Inorganic chemistry

 

作者: A. J. E. Welch,  

 

期刊: Annual Reports on the Progress of Chemistry  (RSC Available online 1945)
卷期: Volume 42, issue 1  

页码: 63-91

 

ISSN:0365-6217

 

年代: 1945

 

DOI:10.1039/AR9454200063

 

出版商: RSC

 

数据来源: RSC

 

摘要:

INORGANIC CHEMISTRY.I. INTRODUCTION AND GENERAL.THIS year's Report on Inorganic Chemistry provides a survey of five specificfields which appear to be dominant a t the present time, and in which rapiddevelopments appear likely in the near future. Exhaustive discussion ofthe large number of available recent papers has been avoided, attentionbeing confined largely to papers of fundamental interest, and to work show-ing notable advances in t'echnique. A few important papers which cannotappropriately be included elsewhere are discussed below.The supposed effects of drying on the physical properties of liquids havefrequently been the subject of discussion, and the appearance of yet anotherpaper proving the invalidity of earlier observations is of considerableinterest. It has been shown that dinitrogen tetroxide, N204, reacts withphosphoric oxide a t room temperature, oxygen and the compound P205,2N0being slowly formed; the liberation of oxygen accounts for the supposedincrease of vapour pressure stated by previous workers to result fromintensive drying of the tetroxide.A small fraction of the liquid tetroxide,distilled from the main bulk after reaction with phosphoric oxide, gavenormal vapour pressures when examined under " dry " conditions. Di-nitrogen trioxide also reacts with phosphoric oxide with formation ofP20,,2N0, and the consequent enrichment of the vapour in N204 =+ 2N0,accounts satisfactorily for the rise of vapour density previously ascribed toa drying process. These observations, together with those of other workerswho have examined other physical properties, lead to the general conclusionthat physical properties of a material are not influenced by intensive drying.A useful contribution to the chemistry of solutions in non-aqueoussolvents is provided by a long paper by G.Jander and G. S ~ h o l z , ~ dealingwith solutions in anhydrous hydrogen cyanide. The feeble dissociation ofliquid hydrogen cyanide itself may be represented by 2HCN (H,CN)+ +(CN)-, so that " acid analogues " in the hydrogen cyanide solvo-system, asin the water system, are substances affording solvated hydrogen ions." Base analogues " in hydrogen cyanide, however, are compounds givingcyanide ions on dissociation. The degree of dissociation of typical acidand base analogues (the latter largely comprising substituted ammoniumcyanides formed by dissolving amines in hydrogen cyanide) have beendetermined by cryoscopic measurements, and the production of salts bytheir " neutralisation " reactions studied.Solvolysis reactions of salts andacyl halides are also described. " Amphoteric " behaviour is typified by atitration with liquid hydrogen cyanide solutions of ferric chloride and tri-ethylammonium cyanide ; the first additions of base cause precipitation ofsolvated ferric cyanide, but excess of the base dissolves the precipitateE. M. Stoddart, J., 1945, 448.2. physikal. Ghem., 1943, 192, 163.a Idem, J . , 1938, 145964 INORGANIC CHEMISTRY.slowly, triethylammonium ferricyanide being formed in solution.Sulphurtrioxide behaves as a potential acid analogue in hydrogen cyanide, with whichi t appears to combine to form a tribasic acid, S(OH),(CN),. Sulphuric acidis thought to give S(OH),(CN),.Continued interest in the simpler derivatives of phosphine and arsine isevident. A simple preparative method for the monosodium derivatives,NaPH, and NaAsH,, has been described-; an ethereal solution containingphosphine or arsine in substantial excess is treated with a solution of sodiumtriphenylmethyl, the required sodium derivative being obtained as a whiteprecipitate. If the hydride is led into a solution of sodium triphenylmethyl,reaction proceeds beyond formation of the monosodium compound, and theproduct is heavily contaminated with Na,PH and Na,P (or correspondingarsenic compounds).The pure monosodium compounds are spontaneouslyinflammable in air, the arsine derivative being notably less stable than itsanalogue. Both compounds afford easy access to alkyl-phosphines and-arsines by treatment with corresponding alkyl halides. Evidence wasobtained for the existence of a magnesium compound, MgBrOPH,, formedby the action of phosphine on ethylmagnesium bromide, but the productwas not isolated. Excess of arsine, bubbled through a solution of lithiumin liquid ammonia a t - 70°, affords LiAsH2,4NH, as a yellowish-whitecrystalline solid ; on warming, this decomposes successively intoLiAsH,,2NH3 and Li,AsH. The thermal decomposition reactions of thecompounds KPH,, NaPH,, and LiPH,,4NH3 have also been studied.6A recent paper on the preparation and properties of chromous iodide 7is noteworthy on account of the elegance of the techniques described forthe manipulation of this substance, which is highly reactive towards airand water.The iodide is prepared by the action of iodine vapour onmetallic chromium a t 700-850" ; it forms a brownish-red crystalline sublim-ate. The colour was suspected to be due to adsorbed or dissolved iodine(the chloride and bromide both being colourless), but since it remainedunchanged in the presence of mercury vapour a t 500" it was finally deducedto be that of chromous iodide itself. The manipulative techniques adopted,including that used for the pyknometric determination of the density ofthe solid by immersion in xylene, are described in detail in the originalpaper.Reactions involved in the oxidation of cobaltous and manganoushydroxides by air have recently been described.Cobaltous hydroxide,precipitated by sodium hydroxide from cobaltous chloride solution, inpresence of the mothor-liquor containing excess of alkali, is not oxidised toany considerable extent by a current of air at room temperature. If thesolid product is washed, air-dried, and heated in air to expel combined water,oxidation proceeds much farther during the heating, and the oxygen contentH. Albers and W. Schuler, Ber., 1943, 76, 23.C. Legoux, Bull. SOC. chim., 1940, 7, 549.P. Fireman, J . Amer. Chern. Soc., 1945, 67, 1447.Idem, ibid., p. 545. ' F. Hein and G.Biihr, 8. anorg. Chem., 1943, 251, 241WELCH : INTRODUCTION AND GENERAL 65may a t one stage exceed that corresponding with the oxide c0304 ; expulsionof the whole of the water, however, is accompanied by liberation of thisexcess of oxygen, the final product being anhydrous Co304. The oxidationof cobaltous hydroxide in the boiling mother-liquor proceeds much fartherthan a t lower temperatures, particularly in presence of a substantial excessof alkali, and the air-dried oxidised material may contain more oxygenthan corresponds with Co,O,; this oxide is again obtained after expulsionof water. Oxidation of a suspension of freshly precipitated manganoushydroxide a t the boiling point affords mangano-manganic oxide, Mn304,directly; a t room temperature a product is obtained containing rathermore oxygen than this oxide.The stability relationships of ferrous hydroxide have been clarified by adetermination of its heat of combustion to or-ferric oxide.s The valueobtained shows that the hydrogen pressure associated with the equilibrium3Fe(OH), + Fe30, + H, + 2H20 is considerable, even at room temper-ature ; ferrous hydroxide is therefore inherently unstable with respect tohigher iron oxides, even in the absence of oxygen.The freshly preparedhydroxide is white, but a greenish tinge is developed on washing, owing toreduction of water. The dry material is pyrophoric, slow oxidation afford-ing ferroso-ferric oxide; a- and y-ferric oxide are formed by oxidation a thigher temperatures.Some recent work on calcium polysulphides lo merits brief reference.When a suspension of calcium sulphide and sulphur is heated, hydrogensulphide is evolved, and calcium polysulphides (CaSIL, where n is between4 and 5), calcium hydrosulphide, and calcium thiosulphate are formed, bythe following reactions : 2CaS + 2H20 4 Ca(SH), + Ca(OH),; Ca(SH), + (n - 1)s _c, CaS, + H2S ; Ca(OH), + 2H2S 4 Ca(SH), + 2H,O ; CaS, + 3H20 -+ CaS203 + 3H2S.Reaction of polysulphide with oxygen alsoproduces thiosulphate : CaS, + 1-50,---+ CaS,03 + (n - 2)s. Evapor-ation of solutions of polysulphides affords oxysulphides of the typeCaS3,2Ca(OH),.Some reviews of particular interest which are now available cover thefollowing topics : the atomic-weight determinations by 0. Honigschmidand his collaborators during the past thirty years; 11 the chemistry of thealkali metals; l2 the chemistry of indium; l3 the inter-relationships of thechemistry of carbon and silicon; l4 reactions in silicate systems containingwater ; l6 the effect of purity on the properties of metals ; J.6 and outstandingrecent developments in inorganic chemistry genera1ly.l'R. Fricke and S.Rihl, Naturwiss., 1943, 31, 326; 2. anorg. Chem., 1943,251,414.lo A. A. Sanfourche, Bull. SOC. chim., 1943, 10, 472.l1 0. Honigschmid, Angew. Chem., 1940, 53, 177.l2 W. Klemm, Chem.-Ztg., 1940, 64, 253.l3 F. Ensslin, Chemie, 1942, 55, 347.l4 R. Schwarz, ibid., 1943, 56, 268.lS W. Noll, ibid., 1944, 57, 90.l6 F. Weibke, Angew. Chem., 1940, 53, 313.l 7 W. Klemm, Chenaie, 1943, 56, 1.REP.