摘要:

2182 J.C.S. DaltonReaction between Hexacyanoferrate(i1i) Ion and Nitric and Nitrous Acids.Part 1 1 2 Nature of the IntermediateBy John C. Bates, Keith M. Davies, and Geoffrey Stedman," Chemistry Department, University CollegeThe structure of an intermediate formed by reaction of nitrogen dioxide with hexacyanoferrate(ii1) ion i s deducedto be a bridged binuclear complex, [Fe,(CN),( N0),I3-, on the basis of stoicheiometric, i.r., and Mossbauerevidence. A linear nitrosyl bridge is suggested. The other products, ammonia, carbon dioxide, and cyanogenall suggest that nitrogen dioxide oxidises a cyanide ligand to cyanate, which subsequently decomposes. Theintermediate is formed by a disproportionate reaction between two of the [Fe(CN)6]3--N0, adducts.of Swansea, Singleton Park, Swansea SA2 8PPIN the previous paper we showed that in 6 mol dm-3nitric acid containing some nitrous acid there was a reac-tion with hexacyanoferrate(II1) ion which can be sum-marised as follows:H+ + HNO, + NO3- + H,O + N,O, =+ ZNO,fast fastslowNO, + [Fe(CN),I3- 4 Intermediate + $HCK -C.iow Infinity productBefore the details of the mechanism can be discussed it isnecessary to know the structure of the intermediateformed.EXPERIMENTALMost of the experimental details have been dealt with inthe previous paper.l When precipitates were produced bythe direct addition of silver(1) nitrate solution to thequenched reaction solution, they formed as a rather stickymaterial that, after washing and drying, gave poorlyresolved X-ray powder photographs.Much more satis-factory results were obtained by slowly adding the AgNO,and intermediate solutions to a large volume of water, withstirring. We also found that, when large amounts of ureawere used to quench the reaction solution, the precipitatesgave much less satisfactory diffraction photographs thanwhen precipitation was carried out without quenching.Diffraction photographs were taken on a Guinier de-Wolffcamera (no. 11) using iron radiation. For the early work,air-dried samples were used, because on drying a t 100 "Cin an oven a distinct smell of hydrogen cyanide was detected.1.r. spectra were recorded with KBr-disc and Nujol-mulltechniques using a Perkin-Elmer 225 spectrometer for high-resolution work. Attempts to obtain Raman spectra failedbecause of decomposition in the laser beam.Mossbauerspectra were recorded by the P.C.M.U. section, Harwell.We also used the help of their interpretive service.Analyses.-Silver analyses were carried out by dissolvinga weighed sample in 0.5 mol dm-3 potassium cyanide solu-tion, and electrolysing with platinum-gauze electrodes withmagnetic stirring. The electrode was washed, dried, andweighed. Carbon and nitrogen analyses were carried outon an F and M 185 analyser, normally used with organicsamples. A somewhat higher combustion temperature wasused, 1 080 instead of 1 020 "C, and some variations in thecombustion catalyst were tried, including addition ofvanadium penta-oxide. However the use of standardsamples of known composition showed that the techniquehad limitations, and could be relied on only to give com-parisons with known standards; absolute analyses forcarbon and nitrogen were probably low.Iron(m) in iron(m)hydroxide was determined by dissolving the precipitate inperchloric acid, adding thiocyanate, and measuring theabsorption a t 460 nm. Standard solutions of iron were usedto calibrate the method. Hexacyanoferrate(I1) ion wasdetermined by oxidising to hexacyanoferrate(Ir1) , andmeasuring the absorption at 420 nm.RESULTSFrom the kinetic experiments described in the previouspaper, it is possible to find conditions under which [Fe(CN)J-ion is quantitatively converted to the intermediate, and nosignificant amount of decomposition to the infinity productoccurs. When such solutions were treated with urea todestroy the excess of HNO,, and then with excess ofAgNO,, a green-yellow precipitate was obtained. Filtration,and back titration of the excess of silver, showed that therewas a 2 : 1 silver : iron ratio based on the original [Fe(CN),3-].Four separate determinations gave a mean ratio