric perchlorate, chloride, and bromide a t different acidities,ionic strengths, concentrations, and temperatures. These observationsenabled the authors to determine the extinction curves and the associationconstant for the first three ionic complexes FeC12 +, FeCl,+, and FeCl,.All thespecies possess intense bands in the ultra-violet which extend into the visibleand contribute to the colour of ferric chloride solutions. Their associationconstants, at 26.7" and an ionic strength p = 1, are given by the expressionsKPeC12+ = [FeC12+]/[Fe3 k][Cl-] = 4.2, KPeCl,+ = [FeC12+]/[FeC12 t][Cl-] = 1.3,KF,.Cl, = [FeCl,]/[FeCl,+][Cl-] 0.04. Of these constants, KFeC12t can beconsidered as fairly accurately determined, but the other two are only roughestimates.The absorption curve of FeC12+ can be obtained practically free from thoseof the higher associates [although " contaminated " by the extinction due tothe products of hydrolysis Fe(OH),+ and Fe(OH),+] by using a low, constantconcentration of hydrochloric acid, and comparatively large concentrationsof Fe(ClO,),. If HClO, and NaC10, are added in quantities required tomaintain constant acidity and ionic strength, the resulting curves show theaverage extinction coefficients as a function of the ferric-ion concentration.The curve8 can be represented by a formula for a single association equili-brium and thus allow the calculation of the extinction coefficient of FeC12 + .(ii) Ions derived from the Interaction of Oxy-acids and Hydrogen Ions.--Itis well known that certain reactions, particularly many oxidation reactions in7 7 2.phyaikal. Chem., 1936, A , 178, 29. J . Amer. Chem. SOC. 1942, 64, 335WEISS : SOME INTERMEDIATE (!OMPOIJNDS I N INORGANIC REACTIONS. 73solution involving oxy-acids, are dependent on an acid medium and are oftengreatly accelerated by hydrogen ions.At present only relatively fewpositive ions, e.g., those derived from nitrous and nitric acid, have beeninvestigated sdiciently thoroughly by chemical and physical methods fortheir existence to be regarded a fully established.In many other cases certain conclusions regarding the presence andstability of such ions can be drawn from kinetic measurements. However,in this group, there is relatively little additional evidence available at present.Nevertheless, there is, perhaps, some justification in discussing a few of theseions, even if it only results in providing a summary of some revelant facts.In 1909, from observations basedon cryoscopic measurements in sulphuric acid solutions A.Hantzsch 79suggested that nitrosyl ions were present in solutions of the so-called" nitrosylsulphuric acid " as nitrosyl sulphate, to which he gave the formulaNO+,HSO,-. Later, with K. Berger,so he took up this problem and preparednitrosyl perchlorate in a fairly pure form and attributed to it the ionicstructure NO+,C10,-, since it behaved as an electrolyte in nitromethanesolutions. Subsequently, their opinion was supported by the study of theRaman spectra of nitrosyl perchlorate and of the sulphate (in solutions ofsulphuric acid), and the Raman shift 2330 cm.-l which appears very pro-minently in these spectra was assigned to the NO+ ion.81 Recently, L. J.Klinkenberg has confirmed this by X-ray analysis of the crystal structuresof nitrosyl perchlorate and nitrosyl borofluoride.In these crystals the NO +ion was found to be about the same size as the H,O+ ion and smaller thanthe NH,+ ion.Nitrosyl perchlorate can be preparediby the action of dinitrogen trioxideon a 707' perchloric acid solution, N,O, + 2HC104 = 2NO,C10, + H,O,and drying over phosphoric oxide. According to E. Wilke-Dorfurt andG . Balz,a nitrosyl borofluoride (N0,BF4) can be made similarly by treatinga concentrated solution of borofluoric acid with liquid trioxide.A. Hantzsch also has pointed out that nitrous acid does not existat, all in acid solutions as the equilibria NO,- +LH+ HNO, andHNO, + H+ =+= H,NO,+ + NOL + H,O are displaced in favour of theNO+ ion. It is noteworthy that, according to these equilibra, the con-centration of the nitrosyl ion in solution is given by [NO +] oc [NO,-][H+I2, andthe appearance of this expression in kinetic equations suggests the possibleintervention of NO+.It is also to be expected that nitrosyl ions play animportant part in the mechanism of the decomposition of nitrous acid inaqueous solutions.Nitrous Acid .- the nitrosyl ion (NO+)..79 Z . physikal. Chent., 1909, 65, 57.go Z . anorg. Chem., 1930, 190, 321.a1 W. R. Angus and A. H. Leckie, Proc. Roy. SOC., 1935, A , 149, 327 ; A , 150, 615 ;s2 L. J . Klinkenberg, Rec. Trav. chim., 1937, 56, 749.f'taw. Paraday SOC., 1935, 31, 958.2. anorg. Chenz., 1927, 159, 197; see also G. Balz and E. Mailander, ibid., 1934,Zl7, 161.c 74 TNORC: ANIC CHEMISTRY.Nitric acid : the nitracidium ion (H,NO,+) and the nitronium ion (NO,+).In his work on the optical absorption of mixtures of nitric and sulphuricacids, A.Hantzsch 84 found that the absorption bands in the ultra-violet, dueto the nitrate ion and to the undissociated nitric acid, were absent from thesesolutions. As a result of this and from his work on the electrical conductivityof these mixtures, he concluded that the nitric acid is largely convertedinto nitracidium ions H,NO,+ and possibly H,NO,’ formed accordingto (i) HNO, + H,SO, = H2N03~ + HSO,- and (ii) HNO, + 2H,SO, =H,NO,’ He also claimed to have isolated the correspondingnitracidium perchlorates ( H2N03 1 )( ClO,-) and (H,NO,I )( ClO,-), and hecarried out migration experiments on these systems and on nitracidiumperchlorates dissolved in nitromethane solutions.All these indicated thatthe nitric acid is present as a positive ion and that in pure nitric acid onepresumably has the equilibrium 2HNO,=+ H2N03 + NO,-. Recentinvestigations discussed below have shown that in anhydrous nitric acid andother anhydrous media H,NO,- is in fact dehydrated to NO,I according tothe equilibrium : H,NO, - NO, + H,O. The nitronium ion NO,+ is ofconsiderable interest, as i t has been recognised that it functions as theprimary nitrating agent in the mixed acids, as was suggested by H. Euler 85and P. Walden 86 and later by C. C. Price.8’ More recently, this workhas been taken up by C. K. Ingold 88 and G. M. Bennett 89 and theirrespective collaborators and by F.H. Westheimer and M. S. Kharasch.wThe evidence of the formation and existence of NO,’ has been carefullyreinvestigated and re-examined by these authors and who find that it isformed in the mixture of nitric and sulphuric acids : NO,*OH + 2H,SO, =NO, . + OH, + ZHSO,-. This is analogous to the formation of NO’ whennitrous acid is dissolved in sulphuric acid : NO*OH + ZH,SO, = NO +OH, ’ + ZHSO,-. In an equimolecular mixture of water and sulphuric acid,nitric acid is present mainly in the form of NO,*OH. An excess of waterconverts it into NO,- ions but an excess of sulphuric acid prodiices NO, ions.That the nitric acid is in a different form is obvious also from the very lowvapour pressure of nitric acid over these solutions.Extensive studies of theRaman spectra of mixtures of nitric and sulphuric acids have been made byJ. Ch6din,g1 who showed that these spectra are characterised by two prominent(polarised) lines at 1050 and 1400 cm.