INORGANIC CHEMISTRY.1, INTRODUCTION AND GENERAL.THE general arrangement of this Report is the same as in past years. Twotopics have been selected for special treatment. The first deals with recentdevelopments in the chemistry of the carbonyls and nitrosyls. Though thissubject was discussed in the Annual Reports for 1934 much new researchhas been published since, many of the developments being based on applic-ation of high-pressure technique to the production of carbonyls. The secondtopic, isotope exchange in inorganic chemistry, is one which is of growingimportance in several fields. Radioactive isotopes of all the elements exist,many of them being of suitable half-period for exchange work, while usefulmethods of separating non-radioactive isotopes have also been worked out.The potential value of this method of studying problems in inorganicchemistry is, therefore, very great.The first section of the Report deals with a comparatively small number ofrecent investigations in scattered fields.No detailed discussion is given ofpapers on complex salts, but attention may be directed to the BakerianLecture on Stereochemical Types and Valency Groups 1 and to the LiversidgeLecture on Complex Formation.2 These two lectures present a systematicbasis for the formulation of a wide range of compounds and are a contributionwhich is of major importance in the development of the theory of valency.A further publication of wide interest is the Report of the Committee for theReform of Inorganic Chemical Nomenclature.3 The recommendations madecover specifically many points which at present are confusing, though in themajority of cases they call for no departure from common practice.A compound of xenon with phenol having the composition Xe,2PhOH hasbeen prepared by B.A. Nikitin.4 This has a dissociation pressure of 760mm. at 4O, its stability being approximately 0.3 of that of the compoundH,S,SPhOH. This work follows earlier investigations by the same authorin which indications of compound formation by radon were obtained.5Brief reference was made in the Annual Reports for 1940 to an entirelynew group of boron hydride derivatives, the metallo-borohydrides. Detailsof fhe preparation and properties of three substances of this type werepublished at the end of 1940.6 The table on p.66 shows their formulze,melting points, and boiling points, together with data for diboranc.N. V. Sidgwick and H. M. Powell, Proc. Roy. Soc., L940, A, 176, 153.N. V. Sidgwick, J., 1941, 433.W. P. Jorissen, H. Basseft, A. Damiens, F. Fichter, and K. Remy, J., 3940,Compt. rend. Acad. Sci. U.R.S.S., 1940, 29, 571.H. I. Schlesinger, R. T . Sanderson, and A. B. Burg, J . Amer. Chem. SOC., 1940,62, 3421; A. B. Burg and H. I. Schlesinger, ibid., p. 3425; H. I. Schlesinger andH. C. Brown, ibid., p. 3429.*1404; J . Amer. Chern. SOC., 1941, 63, 889.Ibid., 1939, 24, 565.REP .-VOL. XXXVIII . 66 INORGANIC CHEMISTRY,Formula. BzH'a. A1B,Hl2. BeBS,. LBH ,M. p. ..................... -165.5" - 64.5" 123' 275"B. p...................... - 92.5" 44.5" 91.3" decomp. 275"1'. p. at 0", mm. ......... very high 119 0.5 < 10-6The first of these compounds, aluminium borohydride, was prepared byheating diborane with trimethylaluminium at 60" for several hours. In thepreparation of the beryllium compound an unstable intermediate productof the approximate formula CH,*Be*BH, was first produced by the inter-action of dimethylberyllium and diborane. Subsequent reaction of thiswith further diborane gave beryllium borohydride, BeB,H,, together with anon-volatile by-product of the empirical formula BeBH,. The lithiumborohydride was prepared directly by the reaction of ethyl-lithium and di-borane. The only other major products in these reactions were the boronalkyls.The technique employed throughout was similar to that used inearlier investigations of the boron hydrides by Schlesinger and his co-workersand was based largely on methods developed by A. Stock.There are interesting differences in the physical properties of the threeborohydrides. For instance, the relatively high melting point and lowvolatility of the lithium compound, together with its insolubility in benzene,suggest a t once that the bonding is polar in character. The volatility ofaluminium borohydride and its ready solubility in benzene show that it isprobably non-polar and may be classed with diborane itself. The berylliumcompound is intermediate in type, but the indications are that it is moreakin to lithium borohydride.This difference in properties is borne out by the chemical behaviour, forlithium borohydrids, unlike diborane, is stable in dry air a t ordinary tem-peratures whereas the aluminium compound is spontaneously inflammable,as is diborane.The beryllium compound also reacts very readily with air.The reactions of these various molecules with trimethylamine again exhibita striking gradation. Diborane and trimethylamine react at temperaturesdown to - looo, yielding BH,,N(CH,),. Aluminium borohydride reacts a t-80" to 25*, and although the changes are complex and have not been fullyelucidated, the crystalline compound AlB,H,,,N(CH,), has been isolatedfrom the products of reaction of equimolecular quantities at - 80'. Whenheated to 100' this breaks up and borine trimethylamine, BH3,N(CH,),, hasbeen identified as one of the products.At -80" compounds containing ahigher molecular proportion of trimethylamine are formed. Berylliumborohydrids forms the addition compound BeB,H,,N( CH,), with trimethyl-aminc and this reacts further with the amine according to the equationBeB,H,,N(CH,), + N(CH3), + BeBH5,N(CH3), + N(CH,),,BH,Lithium borohydride, on the other hand, does not react with trimethylamine,a fact which emphasises the difference between this compound and thealuminium and beryllium borohydrides.The ionic character of lithium borohydride is further supported by therapid reaction of a benzene solution of aluminium borohydride with ethylEMELBITS : INTRODTTCTTON AND C~EENERAT,. 67lithium to produce lithium borohydride as a white precipitate.This appearsto be an ionic double decomposition. The suggestion is made that all threeof the borohydrides contain the [BH,]- ion, but that i t undergoes a certainamount of distortion. This is a minimum in the lithium compound, and isgreatest under the influence of the smaller and more highly charged alumin-ium ion, with the result that the aluminium compound approximates to thecovalent type. The configuration of AIB,H,, has been determined byelectron diffraction,' and it is found that the aluminium atom and the threeBH, groups are coplanar with angles of 120" between the A1-B bonds.The compounds A1(BH4),, Be(BH,),, and LiBH, have each a two-electrondeficit per BH, group and the B-H distances are greater than those expectedfor electron-pair bonds.They are equal, in fact, to the distances in di-borane, tetraborane, and pentaborane, and there is little doubt that theconstitution of the borohydrides can be considered in terms of the theory ofresonance in the same manner as that of other boron hydrides.A number of reactions of the borohydrides have been studied bySchlesinger, Burg, and their co-workers. All are readily hydrolysed. Theyalso react with hydrogen chloride at - SO", producing hydrogen, diborane,and the metal chloride. The aluminium compound was shown to undergoa complex reaction with ammonia, and it also reacted with dimethyl etherto form the compound AlB3H1,,0(CH,)2, which decomposed a t 50". Lithiumborohydride reacted with methyl alcohol according to the equationLiBH, + 4CH,*OH --+ LiB(OCH,), + 4H2 --+ LiOCH, + B(OCH,),A derivative of the type Li(BR,) was obtained by the interaction ofethyl-lithium with trimethylboron, either in presence or in absence of asolvent.A white crystalline product with the formula ' LiB( CH,),C,H,resulted. It is evident that there are many aspects of these investigationsin which important progress may be expected as the experiments areextended.The association of the alkyls of aluminium, gallium, and indium in thevapour phase has been studied by vapour-density determinations.8 Tri-methylaluminium proved to be anomalous in that the vapour consistsof the dimer a t 70". It dissociates with increase of temperature up to1 SOo, at which point decomposition sets in.Triethylaluminium is 12 yoassociated to the dimer at 150". Trimethylga'llium and trimethylindium werefound to yield monomeric vapours, and the trimethyl and triethyl derivativesof boron are also known to be monomeri~.