of 1.992( 0 = 0.06).A similar experiment on the infinity productgave a buff precipitate, with a ratio of 1-71 : 1. Attempts toprecipitate the intermediate from 6 mol dm-3 nitric acid byaddition of a range of other metal cations were unsuccessful,and the silver precipitate was the only form in which theintermediate was isolated. Our reaction conditions werechosen to be as similar as possible to those used in a standardprocedure for preparing disodium pentacyanonitrosyl-ferrate(2-).The species that \ye describe as the inter-mediate sncl infinity products certainly showed manysimilarities to authentic [Fe(CN),N0I2- ion, but they alsoshowed some distinct differences. We summarise below anumber of the properties of these precipitates.Magnetic Properties.-The AgNO, intermediate wasdiamagnetic.X-Ruy Powder Photographs.-The precipitate for theintermediate gave lines (in order of decreasing intensity byvisual estimation) with d values of 3-009, 3.101, 3.221, 4.119,4.157, and 3.710, together with a number of much weakerlines. Those shown in italics correspond to known lines forsilver(1) cyanide. The infinity product could be obtainedonly as a sticky precipitate, and the dried and powderedmaterial showed no lines other than those due to AgCN.No trace of lines due t o Ag,[Fe(CN),NO] was detected ineither case.I .r . S$ectra.-Both intermediate and infinity precipitatesshowed sharp bands in the cyanide and nitrosyl stretchingregions, similar to those in authentic Ag,[Fe(CN),NO].Under high resolution (KBr discs) the nitrosyl stretchingband was resolved into an intense band at 1 937, with a muchweaker band at 1903 cm-l. The nitrosyl stretch for theinfinity precipitate showed a very broad band at 1 940 cm-l.Part I, K. M. Davies and G. Stedman, preceding paper1974 2183Samples were precipitated from 6 rnol dm-3 nitric acid, andmust have been contaminated with co-precipitated im-purities.Weak bands were observed a t other frequencies,but they did not appear in all samples, and their relativeintensities were variable. We cannot ascribe them withcertainty to our intermediate.Mossbauer Sfiectm.-The silver intermediate precipitategave, a t room temperature, a spectrum which on computeranalysis corresponded to two types of iron nucleus present inequal amounts with the isomer-shift (i.s.) parameters {rela-tive to Na2[Fe(CN),N0],2H20}, (a) 0.07, (b) 0.03 mm s-l, andquadropole splittings (q.s.) (a) 1.59, (b) 1-77 mm s-l. Theinfinity precipitate gave a spectrum corresponding to a singletype of iron nucleus with an is. of 0-07 and a q.s. of 1.64mm s-1. A low-temperature magnetic-field experimentshowed the sign of the electric-field gradient a t the nucleusto be opposite in sign in the infinity product to [Fe(CN),-N0I2- ion.Analyses.-Analyses for silver on a sample of the air-driedsilver intermediate gave 46.3%.Analyses by a similartechnique on authentic samples of Ag,[Fe(CN),NO] andAg,[Fe(CN),] showed that our analyses were ca. 1% low.The cause of this is not known. Analyses for carbon andnitrogen by standard microanalytical procedures met withmany difficulties, presumably due to the formation ofcarbides and nitrides {Found (4 analyses): C, 12.8 & 0.4;N, 18-2 4 0.7. Ag,[Fe(CN),NO] requires C, 13.9; N, 19.4.Found (6 analyses) for silver intermediate: C, 11-2 -)= 1.6;h', 18-0 & 0.7. Found (15 analyses) for silver infinityproduct: C, 11.4 5 1.3; X, 18.3 -)= 0.7%).The errorsquoted arc standard deviations.Properties (f the Intermediate in Solution.-The feature ofthe intermediate in solution which distinguished i t mostclearly from authentic [Fe(CN),N0]2- ion and from theinfinity product was its U.V. absorption spectrum. Theintermediate was characterised by a broad maximum a t 382nm {E 960 dm:; mol-l cm-l, based on the original [Fe(CN),3-]},whereas the infinity product showed negligible absorption inthis region. The ion [Fe(CN),N0I2- has a peak at 395 nm(E 21 dm3 mol-l cni-l). In most other respects measure-ments on solutions of the intermediate gave results verysimilar to those obtained with authentic [Fe(CN),NO] 2- ion.The significant experiments are described below.Solutions of [Fe(CN),N0I2- ion when treated with excessof alkali and potassium cyanide, so that final concentrationswere [OH-] := 0.4 and [CN-] = 0.2 mol dm-3, underwentreaction ( I ) .Experiments with the intermediate showedCY- + [Fe(CN),N0l2- + 20H- +its similarity to [Fe(CN),h'Ol2-, one mole of nitriteion and one mole of [Fe(CN),I4- being formed per mole oforiginal [Fe(CN)J3-. A similar result was obtained for theinfinity product. In the absence of cyanide ion [Fe(CN),-N0I2- and alkali reacted to form bright yellow [Fe(CN),-NO,]4-, the rate law being v = K[Fe(CN),N02-][OH-].