-l, which do not belong to the spectrumof either the nitric or the sulphuric acid molecule. The same line appeared-t 2HSO,-.Ber., 1925, 58, 941 ; see also A. Hantzsch and K. Berger, Ber., 1928, 61, 1328;2. phyYika1. Chein., 1930, A , 149, 161.85 2. angew. Chem., 1922, 35, 580. *’ Chem. Reviews, 1941, 29, 51 ; see also 34. UssanoGitch, Acta Physicochim.8 8 E. D. Hughes, C. K. Ingold, and R. I. Reed, Nature, 1946, 158, 448; E. S.13’ G . M. Bennett, J. C. D. Brand, and G. Williams, J., 1946, 869, 875.‘O J .Amer. Chem. SOC., 1946, 68, 1871.91 Ann. Chim., 1937, 8, 243; . J . Physiqzte, 1939, 10, 445; Mkna. Service china. de8 6 Ibid., 1924, 37, 390.U.S.S.R., 1935, 2, 239.Halberstadt, E. D. Hughes, and C. K. Ingold, ibid., p. 514.Z’E’tat, 1944, 31, 113WEISS : SOME INTERMEDIATE COMPOUNDS IN INORGANIC REACTIONS. 75if dinitrogen pentoxide or phosphoric oxide was added to pure nitric acid.The line of 1400 cm.-l is present also in the spectrum of 100% nitric acid butit is stronger in the mixed acids, its intensity increasing with diminishingwater or increasing oleum content. Chkdin assigned these lines to a specialform of dinitrogen pentoxide since in organic solvents it gave a differentRaman spectrum. Bennet, Brand, and Williams 89 interpreted these resultson the basis of the formation of NO,+, to which they assigned the line1400 cm.-l, comparison with the isoelectronic molecule CO, having shownthat a polarised Raman line is to be expected in this region.C. K. Ingold,D. J. Millen, and H. G. Poole 92 proved definitely that the Raman line a t1400 cm.-l is due to NO,', its this line could be obtained in solutions of nitricacid in perchloric and selenic acid, neither of which possesses a Ramanfrequency in this region, without the line a t 1050 cm.-l, which is actually dueto the NO,- and HS0,- ions.Cryoscopic measurements of nitric acid in sulphuric acid have been carriedout, particularly by A. Hantzsch and a number of other workers. Thequestion has been carefully re-examined by C. K.Ingold et CLZ.,~~ who showedthat nitric acid dissociates in sulphuric acid to produce a nearly fourfolddepression of the freezing point (van't Hoff factor = 3.82). This is inagreement with the formation of according to the above equation.Addition of dinitrogen pentoxide to fuming sulphuric acid also leads to theappearance of the Raman frequency at 1400 cm.-l, which presumably is due tothe reaction N,O, + H2S20, = NO, + HS,O,- + HO*NO,, which is followedby-additional formation of NO," from the nitric acid. Hantzsch had previouslycarried out migration experiments on nitracidium perchlorates in nitro-methane and on solutions of nitric in sulphuric acid, obtaining some evidencethat the nitric acid was present in a cationic complex. Bennett, Brand, andWilliams 89 confirmed this; they determined the distribution of nitric acidafter electrolysis and found a general accumulation of the nitric acid at thecathode and a deficit at the anode.They also showed that this was not dueto cataphoresis effects. As mentioned above, Hantzsch had claimed tohave prepared two nitracidium perchlorates. R. Goddard, E. D. Hughes,and C. K. Ingold 9* were able to isolate salts of the nitronium ion. Theyrepeated Hantzsch's work, employing an improved technique excludingatmospheric moisture and working a t low enough temperatures to preventthe formation of and contamination by nitrosoniurn perchlorate. They wereable to isolate pure NO;' C10,- by fractional recrystallisation from nitro-methane or by treatment of the original product with dinitrogen pentoxide.Nitronium perchlorate has been prepared also by W. E.Gordon and J. W. T.Spinks,95 who obtained a deposit of the composition NClO, by mixing gasstreams of ozone, nitrogen dioxide, and chlorine dioxide. According