~ The monomeric charact'er oftrimethylindium is also shown by electron-diffraction measurements.10The oxychlorides of silicon have been reinvestigated by W. C. Schumband D. F. Holloway.11 These compounds, which hitherto were ill defined,J. Y. Beach and S. H. Bauer, J. Amer. Chem. SOC., 1940, 62, 3440.A. W. Laubengayer and W. F. Gilliam, ibid., 1941, 63, 477.9 A. Stock and F. Zeidler, Ber., 1921, 54, 531.lo L. Pauling and A. W. Laubengayer, J. Arner. Chem. h'oc., 1941, 63, 480.l1 Ibid., p. 275368 INORGANIC CHEMISTRY.were prepared by the action of a mixture of chlorine and oxygen on crystal-line silicon at a dull red heat.The reaction product on fractionation yieldedcompounds of the formulae (SiOCI,)4, Si20C1,, Si,02C1,, Si,O,C1lo, Si,O,Cl,,,Si60,CI,, and Si,0,Cll6. The first was a crystalline solid, the others beingliquids of low volatility, all of which were of the type SLO, - ,C1,, + 2, andwere analogous to the oxybromides of silicon.12Reaction of these oxychlorides with absolute alcohol has so far yieldedthe following esters : Siz0(OC2H5),, Si302(OC2H5),, Si403(OC2H5)lo andSi,05( OC2H5) 14. These are non-inflammable high-boiling liquids whichresist hydrolysis by water at 100'. In an extension of these investigationshexacyclohexyloxydisiloxane, ( C,HllO),Si-O-Si( C6H110),, has been preparedfrom cyclohexanol and the oxychloride Si,0C16.13Attempts to prepare silicon compounds of the type SiR,*SiR, andSiR,*O*SiR, from the corresponding silicon halides by the Wurtz synthesisnormally lead to fission of the Si-Si or Si-0-Si bonds.It has been found,however, that by first preparing the sodium compound from an aryl halideand then allowing it to react with the silicon halide, compounds of the requiredtype may be obtained.14 For example, phenylsodium, prepared from chloro-benzene and sodium, reacts with Si20Br6 to yield (C,H5),Si20, and Si2C16gives hexaphenyldisilane .Polymers derived from methylsilicon oxides have assumed considerableimportance during the past year owing to the possibility of their technicalapplication.The simplest starting materials for preparing such polymersare the methylsilicon chlorides, which have been described by W. F. Gilliam,H. A. Liebhafsky, and A. F. Winslow.15 Dimethylsilicon dichloride andmethylsilicon trichloride are formed by the interaction of methylmagnesiumchloride and silicon tetrachloride in the appropriate proportions. Theproducts were fractionally distilled and the boiling points of the two halidesin an approximately pure state were found to be 69.0-70*2° and 66.2-67.09, respectively. These values are anomalous in that they lie outside therange fixed by the boiling points of tetramethylsilicon and silicon tetra-chloride. IPolymeric methylsilicon oxides may be produced by hydrolysis of eitherthe separated or the mixed alkylsilicon halides.16 In either case the halogencompound, dissolved in ether, was poured on ice.The resulting methyl-silicols dissolved in the ether and remained on evaporation of the solvent asa syrup which could be hardened. With an average of 1-0-1-3 methylgroups per silicon atom, the condensation proceeded a t room temperaturethrough a sticky, syrupy stage to a hard, transparent resin, which becamebrittle when warmed. With a ratio of 1.3-1.5, the product a t room tem-perature was an oily liquid, the viscosity of which increased with risingtemperature dntil at 150-200" gelation occurred in a few hours. The gel12 W. C. Schumb and C . H. Klein, J . Amer. Chem. SOC., 1937, 59, 261.1s W. C. Schumb and D.F. Holloway, ibid., 1941, 63, 2853.l4 W. C. Schumb and C . M. Saffer, jm., ibid., p. 93.16 Ibid., p. 801. l6 E. G. Rochow and W. F. Gilliam, &id., p. 795EMEL~TJS : INTRODUCTION AND GENERAL. 69then set gradually to a transparent, horny resin. With CH, : Si ratios of16-1-9 a gel formed in a few hours at 200°, and this product became brittleonly when heated for several weeks at 200". With a CH, : Si ratio of 1.3-1.5 the initial condensation products were soluble in hydrocarbons or alcohols,although the fully hardened resin was insoluble and infusible. The hardenedresin had a high thermal stability; e.g., when a sample was heated in air fora year a t 200' no perceptible change occurred. At 300' in air surface oxid-ation took place. Samples heated for 16 hours in a vacuum at 550' or for 1hour in hydrogen at 450' suffered only a discolouration.A further investigation on this important group of compounds has beenpublished by J.F. Hyde and R. C. DeL0ng.l' These authors studied thehydrolysis of phenylethyl- , phenylmethyl- , diphenyl- , diethyl-, and dimethyl-dichlorosilanes. The products were liquid except those from the diphenylcompound, the diol from which was a crystalline solid. This when treatedin hot alcoholic solution with dilute ammonia or dilute sodium hydroxidewas converted into a cyclic trimer, which was characterised. When theliquid hydrolysis products of tho compounds containing phenyl groups weretreated at high temperatures with aqueous hydrochloric acid there was agradual increase in viscosity, accompanied by evolution of benzene.Theultimate product was an insoluble resin. The dialkyl derivatives werefound, in general, to be more gel-like than those containing a phenyl group,but they had less physical strength after curing than the latter. All filmsformed from solutions of the various resins lost their tackiness and solubilitywith appropriate baking. The final products resembled films formed bydrying oils, though they were superior in thermal stability.There is at present an insufficient basis for discussion of the constitutionof these polymers, but E. G. Rochow and W. F. Gilliam16 suggest thatthe cross-linked siloxane structure shown below probably represents anessential part.-R2Si-O-R~i-O-R2Si-O- .. .-R,Si-O-RSi-O-Ryi-O- . . .Two new iodo-derivatives of monosilane, vix., SiH31 and SiH,I,, havebeen prepared by the interaction of dry hydrogen iodide and monosilane inthe presence of aluminium tri-iodide, which acts as a catalyst.ls The boilingpoints are 45.4' and 149.5', respectively. Neither of the compounds isspontaneously inflammable in air, although both burn readily. In moist air,simultaneous oxidation and hydrolysis take place. Silyl iodide, SiH,I,reacts with mercury in sunlight, yielding mercurous iodide, silane, anddbilane. The liquid iodide and mercury in absence of sunlight farm anunstable, white, crystalline solid which decomposes into mercurous iodideand silane. There is a similarreaction with zinc. With magnesium in diisoamyl ether there is strong5)This may be mercurysilyl iodide, SiH,*HgI.J.Amer. Chem. SOC., 1941, 63, 1194.H. J. Emelbus, A. G. Maddock, andC. Reid, J., 1941, 36370 INORGANIC CHEMISTRY.evidence that a Grignard compound is produced, and treatment of this withwater gives monosilane. Liquid silyl iodide reacts explosively with silvercyanide, forming silver iodide and a brown polymeric solid. By passing thevapour of silyl iodide over silver cyanide a t room temperature the compoundSiH,CN, in. p. 34", is produced. Silyl iodide reacts with sodium, formingdisilane.The reducing action of sodium, potassium, or calcium on liquid ammoniasolutions of nickel salts has been shown to yield free nickel.19 The metal ispyrophoric and catalyses further reaction between alkali or alkaline-earthmetal and ammonia.Nitroxyl perchlorate, NCIO,, has been prepared by the controlled inter-action of chlorine dioxide and an air stream containing ozone and nitrogenoxides.20 It is a white crystalline solid with a vapour pressure less than 0.05mm. a t room temperature.It decomposes at 120' and reacts violently withmost organic liquids. Reaction with water yields a mixture of nitric andperchloric acids, and the compound is regarded as a mixed anhydride ofthese two acids.The direct oxidation of phosphorus tribromide to the oxybromide bymeans of oxygen is not readily controlled and is liable to occur explosively, inwhich case phosphoric oxide and free bromine are among the products. Ithas been shown, however, that under the catalytic influence of nitrogendioxide and with carefully controlled reaction conditions a smooth reactionmay be obtained.21 The direct preparation of phosphorus-halogen com-pounds from phosphoric oxide and metal halides has also been reported.22When mixtures of phosphoric oxide and calcium fluoride are heated to 550°,phosphorus oxyfluoride is the chief volatile product.In iron vessels acertain amount of reduction to the trifluoride also takes place. Whenmixtures of calcium fluoride and sodium chloride were used, the productswere PF,, POF,, POFzCI, POFCI,, and POCI,. I n these reactions the hithertounknown difluorophosphoric acid, HP02F,, which was probably formed bypartial hydrolysis of phosphorus oxyfluoride, was also isolated.In a group of publications dealing with polyiodides of the alkali metals,T.R. Briggs and his co-workers report extensive phase-rule studies.2, Inthe system sodium iodide-iodinewater they obtained the three polyiodidesNa41,4,13-15H20, Na,I,1,17-19H20, and Na,18,10-11H20. There wasno indication of a solid tri-iodide. Similar studies with rubidium iodideshowed the existence of anhydrous RbI,, and in the system czsium iodide-iodine-water the existence of the two binary compounds, CsI, and CsI,.was established. Two compounds of the formule K13,C6H, and KI,,2C6H,Small amounts of nickel amide may also be formed.l D W. M. Burgess and J. W. Eastes, J . Amer. Chem. SOC., 1941,63, 2674.2O W. E. Gordon and J. W. T. Spinks, Canadian J. Res., 1940, 18, B, 358.21 C.R. Johnson and L. G. Nunn, jun., J . Amer. Cheni. Xoc., 1941, 63, 141.p p G. Tarbutton, E. P. Egan, jun., and 8 . G. Frary, ;bid., p. 1782.23 T. R. Briggs, W. F. Geigle, and J. L. Eaton, J . Physical Chem., 1941, 45,, 59.5;1'. E. Briggs, C. C. Conrad, C. C. Gregg, and W. H. Reed, ibid., p. 614; T. R. Briggs antiS. S. Hubard, ibid., p. 806WELCH : METALLTC CARBONYLS AND NTTROSYLS, $1have been prepared by the addition of potassium iodide to a saturatedsolution of iodine in ben~ene.