-VOL.XLII. 66 INORGANIC CHEMISTRY.2. VALENCY AND THE CONSTITUTION OF SOME INORGANIC MOLECULES.Recent contributions to the theory of valency in which inorganic com-pounds are implicated include an important paper l8 on the occurrence ofthe so-called co-ordinate link. Early valency theory had supported thefollowing structural formuh for nitric acid, phosphorus oxychloride,sulphuryl chloride, and perchloric acid :c1 0 IIII cl\ / O H-O-Cl=O0 I\ O I c l / s ~ o 0H-O-N/< CI--proc1I n later developments of the theory of the chemical bond, co-ordinate linkswere substituted for the double bonds in these formulse, and the followingstructures became widely accepted :? c1c1-P-, 0c1Cl 0c1 0H-O-CI+ 0J.0>s!:0 I1 H-O-N:0Still more recent work has brought to light a number of facts which thesestructures are inadequate to explain.There is good reason for supposingthat the length of a co-ordinate bond should approximate to that of thecorresponding single covalent bond. Evidence has now accumulated thatin structures of the general types shown above, the length of the non-metalt o oxygen bond (briefly termed " oxy-bond ") is considerably less than thatof a single bond of the corresponding type, and is in most cases approximatelyequal to, or even less than, the corresponding double-bond length. Theavailable bond-length data, in fact, suggest strongly that oxy-bonds are notco-ordinate links, but are more accurately represented as double covalentbonds. In their recent paper, Sutton et al.have made a careful analysis ofdipole-moment data (including results of ingeniously planned measurementson a number of compounds not previously studied), and of relevant thermaldata, for compounds containing oxy-bonds with phosphorus, sulphur,selenium, and chlorine; this analysis leads to the same general conclusionas considerations of bond length, and there can be little doubt that co-ordinate links are of much less frequent occurrence than has often beenassumed. Clearly the first, and older, series of structures shown abovemust be accepted in preference to the second series.Some other conclusions reached in the same paper merit brief reference.It appears to be generally true that abnormally strong bonds are abnormallyshort.Co-ordinate links, when they do occur, are relatively weak; thisoccurrence is sometimes favoured by the resonance-energy increase of asystem into which they bring the possibility of more component structures,G . M. Phillips, J. S . Hunter, and L. E. Sutton. J.. 1946. 146WELCH : VALENCY AND CONSTITCTION OF SOME INORGANIC MOLECULES. 67as in the case of ozone, described as a resonance hybrid to which the fourstructures shown below are the principal contributors.O N 0 O 0(Predominant.)+ +Some light is also cast on the chemistry of the halogen oxy-acids, whichapparently exist because the atoms of halogens other than fluorine are ableto utilise d orbitals to form double bonds with oxygen; fluorine is unlikelyto give oxy-acids other than hypofluorous acid, the existence of whichseems possible.Additive compounds of the boron halides are among the simplest examplesof molecules clearly involving donor-acceptor (or co-ordinate) linkings, andtheir study allows the general properties of such linkings to be examined ina relatively simple molecular environment.The compound of dimethylether and boron trifluoride has already been studied in considerable detail,2and the work has now been extended by the recording of some thermal andother physical properties of the 1 : 1 additive compounds of methyl cyanidewith boron trichloride and trifl~oride.~ Both these substances dissociatecompletely into their components in the vapour state, although cryoscopicmolecular-weight determinations on benzene solutions of CH,*CN,BF, reveallittle dissociation.Detailed discussion of the valency relationships in thesecompounds awaits the results of X-ray examination of their structures inthe solid state.The revival of “bridged” structures for diborane and other boronhydrides4 has renewed active interest in this group of compounds, thestructures of which have so far defied entirely satisfactory elucidation. Ina recent contribution on this ~ u b j e c t , ~ a new type of chemical bond ispostulated to occur between boron atoms in the hydrides ; this “ protonateddouble bond ” comprises a double bond of the normal covalent type, withtwo protons embedded in the electron cloud between the bonded atoms;these atoms bear single negative charges, compensatingH - H for the charges on the protons (I).An alternative formul-ation of this bond type in terms of resonance, consistent‘H with the views of Longuet-Higgins and Bell: is possible;moreover, the bond bears many of the essential featuresof the mode of linking proposed by E. Wiberg for the boron-boron bond indiborane, and discarded in favour of earlier resonance structures. The diboranemolecule may be visualised, according to Pitzer’s views, as an ethylenea A. W, Laubengayer and G . R. Finlay, J. Amer. Chem. Xoc., 1943, 65, 884; S. H.Bauer, G. R. Finlay, and A. W. Laubengayer, ibid., p. 889; 1945, 67, 339; Ann.Reports, 1943, 40, 90.\BHfB’(1.)*/ =+A. W. Laubengayer and D. S. Sears, J . Amer. Chem.SOC., 1945, 67, 164.4 See, e.g., H. C . Longuet-Higgins and R. P. Bell, J., 1943, 250; Ann. Reports,6.K. S. Pitzer, J. Amer. Chem. Soc., 1945, 67, 1126.1943, 40, 62.Ber., 1936, 69, 281668 INORGANIC CHEMISTRY.molecule in which two protons have been removed from the carbon atomnuclei and inserted in the spread-out electron cloud of the carbon-carbondouble bond ; the same moleculaz orbitals are said to be appropriate in bothcases, although in diborane the electron cloud is more concentrated roundthe protons. The general properties of the boron-boron linking are satis-factorily accounted for on this basis, and an extension of the theory leads togeneral structural principles which are in accord with the somewhat irregularformulation of the higher boron hydrides.The available physical data forthe hydrides are claimed also to support the principles enunciated. Thereis an evident need for more extended and accurate physical measurementson the boron hydrides (particularly the higher hydrides), and until suchmeasurements provide a secure basis for discussion, speculation on finedistinctions of bond type seems unwise. I n the case of diborane it doesseem clear that an ethane-like model must be discarded in favour of someform of bridged structure, but it is doubtful if information available atpresent justifies detailed discussion of the exact nature of the bridge.Other “ bridged ” structures must clearly exist in the various types ofpolynuclear complexes, in many of which the mode of linking between themetal-atom nuclei is not clearly established. Interesting suggestions regard-ing the constitution of binuclear complexes have recently been made.’Up to the present, halogen bridges such as those in the dimeric moleculesof the aluminium halides have usually been represented in the form (11),although the occurrence of some form of resonance has been tacitly assumed.It is likely that this resonance involves structures of type (111) or (IV); 8(11.) (111.) W.)such structures would give higher interatomic distances for the M-X bridgebonds than normal M-X bonds not incorporated in the bridge, a deductionsupported by the Al-Cl distances in dimeric aluminium chloride.9 Theinteresting feature of structures of type (IV) is that they allow atoms withodd numbers of electrons to attain even-numbered electron configurationswithin the complexes; instead of two metal atoms of valency n, eachcomplex contains one of valency n - 1 and one of valency n + 1, resonancebetween the two possible alternative structures of type (IV) preventingspecific association of either of these valencies with one or other of themetal atoms.Odd-electron atoms bridged by structures of types (11) and(111) would introduce two unpaired electron spins into the complex andrender it paramagnetic; the fact that many complexes in which thesestructures are permissible are actually diamagnetic is accounted for iftype-(IV) structures are assumed.7 K. A. Jensen and R. W. Asmussen, 2. anorg. Chem., 1944,252, 234.8 Cf.ref. (7) and K. A. Jensen, “ Om de koordinativt firegyldige metallers storiokemi,”9 K. J. Palmer and N. Elbott, J. Arner. Chern. Soc., 1938, 60, 1852.Copenhagen, 1937, p. 58WELCH : VALENCY AND CONSTITUTION OF SOME INORGANIC MOLECULES. 69Bridges formed by nitro-, carbonyl, and cyanide groups are thought toinvolve resonance structures (V)-(VII), the second structure of (VI) in-volving a subsidiary resonance shown in (VIII). It is interesting to notethat (VIII) accounts for the observed structure and diamagnetism of theiron carbonyl Fe,(CO), without a direct bonding of the two iron atoms.1°(VII.) (VIII. )Clearly all these proposed bridge structures must still pass the more stringenttests of quantum-mechanical treatment and of physical measurements, butthey offer interesting lines of development in the structural chemistry ofpol ynuclear compounds.The constitution of the compound K,Ni(CN),, prepared by the action* * - -N :C: N : : C : :Ni: : C : : N :.. ....* . * * - - ..- * -: NC+: N : : C : Ni : C : : N :C+N :(XI.).. ..- .. + .. 4-+ . * -....* . -..r- ..N .. . . r i 1- ..C: N : : : C : Ni: C : : : N :CN(X.1a . 4-.... .. ....of excess of potassium on K,Ni(CN), in liquid- ammonia solution,ll presents an interestingproblem. It has recently been pointed out l2that the radical Ni( CN)a- is isoelectronic withnickel carbonyl ; the canonical structures(1X)-(XI) are therefore advanced. The double-bonded structure (IX), in which the nickel atombears no formal charge, is thought to be themost satisfactory, and consequently to con-tribute to the greatest extent to the supposedresonance structure of the ion.10 Cf.H. M. Powell and R. V. G. Ewens, J . , 1939, 286; Ann. Reports, 1939,36,167.l1 J. W. Eastos and W. M. Burgess, J . Amer. Chena. SOC., 1942, 64, 1187.l2 C. L. Deasy, ibid., 1945, 07, 16270 INORGANIC CHEMISTRY.The resonance structures (XII) and (XIII) are proposed for the Ni(CN);-ion in K,Ni(CN),, similarly prepared by the action of potassium, in liquidammonia solution, on K,Ni(CN), in excess. Like structure (IX) for the. - - 2- [: k : : C : : N: .. : : C : : N :I [ : N : : : C : Ni: .. C : : : N :Iz-(XII.)previou..CN :-.. ....II CiN c.... ....(XIII.)compound, structure (XII) is preferred (and presumably ontributesto the larger extent to the resonance structure) because the electropositivenature of the nickel atom should result in relatively low stability for struc-tures in which nickel bears a negative formal charge. The developmentof the structural chemistry of these complexes, and of other related onesthat may yet be isolated, is awaited with interest.3. SOME ASPECTS OF THE CHEMISTRY OF COMPLEX COMPOUNDS.I n the 'study of the chemistry of complex compounds a distinct changeof emphasis is noticeable in recent work; although new types of complexand new co-ordinating groups are still being examined, much more attentionis now being given to the quantitative aspect of complex formation andstability, and data are accumulating which should soon lead to a morecomprehensive understanding of complexes generally, and of the valencyrelationships governing their structure.A careful quantitative investigation of the stability of chelate complexesof bivalent copper A group of 21 complexesof copper with o-hydroxy-aromatic aldehydes (substituted salicylaldehydes)and p-diketones was selected for study.The method adopted was to carryout electrometric titrations with alkali on solutions (in 50% dioxan-water)containing known concentrations of copper, chelating agent, and excessacid. From the pH and concentration values, equilibrium constants werecalcula.ted for the general reactionhas given interesting results.