toIngold et U Z . , ~ ~ the compound NO,C10, has a low vapour pressure, scarcely9 2 Nature, 1946, 158, 480.O 3 R. J. Gillespie, J. Graham, E. D. Hughes, C . K Ingold, and E. R. A4. Peeling,94 Ibid.ibid.95 Cunadian J . Res., 19.10, B, 18, 93976 INOROANIO CHEMISTRY.fumes in air, and dissolves in water with slight liberation of heat. Apartfrom analysis, the constitution of the solid salt was confirmed by its Ramanspectrum, which consists of the known spectra, of NO2+ and ClO,-. Ingoldet aZ.g3 also established the formation of NO2+ in sulphuric acid solutions ofthe nitrogen oxides N20, and N204.These oxides produce an approximatelysix-fold depression (i = 5-85) of the freezing point, corresponding to theequations (i) N,06 + 3HiS0, = 2N0,. + H30+ + 3HS04-, (ii) N,O, +3H,SO, = NO,' + NO- + H30+ + 3HS04-. The authors were able toconfirm this by the Raman spectra of these solutions. J. Ch6ding1had already obtained the lines at 1400 and 1050 cm.-l from solid nitrogenpentoxide. On the basis of this Ingold et aLg2 suggested an ionic structureN0,+N03- for this compound. There can be hardly any doubt that,apart from nitration reactions the ions NO 1- and NO,+ are also of importancein the mechanism of other reactions involving nitrous and nitric acid.Forinstance, the behaviour in acid solutions can be represented by the simplemechanism HO*NO + H+ + H,O + NO and NO $- NO3- + NO+N03-of other nitrates of weak bases : NO NO3- + H20 = HNO, + H + NO,-.Hydrogen Peroxide.-A. Simon and F. F6her 96 suggested some time agothat some hitherto undescribed forms might be present in solutions ofhydrogen peroxide. They were unable, however, to get any clear evidencefrom a study of the Raman spectra which they investigated up to con-centrations of 99.5% hydrogen peroxide. The Raman spectrum of hydrogenperoxide still lacks a satisfactory interpretation. From a study of the Ramanspectrum of deuterium peroxide,97 i t has been concluded that the lines3395 and 1421 cm.-l in the H,O, spectrum are connected with O-H vibra-tions, and these lines, which become weaker on dilution, may have somebearing on this problem.Kinetic evidence rests on the accelerating effect of hydrogen ions on someof the reactions of hydrogen peroxide.This can be interpreted on theassumption that in the presence of hydrogen ions, hydrogen peroxide isconverted to some extent into peroxide acidium ion, H30,+ and posHibly OH Iaccording to the equilibria : 98A -- N204, and the hydrolysis of dinitrogen tetroxide falls into line with that'Equilibrium const.H,O,+H+ z$ H30,+ (K,) . . . . (4H302+ z$ H20+ OH+ (Kb) . . . . (PIH,02 e OH++OH- ( K J . . . . (Y)These equations are closely analogous to those for the H20 molecule (H,O,breplacing H30b and OH+ instead of H+) and they suggest that in con-centrated solutions of hydrogen peroxide the equilibrium 2H,02 + H302+ +H0,- may exist.The rate of the reaction between hydrogen peroxide and iodide ions inacid solution, leading to formation of iodine, can be represented by the9 6 2.Elektrochern., 1935, 41, 290.98 J. Weiss, unpublished results.9 7 F. FehBr, ibid., 1937, 43, 663WEISS : SOME INTERMEDIATE COMPOUPJDS IN INORGANIC REACTIONS. 