[Fe(CN),N0I2- + OH- --P [Fe(CN),(NO2H)I3- +[Fe(CN),I4- + 50,- + H20 (1)k OH-slow fast[ Fe (CN) ,NO,] 4- ( 2)The intermediate behaved similarly. In alkali a yellowcomplex was formed, ~ 4 0 0 2 600 dm3 mol-l ern-]-. {For[Fe(CN),NO2l4- under our conditions we found €400 3 220dm3 mol-l cm-l; Swinehart and Rock reported €400 to be3 075 tlm3 rnol-1 cm-l.) The rate constant K for the inter-mediate was 0.149 compared to 0.285 dm3 mol-1 s-1 for[Fe(CN),N0I2- ion under our conditions, I = 0.3 mol dm-3and 25 "C.Individual runs with a large excess of hydroxidegave excellent pseudo-first-order plots over a t least threehalf-lives. A marked difference between the intermediateand [Fe(CN)5NO]2- ion showed up when the reaction wasfollowed for longer periods of time. Although Fe[(CN),-Nod4- underwent slow hydrolysis through dissociation ofthe nitrite ion and formation of [Fe(CN),H20I3-, in thepresence of excess of nitrite ion in solution this dissociationwas supressed and the yellow solution was stable.In someexperiments we could detect no change over 36 h. Solu-tions of the intermediate in alkali, with excess of nitrite ion,were much less stable. A cloudiness appeared in the solu-tion shortly after formation of the yellow colour was com-plete, and on standing the yellow colour faded and a reddishbrown precipitate of iron(II1) hydroxide appeared. Reac-tion was complete within 24 h at room temperature. Theother product identified was [Fe(CN),I4-. The results aresummarised below where [Intermediate], represents theinitial intermediate concentration {based on the initial[Fe(CN),3-]). The main problems in the Fe(OH), analysisProducts of decomposition of the intermediate in alkalilo4 [Intermediate], 103[X0,-]mol dm-3 mol dm-3 yo Fe(OH),, :,b [Fe(CN),]*-2 5 41.2 45.54 10 41-0 44.08 20 47.0 50.616 40 17.9 64.516 a 40 21.0 65.032 80 12.9 53.70 [OH-] = 0.08 mol dm-3; in all other experiments [OH-] =0.4 mol dm-3.arose from its colloidal nature, and we doubt if the analysesare better than f5%.In most cases our analyses accountfor over 80% of the original [Fe(CN)J3- ion and our bestresult accounts for 97.6%. A similar experiment carriedout on the infinity solution yielded 43% Fe(OH), and 57%[Fe (CN) 6] 4-.In an attempt to obtain information about the charge onthe intermediate, the relative rates of uptake of the inter-mediate and [Fe(CN),I3- ion on to an anion-exchange resin,Amberlite I.R.A.400 in the nitrate form, were measured.Solutions containing both [Fe(CN),I3- ion and the inter-mediate were shaken with the resin, and their disappearancecalculated from absorbance measurements a t the maximaa t 420 and 380 nm respectively. Measurements were madeover a short enough period that decomposition of the inter-mediate to the infinity product could be neglccted. Therates of uptake were similar, that of [Fe(CS),I3- ion beingslightly faster. Attempts to obtain the same informationby studying rates of elution from an ion-exchange resin wereunsuccessful because of decomposition of the intermediateto the infinity product.DISCUSSIONWe shall begin by discussing the structures of the inter-mediate and infinity products. When these are estab-lished, at least tentatively, we shall attempt to interpretthe kinetic results described in the previous paper.It isclear that although we have deliberately used conditionsJ. H. Swinehart and P. A. Rock, Inorg. Chenz., 1966, 5, 5732184 J .C.S. Daltonas close as possible to those of Brauer's preparativemethod, neither our intermediate nor our final product