~4In a study of the action of chlorine on hydroxides of lithium and potassiumin presence of iodine, R. K. Bahl and S. Singh 25 have shown that the productfrom a boiling solution of iodine in lithium hydroxide is Li,I,011,2H,U,whereas potassium hydroxide under similar conditions gives KIO,.Whenchlorine is passed into hot solutions of barium or strontium hydroxide, or ahot suspension of calcium hydroxide, containing dissolved iodine, the iodatesBa(I03),,H20, Sr(103)2,H,0 and Ca(103)2,H20 are precipitated, no periodatebeing formed. €3. J. E.2. METALLIC CARBONYLS AND NITROSYLS.Since the chemistry of the carbonyls and nitrosyls was last discpssedin these Reports1 a number of notable developments in this field havebeen reported. Some of these have already been reviewed,29 but a generalsurvey including details of the most recent advances is opportune.The known metallic carbonyls, carbonyl hydrides, and carbonyl halidesare listed in the table on p. 72. Of these compounds the most accessibleare nickel tetracarbonyl and iron pentacarbonyl, both liquids under normaltemperature and pressure conditions, which are obtained on the com-mercial scale by the direct action of carbon monoxide on the finely-dividedmetals.2, Nickel tetracarbonyl is unique in that it is prepared in excellentyield by this direct method without the use of high pressures ; " dry " methodsfor the preparation of the other carbonyls, including iron pentacarbonyl (fromwhich the other iron carbonyls 4 and the tetracarbonyl hydride are prepared),require the use of high-pressure technique, and it is to the systematicapplication of this technique that many of the recent advances are due.The high-pressure apparatus and methods used by W.Hieber andhis collaborators in their more recent studies on the carbonyls havebeen described.5 The rotating autoclave and auxiliary apparatus aredesigned to withstand pressures up to 350 atm.The chief constructionaldifficulty lies in the choice of material for those parts of the apparatuswhich come into contact with carbon monoxide under pressure; alloyscontaining iron, nickel, etc., cannot be used since these metals combinereadily with carbon monoxide to form their carbonyls. Copper-silveralloys have been found to be most suitable for the autoclave lining; purecqper can be used a t temperatures up to 200°, but above this temperature24 J. A. Fialkov and A. B. Polischtschuk, Ber. Inst. Chem. Ahad. Wiss. Ukrain.,25 J . Indian Chem. SOG., 1940,17, 167, 397.W. Wardlaw, Ann. Reports, 1934, 31, 99.W.Hieber, 8. Ekktrochem., 1937, 43, 390.A. 9. Blanchard, Chem. Reviews, 1937, 21, 3 .1940, 7, 95.4 E. Speyer and H. Wolf, Ber., 1927, 60, 1424 [J?e,(CO),];6 W. Hieber, H. Schulten, and R. Marin, ibid., 1939, 240, 261.W. Hieber, 2. anorg.Chem., 1932, 204, 165 ([Fe(CO),]3)72 IXORQAMC CHEMISTRY.co 27Ni 28the various components become welded together and the apparatus cannotbe dismantled. The high-pressure technique used in the preparation ofthe ruthenium carbonyls has also been described.6Metaltic Carbonyls, Carbonyl Hydrides, and Carbonyl Halides.Carbonyl CarbonylElement. At. no. Carbonyls. hydrides. halides.Group I ' {Z 2979~t$o j3X4iPt(CO)X,'cu(c0jx"Au( C0)XjNotea :(a) In the table X represents chlorine, bromine, or iodine; metallic carbonyl(b) Compounds to which references are not appended are referred to in the text of0 Details not yet available; cf.W. Hieber, Z. anorg. C?mn,, 1941, 248.c W. Manchot and J. Konig, Ber., 1924, 57, 2130.d W. Manchot md E. Enk, ibd., 1930, 63, 1635.8 W. Mrtnchot and J. Konig, &id., 1925, 58, 2173.f Idem, ibid., 1926, 59, 883.0 Reported by W. Hieber; work not yet published.h W. Manchot and J. Konig, Ber., 1925, 58, 229.4 P. Schutmnberger, Bull. SOC. chim., 1868, [ii], 10, 188; Ann. Chim. Phys., 1868,j W. Manchot and H. Gall, Ber., 1925, 58, 2175.The direct preparation of cobalt tetracarbonyl from cobalt and carbonmonoxide is inconvenient, since carefully reduced cobalt must be employed.The methods of Schubert and of Coleman and Blanchard 7 use readilyaccessible materials, but give relatively small yields.These disadvantagesW. Manchot and W. J. Manchot, 2. Anorg. Chem., 1936, 226, 385; cf. alsoW. Hieber and H. Fischer, D.R.-P. 695,689 (1940), for ruthenium carbonyls.fluorides have not yet been reported.this Report.Idem, ibid., 1931, 201, 329.[iv], 15, 100; 1870, [iv],2l, 350. *7 Cf. p. 76, refs. (18) and (19)WELCH : METALLIC CARBONYLS AND NITROSYIS. 73are overcome in Hieber's method,5 in which cobalt sulphide is heated a t200" with carbon monoxide under 200 atm. pressure; the reaction proceedsquantitatively according to the equation 2CoS + 8CO + 4Cu + [CO(CO)~], 3-2Cu,S. The copper required in this reaction is obtained from the auto-clave lining, or from copper powder added to the cobalt sulphide.Thecarbonyl is extracted from the solid products with an organic solvent.If water is present in the reacting materials, considerable yields of thevolatile cobalt tstracarbonyl hydride, HCo(CO),, are obtained ; thishydride can also be prepared by high-pressure synthesis from (a) cobaltsulphide, carbon monoxide, and hydrogen ; ( b ) cobalt tetracarbonyl andhydrogen; ( c ) cobalt, carbon monoxide, and hydrogen; or ( d ) cobalthydride (CoH,) and carbon monoxide.The action of carbon monoxide at high pressures on cobaltous halides inpresence of certain other metals also affords cobalt tetracarbonyl, or, inpresence of water or hydrogen, the tetracarbonyl hydride.6 The conditionsand mechanism of formation of the tetracarbonyl by this process havebeen studied in some detail.8 Cobaltous iodide is completely converted intothe carbonyl by heating a t 150' for 15 hours with carbon monoxide under200 atm.pressure; the bromide and chloride give small yields of the car-bony1 at 200-300", but the fluoride does not react. The halogen is eventu-ally eliminated from the reaction by combination with copper in the auto-clave wall : 2C01, + 4Cu + 8CO + [Co(CO),], + 4CuI. This mode ofreaction persists even if direct contact between the cobaltous halide andthe autoclave is prevented by a glass sleeve, showing that some volatileintermediate is involved. In the case of cobaltous iodide this intermediateis probably Co(CO)I,, which can be obtained as brownish-black crystals bythe action of carbon monoxide at 100400 atm.on the anhydrous iodide,a t room temperature ; Co(CO)I, has a high dissociation pressure of carbonmonoxide and decomposes in a few seconds at normal temperatures andpressures. The yields of cobalt tetracarbonyl obtained from cobaltousbromide or chloride and carbon monoxide are considerably increased bymixing a metal powder with the halide in order to facilitate elimination ofthe halogen. The effectiveness of the metals tried decreases in the ordercopper, silver, platinum, gold; the heats of formation of the appropriatehalides also decrease in this order, showing that a controlling factor in theformation of the carboxayl is the reaction between the solid cobaltous halideand the admixed metal, e.g., CoBr, + 2Ag-> Co + 2AgBr. If argon issubstituted for carbon monoxide, reactions of this type give equilibriummixtures containing relatively little cobalt ; this indicates that carbonmonoxide plays an essential part in the reaction between the metal andthe halide, and does not merely combine with free cobalt formed in anindependent reaction.The use of a more electropositive metal, such aszinc or cadmium, leads to the formation of a crystalline, appreciably volatile" mixed carbonyl " of the metal and cobalf, probably to be regarded as aderivative of cobalt carbonyl hydride, e.g., Zn[Co(CO)&. It is stated 8* W. Hieber and H. Schulten, 2. a m g . Ckern., 1939,243, 14654 INORGANIC CHEMISTRY.that reactions corresponding with those described above also occur withnickel and iron halides.Cobalt tricarbonyl, [Co(CO),],, is readily obtained by heating the tetra-carbonyl a t about 52O.9Although a dicarbonyl chloride of iridium, Ir( CO),Cl,, has been knownfor some time,l0 simple carbonyls of iridium have not been described untilrecently.The tricarbonyl, [Ir(C0)3]n, has been prepared by the action ofcarbon monoxide at 200 atm. pressure on a trihalide of iridium.11 Thehalogen is again eliminated by combination with copper from the autoclavelining :IrCl, + ~ C U + 6CO --+ Ir(CO), + 3CuC1,COPolymerThe chloride, bromide, and iodide react in this way at 140°, 110-120°,and 90-lOO', respectively ; conversion of the trichloride into the carbonylis complete after 24--48 hours at 140'.The product can be sublimed incarbon monoxide a t 200-210". In this reaction, actual addition of a halo-gen-absorbing metal, such as copper or silver, is avoided, since immediatereduction of the iridium halide to metallic iridium would occur on heatingto the requisite temperature, and the iridium would not react with carbonmonoxide under the conditions given. If complex iridium halides, such asK21rBr6, are used instead of the trihalides, addition of copper or silver is,however, necessary if the halogen is to be displaced completely; K,IrBr6and carbon monoxide, without added metal, give a new carbonyl halide,Ir(C0)3Br. Although the tricarbonyl is the main product of the reactionof iridium trihalides and carbon monoxide, the complex halides (withcopper or silver) give mixtures of the yellow tricarbonyl and a greenish-yellow crystalline tetracarbonyl, [Ir( CO),], ; the mixtures can be separatedby extraction with carbon tetrachloride or ethyl ether, in which the tetra-carbonyl is slightly soluble, or by fractional sublimation, the tetracarbonylsubliming at about 160' in carbon monoxide a t normal pressure.Iridiumtetracarbonyl is very readily converted into the tricarbonyl, which is suffi-ciently stable to resist attack by concentrated acids, dilute alkalis, or freehalogens at room temperature. The molecular weights of the iridiumcarbonyls could not be determined owing to their low solubility; theproperties of the tetracarbonyl are consistent with the