7’7kinetic equation : 99 rate CC [H202][I-](1 + k[H ‘I}. This suggests two parallelinitial processes, wiz., (a) between H202 and I- and ( b ) between H302+ and I-.From this i t is possible to get a rough idea of a set of pos-sible values forK,, Kg, and K,. One obtains, for instance, with certain plausible assump-tions K , - Kg - 10-2, K , - 10-l8 (at 25’).A study of the Ramangpectra and of absorption spectra of hydrogen peroxide in the presence ofstrong acids might give some further information on these points.Chromic Acid-The rate of oxidat,ions by chromic acid in solution, whichhave been investigated by a number of authors, depend on the hydrogen-ionOH + concentration. More recently a very carefulinvestigation has been carried out by F. H.Westheimer and A. Novick,l who studied theoxidation of isopropyl alcohol in acid solution. [ lfe----j;Me *n ] Taking into account the equilibrium Cr,O;- +H,O + 2HCrO,-, they found the followingequation for the rate of the reaction : rate oc [HCr0,-][H+12[C,H7*OH], andmggested an activated complex of the forin shown in the inset.This can be represented also by the forination of a positive ion according10 H2Cr0, + H ‘ =+ H,CrO,l, or possibly of its dehydration productHCrO,’ , followed by the process H3CrO,I + C,H,*OH ---+ H,CrO, +CH,*(JOH)*CH, + H &, with the formation of a quinquevalent chromiumcompound (H,Cr04 or HCrO,).There can be hardly any doubt that, ingeneral, the conversion of the sexavalent into the tervalent chromium saltproceeds in stages. It has been found that manganese dioxide is formedif the reaction is carried out in the presence of manganous salt, and aschromic acid does not bring about this oxidation, the authors concludedthat an intermediate labile quinque- or quadri-valent chromium compound isformed which does so. Similar evidence for the intermediate formation ofthese compounds has been presented by other authors.2Permnganate.-It is possible that the ion H2Mn0,” is present in acidsolutions of permanganate which, as is well known, behave differently fromneutral or alkaline solutions in oxidation processes.There is perhaps somesignificance in the fact that permanganate in concentrated sulphuric acid orin oleum gives dark green or blue solutions, and it has been suggested thatin these solutions (MnO,),SO, might be present which could also imply thepresence of the cation MnO, +.Perchlmic Acid .-Certain physical properties indicate that perchlor i cacid exists in the aci-form, HClO,, and the pseudo-form, HO-ClO,, a i dalso that in highly concentrated solutions an acidium salt is formed :ZHClO, *( H,ClO, ) + (C104-).According to R. Fonteyne? the Raman/H--°Cr-oH(a) J . A. Christiansen, 2. physikal. Chem., 1925, 117, 433, 448; ( b ) J. Brode,F. H. Westheimer and A. Novick, J . Chem. Physics, 1943,ll. 506.C. Wagner, 2. anorg. Chem., 1928, 168, 279; V. F. Stefanoskii and A. M. Zanko,T. E. Thorpe and E. J. Hambly, J., 1888,53, 175, 182. ‘ Nature, 1936, l a , 886.ibid., 1901, 37, 257; 1904, 49, 208.Acta Physicochirn. U.S.S.R., 1938, 9, 635; R. Lang, Mikrochim. Acta, 1938, 3, 11378 INORGANIC CHEMISTRY.spectrum of absolute perchloric acid is different from that of the aqueoussolution; he also attributed the line at 422 cm.-l to the presence of anacidium salt. A. Simon has pointed out that this cannot be justified on theevidence of the Raman spectrum alone, although it is clear that there is aprofound change in the Raman spectrum of perchloric acid on dilution.According to the more recent work of R.Fonteyne and 0. Redlich,E. K. Holt, and J. Bigeleisen,' Raman spectra certainly confirm theexistence of HOC10, (symmetry C3J when the concentration exceeds 73 ?(, .Raman lines found for the C10, group are also present in the spectrum ofdichlorine heptoxide which points to the ionic structure ClO, ' ClO,-.Recent investigations 8 of concentrated solutions of perchloric acid haverevealed ultra-violet absorption in the region of 2 2 0 0 ~ . which cannot bedue to the C10,- ion as this does not absorb appreciably above ca. 1900 A.gThis increased absorption must be due either to the undissociated acid orpossibly to the ClO,+ ion.Periodic Acid.