is[Fe(CN),NOl2-. The U.V. spectrum of the intermediatein solution is different, and the colours of the silverprecipitates differ from that of authentic Ag,[Fe(CN),-NO] precipitated under identical conditions, and, moredamning, the X-ray powder-diffraction photographs aredifferent. Thus, although there are undoubtedly manysimilarities to [Fe(CN)5N0]2-, we are looking for twodifferent species.The i.r. spectra clearly show thepresence of an NO+ ligand in both intermediate andinfinity products. This is confirmed by the fact thattreatment with alkali and cyanide releases one mole ofnitrite per mole of original [Fe(CN),I3- ion. Thesimilarity in the spectra of the yellow species formed ontreating the intermediate and [Fe(CN),N0]2- ion withalkali, and the similarity in the rates a t which thesespecies are formed argues in the same direction. The factthat reaction of the intermediate with alkali and cyanidegives a quantitative yield of [Fe(CN),I4- ion shows thatwe have a species with iron in the oxidation state 11.This is also consistent with the silver salt being dia-magnetic.The Rilossbauer spectra of the silver salt ofthe infinity product show that this too is diamagnetic,and presumably iron(r1). The i.s.'s observed in theMossbauer spectra are in the range known for iron(1r)compounds, though this is not a very conclusive piece ofinformation as the ranges for iron-(II) and -(HI) overlap.3The stoicheiometric determinations show that whenthe intermediate is precipitated as its silver salt, twomoles of silver ion are consumed per mole of original[Fe(CN),I3-. When precipitation takes place 0-5 moleof hydrogen cyanide is released per mole of [Fe(CN),I3-ion, and blank experiments showed that this should bequantitatively precipitated in 6 mol dm-3 nitric acid,even though the pK, is 10. The X-ray powder photo-graph shows the characteristic line for AgCN. Hence itwould appear that 1.5 moles of silver are used in pre-cipitating the intermediate.The most valuable piece ofinformation on the structure of the intermediate comesfrom Mossbauer spectra, which show that the silver inter-mediate contains two types of iron nucleus in a 1 : 1 ratio.The i s . and q.s. values are similar to those for authenticAg,[Fe(CN),NO]. This suggests that either we haveprecipitated two different compounds of iron(II), or thatwe have an unsymmetrical binuclear complex. The factthat the yield of HCN is ca. 50% when formation of theintermediate is complete suggests that there are twodifferent types of iron present in equal amounts. Thefact that there seems to be a limiting yield of 50%Fe(OH), and 50% [Fe(CN),I4- in the decomposition of theintermediate in alkali points in the same direction.Attempts to separate the intermediate into fractions bypartial precipitation with AgNO, were unsuccessful,though this could be due to a failure in our technique;N. Greenwood and T.C. Gibb, ' Mossbauer Spectroscopy,'Chapman Hall, London, 1971, p. 91.A. Haim, D. A. Dows, and W. K. Wilmarth, J . Inorg.Nuclear Chem., 1961, 21, 33.R. S. Murray, personal communication.thus we were not able to separate the cyanide from theintermediate either. A very successful method forseparating cyanide complexes and determining theapproximate molecular weight is by passage through aSephadex column. However, the instability of the inter-mediate rendered experiments of this type inconclusive.We did not find evidence at any stage for fractionation ofthe intermediate into two compounds.A final argumentis that we are quite unable to suggest two species, bothsimilar to but different from [Fe(CN)jN0]2-, that coulddecompose by a first-order reaction to give anotherproduct again similar to, but different from, [Fe(CN),-N0I2-.We are forced to the conclusion that our intermediateis a binuclear species. The fact that the silver to ironratio is 3 : 2 suggests a charge of 3-. This is consistentwith the results of the ion-exchange experiment, thoughexperiments were not carried out with a wide enoughrange of anions to make this a strong piece of evidence onits own. As there is one nitrosyl ligand per atom of iron,a formal