1976 1293Kinetic and Equilibrium Properties of Pentacyano( 3,5-d imethylpyrid ine)-iron(ii) and Related Anions in Mixed Aqueous Solvents tBy Michael J. Blandamer, John Burgess,. and Robert 1. Haines, Chemistry Department, University ofThe kinetic pattern for reaction of [Fe(CN),(3.5-Me,py)l3- and [Fe(CN)5(3-CNpy)]3- with a range of incorninggroups, in water and in 40% glycol and 40% t-butyi alcohol, IS consistent with the operation of a D mechanism underall the conditions studied. Rate constants for the reaction of [Fe(CN),(3,5-Me2py)l3- with cyanide ion havebeen determined in binary aqueous mixtures containing up to 40% by volume of methanol, ethanol, t-butyl alcohol,glycerol, or tetrahydrofuran. Correlation of these rate constants with Grunwald-Winstein Y values gives a valueof M of -0.1 ; this small value reflects the low sensitivity of the rate to solvent composition.Free energies ofactivation determined from the observed rate constants have been plotted against excess Gibbs free energies ofmixing (GE) for the respective solvent mixtures. The resulting pattern indicates how far the thermodynamiccharacteristics of mixing of the solvent are reflected in the kinetics of this substitution reaction. The relationbetween the relative stabilities of [Fe(CN),(3,5-Me2py)]3- and its pyridine analogue, and solvent mixture GEvalues, has also been examined. The solvatochromic behaviour of the charge-transfer band of [Fe(CN) 5(n/-methyl-pyrazinium)12- has been investigated, and contrasted with that of organic analogues.Leicester, Leicester LE1 7RHTHE majority of the numerous investigations of kineticsof substitution at inorganic complexes in mixed aqueoussolvents have been of reactions with a dissociative inter-change ( I d ) mechanism. In the present paper wereport the results of a study of the kinetics of sub-stitution at complexes of the type [Fe(CN),LI3-, whichproceeds by a limiting s ~ l ( D ) mechanism, in somemixed aqueous solvents.We compare the reactivitytrends for these reactions with those for I d aquation oftypical cobalt (111) amine halide complexes, and in-vestigate whether there is any connection between thevariation of kinetic parameters with solvent com-position and the excess molar Gibbs free energies ofmixing of the respective solvent mixtures.The relative stabilities of pairs of complexes[Fe(CN),LI3- (L = an alkylpyridine and pyridine itself)are known to depend on solvent composition. Wereport how these relative stabilities depend on solventcomposition for several series of binary aqueous mixtures,and compare these trends with reactivity trends forthese complexes.These studies indicate how a com-parison of both kinetic and equilibrium quantitiesdescribing the behaviour of solvents in solution on theone hand, and thermodynamic parameters characterisingthe respective binary aqueous mixtures on the other,can prove fruitful and informative.EXPERIMESTALReagents.-The salts Na,[Fe( CN) ,( 3,5-Me;py)], $ Na,-[Fe(CN),(3-Clpy)],$ and Na3[Fe(CN),(3-CNpy)] were pre-pared by published methods ; they analysed satisfactorily.Potassium cyanide, potassium thiocyanate, potassiumnitrate, pyrazine, and thiourea were AnalaR grade materials.The salt Na,[Fe(CK),(NH,)] (Hopkin and Williams) andNN-dimethyl-p-nitrosoaniline (Ralph N.Emanuel) wereused as supplied. N-Methylpyrazinium iodide was pre-pared from pyrazine and methyl iodide., Methanol wasdried over magnesium and iodine and then distilled;tetrahydrofuran and dioxan were freed from peroxides bypassing down a column of activated alumina; other7 No reprints available.f py = Pyridine, 3,5-Me2py = 3,5-dimethylpyridine, etc.9 K = K,~/l<,p from ref. 3.organic solvents were of the best commercially availablegrade and were used as received.Kinetics.