dimeric formula,[Ir(C0),I2 (indicated by analogy with the cobalt compounds), whereas thetricarbonyl is clearly more complex ([Ir(CO),],?) since it is less volatile andless soluble.The formation of cobalt tetracarbonyl hydride from cobalt compounds,carbon monoxide, and substances containing hydrogen has been notedabove. Apparently an iridium carbonyl hydride, probably HIr( CO),, isformed under similar conditions, for the action of carbon monoxide oniridium trihalides in presence of water or hydrogen affords a volatile andvery unstable substance containing iridium ; this gives a colourless solid* W.IZieber, F. Muhlbauer, and E. A. Ehmann, Ber., 1932, 65, 1090.lo W. Manchot and H. Gall, &id., 1925, 58, 232WELOH : METALLIC CARBONYLS AND NITROSYLS.75compound with mercuric chloride solution.11 It appears, therefore, that aclose analogy exists between the carbonyl derivatives of cobalt and iridium.The hexacarbonyls of chromium, molybdenum, and tungsten were firstprepared in quantities sufficient for analysis by the action of carbon mon-oxide and a Grignard reagent on the anhydrous halides of the metals.12The yields obtained by this method are small. Molybdenum and tungstenhexacarbonyls are better prepared by treating the reduced metal withcarbon monoxide (at 200 atm. and about 225') in presence of another metal(e.g., copper or iron) ; 13 these carbonyls have also been obtained by treatingthe chlorides or bromides of the metals, or corresponding complex salts(e.g., K3MoC16, K3W,C1,), with carbon monoxide under pressure, in presenceof a metal to remove the halogen.14 Chromium hexacarbonyl apparentlycannot be prepared by either of these methods, and only the original methodof Job12 is available in this case.The hexacarbonyls of the Group V Imetals, which are colourless crystalline solids, are considerably more stablethan the simpler carbonyls of the Group VIII elements, probably becausethe attainment of the effective atomic number of a rare gas by the centralmetal atom coincides with the formation of the inherently stable six-co-ordinate complex.Brief reference has already been made to the tetracarbonyl hydrides ofiron, cobalt, and iridium; it appears likely that a similar compound ofruthenium [presumably H,Ru( CO),, maintaining the analogy betweenruthenium and iron] also exists, since ruthenium pentacarbonyl reacts withalkalis to give fairly stable solutions with strong reducing properties,6 justas iron pentacarbonyl reacts to give strongly reducing solutions of the irontetracarbonyl hydride.Iron and cobalt tetracarbonyl hydrides have been isolated and studiedin some detail; they are very volatile liquids, stable only a t temperatureswell below 0'.The free iron compound was first obtained by Hieber andhis collaborators15 by the action of an aqueous solution of a base (pre-ferably barium hydroxide) on iron pentacarbonyl, followed by acidificationof the solution. A similar reaction with cobalt tetracarbonyl affords freecobalt tetracarbonyl hydride.1691'Several reactions have now been studied in which cobalt tetracarbonylhydride derivatives are obtained by the action of carbon monoxide onsolutions (or suspensions) of cobalt salts.These reactions take place a tl1 W. Hieber and €3. Lagally, 2. anorg. Chem., 1940, 245, 321.l2 A. Job and A. Cassal, Cornpt. rend., 1926,183, 58, 392; Bull. SOC. chim,., 1927, [iv],41, 1041; A. Job and J. Rouvillois, Compt. rend., 1928, 187, 564; W. Hieber andE. Romberg, 2. anorg. Chem., 1935, 221, 321.Is I. G. Farbenindustrie A.-G., F.P. 708,379, 708,260 (1930) ; B.P. 367,481 (1930);D.R.-P. 547,025 (1931).l4 W. Hieber et aE., details not yet published.l6 W. Hieber and F. Leutert, Naturwiss., 1931, 19, 360; Ber., 1931, 84, 2832; 2.18 W. Hieber, Angew.Chem., 1936, 49, 463.l7 W. Hieber and H. Schulten, 2. anorg. Chem., 1937, 232, 17, 29.anorg. Chem,., 1932, 204, 145; W. Hieber and H. Vetter, ibid., 1933, 212, 14576 INORGANIC CHEMISTRY.normal pressures, and provide means of preparing cobalt tetracarbonylhydride and the tetracarbonyl (which is readily obtained by decompositionof the hydride on warming) with readily available apparatus and materials.Unfortunately, the reactions are slow and accumulation of moderate quan-tities of the products is tedious. An alkaline solution of cobalt chloridewill not absorb carbon monoxide, but on addition of cysteine to this solutionthe gas is absorbed, and cobalt tetracarbonyl hydride is liberated on acidi-fication of the resulting liquid.18, l9 This reaction probably occurs by theintermediate formation of complexes of cobalt and cysteine containingcarbon monoxide; such complexes have not yet been isolated, althoughiron compounds of a corresponding type are known.18 A mechanism hasbeen worked out for the above reaction,18 involving “ disproportionation ”of a bivalent cobalt complex to a cobalt tetracarbonyl hydride derivative(presumably the alkali salt) and a tervalent cobalt complex; the latter isreduced by carbon monoxide in a subsequent stage of the reaction. Cer-tain other substances, notably tartrates, can replace cysteine in this re-acti0n.1~ Nickel tetracarbonyl is readily prepared by the action of carbonmonoxide on an alkaline suspension of nickel cyanide or sulphide,20-22 andcarbon monoxide is also absorbed by alkaline suspensions of cobalt cyanideor sulphide; cobalt tetracarbonyl or the hydride has not, however, beenisolated from the reaction products, although cobalt nitrosyl carbonyl,Co(CO),NO (see below), has been obtained by passing nitric oxide into thesolution.22 An interesting addition compound of cobalt tetracarbonyl,[Co(CO),],,EtOH, has been prepared by the action of carbon monoxide onan alcoholic solution of cobaltous chloride and potassium ethyl anth hate.^^In all these processes it appears likely that intermediates containing carbonmonoxide are formed, and a further study of the reactions involved maylead to an interesting new branch of carbonyl chemistry.It is noteworthythat linkings between metal and sulphur atoms are, for some reason notyet understood, particularly susceptible to reaction with carbon monoxide.The tetracarbonyl hydrides of iron and cobalt form an interesting seriesof crystalline derivatives with complex cations containing ammonia oro-phenanthroline (phenan) .24 Typical members of the series arewhich are prepared by addition of a solution containing the appropriatecomplex cation to an ammoniacal solution of the carbonyl hydride.Theammonia compounds are stable in the absence of air, but react readilywith ammonia, pyridine, etc., and slowly with methyl alcohol; the o-phen-anthroline compounds are stable in air and notably less reactive. With18 M. P. Schubert, J . Amer. Chem. SOC., 1933, 55, 4663.1’ G. W. Coleman and A. A. Blanchard, ibid., 1936, 58, 2160; cf.also A. A. Blan-2o W. Manchot and H. Gall, Ber., 1929, 62, 678.s1 M. M. Windsor and A. A. Blanchard, J . Amer. Chem. SOC., 1933, 55, 1877.2* A. A. Blanchard, J. R. Rafter, and W. B. Adams, ibid., 1924, 58, 16.2s W. Hieber, Angew. Chem., 1936, 49, 463.24 W. Hieber and E. Fack, 2. anorg. Chern., 1938,238, 83.“i(NH,),I[Co(CO),I,, [Mn(NH,),l[HFe(C0)412, and Cco(Phenan),l[Co(CO),l,,chard and P. Gilmont, ibid., 1940, 62, 1192WELCH : METALLIC CARBONYLS AND NITROSYLS. 77zinc, cadmium, and copper salts the ammoniacal solution of iron tetra-carbonyl hydride reacts to give compounds of a different type, &.,Zn(NH,),,Fe(CO),, Cd(NH3),,Fe(C0),,25 and Cu,(NH,),,Fe(CO),, in whichboth hydrogen atoms are displaced from the carbonyl hydride.The hex-ammine and o-phenanthroline compounds are regarded by Hieber as truesalts of iron tetracarbonyl hydride, since they are shown by electrical con-ductivity measurements to dissociate in methanol and acetone solutions.The zinc, copper, and cadmium compounds are non-electrolytes, however,and these are considered to be polynuclear complexes which are possiblysimilar in constitution to substituted carbonyls, such as Fe2( CO),( C5H,N),.Conductivity measurements also show that the free iron and cobalt tetra-carbonyl hydrides are dissociated in pyridine solution.The table given on p. 72 shows the large number of carbonyl halidesnow known to exist. The properties and reactions of the iron compoundshave been studied in considerable detail.2s At low temperatures ironpentacarbonyl adds free halogens to give unstable compounds of the typeFe(CO),X,; these decompose at temperatures varying from - 35' to Oo,more stable tetracarbonyl derivatives, Fe( C0),X2, being formed. Thetetracarbonyl iodide is also formed on treating iron tetracarbonyl hydridewith iodine, or by the action of carbon monoxide under high pressure onanhydrous ferrous iodide ; 27 the kinetics of the latter reaction show certainunusual features." Mixed " iron tetracarbonyl halides, Fe(CO),IBr andFe(CO),ICl, have now been prepared 2s by the action of iodine monobromideor monochloride on iron pentacarbonyl ; these compounds decomposereadily into mixtures of the two simple carbonyl halides. As expected,the " mixed " halides possess, in general, properties intermediate betweenthose of the two corresponding simple carbonyl halides.The possible reactions of iron pentacarbonyl with halides of metals intheir higher states of oxidation have recently been examined.28 The fir$possible type of reaction, in which part of the carbon monoxide is displacedfrom the carbonyl and oxidised to carbon dioxide, occurs with mercuricchloride in aqueous solution : z9Fe(CO), + 2HgC1, + H20 --+ Fe(CO),,Hg2C12 + CO, + 