\c-0- 'c-07-C/ +*cu++ * -cH ' ::cu=,=O %= O.''where cu represents one equivalent of copper, and the hature of the bondsshown as dotted lines is open to discussion.With a 4-co-ordinate complextwo equilibrium constants, relating to the successive attachment of twochelate groups, must apply, but in the absence of data sufficiently accuratefor a precise comparison of these constants, an average value was derived1 M. Calvin and K. W. Wilson, J . Amer. Chem. SOC., 1945,67, 2003WELCH : SOME ASPECTS OF THE CHEMISTRY OF COMPLEX COMPOUNDS. 71for each chelating agent.the chelate grouping, formulated as follows :The dissociation of the acidic hydrogen atom of\C-0 \C-0--c/ \H A T --C/ + H+\C= 0.:' \c=o/ /shows a formal resemblance to the dissociation of the chelated metal com-plex, and a comparison of the respective equilibrium constants shows .thatthe two processes are approximately parallel, so long as similar groups areattached to the three-carbon system of the chelate ring.When the attachedgroups differ in type (e.g., in acetylacetone and salicylaldehyde, the doublebond in the latter case forming part of an aromatic ring), the relationbetween the two constants is abnormal. It appears from this result thatthe " enolate resonance," between (I) and (11), which is largely affected bythe environment of the three-carbon system, plays a much more funda-mental part in the bonding of copper in the complex than it does in thebonding of hydrogen in the free chelating agent. It is suggested that thebonding in the complex involves participation of low-lying vacant orbitalsof the metal atom, homopolar metal-oxygen bonds being formed; onepossibility is the formation of a completely conjugated six-membered ring,as in (111).A promised further paper dealing in more detail with thevalency questions involved in this type of structure is awaited with interest.The reactions of a series of amines with the cis- and trans-forms of thecobalt complex [Co en2C1,]C1 (en = ethylenediamine) have been examinedin detail., I n aqueous solutions of this salt, amines with strong donorproperties generally displace one of the chlorine atoms from the complex,affording salts of the [Co en,(NH,R)CI]" ion; substitution by the amine ofthe second chlorine atom was not observed in any of the cases examined.Some weakly basic amines do not enter the complex to give stable salts, butcause disproportionation to [Co en,]"' salts, or rearrangement from thetrans- to the cis-form.In isolated cases, products of other types, including[Co en,(NH,R)OH]Cl,, [Co en,(H,O)OH]CI,, and [Co en,(OH)CI]Cl, are ob-tained. A survey of the observed reactions leads to the important generalconclusion that the type of complex ion formed by a particular amine isnot primarily dependent on its basicity ; this suggests that resonance effects,such as those already discussed above for chelate complexes of copper, mayplay a part in determining the stability of the amine complexes of cobalt.a J. C. Bailar, jun., and L. B. Clapp, J .Amer. Chem. SOC., 1945, 67, 17172 INORGANIC CHEMISTRY.Study of the equilibria involved in the formation of metal-amminecomplexes has usually been limited to determination of equilibrium con-stants for overall reactions of the general type M + nA =+ [MAn], theindividual constants for the successive additions of co-ordinated addendahaving been neglected. J. B j e r r ~ m , ~ emphasising the importance of theseindividual " formation constants " (distinguished from the " complexityconstant " for the overall reaction), has shown how they may be derivedfrom accurate pH data. Bjerrum's method has now been applied* tocomplexes of cadmium, zinc, nickel, and cupric copper with ethylenediamineand propylenediamine, and of silver with ethylamine and diethylamine.Cadmium, zinc, and nickel all form complexes with three molecules ofdiamine per metal atom, but with copper only two diamine molecules enterthe complex.The silver complexes contain not more than two monoaminemolecules ; data for other silver complexes containing one amine moleculefail to reveal any parallelism between the formation constants and thedissociation constants of the amines, the participation of specific valencyeffects again becoming apparent.The use of surface-active catalysts in reactions involving complexes hasuntil recently received little attention,6 but there is now evidence thatactivated carbon may be employed with considerable practical advantagein the preparation of certain cobaltammines ; hexamminocobaltic chloride,[CO(NH,)~]CI,, previously prepared by relatively cumbersome methods, maybe obtained from cobaltous chloride, ammonia solution, ammonium chloride,and air at atmospheric pressure and temperature, if activated carbon isemployed as a catalyst.Several useful catalysed reactions of cobalt andchromium complexes have been examined more recently. The catalystsselected, vix., activated carbon, Raney nickel, and silica gel, are all effectivein promoting formation of bonds between nitrogen and chromium or cobalt,silica gel being less efficient than the other two, and carbon being preferableto Raney nickel on practical grounds. The reaction [Co en2C1,]C1 + 2NH,(aq.) + [Co en2(NH,),]C1, is normally slow and incomplete, the mainreaction product being [Co en,(NH,)Cl]Cl, ; in presence of charcoal thereaction proceeds rapidly to completion.The preparation of [Cr en,]Cl, isusually difficult because of the need to use anhydrous materials, watermolecules co-ordinated to chromium resisting displacement by amines underordinary conditions ; in presence of charcoal, however, the complex isreadily formed from hydrated chromic chloride and aqueous ethylenediamine.Charcoal also has a very marked catalytic action on the interchange ofammonia and nitro-groups in cobalt complexes ; thus [Co en,(NH,)Cl]"3 " Metal Ammine Formation in Aqueous Solution," Copenhagen, 1941.4 G. A. Carlson, J. P. McReynolds, and F. H. Verhoek, J . Amer. Chem. SOC., 1945,67, 1334.6 For earlier work, see R. Schwarz and W. Krlinig, Ber., 1923, 56, 208; N.Shilovand B. Nekrasov, 2. physikal. Chem., 1925, 118, 79; B. Nekrasov, J. Russ. Phys.Chem. SOC., 1926, 58, 207; I. I. Shukoff and 0. P. Shipulina, liolloid-Z., 1929, 49, 126.6 5. Bjerrum, ref. (3).7 J. C. Bailar, jun., and J. B. Work, J . Amer. Chem. SOC., 1945, 67, 176WELCH: SOME ASPECTS OF THE CHEMlSTRY OF COMPLEX COMPOUNDS. 73reacts smoothly with sodium nitrite, the [Co en,(NH,)(NO,)]" formed a troom temperature reacting further, on warming, to give [Co en,(NO,),]'.Even [Co(NH,),]"' reacts with sodium nitrite on heating the solution withcharcoal, and [Co(NH,),(NO,),] is formed ; the same product is obtainedby the catalysed reaction of [CO(NO~)~]"' or [Co(NH,),(NO,),] with liquidammonia, indicating that the non-electrolytic [Co(NH,),(NO,),] is the stableend-product of ammonia and nitro-group interchange in this series of com-plexes.I n the original paper several other catalysed reactions are con-sidered, and evidence is presented that the use of charcoal in reactions ofoptically active cobaltammines does not influence the sign of the rotationof the products. It is evident that the judicious use of catalysts mayconsiderably facilitate access to complex compounds previously preparedby laborious methods, and thus aid the wider development of theirchemistry.The constitution of the group of complexes typified by the carbonato-tetra- and -penta-amminocobaltic ions has been uncertain since the time ofWerner. At first, the latter complex, [CO(NH,)~CO,]*, was thought tocontain a carbonato-group anchored to the cobalt atom by a single bond,one available valency of the CO, group being left free.Later, some evidencewas obtained that known carbonato-pentammine salts all contained onemolecule of water as part of the structure of the complex ion, and Wernerthen proposed the peculiar " betoxine " formula (IV) forOH these salts. The carbonato-group was thought to be abicarbonato-group attached to a basic hydroxyl groupX outside the complex. I n the original or some modifiedform this structure was accepted for a considerable time.Recently it has been pointed out that the complex ion may be regarded asanalogous to a zmitter-ion, such as that of glycine, the only formal differencebeing that the cobalt complex bears a net positive charge :O.CO,H [ C<NHd.](IV.)'0-TO-H 0-E-0-0 0of. +NH,*CH,*CO,H += +NH,*CH,*CO*O- + HSThis structure obviates the need for a molecule of water of constitution.It has now been confirmed that this water is, in fact, not essential to thestability of the complex, the hydrate of carbonatopentamminocobalticnitrate undergoing dehydration, under suitable conditions, without decom-position.8 In the original paper, considerable physical evidence is adducedfor the dipolar-ion structure of the carbonatopentamminocobaltic ion, andthis type of structure is considered to exist in other complexes of generallysimilar character.The application of the spectrophotometer in the identification and study* A.B. Lamb and K.J. Mysels, J . Amer. Chem. Soc., 1945,67, 468.c 74 INORGANIC CHEMISTRY.of coloured complexes in solution is illustrated by recent work on cericsulphate compIex ions.g The equilibrium involved in the formation of asingle species of complex may be represented, in general, as mA + nB[A,B,]. Spectrophotometric measurements on solutions containing vary-ing concentrations of A and B enable the light absorption of the [A,B,]complex to be determined accurately, and although the actual concen-tration of A,B, is not directly deducible unless the extinction coefficient isknown, values of m and n can be derived from the variation in opticaldensity with the concentrations of A and B. Measurements with solutionsof ceric perchlorate and sodium sulphate show that, a t concentrations upto 0 .0 1 ~ ~ a complex is present containing one ceric ion and one sulphateion ; in more concentrated solutions other complexes, containing moresulphate, appear to exist. I n this case the results obtained themselvesallow an estimate to be made of the extinction coefficient of the complex,from which concentration values, and values of the equilibrium constantfor the dissociation of the complex, can be derived. Although the inform-ation they give is sometimes limited, physical methods for the study ofdissolved complexes are of considerable value, and the use of the spectro-photometer is a welcome addition to the available experimental techniques.Similar photometric methods, supplemented by pH determinations andconductometric and polarographic measurements, have been used in a studyof the t'artrate and citrate complexes of nickel and bivalent copper.1° Inall cases the complexes contain metal atoms and organic acid groups in a1 : 1 ratio; the tartrate and citrate complexes, respectively, are consideredto have structures (V) and (VI), or corresponding six-co-ordinate structureseach containing two more water molecules.Both types of complex behaveas monobasic acids, the addition of alkali resulting in removal of an acidicr 1hydrogen atom (thought to be that shown in heavy type in the formula),and increase of the net negative charge on the complex by one unit. Someof the complexes show further reactions with alkali at high pH values.So little is known about the behaviour of complexes in electrode reactionsthat a recent paper on the polarographic reduction of cobaltammines l1merits brief reference.A wide range of complexes containing co-ordinatedammonia, ethylenediamine, water, chlorine, and hydroxyl, nitro-, andnitrito-groups, was studied, 0 . 