-Kinetic runs were conducted in 1 cm silicacells in the thermostatted cell compartment of a UnicamSP 800A recording spectrophotometer.Rate constantswere computed (PDP 11) from the observed x-ariation ofoptical density using a standard least-mcan-squaresprogram.Equilibrium Studies.-Values for the equilibrium constantK $ for reaction (1) were obtained by a mcthod fullydescribed eIsewhere.3 We allowed a period of 3 h forequilibration ; results were calculated from optical densitiesCFe(CWpy)13- + 3,5-MezPy[Fe(CN),(3,5-Me2py)l3- -: PY (1)measured a t 650 nm. For a selection of reaction mixtures,optical densities were monitored a t intervals to check thatthe 3-h period was long enough for attaininent of equili-brium, but not so long that significant decomposition wasdetectable.RESULTSAll the kinetic runs were in the presence of a n excess ofthe incoming ligand.Under these conditions all the runswere first order in iron(I1) complex concentration up to atleast three half-lives. For the reactions of [Fe(CN),-(3, 5-Me2py)]3- with thiocyanate, thiourea, pyrazine, andN-methylpyrazinium, isosbestic points were observed at431, 515, 393, and 456 nm respectively. In these cases therate of formation of the product was confirmed t o be thesame as the rate of disappearance of the starting complex(cf. Table 2).First-order rate constants for reactions of [Fe(CN),-(3,5-Me2py)l3-, [Fe(CN),(3-CNpy)I3-, and [Fe(CN),(3-C1-py)I3- with cyanide ion, in water and in 4076 ethyleneglycol or 40% t-butyl alcohol, are reported in Table 1.First-order rate constants for reactions of the 3, 5-RIe2pyand 3-CNpy complexes with a range of incoming ligsncls arein Table 2.These rate constants were, except where statedotherwise, determined by monitoring the disappearance ofthe starting complex.H. E. Toma and J. M. Malin, Inorg. Chem, 1973, 12, 1039.C. T. Bahner and L. L. Norton, J . Amei.. Chenz. Soc., 1950,D. P. Biddiscombe and E. F. G. Herington, A~zalyst. 1956,72, 2881.81, 711J.C.S. DaltonTABLE 1First-order rate constants, hobs., for reaction of [Fe(CN),(X-py)13- anions with cyanide ion, at 298.6 K and an ionicstrength of 0.024 mol dm-3 (maintained with potassium nitrate) alo3[ KCN] /mol dm-37X Solvent 0.6 1.2 1.8 2.4 3.0 3.6 4.2 4.8 6.0 A 7.2 8.4 9.0 9.6 12.0 15.0 18.0 21.0 2 4 2I 13,5-Me2 Water 0.97 1.15 1.23 1.26 1.29 1.24 1.35 1.33 1.30 1.33 1.24 1.35 1.3240% 1.55 1.71 1.76 1.83 1.68 1.67 1.63 1.943-CN Water 1.74 1.94 1.93 2.12 2.09 2.00 2.15 2.20 2.25 2.24 2.5440% 2.40 2.44 2.47 2.41 2.44 2.42 2.49 2.46 2.58 2.61 2.60 2.7640% 0.22 0.80 1.11 1.32 1.40 1.23 1.35 1.43 1.35 1.38 1.42 1.39103k/s-1ButOH bButOH bglycol b3-C1 Water 1.64 1.64 1.75 1.76 1.93 1.88 2.05 2.00 2.15 2.28 2.432.31 2.38 2.40 2.55 2.61 2.45 2.46 2.57 2.56 2.49 2.73 2.75a Runs were monitored at 369 nm for the 3,5-Me,py complex, at 403 nm for the 3-CNpy complex, and at 382 nm for the 3-Clpy40zutOH bcomplex.b Solvent compositions are expressed in volume percentages.TABLE 2First-order rate constants, 103K0bs,/s-1, for reactions