2HC1The compound Fe(CO),,Hg,Cl, should probably be regarded a8 a doublesalt of the mercury derivative of iron tetracarbonyl hydride, HgFe( CO),,HgCl, ;the simple mercury derivative, HgFe(CO),, is obtained by using mercuricsulphate instead of the chl0ride.2~ The reactions of iron pentacarbonylwith stannic chloride and antimony pentachloride are of a second type inwhich carbon monoxide is displaced but not oxidised :Fe(CQ), + SbCl,+ Fe(CO),SbCl, + COFe(CO), + SnC14 4 Fe(CO),SnCl, + CO26 Cf.also F. Feigl and P. Knunholz, 2. anorg. Chem., 1933, 215, 242.26 W. Hieber et al., ibid., 1930, 190, 192, 215; 1931, 201, 329.27 W. Hieber and H. Lagally, ibad., 1940, 245, 305.28 W. Hieber and A. Wirsching, ibid., p. 36.2s H. Hock and H. Stuhlmann, Bey., 1928, 61, 2097; 1929, 62, 431, 269078 INORGANIC CHEMISTRY.The new compounds obtained are shown by their reactions to contain ironin the bivalent condition, with tervalent antimony or bivalent tin. Theyare more stable than the iron tetracarbonyl halides to which they appear tobe related; the antimony compound is dissociated in benzene or nitro-benzene solution :Fe(C0)4SbC15 =+ Fe(CO),Cl, +but the tin compound shows a normal molecularas a binuclear complex : c1 C1(CO),Fe/ ‘Sn’‘Clf ‘c1SbC1,weight and is regardedWith other halides (e.g., ferric and cupric chlorides) both the above typesof reaction appear to occur together, both carbon monoxide and carbondioxide being evolved ; additive compounds analogous to those justdescribed have not, however, been isolated.Iron pentacarbonyl shows a strong tendency to displace halogens fromlion-metallic halides, as the following reactions show : 28Fe(CO), + 2CC14 --+ C,CI, + FeCI, + 5COFe(CO)5 + SO,CI, + SO, + FeC1, + 5COIn each case an iron carbonyl halide is probably formed as an intermediateproduct.The corresponding reaction with thionyl chloride is more com-plex, and affords evidence for the existence of an iron dicarbonyl chloride,Fe(CO),CI, ; this compound, a reactive, brown, crystalline solid, is isolatedby using iron tetracarbonyl iodide instead of iron pentacarbonyl :Fe(CO),I, + 2SOC1, + Fe(CO),Cl, + I, + 2CO + SO, + SC1,The tetracarbonyl bromide reacts in a similar manner. It is noteworthyt,hat a ruthenium dicarbonyl chloride, Ru(CO),CI,, analogous to the newiron derivative, has been known for some time.31Although a number of iron carbonyl derivatives containing co-ordinatedsubstituents (amines, etc.) contain from one to four molecules of carbonmonoxide per atom of iron,S2 the dicarbonyl chloride discussed above is thefirst known example of a simple iron carbonyl halide with a CO : Fe ratioof less than 4.Further investigation33 has shown that a similar com-pound, Fe(CO),I,, is formed as a product of the thermal decomposition ofiron tetracarbonyl iodide in an inert atmosphere ; other products, obtainedunder different conditions, are Fe(CO),I and the hitherto unknown iodideof univalent iron, FeI. Fe(CO),I, is a reddish-brown solid; its alcoholicsolution gives dark red Fe(CO),I,,phenan with o-phenanthroline, and greenFe(CO),I,(C,H,N) with pyridine. Fe(CO),I and FeI are extremely un-stable; the latter is it bright red solid which reduces silver nitrate in acid3u A. Mittasch, 2. angew. Chem., 1928, Q1, 827.3 1 11‘.Manchot and J. Konig, Ber., 1924, 57, 2130.33 IfT. Hieber and G. Bader, 2. aaorg. Chem., 1930, 190, 193.33 W. Hieber and H. Lagally, ibid., 1940, 245, 295WELCH : METALLIC CARBONYLS AND NTTROSYLS. 79solution to metallic silver and reacts with water to form ferrous hydroxideand hydrogen. The molecular weights of these new compounds have notbeen determined, and the formuke given are empirical.Brief reference has been made above to a tricarbonyl halide of iridiumof the type Ir(CO),X; the other halides of this type have now been pre-pared and e~amined.~4 When dry carbon monoxide a t atmospheric pres-sure is passed over the monohydrate of the iridium trihalide, IrX,,H20, atabout 150°, a mixture of Ir(CO),X, Ir(CO),X,, and [Ir(CO),], sublimes on tothe cooler parts of the apparatus.The dark brown, crystalline tricarbonylhalides sublime on heating the mixture a t 115', and the dicarbonyl com-pounds, which are colourless or yellow, sublime at 150'. Although thedicarbonyl halides are unstable and decompose on exposure to air, thetricarbonyl compounds are stable. It is interesting that halides of thetype Ir(CO),X are not obtained from anhydrous iridium halides and carbonmonoxide under the conditions described, although small yields of Ir(CO),X,have been prepared from these reactants; 35 the necessity for the presenceof water of hydration provides an interesting parallel to the use of methylalcohol vapour as a catalyst in certain reactions in which carbonyl com-pounds are f0rmed.M It is noteworthy that in the reaction just describediridium tricarbonyl is prepared, although in small yield, without the use ofhigh-pressure technique.Hieber considers that formation of the iridiumcarbonyls, at both normal and high pressures, occurs by successive dis-placement of the halogen atoms in the halide by carbon monoxide.Until recently, attempts to prepare carbonyls of the metals of Group VII(manganese, masurium, and rhenium) were unsuccessful, although theexistence of well-defined carbonyl derivatives of Group VI and Group VIIImetals led to a careful search for related compounds of the interveningGroup VII elements, particularly manganese. The recent preparation ofrhenium pentacarbonyl halides 3' is therefore of special interest. Thesecompounds, which have the formula Re(CO),X, are prepared by heating achloride of rhenium (ReC1, or ReCl,), or potassium rhenibromide or rheni-iodide (K,ReX,) mixed with copper powder, in carbon monoxide a t 200 atm.and 200-230'.Under these conditions the rhenium compounds are con-verted completely into the pentacarbonyl halide after about 30 hours; thepentacarbonyl derivatives are extracted from the products with benzeneor light petroleum. The new compounds are stable, odourless, colourlessor yellow crystalline solids which sublime unchanged in carbon monoxidebut are decomposed on heating in air at about 400'. They are soluble inorganic solvents but insoluble in water. The rhenium pentacarbonylhalides are similar in character and stability to the hexacarbonyls of theGroup VI metals.3J W.Hieber, H. Lagally, and A. Mayr, 2. anorg. Chem., 1941, 246, 138.36 W. Manchot and H. Gall, Ber., 1925, 58, 232.36 E.g., in formation of Pd(CO)Cl,, W. Manchot and J. Konig, ibid., 1926, 59,37 W. Hieber and H. Schulten, 2. anorg. Chem., 1939, 243, 164-88.380 INORGANIC CHEMISTRY.The unusual facility with which carbon monoxide reacts with certainmetallic compounds containing sulphur directly bound to the metal atomhas been noted above. This has led to a study of the chemistry of sub-stituted carbonyls containing various groups attached to the metal atomby one or more atoms of sulphur; such compounds might occur as inter-mediates in the reactions in question. The tetracarbonyls of iron andcobalt react readily with ethyl- and phenyl-thiol to form stable, well-crystal-l i d , red derivatives,s7a> 3s [B’e(C0)3SEt],,39 [Co( CO),SEtJ,, Fe( CO),SPh,and Co( CO),SPh, with elimination of carbon monoxide and hydrogen.These are typical non-polar compounds.Diphenyl disulphide gives thesame phenyl compound with iron tetracarbonyl, without elimination ofhydrogen. Mercaptobenzthiazole in solution in light petroleum reacta withiron tetracarbonyl to give a red crystalline compound, Fe,(CO),,<>CH,,which is also formed in very small quantities from parathioformaldehyde,( CH,S),. a-Thionaphthol readily gives a dimeric compound analogous toFe(CO),SPh, but P-thionaphthol reacts with difficulty to form a derivativewhich is at least trimeric.Attempts to prepare chelate derivatives fromiron tetracarbonyl and thiosemicarbazide, thiocatechol, thioacetamide, orthiosalicylic acid were unsuccessful ; a chelateiron carbonyl halide derivative (I) containing co-CH,- r ordinated sulphur is, however, kn0~n.40 Alkyl I ‘Fe(CO),X, sulphides in light petroleum solution react withCH2-YZ iron tetracarbonyl to form compounds of the typeE t (I.) Fe(CO),SR,; some Fe(CO),SR is also formed, itsamount increasing as the temperature is raised ;iron pentacarbonyl gives similar products. Dimethyl disulphide and thetetracarbonyl afford [Fe(CO),SMe],, which is not conveniently preparedfrom the more volatile methyl sulphide. Amine-substituted carbonyls suchas Fe,(CO),(C5H5N), with phenylthiol give Fe(CO),SPh, but no amine-substituted derivatives of a similar type are formed; the compoundsFe(CO)SPh,phenan and Co( CO),SEt,phenan are formed, however, on treat-ing Fe(CO),SPh and [Co(CO),SEt], with o-phenanthroline.The hexacarbonyls ofthe Group VI metals do not react, but the substituted compoundMo( CO),( C5H5N), gives a dark brown crystalline derivative,Mo(CO),(C5H5N)SPh,showing that the partly substituted carbonyls may be more reactive thantheir parent compound^.^^Little reference has been made above to the very numerous substitutedcarbonyls and carbonyl halides in which part of the carbon monoxide isreplaced by amine, alcohol, or other groups; these are discussed in theNickel tetracarbonyl does not react with thiols.37a W.Hieber and P.Spacu, 2. amrg. Chem., 1937, 233, 353.38 W. Hieber and C. Scharfenberg, Ber., 1940, 78, 1012.39 H. Reihlen, A. Gruhl, and G. von Hessling, AnnaEen, 1929, 472, 268.40 W. Hieber, 2. a w g . Chem., 1931, 201, 329WELCH : METALLIC CARBONYLS AND NITROSYLS. 