-It has been found lo that in the reaction between periodicacid and iodide ions in acid solutionthe rate is given by : rate OC [I04-1[H 'I2[I-1,which suggests the formation of an acidium ion according to : (i) 10,- +H r =+ HIO,, (ii) HIO, + H t + H2104', followed by : (iii) H2104+ +I- .+ H,IO, + I, (iv) H2IO4+ 10, + H,O, (v) 10, + I- --+ 10,- + I,leading first of all (in two steps) to the formation of iodate.Ramanspectra of solutions of periodic acid are not available at present. Recentexperiments l1 show that aqueous solutions of periodates and periodicacid show strong absorption in the ultra-violet from 2300 A. onwards.Solutions of periodic acid in concentrated sulphuric acid do not absorbnearly as strongly.This could he due either to the undiasociated acidor to the formation of H,IO,' or 10,' in these solutions, which is suggestedby the somewhat analogous behaviour of nitric acid in concentrated sulphuricacid. It is possible also that the H4105+ ion is present in the mesodiperiodatesIodic Acid.-The Raman spectrum of iodic acid l2 contains a greatnumber of lines, but the experimental evidence is inadequate for itsinterpretation, although several authors believe that iodic acid is polymerisedin solution. H,IO,i- + 10,-exists. The formation of an ion H,IO,+ or 10, is suggested by kineticevidence on the oxidation by iodate in acid solution which is found l3 to beproportional to [H+]., A similar state of affairs seems to exist with regard(H,IO,T ,IO,- ).It is possible that the equilibrium 2HI032. anorg.Chenz., 1938, 239, 329.Nattiurwetensch. Tijds., 1939, 21, 6 ; 1938, 20, 275.J . Anaer. Chem. Xoc., 1944, 66, 13.H. Fromherz and W. Rlenschik, 2. physikal. Chem., 1929, B, 3, 18; H. Ley and* J. Weiss (in preparation).B. Arends, ibid., 1929, B, 6, 240.lo E. Abel and R. Siebenschein, ibid., 1928, 130, 631.l 1 ,J. Weiss (in preparation).l2 R. Fonteyne, A'atzitcrwete,i.srh. IT;'&., 1'339, 21, 141; C. S. Rao, Current Sci.,l3 S. Dushman, J. PhysicaZ Chenz., 1904, 8, 453.1942, 11, 429WEISS : SOME INTERMEDIATE COMPOUNDS IN INORGANIC REACTIONS. 71)t o bromic acid, where i t is found that the rate of reaction between bromateand iodide ions in acid solutions l4 is proportional to [Br0,-][H+]2[I-]2,suggesting the primary process H2Br0,1 + I- ---+ H,BrO, + I which isbased on the formation of H,BrO,T or BrO,+ ions. The interaction of theundissociated molecules (ion-pairs) cannot, of course, be excluded.Hypobromic Acid and the Br+ Cation.-The HBrO molecule occurs in theniechanism of a number of reactions in solutions involving Br, or Br .Recently, C.N. Hinshelwood 15 has made the important and far-reachingsuggestion that many of these reactions could be explained more satisfactorilyby the action of the Br+ cation formed thus :Br, Br- + Br- . . . . . . . . (2, 1)It is clear that, as for instance in the case of nitrous acid, the exclusivepresence of HBrO in acid solutions is very unlikely, for the equilibriumHBrO+H & Br'-+H,O .. . . (2, 1 4is probably shifted considerably towards tho right.alkaline solutions, where BrO- is formed by the reactionThis is different illHBrO -+- OH- =+= Br0- -+ H,O . . . (2, l b )Aocording to Hinshelwood, the initial reaction between bromine and oxalicacid can be represented by the simple electron-transfer process, HC,04- +Rr -+ HC,O, + Br, which accounts very wcll for the experimentalfacts. l6Similarly, in the reaction between bromine and hydrogen peroxide,H,O, + Br, = 2HBr + O,, the initial primary process can be representedbyleading to the formation of HO, which, according to (1, 4), is capable ofgiving molecular oxygen and hydroxyl radical.H0,-+ Br'+HO,+ Br . . . . (2,2)The latter can then react :Br-+OH+Br+OH- . . . . . (2,3)followed by 2Br = Br, . . . . . . . ( 2 , 4 )One can show easily that the mechanism involving the reactions (2, l ) ,( 2 , 2), (2, 3), (2, 4), and (1, 4) leads to the correct kinetic equation. On theassumption that the equilibrium (2, 1) is practically always established oneobtains the equation for the rate of the reaction which is in agreement withthe experimental facts.l7Hypoiodous Acid and the I Cation.