1 ~ and N-potassium sulphate, and 0.1N-SOdiUmacetate being used as supporting electrolytes. I n every case reduction takesplace in two successive steps; the half-wave potential of the first dependson the complex used, but the second step has a substantially constant@ R. L. Moore and R. C. Anderson, J . Amer. Chem. SOC., 1946,67, 167.lo M. Bobtelsky and J. Jordan, ibid., p. 1824.l1 J. B. Willis, J. A. Friend, and D. P. Mellor, ibid., p. 1680WELCII: SOME ASPECTS OF THE CHEMISTRY OF COMPLEX COMPOUNDS. 75potential through the whole range.The first step is associated with thereduction of cobalt from the tervalent to the bivalent state, with consequentdisruption of the complex, and formation of [Co(H,O),]", and the second stepwith reduction of the aquated cobaltous ions to the metal. The half-wavepotential of the first step is related to the stability of the complex, and it isclear that polarographic measurements may supply yet another techniquefor studying stability in complexes generally.Studies on the preparation of new complexes include a paper on thehydrazine complexes of chromous iodide,12 which were investigated in thehope of stabilising the bivalent state of chromium by co-ordination.Earlierworkers l3 had claimed the existence of a compound formulated as[Cr(N2H4),]I,, stated to be stable in air, but repetition of their preparativemethods gave only a basic, polynuclear compound of evidently complicatedconstitution. Treatment of anhydrous chromous iodide with anhydroushydrazine was found to give a viscous, cherry-red solution which appearedto be stable in dry air, but reacted vigorously with moisture. Attempts toremove excess of hydrazine from the solution by distillat>ion in a high vacuumgave a solid residue, thought to be inhomogeneous, of the approximatecomposition CrI2,3N,H,. Removal of the hydrazine over sulphuric acid ina vacuum afforded a new, stable complex, [Cr(N,H,),]I, ; prolonged keepingover sulphuric acid, after preliminary evaporation of most of the excess ofhydrazine, gave a solid of the composition CrI,,4N,H4.Evidence is adducedthat [Cr(N,H,),]I, is correctly formulated with six-co-ordinate chromium,the hydrazine groups being attached by single bonds, but no indication isgiven of the constitution of the other compounds. Treatment of theviscous solution with alcohol gave a rose-coloured solid formulated asCr,(OEt),I,(N,H,),, which was not further examined.Complexes of rhodium continue to afford a fruitful field of study toF. P. Dwyer and R. S. Nyholm, their latest communication l4 dealing withcomplexes of tervalent rhodium halides with ethyl sulphide. The chloride,bromide, and iodide of the type RhX3,3Et,S have all been prepared byheating the trihalides with an ethyl-alcoholic solution of ethyl sulphide ;co-ordination is sluggish, however, and the complexes dissociate readily.An apparently binuclear compound, ( RhBr3,2Et,S),, was occasionallyisolated in addition to RhBr3,3Et,S.Potassium argentocyanides and cuprocyanides, and the correspondingcomplex acids, have received some attention.15 The compoundK3Ag(CN),,3H,0, in 0-Oh-solution, is completely dissociated into KAg(CN)and potassium cyanide, an interesting confirmation that covalencies greaterthan two are not favoured by univalent silver.A potentiometric titrationmethod indicates that HAg(CN),, HCu(CN),, and HCu,(CN), are all relativelystrong acids. Like H,Ag( CN),, the acid H,Cu(CN),, corresponding withF. Hein and G.Biihr, 2. anorg. Chern., 1943, 252, 55.W. Traube and W. Passarge, Ber., 1913, 46, 1505.14 J. Proc. Roy. Xoc. N.S.W., 1945, 78, 67.16 (Mlle.) J. Brigando, Compt. rend., 1942, 214, 90876 INORGANIC CHEMISTRY.the well-known potassium salt, K3Cu(CN),,3H20, is unstable in aqueoussolution, even a t 0"; it is thought to dissociate successively into H,CU(CN)~and HCU,(CN)~.Further development of the chemistry of carbonyls by W. Hieber andhis collaborators is shown by work on carbonyl halides of osmium.16 Thetetracarbonyl dichloride, [Os(CO),Cl,], is prepared by heating osmium tri-chloride with carbon monoxide a t 120-160°, under 200 atm. pressure; itis a solid, purifiable by sublimation, and is relatively unreactive towardswater and hydrochloric acid.Similar reactions with Os2Br9 and an oxy-iodide of osmium, respectively, give the corresponding bromide and iodide,the iodide being distinctly more reactive than the chloride. The compound[Os( CO),Br],, a corresponding iodide, and osmium tri- and di-carbonyldi-bromides and -iodides, are obtained by appropriate change of conditions,or by thermal degradation of the preceding compounds.4. ORGANOSILICON COMPOUNDS :. VOLATILE HALIDES AND THEIRDERIVATIVES,Continued interest in organosilicon compounds lends importance to anentirely new method of preparing simple alkyl and aryl silicon halides.1I n the past these compounds have almost always been prepared by theaction of Grignard reagents on silicon halides. The new method consists inpassing the vapour of an alkyl or aryl halide (RX) over heated silicon,generally in the presence of a catalyst, or in heating the liquid halide withsilicon and catalyst in an autoclave. The product comprises a mixture ofalkyl or aryl silicon compounds, the composition of which is stronglydependent on the reaction conditions.Under conditions favourable tohigh yields the principal reaction is 2RX + Si --+ R,SiX,, the dialkyl(or diaryl) silicon dihalide being the predominant product ; small quantitiesof RSiX, and R3SiX, and even of R,Si and Six,, are also formed. Whenconditions are such that pyrolysis of free hydrocarbon radicals is favoured,the products include incompletely substituted monosilanes of the typeSiHRX,. Metallic copper appears to be the best catalyst in the preparationof alkyl silicon halides, whereas silver is preferable in the case of aryl com-pounds ; copper may be incorporated as a silicon-copper alloy, by reductionof added cuprous chloride by the silicon itself, or by the use of sintered,compacted pellets of copper powder and finely powdered silicon.I n a subsequent investigation of the reaction between methyl chlorideand silicon-copper,2 the function of the metal catalyst has been more closelyexamined. A series of ingeniously devised experiments affords evidencethat cuprous chloride and methylcopper are the primary reaction products ;in a subsequent change the cuprous chloride is reduced by silicon, copperbeing regenerated and an active intermediate containing silicon (possibly a16 W.Hieber and H. Stallmann, Ber., 1942, 75, 1472.1 E. G. Rochow, J . Amer. Chem. Soc., 1945, 67, 963.D. T. Hurd and E. G. Rochow, ibicE., p. 1057WELCH : ORGANOSILICON COMPOUNDS. 77silicon-chlorine radical) being produced. Finally, the intermediate reactswith methylcopper, cuprous chloride, or free methyl radicals until a stablequadrivalent silicon compound is formed. The metal catalyst thus operatesby making the halogen from the organic halide readily available for reactionwith silicon, and by prolonging the life of the alkyl radicals in the formof metal alkyls readily susceptible to reaction with silicon-containingintermediates.affords another example of thesame type of reaction. Chlorobenzene and silicon react only sluggishly inthe absence of a catalyst; the product is a complex mixture containingchlorinated diphenyls as well as some phenylchlorosilanes, the yield of thelatter being very small.Better yields are obtained with silicon-copper,provided that an " aged " alloy, rendered friable by intergranular oxidation,is used ; the hard, brittle, unoxidised silicon-copper gives very poor results.Other silicon alloys with nickel, antimony, platinum, and silver were tried;silicon-silver was found to give the best yields of phenylchlorosilaneswith the minimum of unwanted pyrolysis products. The alloy used con-tained 10% of silver and was prepared by pressing the mixed powdersand heating in hydrogen a t 900"; it was then heated a t 400" in a stream ofchlorobenzene vapour.The products contained almost all the reactedchlorobenzene as phenyltrichlorosilane and diphenyldichlorosilane, in mole-cular proportions of about 1 : 3. Before distillation of the liquid product itwas necessary to remove, by filtration, small quantities of aluminium chlorideformed from aluminium present as impurity in the silicon; even smallamounts of aluminium chloride, present during distillation, catalyse a dis-proportionation reaction of phenylchlorosilanes to silicon tetrachloride andbenzene." Rearrangement " reactions in covalent halides and related compoundswere commented on in last year's Report: and a recent paper gives furtherinformation on these interesting processes, with special reference to com-pounds of Group IV elements.On being heated in presence of potassiumchloride and partly hydrolysed aluminium chloride, chloroform and bromo-form undergo rearrangement to mixtures of the original compounds withthe two possible chlorobromides, the halogen atoms being randomly dis-tributed among the available CH groups. Methylene chloride rearrangesin a similar fashion with methylene bromide or iodide. In neither case wasmigration of hydrogen atoms detected. Silicon trichlorothiocyanate,SiCl,SCN, disproportionates into the tetrachloride and tetrathiocyanatewhen the vapour is passed through a tube at 600°, and the results showthat the equilibrium 4SiC1,SCN 3SiC1, + Si(SCN), is established, theequilibrium constant (in terms of mo1.