81reviews cited 2* and in the references given below.41 A few of thesecompounds, e.g., the o-phenanthroline derivative (11) ofnickel tetracarbonyl, can be formulated without difficultyas compounds in which the metal atom has the sameco4 /r N\/’.\ I electron configuration as in the parent carbonyl. MostCoflNikN,,/ of the substituted carbonyls, however, are more complex,and their structures deserve further study.The “ carbonyls ” of the alkali metals have not beenincluded, since there is no evidence that they are truecarbonyls containing co-ordinated carbon monoxide molecules.Closely related to the carbonyls are the volatile nitrosyl curbonyls ofcobalt and iron, Co(CO),NO 42 and Fe(CO)2(N0),,43 which are the onlycompounds of their class so far discovered; both may be prepared by thereaction of nitric oxide with either of the corresponding polynuclear car-bonyls under appropriate conditions.The cobalt compound is convenientlyprepared by a method involving the cobalt cysteine complexes discussedabove; carbon monoxide is passed into the alkaline cobalt salt solutioncontaining cysteine, which is then acidified; on passing in nitric oxide thenitrosyl carbonyl is evolved as a red vap0ur.1~ Cobalt nitrosyl carbonylcan also be obtained, in a similar manner, from an alkaline suspension ofcobalt cyanide or sulphide in which carbon monoxide has been absorbed.22A nitrosyl carbonyl of nickel has not yet been isolated, although a com-pound Ni(NO),CO might be expected to exist; the action of nitric oxideon nickel tetracarbonyl affords compounds of a univalent radical Ni(NO),e.g., Ni(N0)OH.aReference has been made in the literature and in patents 46 to a volatilecobalt nitrosyl, Co(NO),, but the properties of this interesting compoundhave not been described.Non-volatile “ nitrosyla ” of iron, Fe(N0),,46and ruthenium, Ru(NO), or R U ( N O ) ~ , ~ ~ have also been described, but theformer is probably a hyponitrite, Fe(NO),N,O,, and the latter requiresfurther investigation to clarify its constitution.An interesting new group of cobalt nitrosyl halides, Co(NO),X, hasAI I(11.)41 Work by W.Hieber and his collaborators: (iron compounds) Ber., 1928, 61,558, 2421 ; Sitzungsber. Heidelberg. Akad. Wiss,, math.-nat. KE., 1929, Part 3, 3-9; Ber.,1930, 63, 973; 2. anorg. Chern., 1930, 190, 193; Ber., 1930, 63, 1405; 1931, 64, 2340;2. anorg. Chem., 1931, 201, 329; Ber., 1932, 65, 1082; (nickel and cobalt compounds)ibid., p. 1090; (chromium and molybdenum compounds) 2. anorg. Chem., 1936, 221,337; (tungsten compounds) ibid., p. 349.42 R. L. Mond and A. E. Wallis, J., 1922, 121, 32.43 W. Hieber and J. S. Anderson, 2. anorg. Chem., 1932, 208, 238. For Bornereactions of the iron and cobalt nitrosyl carbonyls, see idem, ibid., 1933, 211,132.44 J.8. Anderson, ibid., 1936, 229, 357; cf. also J. C. W. Frazer and W. E. Trout,J. Amer. Chem. SOC., 1936, 58, 2201.46 W. Hieber, Angew. Chem., 1936, 49, 463; I. G. Farbenindustrie A.-G., D.R.-P.613,400, 613,401 (1932).46 W. Manchot and E. Enk, Anncalen, 1929, 470, 276.47 W. Manchot and W. J. Manchot, 2. anorg. ChRm., 1936,226,41082 INORGANIC CHEMISTRY.recently been described.48 These compounds are prepared by the actionof nitric oxide on the anhydrous cobalt halide a t about 60'. The iodide ismost readily prepared and forms dark brown or black, shining crystals. Inthe case of the bromide or chloride the reaction proceeds to completion onlyin presence of a metal, such as zinc or cobalt, which absorbs the halogen.An alcoholic solution of cobaltous iodide also reacts with nitric oxide,affording the dinitrosyl iodide.The cobalt dinitrosyl halides are extremelystable substances which can be sublimed without decomposition in air,carbon monoxide, or hydrogen a t normal pressures; a t high pressurescarbon monoxide reacts to give cobalt tetracarbonyl and cobalt nitrosylcarbonyl. The compounds are regarded as true nitrosyl derivatives ; theyreact readily with alkali thiosulphate solutions, giving deep green solutionsof [Co(NO)2(Sz03)2]"',49 and also with alkali sulphide solutions ; in the lattercase compounds of the type Co(NO),SM (where M is a univalent metal),analogous to Roussin's red salts, are possibly formed. Organic amines givevarious additive and substitution products of the nitrosyl halides ; pyridinegives CO(NO)I,(C~H~N)~, whereas o-phenanthroline does not displace nitricoxide, but forms Co(NO),I,phenan. The cobalt nitrosyl halides are clearlyrelated to the carbonyl mercaptides described above, and also to the knowncompounds Fe(NO),X,5O Fe(NO),SM (Roussin's red salts), Fe(NO),SR,Co (NO),SR, Ni (NO) SR, et c.Little attention has been paid above to the molecular constitution ofthe carbonyls and nitrosyls, since the general principles governing theirstructures have been adequately dealt with in these Reports 1 and in arecent review ; 52 this review also considers a number of nitrosyl cyanideswhich have not been included in this Report.Special importance attaches,however, to the examination by electron-diffraction and X-ray methods ofthe structures of iron penta~arbonyl,~~ the Group VI metal he~acarbonyls,~4nickel tetracarbonyl, cobalt nitrosyl carbonyl, iron nitrosyl ~ a r b o n y l , ~ ~ ironenneacarbonyl [Fez( CO),] ,56 and the cobalt and iron tetracarbonyl hydride~.~'The structures of the carbonyl hydrides are of particular interest, since thehydrogen atoms are not attached directly to the metal atoms but to thecarbonyl group, M-C-O-H (the lines here do not indicate the character of4 8 W.Hieber and R. Marin, 2. anorg. Chem., 1939, 240, 241.49 W. Manchot et al., Ber., 1926, 59, 2445; 2929, 62, 681.60 W. Manchot and H. Fischer, Diss., Tech. Hochschule, Munchen, 1937. For5 1 H. Reihlen etal., Annulen, 1927, 457, 7 1 ; 1928, 465, 72; 1929, 472, 268; 1930,52 A.A. Blanchard, Chem. Reviews, 1940, 26, 409.63 R. V. G. Ewens and M. W. Lister, Trans. Faraday Xoc., 1939, 35, 681 ; ,4?m64 L. 0. Brockway, R. V. G. Ewens, and M. W. Lister, Trans. Faraday SOC., 1938,6 6 L. 0. Brockway and P. C. Cross, J . Chem. Physics, 1935, 3, 828 ; L. 0. BrockwayLe H. M. Powell and R. V. G. Ewens, J., 1939, 286.6 1 R. V. G. Ewens and M. W. Lister, Trans. Faraday SOC., 1939, 35, 681,other iron nitrosyl compounds, cf. ref. (51).482, 161; W. Manchot and F. Davidson, Ber., 1929, 62, 681.Reports, 1939, 36, 166.34, 1350.and J. S. Anderson, Trans. Faraday SOC., 1937, 33, 1233REES : ISOTOPE EXCHANGE IN INORGANIC CHEMISTRY. 83the bonds). The hydrides may therefore be formulated as Co(CO),COH andFe(CO),(COH),; the valency relation of the COH groups to the metalatoms appears to be analogous to that of the nitrosyl groups in the nitrosylcarbonyls.A. J. E. W.3. ISOTOPE EXCHANGE IN INORGANIC CHEMISTRY.In recent years considerable use has been made of partly enriched non-radioactive isotope mixtures and also of radioactive isotopes in the elucid-ation of many of the problems of inorganic chemistry. Much of the workhas been concerned primarily with problems of a more or less physicalnature, but in many cases the bearing on inorganic chemistry has beenconsiderable, and t,his is here reviewed.Isotopic Concentration by Methods of Chemical E’xchnge.Of the various methods available for the concentration of isotopes intheir mixtures, those dependent on gaseous diffusion, fractional distillation,and fractional electrolysis have been extensively employed and are alreadywidely known.The more recent application of thermal diffusion in gasesand liquids to the separation of isotopic mixtures was reviewed in lastyear’s Report.1 However, apart from sections devoted to it in generalreviews on methods of isotopic separation,, no account of the chemicalexchange methods of separation has yet appeared.That chemical exchange reactions could be employed in the practicalseparation of isotopes was first suggested by H. C. Urey and L. J. Grieff.3The existence of equilibria in exchange reactions and the experimentalverification that the constants of these equilibria can be evaluated bystatistical methods, permit the evaluation from spectroscopic data of equi-librium constants and enrichment factors for possible exchange reactions.In the case of exchanges involving only one isotopic pair the enrichmentfactor is defined as (X,/N2)/(nl/n2), where N , and N , are the numbers oflight and of heavy atoms in the one compound, and n, and 72, the corre-sponding numbers for the other.Provided that the assumption that theatoms distribute themselves statistically among the various molecules ismade, and that the fundamental molecular constants for the two isotopicspecies are available, then the calculation of this enrichment factor can bemade with reasonable accuracy. On the basis of their calculations, it wassuggested by Urey and Grieff that isotopes of lighter elements might beseparated by chemical exchange methods and that the equilibria might beestablished in liquid-gas two-phase systems by a counter-current operation.Theoretical calculations of the overall enrichment factors to be expectedin scrubbing columns of a theoretical design were made.These calculationsP. 153.