of [Fe(CN),(3,5-l'~le,py)]~- and [Fe(CX)5(3-CNpy)]3- anions with amol dm-1 and I = 0.20 mol dn1-3 range of incoming ligands (L) in aqueous solution a t 298.6 I<; initial [complex](KNO,) except where otherwise indicateda[Ll -7 I crnol dm-3 CN- SCN- tu 4,4'-bipy im PZ mpz+[Fe(CN) 5( 3, B-Me,py)] 3- 0.024 b 1.32 0.30 1.2, (1.30) 1.3, (1.4,) 1.8, (1.7,) e0.024 2.5 2.50.04 2.10.05 2.40.06 2.00.10 2.4,0.14 2.38 2.1 2.60.15 2.40.20 2.3, 2.0 2.5[Fe(CN) 5(3-CNPY)l 3-0.16 2.4,a tu = Thiourea, im = imidazole, pz = pyrazine, and mpz+ = N-methylpyrazinium.b I = 0.024 mol ~ l r n - ~ (K[NO,]). c From + d[produ ct]/dt. d 1 = niol drn-,.TABLE 3First-order rate constants, Klirn., for the reaction of [Fe(CN),(3,5-Me,py)I3- with cyanide ion ([KCXI = 0.024 mol dm-3) at298.6 K in ranges of binary aqueous mixtures (mole fraction x2 of organic component), and respective excess molarGibbs free energies of mixing, GEVolume per cent of organic co-solvent20 30 40*- - - - - T o m v h-- - r >10103kli,.GE -- GEx 2 s-1 J mol-1 x2 s-1 jGziI . 103ka,. GE 10'him. - Co-solvent x2 s-1 J Y z F X 2 S-1Methanol 0.047 1.36 116 o.ioo 1.45Ethanol 0.033 1.45 107 0.072 1.65t-Butyl alcohol 0.021 1.41 129 0.046 1.60Glycol 0.044 1.28 -101 0.094 1.28Glycerol 0.031 1.19 -84 0.067 1.11Tetrahydrofuran 0.053 1.85The dependence of the limiting rates for cyanide sub-stitution, at relatively high cyanide concentrations, on thenature of the mixed solvent is shown in Table 3.Forsolvent mixtures for which the dependence of the observedfirst-order rate constants, Kobs., on cyanide concentrationwas established over a range of cyanide concentrations(i.e. values in Tables 1 and 2), hli,. values were estimatedfrom plots of l/Kob,. against l/[CN-] (see below). For othersolvent mixtures, Klirn, values in Table 3 are mean rateconstants determined at the maximum cyanide concen-tration, 0.024 mol dm-3. Values for the equilibriumR. P. Mitra, B. K. Sharma, and S. P. Mittal, J . Inovg.Nucbar Chem., 1972,34, 3919.L. Dozsa, I. Szilassy, and M. T. Beck, Magyar Ktfm. Folyciirat,1973, 79, 45.21 1 0.i60 1.49 ~ 284 0.229 1.61 335224 0.117 1.84 349 0.171 2.10 479269 0.076 1.85 419 0.113 2.21 576- 197 0.151 1.19 -288 0.217 1.10 -371- 123 0.110 1.14 -274 0.161 1.05 -367343 0.087 2.28 532 0.129 2.55 728constant for reaction (1) above in series of binary aqueousmixtures are reported in Table 4.DISCUSSIONKinetics of Sahtitution at [Fe(CN),LI3- Anions.-There have been many kinetic studies of substitutionat complexes of the pentacyanoferrate(I1) type,[Fe(CN),L]"-, in aqueous solution.These include :aquation of [Fe(CN),(N0)I2- (ref. 4) and reactions ofthis anion with ammonia, hydroxylamine, and hydra-zine; reaction of [Fe(CN),(ONPh)]3-,6 of [Fe(CN),-D. PavloviC I. Murati, and S. ASperger, J.C.S. Dalton, 1973,6021976 1295( SO3)],-,’ of [ Fe (CN) ,(py ) ] 3-, * and of [Fe( CN) 5(3-CNpy)]3-(ref. 9) with cyanide; and of a series of complexesTABLE 4Dependence of the relative stabilities, K , of [Fe(CN),-(3,5-Me,py)I3- and [Fe(CK),(py)]3- for equation (1) onsolvent composition at 298.6 KSolvent x2 K GE/J mol-lMethanol 0.047 0.98 1160.100 0.92 2110.160 0.81 2840.229 0.80 335Ethanol 0.033 0.89 1070.072 0.84 2240.117 0.77 3490.171 0.70 479t-Butyl alcohol 0.021 0.8 1290.046 0.6 2690.076 0.5 4190.113 0.4 5760.05 0.94 - 1400.08 0.92Tetrahydrofuran 0.05 0.67 340Glycerol 0.03 0.94 - 54[Fe( CN),Lj3-In all these1with