-fractions) being 0.11 ; the reality ofthe equilibrium was verified by a rearrangement reaction with silicon tetra-chloride and tetrathiocyanate, from which a mixture of products of closelyThe synthesis of phenylchlorosilanesE.G. Rochow and W. F. Gilliam, ibirE., p. 1772.Ann. Reports, 1944, 41, 88.G . S. Forbes and H. 33. Anderson, J . Amer. Chem. Soc., 1945, 67, 191178 INORGANIC CHEMISTRY.similar composition was obtained. Partial rearrangement of silicon iso-cyanate and thiocyanate was obtained after heating in a sealed tube a t 140"for 40 hours, the product apparently containing the new compound silicontriisocyanatothiocyanate, Si(NCO),SCN ; isolation of this compound in apure state is considered possible, in spite of relatively rapid rearrangementon distillation. Evidence was obtained for the occurrence of germaniumtrichloroisocyanate, GeCl,NCO, in the products of rearrangement of german-ium tetrachloride and tetraisocyanate in a hot tube a t 500"; this newcompound decomposes extensively on distillation.When chlorine orbromine is passed over stannous oxide or fluoride a t 550-600", the halogen-ation of the stannous compound is followed by rapid disproportionationwhich prevents isolation of stannic oxyhalides or mixed halides : 2Sn0 +2Br, --+ SnO, + SnBr,, 2SnF, + 2C1, --+ SnF, + SnC1,. I n general, itis evident that rearrangement reactions become increasingly more facilewith carbon, silicon, germanium, and tin tetrahalides and pseudo- halides,ease of rearrangement also increasing from chlorides to iodides. Rearrange-ment is favoured, in fact, by increasing atomic volume of the atoms con-cerned, and by decreasing electronegativity of the halogen or pseudo-halogen group.The isolation of chlorothiocyanates of phosphorus and silicon, and ofphosphorus chloroisocyanate, provides further examples of mixed pseudo-halogen derivatives of non-metals.The compounds SiCl,SCN, POCl,SCN,and PC1,NCO were prepared by the action of silver thiocyanate (or iso-cyanate) on the corresponding chlorides, in benzene or carbon disulphidesolution. Definite evidence was obtained for the existence of PCl(NCO),,but very rapid rearrangement makes its isolation in the pure state difficult(see below). It is noteworthy that the mixed halogeno-thiocyanates con-taining more than one SCN group in the molecule have not been obtained;since substitution of one SCN for chlorine raises the boiling point of thecompound by about 64", it seems likely that introduction of two or moreSCN groups would ,raise the boiling point to such an extent that rapidrearrangement would prevent detection of the compounds by any simpledistillation process.I n general, chlorothiocyanates are more rapidly pre-pared, but less stable to rearrangement, than chloroisocyanates. Mixedhalides or pseudo-halides of silicon are more stable than those of tervalentphosphorus, these, in turn, being less stable than corresponding phosphoryland thiop hosphor yl compounds.Other derivatives of the same type, described more recently, includeailicon trimethoxythiocyanate, Si(OMe),SCN,7s phosphorus dichlorothio-cyanate, PCl,SCN,8 and phosphorus chlorodiisocyanate, PCl(NC0),.8 Thefirst compound is prepared by the action of methyl alcohol on silicontetrathiocyanate, and is fairly stable, although slow deposition of a soliddecomposition product occurs on long storage.No evidence was obtainedfor the production of other methoxythiocyanates in the same reaction,H. H. Anderson, J. Amer. Chem. SOC., 1945,67, 223.Idem, ibid., p. 2176. 1 Idem, ibi&., p. 869WELCH : ORGANOSILICON COWOUNDS. 79although all three possible methoxyisocyanates are known ; this confirmsthat introduction of more than one thiocyanate group into the same mole-cule is difficult. The phosphorus compounds are prepared and isolated bythe action of phosphorus trichloride on silver thiocyanate or isocyanate,followed by special distillation procedures designed to minimise rearrange-ment reactions among the products.Brief mention has been made8 of the existence of the three mixedsilicon isocyanatothiocyanates, prepared by thermal rearrangement reactionsof silicon tetraisocyanates and tetrathiocyanate.Clearly, the pseudo-halide chemistry of silicon and phosphorus, in particular, is rapidly beingmapped out, and extension of the field to other elements should soonfollow.Recent, work on the fluorination of volatile inorganic halides by theconventional Swarts method has resulted in the isolation of a new siliconoxyfluoride, Si,OF,, and of two oxyfluorochlorides, Si,OF,Cl, and Si,OF,Cl,.loThe reaction of hexachlorodisiloxane with antimony trifluoride yielded allthree compounds, together with traces of other products not yet identified(possibly Si,OF,Cl and Si30,F,), and silicon tetrafluoride. The three com-pounds have been fully characterised by their physical properties ; hexa-fluorodisiloxane is gaseous under normal conditions, its boiling point being- 23.3". The compounds all undergo hydrolysis by water or alkalinesolutions ; an attempt to prepare an oxyfluosilicate by passing hexafluoro-disiloxane into potassium fluoride solution was unsuccessful, although " awhite precipitate of no definite composition " is said to have been obtained.The oxyfluorochloride Si,OF,Cl, is thought to have the constitutionF,Si*O*SiCl,, as there are reasons for supposing that once fluorination ofone silicon atom has been initiated, substitution by fluorine is completedon that atom before the second silicon atom is attacked.Pure vanadium tetrachloride has recently been prepared l1 by passingchlorine over ferrovanadium (containing 90% of vanadium) a t 200", andfractionally distilling the product through an all-glass column.Accuratevalues of the principal physical constants have been determined. Purevanadium trichloride was prepared from the tetrachloride by heating a t140" for one week, in a stream of carbon dioxide. The trichloride wasused, in a specially devised all-glass apparatus, for a study of the equi-librium 2VCl,(s) + Cl,(g) =+ 2VCl,(g), for which the mean value of Kp a t160" was found to be 1480 (pressures in mm.).Measurements of the freezingpoint of solutions of vanadium tetrachloride in carbon tetrachloride revealextensive polymerisation to double molecules, which is to be expected inview of the odd number of electrons in the VCl, molecule; the equilibriumconstant of the reaction V,Cl, T- WCl, a t - 25" is 18-2 x 1V (con-centrations in g.-mol. per 1000 g. of carbon tetrachloride). An electron-O G. s. Forbes and H. H. Anderson, J . Amer. Chenz. SOC., 1944, 66, 1703; Ann.10 H. S. Booth and R. A. Osten, J. Amer. Chem. SOC., 1945, 67, 1092.l1 J. H. Shone and M. G. Powell, ibid., p. 76.Reports, 1944, 41, 8980 INORGANIC CHEMISTRY.diffraction study of the vanadium tetrachloride molecule 12 in the gasphase shows that the structure is tetrahedral, the configuration not beingaffected, apparently, by the presence of the odd, unshared electron.The hydrolysis of sulphur chlorides gives rise to a complex mixture ofproducts, including sulphur dioxide, sulphur, hydrogen sulphide, polythionicacids, sulphuric acid, and hydrogen chloride; the mechanism by whichthese products are formed has recently been discussed.13 The course of thehydrolysis reaction, and the mode of reaction of sulphur chlorides withpotassium iodide solution, lend support to the view that the initial step inthe hydrolysis reaction involves formation of hydrogen sulphide (or di-sulphide) and hypochlorous acid : SC1, + 2H,O --+ H,S + 2HC10 ; S,C1, + 2H,O --+ H,S, + 2HC10.In subsequent reactions the hypochlorousacid oxidises the hydrogen sulphide or disulphide to the other productsfound experimentally.If such a mechanism can be established, sulphurmust be regarded as the electronegative constituent in sulphur chloridemolecules, and these compounds must be formally considered as chlorinesulphides. Certain reactions of organic derivatives of sulphur chlorides areadvanced in support of the proposed formulation.Considerable attention is being given to the chemistry and structuralrelations of the numerous additive compounds of volatile inorganic halides,particularly those of boron. In a recent study of the acceptor propertiesof the boron atom in boron trichloride, the systems SO,-BC1, and H,S-BCl,have been examined by thermal analysis,14 the apparatus used incorporatingimprovements in a design described earlier.15 Sulphur dioxide and borontrichloride were found to be immiscible a t low temperatures (- 78" andbelow), and no evidence of additive-compound formation was obtained.With hydrogen sulphide the freezing point-composition curve showed abroad maximum a t about 50 mols.-% of boron trichloride, indicating theexistence of a compound H,S,BCl,, melting a t -35.3".Trimethylamine oxide, NMe,O, which has strong electron-donor pro-perties, also forms a very stable boron trifluoride adduct, Me3N0,BF3,16which is soluble in water without decomposition.It is also stated thattrimethylamine oxide forms additive compounds with silicon tetrachlorideand phosphorus trichloride, the formulze of which have not yet been estab-lished ; valency relationships in these compounds should be of considerableinterest.The boron halide adducts referred to above and on p.67 appear to becorrectly formulated with a donor-acceptor bond between the boron andthe oxygen, nitrogen, or sulphur atoms; the existence of a class of borontrihalide additive compounds of quite different constitution is therefore ofconsiderable interest. It has been shown l7 that the compound CH,*COF,BF,,12 W. N. Lipscomb and A. G. Whittaker, J . Amer. Chem. SOC., 1945,67, 2019.l3 H. Bohme and E. Schneider, Ber., 1943, 76, 483.14 D. R. Martin, tJ. Amer. Chem. SOC., 1945, 67, 1088.l5 H. S . Booth and D. R. Martin, ibid., 1942, 64, 2198; Ann. Reports, 1942, 39, 92.l6 A.B. Burg and J. H. Bickerton, J . Amer. Chem. SOC., 1945, 67, 2261.1' F. Seel, 2. anorg. Chem., 1943, 250, 331WELCH : SOME HETEROGENEOUS EQUILIBRIA. 