-This case is similar to that of bromine,and the equilibria corresponding to (2, l ) , (2, la), and (2, Ib) are probablystill further shifted towards I . There is also some other chemical evidencefor the I - ion. L. Birkenbach and J. Goubeau l8 have suggested the inter-l4 A. A. Noyea and W. 0. Scott, 2. physikal. Chew., 1895, 18, 118.l5 J., 1947, 694.l6 R. 0. Griffith, A. bIcKeown, and ,4. G . Winn, Trans. Faraday SOC., 1932, 28, 107.W. C. Bray, Chem. Reviews, 1932, 10, 161. Ber., 1932, 65, 39580 INORGANIC CHEMISTRY.mediate formation of I +,ClO,- if iodine is treated with silver perchlorate.This compound is a very effective agent for iodination in the benzene nucleus.I. Masson and C. Argument l9 have suggested that Chrbtien's yellowsulphate should be represented as (IO),SO, and that IO+ cations are present.According to I. Masson,20 these can react in solution in the presence of iodineas follows : (i) 10' + H,O =+= 13+ + 20H- (ii) 13+ + I, 3I+. Intro-duction of I+ into the mechanism of some of the reactions of iodine leads tointeresting results.The older work of J. R. Roebuck 21 on the oxidation of arsenious acid byiodine is not capable of a straightforward interpretation on these lines as theexperiments were not carried out at well-defined hydrogen-ion concentrationsand it is not clear whether H,AsO, or H,AsO,- enters into the mechanism.22The reaction between hydrogen peroxide and iodide ions in the absenceand in the presence of iodine has been studied by a number of w ~ r k e r s . ~ ~ @ The existing theories, which are all based on the intermediate formation ofHIO, lead to complex rate equations which are difficult to reconcile withcertain experimental facts.= Recently, it has been found25 that theessential features of the mechanism of this reaction can be represented by thefollowing six simple processes :Rate constantsH,O,+I-+I+OH- +OH . . . * (k1)21 ---+ I,123,1+ +I- . . . . (kII, 6 1 )H 0 2 - + I - - 3 H 0 , + I . . . . . . (kIII)I-+OH + I + O H -+ +H,O, + HO, ---+ 0, + H,O + OHI n strongly acid solution the hydrogen-ion catalysed, primary process,which has been discussed in connection with the formation of H,O,+, mustbe added. However, this is not an essential point in the mechanism as awhole. The above equations lead to the following simple expressions forthe formation of iodine and oxygen in the stationary state := k,[H,O,][I-] - 1 d[O,I dtThe two well-known limiting cases are easily obtained as follows : (a) Inacid solutions only iodine is formed (without evolution of oxygen), ie.,J., 1938, 1702. 2o Ibid., p. 1708.21 J . Physical Chem., 1908, 6, 365; 1905, 9, 727.22 Cf. P. Goldfinger and H. D. Graf von Schweinitz, 2. phyaikal. Chem., 1932,23 Cf. E. Abol, ibid., 1928, 138, 16.26 H. A. Liebhafsky, ibid., 1931, A , 155, 289.B, 19, 219.26 J. Weiss, unpublished reaultsWEISS : SOME INTERMEDIATE COMPOUNDS I N INORGANIC REACTIONS. 81d[O,] /dt = 0, whence d[I,]/dt = EI[H202][I-] (neglecting the acid-catalysedparallel process H302+ + I- + I + H,O + OH). ( b ) In neutral solutionsthe case of " pure catalysis " can be realised, i.e., d[12]/dt = 0, and one obtainsfrom the above equations for the rate of oxygen evolution : d[O,]/dt =J. W.R. E. DODD.P. I,. ROBINSON.J. WEISS