the N-methylpyrazinium cati0n.lreactions, the kinetic evidence indicates3 400 800( 1 /[KCN] I / dm3 mol-’Dependence of the first-order rate constant (Aobs.) oncyanide concentration (a) and of the reciprocal of the first-order rate constant on the reciprocal of the cyanide concentra-tion ( b ) for the reaction of [Fe(CN),(3-CNpy)13- with cyanideion in aqueous glycol (40% glycol by volume) a t 298.6 I<FIGURE 1the operation of a limiting S N ~ , or D, mechanism[equations (2) and (3)].klks[Fe(CN),L1]3- -- [Fe(CN),]3- + L1[Fe(CN),]3- + L2 &.[Fe(CN),L2I3-(2)(3)k,k.The operation of the D mechanism for several sub-stituted pyridines as leaving groups has been demon-strated, for aqueous solutions.1 We have examined in0V0 0 0-0A[ L ] /mol dm-3FIGURE 2 Dependence of the first-order rate constant (Aoba.) onthe nature and concentration of the incoming ligand (L) forsubstitution a t [Fe(CN),(3,5-Me,py)I3- in aqueous solution at298.6 K.L = SCN- (A), thiourea (A), CN- (O), pz (V), andmpz+ ( 0 )detail the kinetics of the reaction of [Fe(CN),(3-CNpy)I3-with cyanide in 40% glycol solution, and of the reactionsof this anion and of [Fe(CN),(3,5-Me,py)I3- with arange of nucleophiles in aqueous solution in order totest the operation of the D mechanism here for thesespecific pyridine derivatives in mixed aqueous solutionas well as in aqueous media. This mechanism is con-firmed by the characteristic curve for the dependence ofkobs. on cyanide concentration [Figure l ( a ) ] and thelinear dependence of l/Kobs. on reciprocal cyanideconcentration [Figure l(b)] 7 for the reaction of the3-CNpy complex in 40% glycol. Similarly the kineticpattern in Figure 2, with the characteristic dependenceof kobs. on incoming ligand concentration but independ-ence of Klim.on the nature of the incoming ligand,confirms the operation of a D mechanism for sub-stitution at [Fe(CN),(3,5-Me,py)l3-. The slightly higher2. BradiC, D. PavloviC, I. Murati, and S. ASperger, J.C.S.Dadton, 1974, 344.8 B. Jezowska-Trzebiatowska, A. Keller, and J. Ziolkowski.Bull. Acad. polon. Sci., Sky. Sci. chim., 1972.20, 649; A. D. Jamesand R. S. Murray, J.C.S. Dalton, 1975, 1530.2. BradiC, M. PribaniC, and S. ASperger, J.C.S. Dalton, 1976,3531296 J.C.S. Daltonkljm. value for the reaction of this complex with theN-methylpyrazinium cation is consistent with a similarobservation for reactions of the 3-CNpy c~mplex.~ Inthis case the irregularity may be ascribed to the conse-quences of electrostatic interaction between the nega-tively charged complex or intermediate and the positively- 2.0/-1.0 2 .0 3.0YGrunwald-Winstein plot for the reaction of [Fe(CN),-(3,5-Me,py)I3- with cyanide ion. Co-solvents: (A), MeOH;(O), EtOH; (V), ButOH; (A), glycol; (w), waterFIGURE 3charged incoming group. The very different behaviourof thiocyanate as entering group (Figure 2) indicatesthat it competes only feebly with the leaving 3,5-Me2py;indeed these reactions go to equilibrium rather than tocompletion.Rates of substitution at these [Fe(CN),LI3- anions areremarkably insensitive to the composition of the mixedaqueous solvent. In this they resemble the complex[Fe(CN),(bipy),] lo (bipy = 2,2’-bipyridyl).The de-pendence of reactivity on solvent composition is illus-trated, for the 3,5-Me2py complex, by the form of aGrunwald-Winstein mY plot l1 in Figure 3. Thecorrelation lines corresponding to the three organicco-solvents