81formed by direct combination of acetyl fluoride and boron trifluoride, is tobe regarded as the acetyl salt of fluoboric acid, [CH,*CO]+[BF,]-; this con-stitution is established by electrical conductivity measurements on liquidsulphur dioxide solutions of the compound, and by a number of its chemicalreactions. There is also crystallographic evidence that the compoundNOF,BF3, previously regarded as a molecular compound, should be classifiedas a nitrosyl borofluoride, [NO]+[BF4]-.18 A similar constitution has beenassigned l9 to additive compounds of antimony pentachloride with acetyland benzoyl chlorides, formulated as [R*CO]+[SbCl,]-.Fully ionic charactercannot, however, be assumed for the compound NOCl,SbCl,, which appearsto be a resonance hybrid involving contributions by both ionic and molecular-compound types of structure. The compound SOCl2,2SbC1, has little ioniccharacter, its solution in liquid sulphur dioxide having a relatively very lowelectrical conductivity. In the original paper l9 interesting details are givenof the technique used in the manipulation of the compounds, and oftheoretical considerations bearing on their unusual constitution.Solvates formed by titanium and stannic tetrahalides with liquid sulphurdioxide have been examined.20 Titanium tetrachloride and tetrabromideand stannic bromide all give solid solvates containing half a molecule ofsulphur dioxide per halide molecule.Each of the three systems gives apair of liquid layers a t temperatures slightly above the quadruple point,and in each case the denser of these layers corresponds in composition withthe solid solvate. Compounds of boron trichloride and silicon tetrachloridewith sulphur trioxide, having the compositions BC13,2S03 and SiCl,,SO,,respectively, have also been described ; 21 they are white crystalline solidsprepared by adding liquid sulphur trioxide to the cooled chlorides.5. SOME HETEROGENEOUS EQUILIBRIA.Applications of phase-rule methods to systems involving inorganic com-pounds of special interest continue to yield fruitful results. The selectionof systems discussed below is intended to show the trend of recent inves-tigations, and numerous studies on binary and ternary systems of well-known types have not been included.Developments in the investigation of relatively simple systems are wellillustrated by recent work on part of the iron-oxygen system, in which thestability relationships of wiistite, the non-stoicheiometric iron-oxide phaseapproximating in composition to ferrous oxide, have been examined inconsiderable detai1.l The general method adopted, frequently used forsystems of this kind, consisted in determining a t known temperatures theratios of carbon dioxide to carbon monoxide in equilibrium with a par-ticular pair of solid phases (corresponding, e.g., with establishment of the18 I.L. Klinkenberg, Chem.Weekblad, 1938, 35, 197.lD F. Xed, 2. anorg. Chem., 1943, 252, 24.2o P. A. Bond and W. E. Belton, J . Amer. Chem. SOC., 1945, 67, 1691.21 G. P. Lutschinski, J . Gem. Chem. RUSG., 1941, 11, 884.1 L. S. Darken and R. W. Gurry, J. AWT. Chem. SOC., 1946, 67, 139882 INORGANIC CHEMISTRY.equilibrium FeO + CO Fe + CO,), or with wustite alone; since wiistiteexists as a homogeneous solid phase over appreciable ranges of composition,carbon dioxide and monoxide mixtures of corresponding composition rangescoexist with wiistite -as the only solid phase. The measurements a t theiron-wiistite phase boundary were made by holding a strip of pure iron ina known small temperature gradient, in a stream of carbon monoxide anddioxide containing the gases in an accurately predetermined ratio ; a tequilibrium the strip showed a sharp boundary between oxidised and reducedzones, the position of this boundary in the temperature gradient showingaccurately the temperature at which the gas mixture used was in equi-librium with wiistite and iron.A generally similar method was employedfor measurement with wustite and magnetite as solid phases. Data for thecomposition of wustite itself, within the homogeneity range, were obtainedby passing carbon monoxide and dioxide, in predetermined proportions,over ferric oxide or pure iron held a t an accurately known temperature,and analysing the solid phase after attainment of equilibrium ; establish-ment of true equilibrium was verified by approach from both sides, the pairsof results showing excellent agreement.The wustite composition data foreach of the temperatures used (1 100-1400") were plotted against theC02/C0 ratio, and the usually linear plots extrapolated to the CO,/COvalues corresponding with equilibrium a t the wiistite-iron and wustite-magnetite phase boundaries; by this means composition data for theboundaries of the field of existence of wustite were secured, and an importantsection of the iron-oxygen phase diagram was accurately constructed. Impor-tant thermodynamic properties of the system were also derived. A notablegeneral conclusion reached in the course of this work is that thermal diffusionmay vitiate accurate measurements of gas composition in heterogeneousequilibria, unless careful precautions are taken.Systems involving vanadium oxides and typical slag-forming oxides haverecently been studied2 by methods developed largely by R.Schenck andhis collaborators. I n general, a metal oxide undergoing decomposition orreduction gives rise to a system containing two solid phases (oxide + loweroxide, or oxide + metal), in which the equilibrium oxygen pressure is fixedat a given temperature, in accordance with the phase rule. When the oxideis completely converted into a lower oxide, and thus disappears from thesystem, it is replaced by another solid phase comprising the reduction productof the lower oxide; a t this point an abrupt decrease of equilibrium oxygenpressure indicates the changed composition of the solid phases present, andexperimental observation of the oxygen pressure change usually enablesthe stoicheiometric composition of the individual phases to be deduced.Ifsolid solutions or non-stoicheiometric phases occur as intermediate steps inthe reduction, the corresponding oxygen pressure changes are gradualinstead of abrupt, and they enable an estimate to be made of the com-position ranges of stability of the phases in question. It is usually incon-venient to measure oxygen pressures directly, and the reduction is usually2 J. KlBrding, 2. anorg. Chem., 1944. 252. 190WELCH : SOME HETEROGENEOUS EQUILIBRIA. 83effected by carbon monoxide or hydrogen, the CO,/CO or H,O/H, ratio inthe equilibrium gas providing a direct measure of the oxygen pressure ofthe oxide system.If a '' foreign oxide," not involved in the direct reduc-tion equilibria, is added to the solid phase, the oxygen pressure may remainunchanged, or it may increase or decrease; if it increases, formation ofmixed crystals cr of a solid compound between the foreign oxide and thelower oxide is indicated, but if it decreases, compound or mixed-crystalformation involves the higher oxide undergoing reduction. Clearly if bothoxides form compounds or mixed crystals with the foreign oxide, thedirection of oxygen-pressure change depends on the relative stability of thetwo compounds. The study of equilibrium oxygen pressures over oxidesystems is thus capable of affording detailed information on the chemical(and thermodynamical) relationships of the solid oxides.I n the case of the vanadium oxides, reductions of the pentoxide inpresence of calcium, magnesium, and manganous oxides, silica, alumina,and calcium silicates and aluminates were carried out at 600-700".Theexistence of three stable calcium vanadates, CaO,V,O,, 2Ca0,V,05, and3CaO,V,O,, was confirmed, and their stabilities were compared ; a new solidphase, CaVO,, shown by X-ray analysis to have a perovskite structure, wasfound. Minor effects, corresponding probably with mixed-crystal formation,were observed with the other oxides, but no other stable compounds wereidentified. It is concluded that calcium vanadate formation dominatesslagging processes in which vanadium oxides participate.A thorough study has recently been made of the equilibrium reactionbetween anhydrous ferrous chloride and hydrogen sulphide in the temper-ature range 340460".3 As expected, the principal equilibrium is FeCl, +H,S =+ FeS + 2HC1, which is established relatively quickly, but there isa further slow reaction, FeS + xH,S FeS, +s + xH,, in which theferrous sulphide phase gains sulphur in excess of the stoicheiometric amount,forming pyrrhotite.This secondary reaction accounts satisfactorily for theproportion of hydrogen found to occur in the gas phase a t equilibrium. Inthe temperature range studied, the composition of the equilibrium sulphidephase approximates to FeS,.,,.The interconversion of the oxides and sulphides of iron, important inthe usual process for removing hydrogen sulphide from fuel gases, hasrecently been studied in some detail.* The initial product of the reactionof hydrogen sulphide with any form of ferric oxide (hydrated or otherwise)is an unstable monohydrate of ferric sulphide, Fe,S3,H,0 ; a t temperaturesabove 20" and in presence of hydrogen sulphide, this decomposes into amixture of FeS, and a highly magnetic sulphide formulated as Fe8Sg (i.e.,FeS,.,,).When Fe,S,,H,O, Fe8Sg, or precipitated ferrous sulphide, in theform of a moist solid, is oxidised by oxygen at temperatures below 50",or-Fe,O,,H,O is obtained; oxidation of FeS, is difficult. In aqueous sus-pensions the course of the reaction is evidently different, for oxidation ofJ. J. Lukes, C. F. Prutton, and D. Turnbull, J . Amer. Chem. SOC., 1945, 67, 697.R.H. Griffith and A. R. Morcom, J., 1945, 78684 INORGANIC CHEMISTRY.suspensions of the ferric sulphide hydrate or of precipitated ferrous sulphideaffords y-Fe,O,,H,O. All the sulphides give a mixture of ferric sulphatewith an oxide (mainly y-Fe,O,) on oxidation a t high temperatures. Therates of the various reactions, and their dependence on the porosity of thesolid materials, have been studied, but data for the various equilibriumconstants have not been obtained so far.Phase relations between lead sulphide, lead monoxide, and lead sulphateare of practical interest in view of their importance in the smelting of leadores, and certain features of the Pb-O-S system have been examinedr e ~ e n t l y . ~ The equilibrium sulphur dioxide pressure for the reaction PbS +2Pb0 += 3Pb + SO,, which has previously been measured only over lowertemperature ranges, reaches one atmosphere at 920".This reaction, andthe reaction PbSO, + PbS =+ 2Pb + 2SO,, proceed a t appreciable speedonly a t temperatures a t which the mixtures are molten (i.e., above about800" and 850", respectively). At and above 920", a t atmospheric pressure,formation of metallic lead is very rapid, because the system has high fluidityand the dissociation pressure for the first reaction equals or exceeds that ofthe atmosphere. It appears that reduction of lead sulphate to the metalby interaction with the sulphide is necessarily preceded by formation of abasic sulphate, the reaction in fact proceeding in two stages : 7PbS0, +PbS 4(Pb0,PbS04) + 4SO,, 4(Pb0,PbS04) + 6PbS =$ 14Pb + 10S0,.Although lead sulphide and lead monoxide readily react together, rapidheating allows determination of melting points in the binary system PbO-PbS, which has a simple eutectic melting a t 790".Interesting methods have been used in the study of equilibria betweenchromium halides, in the gas phase, and iron,6 which are of practical im-portance in methods of coating iron surfaces with chromium. The vola-tilities of chromium halides in an inert gas (nitrogen) and in hydrogen-hydrogen chloride mixtures were determined by an entrainment method,and from the results, combined in some cases with previous information,all the essential equilibrium data were derived.The equilibrium constantfor the principal " chromising " reaction, CrC1, (gas) + Fe (mixed crystals)+ FeCl, (gas) + Cr (mixed crystals) is found to be approximately unitya t $79-930".The entrainment method used in this investigation allowsan experimental approach t o many systems which would otherwise bediflicult to study, and a discussion of its application, given in the originalpaper, is of general interest.The chemistry of the carbides of chromium is now much clearer as aresult of carefully planned equilibrium ~tudies.