balancing of charges requires there to be ninecyanide ligands.With 11 ligands in all, this must be asingly bridged complex [Fe,(CN)s(N0),]3- with either abridging cyanide or nitrosyl group. I .I-. spectroscopywould seem the obvious method of distinguishing betweentwo possibilities. A good deal of work lias been done onthe i.r. spectra of linear cyanide bridges, and it has beenreported that they are characterised by a band ca. 100cm-l higher than the normal cyanide stretching4 ,4similar result has been found for [Fe,(CN),,]4-, whichappears to be a doubly bridged species where the cyanidebridges cannot be linear. We have not found any suchband in our spectra. Not very much is known aboutbridging nitrosyl groups.Those that are known bridgein a non-linear manner through the nitrogen atom andhave i.r. bands in the range 1330-1 550 cm-l, verymuch lower than in our spectra. Linear nitrosyl bridgesdo not seem to have been reported. Another ligand iso-electronic to the nitrosonium ion is dinitrogen. A sub-stantial number of complexes with this ligand have beenmade, and it appears that a binuclear complex with alinear - N S - bridge shows a lower frequency for the N-Nvibration than the analogous mononuclear species. Forthe salt [Ru(NH,),(N,)](BF,), the band is at 2 154 com-pared with 2 100 cm-l for [(NH,),Ru-N-N-Ru(NH,),]-(BF4)4. Several other examples could be q u ~ t e d , ~ - ~ allof which show a reduction in frequency in the species witha dinitrogen bridge.The fact that the isoelectronicspecies cyanide ion and dinitrogen show respectively anincrease and a decrease in vibrational frequency on form-ing a bridge is striking, and is doubtless to be interpretedin terms of the relative contributions of o-donor and x-acceptor bonding. We suggest that the weak band a t1903 cm-l in our intermediate is due to a linear6 J. E. Huheey, ' Inorganic Chemistry, Principles of StructureA. D. Allen and C . V. Senoff, Chem. Comm., 1965, 621. * J . Chatt and G. J. Leigh, Chem. SOC. Rev., 1972, 1, 121.M. N. Hughes, ' The Inorganic Chemistry of Biologicaland Reactivity,' Harper and Row, London, 1972, p. 459.Systems,' Wiley, London, 1962, p. 2101974 2185Fe-XEO-Fe bridge, while the other peak at 1 940 cm-l isdue to a terminal linear nitrosyl group, similar to that in1 Fe(CX)5XO]2- ion.Another piece of evidence leading to the same conclu-sion concerns the structure of the infinity product formedby first-order decay of the intermediate.The hlossbauerspectrum shows it to contain a single type of iron nucleus,in an environment tyerj‘ similar to authentic [Fe(CN),-Sol2 - ion. However, the low-temperature magnetic-field Miissbauer spectrum shows the sign of the electric-field gradient at the nucleus to be opposite to that of1 I;e(CN),XO:2 ion. Simple theory indicates that thel~resencc of a structure with cis-nitrosyl groups. \Vc canonl!, find two structures that satisfy the requirements ofthe MiIijsshaucr spectrum, (B) and (C).Either of thesecould bc \-ibualised as being readily formed from a nitro-~!+bridg:cd structure. To iorni these structures from ac.!-anitlc-l,ridEe~l structure would require a much inorecwmples rearrangmient . Several lines of evidenceAnalysis of standard substances gave figures that were1% low. Even assuming that this can be applied to oursilver intermediate precipitate the analysis is still 3%low. There were many problems in the rnicroanalyticaldetermination of nitrogen and carbon, and we wereforced to rely on comparison with known standards,rather than use the absolute figures. As the intermediateshows many similarities to [Fe(CN),N012- ion, and bygood fortune is thought to have the same overall com-position, we used Ag,[Fe(CN),NO] as our standard.Carbon and nitrogen analyses for the intermediate and[I;e(CN),N012- agreed within the rather large standarddeviations, and the same is true for the infinity product.The data are consistent with our postulated composition,but are not good enough to prove it.The general similarity between [-l;e(CN),NO la- ion andour intermediate seems reasonable in the light of struc-ture (A).Reaction with alkali