‘2 N. S . Bayliss and R. W. Pickering, J . Proc. Austral. Cheni. Irist., 1950, 7 , 51;H. C. Urey, Rep. Prog. Physics, 1939, 8, 48.J . Amer. Chem. SOC., 1935, 57, 32184 INORGANIC CHEMISTRY.indicated that the process was favourable for enrichment of certain of therare isotopes.The practical application of this method involves the use of three dis-tillation units in cascade 4* 5 as had been suggested earlier by H.c. Urey,J. R. Huffman, H. G. Thode, and M. Fox.6 If each stage of the cascadeincreases the concentration by a factor p, then the first unit should have aflow p times that of the second, and the second a flow p times that of thethird, the isotopic ratio being increased theoretically by a factor p3, whilstthe total transport should be the same in each unit. The original distillationapparatus5 consisted of three units with an effective column length of115 feet, in sections some 15 feet long. Only the second and the third unitof this apparatus being used for the exchange reaction between gaseousammonia and ammonium nitrate solution, wix.,15NH, (g.) + l411y'H4+ (sol.) + 14NH, (g.) + WH4+ (sol.)( K = 1.023)a 46-fold increase in the isotopic ratio 15N/14N was obtained in the solutionin 2 weeks' operation, the sample containing l4.8Y0 of W .4 With all threeunits in operation, 8.8 g . of 70.6% 15N were obtained in 4 days.5The sulphur isotope 34s has been obtained in 6-Sy0 concentration in theliquid in 7 days, only one unit of the apparatus being used, for the exchangebetween gaseous sulphur dioxide and sodium bisulphate solution.*Exchange has also been observed in the system H,S (g.)-NaHS (sol.), the84S being concentrated in the gas phase.' D. W. Stewart and K. Cohen *have found satisfactory exchange to take place between gaseous sulphurdioxide and an aqueous solution of sodium bisulphite according to theequationThe maximum concentration of 34s was 27% after 21 days at a rate of 0.8 g.of per day.Successful concentration of the heavy carbon isotope 13C has also beenaccomplished by exchange between gaseous hydrogen cyanide and sodiumcyanide solution by a similar m e t h ~ d .~ ~ 10 The possible reactions involvedareH12CN + 13CN' =+ H1WN + 12CN' ( K = 1.026)and HC14N + C15N' HC15N + CI4N' ( K = 1.003)The equilibrium constants quoted have been derived from the knownvibrational fiequenciee of HCN and CN'. 15N is concentrated to a small3450, (g.) + ~ 3 2 ~ 0 ; (sol.) + 3 2 ~ 0 , (g.) + H~~SO; (sol.)4 H. G. Thode, J. E. Gorham, and H. C. Urey, J . Chern. Phyaks, 1938, 6, 296.6 H. G. mode and H. C. Urey, ibid., 1939, 7, 34.* Ibid., 1937, 6, 856.7 H. C. Urey, A. Mills, I. Roberts, H. G. Thode, and J. R. Huffman, ibid., 1939,7, 138.Ibid., 1940, 8, 905.I.Roberts, H. G. Thode, and H. C. Umy, ibid., 1939, 7, 137.10 C. A. Hutchison, D. W. Stewart, and H. C. Urey, ibid., 1940, 8, 532REES : ISOTOPE EXOHANGE M INORGANIC CHEMISTRY. 85extent in the liquid phase, whereas 13C is concentrated to a greater extentin the gas. The average production of 13C was 0.150 g. (in sodium cyanidecontaining 23% of 13C) per day.Base exchange reactions of zeolites with the chlorides of potassium,lithium, and ammonium are the basis of a method of isotopic concentrationof 41K, 7Li, and 15N.11 The concentration of these isotopes is made possibleby the difference in binding of the two isotopic ions with water and withzeolite in each case. Changes of 25% in the abundance ratio 'Li/gLi, andof 10% in 39K/41K and 14NW/14N15N were observed.Exchange Reactions involving Non-radioactive iTsobpes.-Exchange re-actions occurring with non-radioactive isotopes have been investigatedexhaustively in the cases of deuterium and to a less extent of the heavyoxygen isotope lSO and the heavy nitrogen isotope 15N.A very com-prehensive account of the exchange reactions involving deuterium has beengiven by H. C. Urey and G. K. Teal,l2 and they will not be further discussedhere.Interchange of 180 from heavy-oxygen water withsulphate ions was reported by S. C. Datta, J. N. E. Day, and C. K. Ingold 1sto be very slow in neutral solution at I O O O , whereas in alkaline solutionrapid interchange was found a t the same temperature. These workerspostulated a mechanism for the exchange, assuming the active agent to bethe hydroxide ion ISOH'.However, subsequent investigations by E. R. S.Winter, (Miss) M. Carlton, and H. V. A. Briscoe14 have shown that heavy-oxygen water, having an excess density due to l80 of 150-200 yd, doesnot interchange with sulphate in neutral, acid, or alkaline solution at 100'.It was suggested that the interchange observed by Ingold et al. was actuallyan interchange with silicate ions produced by alkali attack upon the glasscontainer. T. Titani and K. Goto 15 reported partial exchange of potassiumbisulphate with H,lsO and this has been confirmed by G. A. Mills,16 whosuggests that the mechanism responsible for the exchange is one of reversibleanhydride formation, H2S04 H,O + SO,, a reaction favoured by acidicconditions.A further investigation l7 confirms that no exchange occursbetween H21s0 and sulphate in neutral or alkaline solution, but that hydrogenchloride catalyses the exchange. The latter result is in good agreement withTitani and Goto, but not with Briscoe et at?., who used sulphuric acid.Furthermore, J. L. Hyde 18 has recently published results which indicatethat the H21s0 exchange with sulphate ions is catalysed by H but not byOH'. No exchange was observed in neutral solution. It appears that themost satisfactory mechanism is one of reversible anhydride formation inacid solutions, as was suggested by Mills.Oxygen exchange.l1 T. I. Taylor and H. C. Urey, J. Chem. Physics, 1938, 6, 429.l2 Rev.Mod. Physics, 1935, 7, 34.l6 J . Amer. Chem. SOC., 1940, 62, 2833.l7 N. F. Hall and 0. R. Alexander, ibid., p. 3455.l8 Ibid., 1941, 63, 873.l3 J., 1937, 1968.l5 Bull. Chem. SOC. Japan, 1939, 14, 77. J., 1940, 13186 INORCt ANTC CHEMISTRY.A similar situation obtains in the case of phosphate, where a basecatalysis has been reported by E. Blumenthal and J. B. M. Herbert l9 andT. Titani, N. Morita, and K. Goto,,O but could not be repeated when stepswere taken to eliminate the possibility of alkali attack on the glasscontainers .I4Chlorate and nitrate have been found by Briscoe et aZ.14 to interchangecompletely in acid solution, but not in neutral or alkaline solution, althoughother workers17 find that addition of acid has little influence on the non-exchange in nitrate.Rapid and complete exchange in neutral solution has been observedin the-case of SiO,”, BO,’, BO,”’, B 0 ”, Cr207”, CrO,”, MOO:’, Woe”,Cost’, MnO,‘, IO;, SeO;’, SO,”, S 2 4 ” , 7 A ~ 0 ~ ” , and AsOi, but not withNO,’, ClO:, ClO,‘, and SeO,”.141161 l7 It is- possible that the exchangemay take place in one of three ways : (i) Direct interchange of oxygen atoms.(ii) Addition, and subsequent removal, of water or hydroxyl ion to theanion with possible exchange.(iii) Formation of undissociated acid byhydrolysis, followed by reversible anhydride formation, as has alreadybeen formulated in the case of sulphate exchange. The experimental dataappear to be adequately explained by the last of these possibilities, for inall cases the rate of 1*0 exchange is related to the acid strength of thecorresponding acid.16Isotope exchange between gaseous oxygen and water vapour on catalyticoxide surfaces has been reported by N.Morita.,l For aluminium oxidesurfaces, it was found that the rate of exchange is almost independent ofthe composition of the gas mixture, indicating that the determining factoris the activated adsorption of water on the oxide surface. The activationenergy of this exchange is calculated to be 18-20 kg.-cals.The exchange of 15N between nitrogen di- andmon-oxide in the gas phase has given a very interesting confirmation ofthe structure of the dinitrogen trioxide molecule, which is known to occurin these mixtures.22Nitrogen exchange.The exchange reaction, wix.,1 4 ~ 0 + 1 5 ~ 0 , 1 5 ~ 0 + 1 4 ~ 0 ,was found to be rapid and only the lower limit of the exchange rate couldbe established.Assuming the intermediate formation of N,O,, the exist-ence of which has been demonstrated by measurements of the equilibriumconstant and by absorption-spectra data, the mere fact of exchange requiresan oxygen bridge in the N203 molecule, such as O=N-0-N=O. Thisstructure is to be expected on account of the symmetry of the resonancestate.Several recent investigations have been directed a t the possibility ofexchange between gaseous and combined nitrogen, in an endeavour toelucidate the mechanism of nitrogen fixation by plants. Exchange wasreported to occur between gaseous nitrogen containing radioactive 13N andeo Bull.Chem. SOC. Japan, 1938, 13, 329.22 E, Leifer, J . Chem. Physics, 1940, 8, 301.lB Trans. Faraday SOC., 1937, 33, 849.21 Ibid., 1940, 15, 47, 298REES : ISOTOPE EXCHANGE IN INORGANIC CHEMISTRY. 87solutions of sodium nitrite, sodium nitrate, hydroxylamine hydrochloride,and sodium hexanitrocobaltiate(II1) by Y. Nishina, T. Iimori, H. Kuto,and H. N a k a ~ a m a , ~ ~ but other workers, using heavy nitrogen gas (15N) 2 4 ~ ~ 5and radioactive nitrogen gas (13N),26 failed to observe any exchange withcombined nitrogen in solution. It appears, therefore, that the fixation ofnitrogen in plants does not occur by direct exchange of gaseous and com-bined nitrogen.The adsorption and nitrogen isotope exchange on iron, tungsten, andosmium catalytic surfaces have been investigated recently by H.S. Taylorand his co-w~rkers.~~~ 28 The exchange may be represented thus :28Nz + 30N2 =p Z28N30N and appears to be of the second order. Smallquantities of hydrogen were found to inhibit the exchange on osmium,and large amounts to suppress it, whereas on iron surfaces the presence ofhydrogen accelerated the exchange. This exchange process apparentlyrequires migration of atoms to positions favourable to exchange; in thecase of osmium, the adsorbed hydrogen (where the adsorption is some 15times greater than that of nitrogen) impedes this migration. The factthat the activation energy of exchange is 50 kg.