~ The existence of threecarbides, Cr7C3, Cr,C, and Cr3C2, is established, the supposed carbide Cr,C2being a mixture of Cr7C, and Cr,C. Equilibrium data, obtained by measure-6 E. J. Kohlmeyer and W. Monzer, 2. anorg. Chem., 1943, 252, 74.6 C.Wagner and V. Stein, 2. physikal. Chem., 1943, 192, 129.7 F. S. Boericke, U.S. Bur. Mines, Rept. Invest., 1944, No. 3747; K. K. Kelley,F. S. Boericke, G. E. Moore, E. H. Huffman, and W. M. Bangert, U.S. Bur. Mines,Tech. Paper, 1944, No. 662WELCH : SOME HETEROGENEOUS EQUILIBRIA. 85ment of carbon monoxide pressures in a closed system, were obtained forthe four reactions 3Cr203 + 13C 2Cr3C, + 9C0, 5Cr203 + 27Cr,C, =i+13Cr7C3 + 15C0, 5Cr203 + 14Cr,C, =$ 27Cr,C + 15C0, and Cr203 + 3Cr,C14Cr + 3C0, and these data enable the principal thermodynamicalproperties of the equilibria and of the participating phases to be evaluated.Low-temperature specific heats for the carbides were also determined. Allthe carbides are found to be stable, at temperatures below their meltingpoints, with respect to decomposition into chromium and carbon, lowercarbide and carbon, or chromium and higher carbide; Cr7C3 is also stablewith respect to Cr,C2 and Cr,C.The measurements allow useful conclusionsto be drawn regarding methods for decarburising ferrochrome.Modern methods of study of metal-oxygen systems are typified byrecent publications on oxides of tungsten * and ~ a n a d i u m . ~ Many of thelower oxides, particularly of transition metals, form non-stoicheiometricphases which are stable within appreciable ranges of composition ; theexperimental problem lies in the identification of the successive phases inthe system, and the establishment of their composition limits of stability.The vanadium oxides were examined by preparing mixtures of vanadiumwith one of its oxides, covering in 15 uniform steps the whole net com-position range between V and VO,.,, the upper limit of investigation.These mixtures were heated in a vacuum at suitable temperatures (1200-1600°), and the products examined, largely by powder X-radiograms, aftercooling.At low oxygen contents the system consists of a solution of oxygenin metallic vanadium, the dissolved oxygen deforming the cubic body-centred lattice of the metal towards a tetragonal structure-an effect show-ing it close analogy to the solid solution of carbon in iron. Experimentaldifficulties prevented determination of the limit of solubility of oxygen invanadium. Preparations having the net composition V00.2 were unusuallyvolatile a t high temperatures, material subliming on to the upper part ofthe alumina crucible and on to the silica tube used for heating.Composi-tions between VO,., and VO,., gave results suggesting the existence of aphase, as yet not clearly identified, lying between the vanadium-oxygensolid solutions and the known “ VO phase,” which exists in the rangeVO,., to VO,.,. At the composition VO,.,, and above, the corundumlattice of V,O, appears. An interesting feature of the system is that theVO phase disproportionates on slow cooling, affording a mixture of vanadium(containing dissolved oxygen) and a higher oxide not yet identified; in thisrespect the VO phase resembles wiistite, which a t low temperatures isunstable with respect to iron and magnetite.Studies of the kind justdescribed illuminate the frequently discordant results of early investigators,who had no knowledge of the widespread occurrence of non-stoicheiometryin solids, and show that there is a large field for the application of modernexperimental methods to quite simple systems.41, 92.8 0. Glemser and H. Sauer, 2. anorg. Chem., 1943, 252, 144; Ann. Reports, 1944,W. Klemm and L. Grimm, 2. anorg. Chem., 1942, 250, 4286 INORGANIC CHEMISTRY.The possible existence of higher oxides of platinum, as yet unexamined,is indicated by the relatively high stability of osmium tetroxide, OsO,, andby the supposed existence of a corresponding iridium compound.1° It issignificant, therefore, that a t high temperatures platinum loses weight a t agreater rate in an atmosphere containing oxygen than in an inert atmosphere,evidently owing to volatilisation of an oxide or oxides.Measurements haverecently been made of the rate of loss in weight of platinum a t 1200" inoxygen a t different partial pressures, and from the results, values of thepartial vapour pressures of possible platinum oxides have been calculated. l1These values give concordant equilibrium constants for the reaction Pt +&no, + PtO, if the three oxides PtO,, PtO,, and PtO, are assumed tovolatilise in different oxygen-pressure ranges (up to 275 min., 275-850 mm.,and above 850 mm., respectively, a t 1200"). Sharply distinct segments ofthe pPtOn-pO, curves correspond with these three pressure ranges.The collection and determination of basic thermodynamic informationfor systems of metallurgical, as well as purely chemical, interest have forsome time been an important feature of the work of K.I<. Kelley and hiscollaborators at the Pacific Experiment Station of the U.S. Bureau ofMines, and Kelley's compilations l2 are of fundamental value to inorganicchemists. Much of the work done relates to heterogeneous equilibria, thelatest publication available l3 dealing with reactions of nitrogen dioxideand the decomposition of the nitrates of manganese, calcium, barium, andaluminium, for which all important data are given. Collection of basicinformation of this kind allows useful general surveys to be made of importantclasses of rea~ti0ns.l~Recent work on ternary systems of conventional type includes aninteresting study of the system Al,0,-S0,-H,0,15 in which phase relation-ships a t relatively low temperatures (down to - 22") were examined.Anew crystalline sulphite, A1,0,,3S0,,xH20, was isolated and characterised.Further data on the same system are available from an investigation of thereaction of superheated water with more basic aluminium sulphites.16Considerable interest also attaches to studies on the solubility in water ofsalts and salt mixtures a t temperatures substantially above inwhich phase changes at the critical temperature of water have beenelucidated.10 L. Wohler and W. Witzmann, 2. Elektrochem., 1908, 14, 106; (Sir) W. Crookes,l1 A. Schneider and U.Esch, 2. Elektrochem., 1943, 49, 55.l2 U.S. Bur. Mines, Bull. Nos. 350 (1932), 371 (1934), 383, 384 (1935), 393, 394l3 K. K. Kelley, U.S. Bur. Mines, Rept. Invest., 1944, No. 3776.l4 See, e.g., a discussion on the reducibility of metallic oxides and sulphides, byl5 W. Fischer and E. Burger, 2. anorg. Chem., 1943, 257, 355.l7 A. Benrath, ibid., 1943, 252, 86.Proc. Roy. SOC., 1912, A , 86, 461.(1936), 406, 407 (1937), 434 (1940).H. J. T. Ellingham, J . SOC. Chem. Ind., 1944, 63, 125.Idem, ibid., p. 369WELCH : " SUB-COMPOUNDS " AND INORGANIC FREE RADICALS. 876. " SUB-COMPOUNDS " and INORGANIC FREE RADICALS.Spectroscopists have for a considerable time assumed the existence,under special conditions, of molecules of extremely simple types, many ofwhich are unfamiliar to chemists on account of the apparently anomalousvalencies associated with their constituent atoms.Most of these moleculesare diatomic; some (such as CO and H,) are known to chemists as stableentities; others (such as CH and OH) have been accepted as intermediateproducts of short life involved in the mechanisms of known chemicalreactions ; many, however, have not been identified as participants innormal chemical processes. The wide variety of diatomic units known bytheir band spectra is evident from works on the spectroscopy of mo1ecules.lMany such units, moreover, are thought to exist as important constituentsof stellar atmospheres, and preliminary calculations have been made on theequilibria governing their coexistence a t the high temperatures existing instars.2 These diatomic molecules usually contain atoms exhibiting abnor-mally low formal valencies, and most of them may be classed (with otherlower-valency molecules of anomalous types, not necessarily diatomic) as( ( sub-compo~nds.~~ Although so many of the possible sub-compounds haveeluded detection and characterisation by chemical means, evidence of theirimportance in chemical processes, particularly at high temperatures, isaccumulating from scattered sources, and brief discussion of the subject inthese Reports is felt to be timely.A review dealing with a selection ofsuboxides has already a~peared.~The existence of a lower oxide of silicon has been assumed by manyinvestigators over a considerable period,* and a commercial material(" Monox '7 marketed early in the present century was regarded as siliconmonoxide.More recently, X-ray examination of this substance has indic-ated that it consists of silica and silicon in intimate admixture.5 Themore important characteristics of silicon monoxide were not realised,apparently, until 1940, when a useful contribution to the subject appeared.It was found that when silicon is heated a t 1450" with the requisite quantityof silica, dehydrated kaolin (A1,03,2Si0,), beryl ( 3BeO,Al20,,6Si0,) , or zircon(ZrSiO,), in an evacuated tube, the silicon and silica are volatilised com-pletely from the reaction zone, and appear in cooler parts of the apparatusas a brown sublimate having the net composition SiO; in the case of thesilicates a residue of alumina, alumina and beryllia , or zirconia, respectively,remains in the heated region.The condensed material has the propertiesof an intimate mixture of silicon and silica, and it is inferred that siliconSee, e.g., G. Herzberg, " Molecular Spectra and Molecular Structure," translatedK. Wurm, Chem.