presumably involves con-version of the non-bridging nitrosg.1 to a nitro-groupCN/ / Nc7CN CN( A )CNfavour (C), but our work on the infinity product is not yetcomplete and we will present the evidence later. Thechoice bvtween (B) and (C) does not affect the presentnrgumerit. Structure (A) is clearly consistent with ourhlossbaucr evidence. The magnitude of the q.s.is thesame in l;eS,Y as it is in cis-FeX4Y2, and does not dependon tlie sign o f the electric-field gradient. Thus both ironnuclei in (‘1) will have :Fe(CN),N0j2--like environments,with similar q.s. and i.s. values, as is observed. The mostdubious feature of this argument is the tacit assumptionthat the nitrosj.1 ligand has tlie same effect on the ironiiuclcus wlictlier it is bonded through nitrogen or oxygen.‘[his is unlikely to be strictly true, but we cannot niakeany otlicr ab3umption at this stage. I t is clear that thenon-bridging nitrosyl group in (A) must be cis to thebridging iiitrosyl; if it were trans then one of the ironnuclei would show a q.s. half that of [Fe(CN)5N0]2 -, ast 1icor.i.I-mclicts for LYLZIIS-F~X~Y, compared to FeX,Y.Jt is also Ixobablc that there will be an effect because theiiitrosyl bridges in (C) must be non-linear. I t would beinteresting to look at the Mossbauer spectra of the knowndinier [F(:,((’N),,,jG- and compare it with [Fe(CX),j4- ion.It ii-o~~l(l lie x-ery desirable to confirm this structure byanalj-tical data on the precipitated solid. Our solid issuggested to be ilg,jFe,(CS),(NO),] +- AgCN. This hasthe same o\-erall composition as Ag,LFe(CN),NO], andshould give identical analyses. The figure for silver was464(7/,, coinparod with a theoretical value of 50-00/, .I-*CNN C t - C N / Fc /NoCNCNinitiallj-. The rate constant was similar to that observedfor the corresponding Ve(CN)5NO]2- reaction; tliefact that it is close to half the value is probably acoincidence.If the product -( N C) ,Fe-N 0-lie (CN) 4-N0,j5- broke down rapidly then we would expect releaseof [Fe(CX)5KO]2- which would react with alkali at a ratesimilar to that of the intermediate to form the ion[Fe(CN),K0214 -. \Ye would certainly not expect thesimple first-order kinetics observed under these con-ditions. I t seems much more probable that the slowreaction actually observed, which leads to formation of[I;e(Ch’),js- ion and Fe(OH), overnight, arises from slowbreakdown to LFe( CX),N0j2- and [Fe(CN),( H,O) XOJ~-.Presumably it is the latter species that overnight hydro-lyses to yield Fe(OH), [oxidised by dissolved dioxygen toFe(OE-I),] and cyanide ion.The ion [Fe(CK),NOi2- inalkali and cyanide gives [Fe(CN),j4- quantitatively, sothe limiting yields should be 500/, [I:e(CN),]4 - and 500,;Fe(OH),. Our results approach this at low initial inter-mediate concentrations, and in the case where we have a9’7.G:/, material the yields were close to these limits. Inthe presence of excess of cyanide ion produced byhydrolysis it may be that there is some rcaction (3).2CX- + j l;e(CN),(H,0)N0,j3- --t[Fe(CN)J- i- KO,- + I€,OPresumably this is why the [Fe(CN),j4- yield is oftengreater than 5076. I t is significant that the experiments(32186 J.C.S. Daltonwith the highest initial intermediate concentration, whichwill produce the highest cyanide concentration byhydrolysis, also produce the lowest yields of Fe(OH), andthe highest yields of [Fe(cN),l4-.The limiting yield,when all the cyanide ion is used in this way, will cor-respond to a [Fe(CN),I4- yield of 100 x (lO/l2) = 83.3%.The highest figure we obtained was 65%.The kinetics, discussed in the previous paper, showedthat the intermediate was formed by a reaction of nitro-gen dioxide with [Fe(CN),13- ion. In the process anatom of iron(II1) is reduced to iron(II), and a nitrosylcation is introduced as a ligand. Simple material balanceshows an atom of oxygen still to be accounted for. Thereaction products include ammonia, carbon dioxide, andcyanogen. This suggests that a cyanide ion is oxidisedto cyanate (or fulminate). Acid hydrolysis of cyanic acidis known to produce ammonia and carbon dioxide, veryrapidly, and cyanogen is the pseudo-halogen derived fromcyanic acid and hydrogen cyanide.We cannot accountquantitatively for all cyanide which we postulate to beoxidised to cyanate ; the final yield of ammonia is 50% ofthe maximum possible value.The intermediate, being a binuclear complex, must bederived from two species formed by initial reactionbetween NO, and [Fe(CN),13- ion. We call this speciesthe ‘ monomer.’ The close agreement between the rateof disappearance of [Fe(CN),I3- ion and the rate ofappearance of the intermediate shows that this ‘ dimeris-ation ’ must be a rapid process; there does not seem to beany build-up of the ‘ monomer.’ This is confirmed bythe observation of an isosbestic point in the U.V.spectra,when the [Fe(CN),I3- peak at 420 nm was being replacedby the intermediate peak at 382 nm. Discussion of thenature of the ‘ monomer ’ is more speculative. It seemsimprobable that the cyanate ion is released from the‘ monomer ’; if it is there seems no obvious reason whythe monomer should not be the ion [Fe(CN),N0I2-. Apossible alternative would be an isomer, a nitrosyl-0-complex as in equations (4)-(6). This assumes thatNO, + [Fe(CN),13- --t [Fe(CN),0N]2- + OCN-2 [Fe (CN) ,ON] ,-L(4)[(NC),Fe-ON-Fe(CK)40N]3- + CN- ( 5 ) 4 [ (NC),Fe-ON-Fe ( CN)4XO]3- (6)‘ dinierisation ’ occurs before the nitrosyl-0 group has achance to isomerise to the normal nitrogen-bonded form.Isomerisation of nitrogen-bonded cyanide to the more-stable carbon-bonded form can occur sufficiently slowly10 J .P. Birkand J. H. Espenson, J . Amer. Chem. Soc., 1968,90,l1 A. W. Adamson, J . P h y s . CIzenz., 1952, 56, 858.1153.to be followed kinetically.1° I t is, however, surprisingif a cyanide ion could be displaced sufficiently rapidly in(5) to occur faster tha.n isomerisation. Possibly a morecogent objection is the fact that the yield of ammonia is50%. This suggests that the two cyanate groups havedifferent fates. The 50% yield could be a chance result,or it inight be due to some disproportionation reaction ofcyanate. Adamson,ll in studying alkaline oxidation ofcyanide by [Fe(CN),13- ion, could only account for 500/0of his product as ammonia.If the cyanate ion is still attached to the monomer, thenit is probably bound to the nitrogen atom of the nitrosylgroup. The ion [Fe(CN),N012- forms adducts withmany species, which are written as [Fe(CN),(NOX)I2-where X is the species in question.Some of these sys-tems have been studied, in detail, and the structuressuggested are of the type -N(O)X. Swinehart l2 andLancaster and Murray l3 have suggested that, in theoxidation of sulphite by [Fe(CN)$ ion, sulphite adds onin a similar fashion to a cyanide ion, producing a species[Fe(CN),(CNSO3)I5-, analogous to the well known adductof SO3,- and [Fe(CN),N0I2. If such an addition were tooccur in our system, the bonding of a cyanate onto acyanide ion might appear to provide a route for formationof cyanogen. However, we can account for all the cyan-ide except that postulated to be oxidised to cyanate, sosuch a reaction cannot be of any importance. Thus wesuggest a structure of the type [ (NC),Fe-N(0)OCN]3-.We also suggest that the two cyanates are lost in separatestages, and may suffer different fates.2[ (NC),Fe{ N (0) OCN)I3- +[(NC),Fe-NO-Fe(CN),( N(0) OCN}]*- + OCN- + CN- (7)[ (N C) ,Fe-N O-Fe (CN) 4N 01 3- - 1- OCN - (8)In view of the variety of functions that NO, has per-formed in this reaction, reducing iron-(111) to -(II),oxidising cyanide ion to the level of cyanate, and intro-ducing a nitrosyl ligand into the co-ordination shell, itwould be interesting to examine the reactions of nitricand nitrous acids with other metal complexes.We thank the S.R.C. and the University of Wales formaintenance awards (to K. M. D. and J . C. B. respectively).The S.R.C. provided a grant enabling the Mossbauer spectrato be run. We also acknowledge helpful discussions withDr. R. S. Murray.[4/560 Received, 20th March, 19741l2 J. Swinehart, J . Inorg. Nucleai! Chem., 1967, 29, 2313.l3 J . M. Lancaster and R. S. Murray, J . Chew. SOC. ( A ) , 1971,2755

ISSN:1477-9226    

DOI:10.1039/DT9740002182

出版商:RSC    

年代:1974

数据来源: RSC