-cals. for iron and 22 kg.-cals.for osmium surfaces indicates a greater lability of nitrogen on the osmiumsurface and a weaker Os-N than Fe-N bond.This is in agreement withthe stability of the known iron nitrides and the non-existence of analogousosmium compounds.Exchange Reactions involving Radioactive Isotopes.-Exchange reactionsbetween radioactive isotopes and inorganic compounds have been recentlydiscussed in an excellent and comprehensive review on artificial radio-activity by G. T. S e a b ~ r g . ~ ~ This section will be entirely supplementaryto Seaborg’s article, indicating only those radioactive exchanges of inorganicinterest which have been investigated in the last two years.L. C. Liberatore and E. 0. Wiig 30 have made a study of the exchangeof radioactive bromine (produced by the bombardment of selenium withprotons) with hydrogen bromide in the gas phase.Calculations indicatethat the postulate of an intermediate cluster of molecules would lead toan exchange rate some 10l2 times too small, whereas a chain mechanismleads to a rate of the same order as that determined experimentally. Freebromine atoms are produced by the reactions (where Br* indicates a radio-active bromine atom)H + BrBr* ---+ HBr $- Br*H + HBr* + H, + Br*a3 J . Chern. Physics, 1941, 9, 571.24 R. H. Burris and C . E. Miller, Science, 1941, 93, 114.25 G. 0. Joris, J . Chem. Physics, 1941, 9, 775.2* T. H. Norris, S. Ruben, and M. D. Kamen, ibid., p. 726.2 8 W. R. F. Guyer, G. G. Joris, and H. S. Taylor, ibid., 1941, 9, 287.28 Chem. Reviews, 1940, 27, 199.30 J . Chem. Physics, 1940, 8, 165.G.G. Joris and H. S . Taylor, ibid., 1939, 7 , 89388 INORGANIC CHEMISTRY.after which equilibrium is attained by the partition of the Br* betweenmolecular bromine and hydrogen bromide :Br* + Br, .+ BrBr* + BrBr* + HBr + HBr* -t BrW. F. Libby,31 however, on the basis of exchange carried out a t reducedradioactive concentrations, suggests that the alternative bimolecularmechanism of intermediate formation of HBr, is correct. Liberatore andWiig in a further publication32 report no exchange of radioactive brominewith gaseous ethyl bromide at room temperature, but on heating to 200-300" rapid exchange was found to set in. The fact that the C1-Br bond isweaker than the H-Br bond would require exchange of bromine with ethylbromide also to occur at room temperature if Libby's mechanism werecorrect. The atomic exchange mechanism requires activation energies of25 and <5 kg.-cab.for ethyl bromide and hydrogen bromide, respectively,the latter exchange therefore being the only one which can proceed at roomtemperature.Phosphoric acid containing radioactive 32P (half-life 14.3 days), in solu-tion with disodium hydrogen phosphate, has been used as an end-pointindicator in the volumetric estimation of cations which form insolublephosphates, e.g., Mg", Ag', Ba**, Pb.', Tho***, and U0,".33It has been shown that no exchange of radioactive 32P occurs betweenortho-, pyro-, and meta-phosphoric acids in acid or alkaline s0lutions.3~Furthermore, by using S2P as an indicator, the completeness of separationof meta- from other phosphoric acids by precipitation with barium ionshas been demonstrated, whereas in the precipitation of pyrophosphatewith Cd" a small amount of co-precipitation of ortho-phosphate occurs.Elementary radioactive sulphur (35S) has been shown 36 to exchangewith sulphur monochloride according to the reactionthe rate of exchange being directly proportional to the concentration of S,molecules.The most satisfactory mechanism involves a slow stages, + 86 + s,, followed by a rapid stage s, + s2c1, + 2S2c1,.Applications to the Chemistry of Complex Compounds.-Several recentinvestigations have demonstrated the usefulness of isotopic exchangereactions in solving problems associated with the chemistry of co-ordin-ation complexes.Up to the present, only radioactive exchange has beenemployed in this field, but with obvious success. A. A. Grunberg andP. M. Filinov36 have used radioactive bromine to demonstrate the freeexchange between the complex anions, tetrabromoplatinate(II), [PtBr,]",and hexabromoplatinate(IV), [PtBr6]", (containing radiobromine) andThis is in accordance with the experimental results.ClSSCl + 35ss, =+ c1s35sc1 + s,31 J . Chem. Physics, 1940, 8, 348.33 A. Langer, J . Physical Chem., 1941, 45, 639.34 D. E. Hull, J . Amer. Chem. SOC., 1941, 83, 1269.36 R. A. Cooley and D. M. Yost, d i d . , 1940, 82, 2474.30 Compt. rend. Acad. Sci. U.R.S.S., 1939, 23, 912.32 Ibid., p. 349REES : ISOTOPE EXOHANGE IN INORUANIC CHEMISTRY.89bromine ions in solution. This work has been extended by A. A. Griin-berg?' Codinnation of the existence of an equilibrium between Br' andthe above platinum complexes was obtained, as the exchange can take placeonly if the complex dissociates with subsequent recombination. The func-tional equivalence of the bromine atoms in the complex anions was demon-strated by the equilibriumK,[PtBr,*] L+ II,[PtBr4*] + Br,*Interchange between tho complexes themselves has also been establishedin the twa following observed exchanges :(i) KdPtBr41 + cis-[Pt(NH,),Br,*](ii) K2[PtBr4] + K,[PtBr,*] + K,[PtBr,*] + K,[PtBr,*Br,]K2[PtBr,*Br2] + [Pt(NH3),Br,]but, on using the radioactive isotopes of platinum and iridium, no exchangewas found to occur between stable complexes such as K,[PtCI,] and[Pt(NH,),Cl,] ; (NH4)2[IrC1,] and [Ir py2 cl,] ; (K€14)3[IrC1,] andH[Ir py2 Cl,,].Cobalt(II1) in certain complex compounds has been found not to ex-change with radioactive Co" ions in solution38 according to the expectedreaction Co" Co"'.The only stable forms of cobalt(II1) occur in cobaltcomplexes, and in aqueous solutions of these a small concentration of Co"'in equilibrium with the complexes is to be expected. However, exchangecould only be observed if (i) the exchange Go" + Co"' were operative,and (ii) the exchange were more rapid than the reduction of Co"' by water.As exchange due to electron transfer is very probable, then the failure toobserve it may be due to decomposition by reduction before measurableexchange can occur.F. A. Long39 has studied the exchange of oxalate ions with the tri-oxalato-compounds of the tervalent metals cobalt, iron, aluminium, andchromium (containing the radioactive 11C isotope). Exchange was foundto occur with potassium trioxalato-ferrate(II1) and -aluminate(III) ,K,[M(C,O,),], but not with the analogous chromium and cobalt compounds.Theoretical considerations indicate that the bonds in various co-ordinationcomplexes can approach either covalent or ionic types fairly closely. Wherestrong s-p-d hybridisation occurs covalent bonding is expected, as is indi-cated in the case of the cobalt and chromium complexes by the values oftheir magnetic su~ceptibilities.~O With aluminium compounds, however,s-p-d hybridisation is improbable, and in the case of the iron complexes,analogous complexes have been shown to possess bonds of the ionic type.Furthermore, the most recent work 40 indicates that Fe(II1) and Al(II1)complexes are not capable of resolution into optical isomers, and this shouldbe possible if the oxalate-central atom bonds were of the ionic rather thanthe covalent type. The resolution of Co(II1) and Cr(II1) complexes is awell-established fact, indicating covalent bond types. This is in complete37 Bull. Acad. Sci. U.R.S.S., SQr. Phys., 1940, 4, 342.38 J. F. Flagg, J. Amer. Chem. SOC., 1941, 63, 557.3* Ibid., 1939, 61, 570; 1941, 63, 1353.40 C. H. Johnson, Trans. Paraday SOC., 1932, 28, 84690 INORGANIC CTIEMTSTRY.agreement with the exchange measurements, where exchange is expectedonly in those complexes involving ionic bond types, such as those of Fe(II1)and AI(III), but not those of Cr(II1) and Co(II1).Exchange has been found to occur between mercuric iodide and am-monium iodide (containing radioactive iodine) when the complex (NH,),HgI,is formed and subsequently decomposed. Similar results have been obtainedin the reactions of ammonium iodide with bismuth iodide and lead iodidethrough the intermediate formation of the complex anions [BiI,]‘ anti[PbI,]”, suggesting that no essential difference exists between normal andco-ordinate covalencies in these complexes .41Although the isotope effect in band spectra of molecules has been wellknown for many years, only one case is on record where the visual colouror light absorption of a compound has been affected to a large extent byisotopic substitution. S. H. Maron and V. K. LaMer 42 have reported achange in the colour of the CH,*CH:NO,’ ion on substitution of the a-hydrogenatom by a deuterium atom. The addition of barium deuteroxide, Ba(OD),,to nitroethane in heavy water yields the colourless ion CH,*CH:NO,’, which,on addition of D,SO,, yields the compound CH,*CHD*NO,. Addition ofBa(OD), to this produces a light yellow colour due to the formation of theion CH,*CD:NO,’, according to the reaction2CH,*CHD*NO, + Ba” + 20D’ ---+ BCH,*CD:NO,’ + Ba” + 2HODAbsorption-spectra measurements of solutions containing this ion show anabsorption starting in the region hh 5000-5200 A. and continuing into theultra-violet. It is possible that similar differences in colour may be foundlater in inorganic isotopic pairs. A. L. G. R.H. J. EMEL~US.A. L. G. REES.A. J. E. WELCH.*I S. Chatterjee and P. Ray, J . Indian Chem. SOC., 1940, 17, 524.4 p J . Chem. Physics, 1938, 6, 299