-Ztg., 1940, 64, 261.C. A. Zapffe, J. Arner. Ceram. SOC., 1944, 2'7, 293.H. N. Baumann, jun., Trans. Electrochem. SOC., 1941, 80, 95.E. Zintl, W. Brauning, H. L. Grube, W. Krings, and W. Morawietz, 2. anorg.by J. W. T. Spinks, New York, 1939, pp. 482-494.4 See reviews and bibliographies in refs. (3), (7).Chem., 1940, 245, 188 INORGANIC CHEMISTRY.monoxide is formed in the vapour phase in the initial reaction, but dispro-portionates rapidly on condensation (2SiO + Si + SiO,).This inabilityof the sub-compound to exist as subh in the solid state is characteristic,and attempts to condense or isolate sub-compounds are generally unsuccess-ful because of very rapid reversal of the equilibria involved in their formation.Recognition of this general property leads to a better understanding ofsub-compound chemistry ; earlier failures to isolate sub-compounds aresatisfactorily explained, and it becomes evident that such materials mayplay an important part in high-temperature processes involving oxidationand reduction.Formation of silicon monoxide is not confined to reactions betweensilicon and silica (or a silicate) ; oxides of niobium and tantalum are reducedon heating with the calculated quantity of silicon, the monoxide " sublim-ing " and leaving a residue of substantially pure niobium or tantalum.6The condensed " silicon monoxide " (silicon and silica) is itself an effectivereducing agent, as might be expected from its content of finely dispersedsilicon; dolomite and a zinc ore give magnesium and zinc, respectively, onheating with the condensed product.Calcium phosphate is reduced tophosphorus, but the yield is poor.Although silicon monoxide is not stable, apparently, as a condensedphase, its existence in dilute solution is indicated by recent work on silicon-oxygen equilibria in molten steel.', Previously the reaction betweensilica and molten iron had been examined on the basis of a single equilibrium,SiO, + 2Fe += 2Fe0 + Si, but a careful re-investigation of the equi-librium under laboratory conditions shows that a consistent equilibriumconstant for this reaction cannot be obtained. If two successive reductionsteps are postulated, SiO, + Fe =+ FeO + SiO, SiO + Fe FeO + Si,the experimental data can be interpreted quite satisfactorily. These datagive, moreover, a considerable amount of preliminary thermodynamicalinformation on reactions involving silicon monoxide ; the standard free-energy change for the reaction SiO, + Si + 2Si0, a t 1600°, is calculatedto be + 10,000 g.-cals., the equilibrium constant a t this temperature being0.067.It is found that a t high oxygen contents silicon present in a molten-steel bath is to a large extent combined as silicon monoxide, a conclusionwhich has an important bearing on certain aspects of equilibria in moltensteel. Clearly, similar considerations may apply to equilibria involvingother alloying and impurity elements in steel, and in metal systemsgenerally.Boron monoxide, BO, appears to exist under similar conditions tosilicon monoxide.9 Boron trioxide is too volatile for its reaction with borona t high temperatures to be examined in simple apparatus.There isevidence, however, for the occurrence of the interesting reaction A1,03 + B-+ 2AlO + BO, material corresponding with the disproportionation pro-7 C. A. Zapffe and C. E. Sims, Iron Age, 1942,149, 29, 34.0 E. Zintl, W. Morawietz, and E. Gastinger, 2. anorg.Chem., 1940, 245, 8.Idem, Amer. Inst. Min. Met. Eng., Tech. Publ. No. 1498 (1942)WELCH : (' SUB-COMPOUNDS " AND INORGANIC FREE RADICALS. 89ducts of aluminium and boron monoxides being obtained as a sublimatewhen alumina and boron are heated in a vacuum a t 1300". At highertemperatures ( lsOoO), a mixture of zirconium dioxide and boron loses boronand oxygen in an approximately 1 : 1 ratio, indicating volatilisation ofboron monoxide, but no zirconium is lost. Aluminium monoxide wasapparently being studied by E. Zintl at the time of his death,1° but nodetails of the work have been made available so far.Detailed attention has not hitherto been given to labile suboxides otherthan those just described, but there is evidence that others exist.3 Theunusual volatility of lower oxides apparently present in the systemstitanium-oxygen 11 and vanadium-oxygen 12 may be accounted for by theexistence of suboxide molecules in the vapour phase.The stability relation-ships and chemistry of sulphur monoxide, which fits well into sub-compoundchemistry generally, have been reviewed recently. l3 Subsulphides appearto merit further study; silicon monosulphide, SiS, is well known,14 andband spectra of other similar species have been 0bserved.l High volatilityof what appears to be a titanium subsulphide has also been noted.15Several sub-halides are known, and although a sub-chloride (CaCl), sub-fluoride (CaF), and sub-iodide of calcium (CaI) were studied 16 as long agoas 1909, little attention has been given to them since.These compounds ofcalcium, although apparently isolable in the solid state by rapid quench-ing, show the characteristic sub-compound property of easy reversion tothe dihalides and free calcium, from which they are prepared.Sub-iodides of cadmium and zinc, presumably CdI and ZnI, are obtainedas greenish-yellow and black powders, respectively, on heating the metalswith the requisite quantity of iodine at 1000" for ten hours, in a steel bomb,and then cooling rapid1y.l'Aluminium monofluoride, AlF, affords an interesting example of a sub-fluoride, to which attention has recently been directed.lB It was earliershown that aluminium in presence of a metal fluoride can be volatilised a t800-1000", a t a pressure of several rnm.,l9 although the vapour pressure ofaluminium in this temperature range is extremely small. Klemm and Voss,suspecting the existence of a volatile aluminium sub-fluoride to which thisunusual volatility of the metal could be attributed, heated suitable mixturesof anhydrous aluminium fluoride and aluminium in close proximity to awater-cooled silica thimble, in an evacuated tube.At temperatures oflo See obituary notice by H. W. Kohlschutter, Ber., 1942, 75, 66.11 P. Ehrlich, 2. anorg. Chem., 1941, 247, 53.l2 Seep. 85.l3 P. W. Schenk, Chem.-Ztg., 1943, 67, 257, 273.l4 F. Wust and A. Schiiller, Stahl u. Eisen, 1903, 23, 1128; W. Hempel and vonW. Biltz, P. Ehrlich, and K. Meisel, ibid., 1937, 234, 97.Haasy, Z . anorg. Chem., 1900, 23, 32.l6 L.Wshler and G. Rodewald, ibid., 1909, 61, 54.17 K. Siddiqi, Current Sci., 1943, 12, 147.l8 W. Klemm and E. Voss, 2. anorg. Chem., 1943, 251, 233.l9 C. B. Willmore, U.S.P. 2,184,705 (1939)90 INORQANIO CHEMISTRY.650-670" (measured outside the evacuated tube, inside which a somewhatlower temperature prevailed on account of the large temperature gradientin the apparatus), a black sublimate was obtained; a t higher temperaturesthe sublimate in the hotter zones of condensation became white, and above750" the whole sublimate was white. The black sublimate became whiteon heating in argon at 600". Both types of sublimate were shown, byX-ray examination, to contain aluminium fluoride, but in the black materialonly one or two lines of the fluoride pattern appeared weakly. The whitesublimate had a composition lying between AIF, and A1F (in a typicalcase, AlFl.,&, the excess of fluorine over that required by the sub-fluorideformula being ascribed to direct sublimation of some aluminium fluoridefrom the reacting mixture. On repeated resublimation of the white materialwith excess of aluminium, the composition progressively approached AIF(e.g., AlFp36, A1F1.22, AIF1.,,, and AlF,.,, in a series of successivesublimations). The black sublimate, however, always contained morefluorine than a white sublimate from the same experiment, and the com-position was not appreciably changed by resublimation with aluminium.From these results the formation of a volatile sub-fluoride, AlF, which dis-proportionates (3A1F =+ All?, + 2A1) on condensation, appears to be fullyestablished. Klemm and Voss consider that the differences between theblack and the white sublimate can be accounted for by different temperaturecoefficients for the processes of A1F formation and AlF, vaporisation,and by condensation of AlF, in a poorly crystallised or vitreous form a tthe lower temperatures. Further investigations, with closer control oftemperatures throughout the system, and measurement of the rate oftransfer of material to the sublimate, appear to be necessary before theseexplanations can be unreservedly accepted.Another indication of the part played by sub-compounds, this time of aquite distinct and novel type, in reactions of practical interest, is affordedby recent work on the transfer of iron from the molten metal to the gasphase.20 It is suggested that reactions between iron oxide and iron carbide,occurring in the metal surface, result in the release of a gaseous materialof composition FeCO, which subsequently decomposes into iron and carbonmonoxide. Further investigation of this supposed compound is clearlynecessary.The few examples cited above illustrate the varied branches of inorganicchemistry which are involved in the interesting field of lower-valency com-pounds. Clearly, many early conclusions relating to supposed compoundsof this type, often abandoned since as incorrect, need to be reviewed again.Many supposed solid " sub-compounds " have been identified as intimatemixtures of elements with compounds exhibiting more normal valencies,but these mixtures may, in fact, be the disproportionation products of truesub-compounds stable a t high temperatures or under other special con-ditions. Clearly, many sub-compound molecules may be correctly regardedas free radicals.20 E. J. Kohlmeyer and H. Spandau, Arch. Eisenhuttenwesen, 1944, 18, 1WELCH : " SUB-CORIPOUNDS " AND INORGANIC FREE RADICALS. 91Mention may be made of a recent investigation on the imine radical,21NH, which is formed in the decomposition of azoimide by non-luminousactive nitrogen; the radical was identieed by its reactions with hydrogenand benzene, in which ammonia and aniline, respectively, are formed.A. J. E. WELCH.2 1 I<. Stewart, Trans. Faraday SOC., 1945, 41, 663

 

点击下载:  PDF (2425KB)



返 回