do not quite coincide; such behaviour fordissociative inorganic reactions has been established forcobalt (111) complexes.l2 The Grunwald-Winstein mvalue for aquation of cobalt(II1) chloride complexes isca. 0.3,12 but for substitution a t [Fe(CN),(3,5-Me2py)l3-is ca. -0.l.t Negative m values have previously beenreported for inorganic substitution reactions, for exampleaquation of the hexabromorhenate(1v) anion, wherem = -0.55.13 Both these negative m values areassociated with reactions in which one of the separatingmoieties is a negatively charged species containingseveral hydrophilic groups, viz.cyanide or bromide1igands.S If indeed it is the solvation of the [Fe(CN)J3-moiety rather than that of the 3,5-Me2py leaving groupwhich determines the solvent variation of reactivity inour system, then it is easy to understand how the ratioof substitution rates in say water and in 40% aqueoust This value was estimated from the results in the ‘ typicallyaqueous ’ (t.a.) solvent mixtures, ignoring results in aqueousglycol (‘ typically non-aqueous negative ’, t.n.a.n.) .Interestingly, rates of substitution at [Co(CN),(OH,)I2-, e.g.by [NJ-, also increase on going from aqueous solution to aqueousethanol or to aqueous t-butyl alcohol.1410 J. Burgess, Chem.Comm., 1969, 1422; J.C.S. Dalton, 1972,203.11 E. Grunwald and S. Winstein, J . Amev. Cltem. SOG., 1948,70,846; I.=. M. Kosower, ‘ Introduction to Physical Organic Chem-istry,’ Wiley, New York, 1968, chs. 2.6 and 2.7.t-butyl alcohol is very similar for the leaving groups3,5-Me2py, 3-CNpy, and 3-Clpy (Table 1) despite theexpected differences in solvation of these three species,as evidenced by their very different solubilities in water.We are currently interested in examining whetherreactivities in substitution reactions in binary aqueousmixtures can be correlated with excess Gibbs freeenergies of mixing of the respective mixtures. We havehad some success in such an analysis of aquation of thetris(5-nitro-l,lO-phenanthroline)iron(11) cation, [Fe(5-NO2phen),l2+, and of the solvolysis of t-butyl ch10ride.l~Binary aqueous mixtures can be classified in threegroups according to their excess molar thermodynamicfunctions of mixing.16 These are the ‘ typicallyaqueous’ (La.) mixtures where GE is positive and isdominated by its entropy component over the enthalpycomponent (1TSEl > IHEl), ‘ typically non-aqueous posi-tive ’ (t.n.a.p.) with positive GE and JHEJ > ITSEI, and‘ typically non-aqueous negative ’ (t .n.a.n.) with negativeGE but again IHEI > ITS”I.17J8 Satisfactory results forsubstitution at the [Fe(CN),LI3- complexes have beenobtained for a range of t.a.mixtures and for t.n.a.n.mixtures containing glycol or glycerol. In Figure 4 ourkinetic results, in terms of the change in the Gibbs freeenergy of activation (6,AGZ) derived from transition-state theory treatment of the experimental rate con-stants, are plotted against the respective GE values.‘Oo0 1FIGURE 4 Relation between 6,AG: and GE for limiting rates ofsubstitution at [Fe(CN),(3,5-Me,py)I3- in binary aqueousmixtures.Co-solvents: ( b ) , MeOH; (O), EtOH; ( V ) ,ButOH; (a), thf (t.a. mixtures); (m), glycerol; and (A),glycol (t.n.a.n. mixtures)The latter were calculated from published thermo-dynamic data (for detailed references see ref. 15) andl2 G. Thomas and L. A. P. Kane-Maguire, J.C.S. Dalton, 1974,1688; J. Burgess and 31. G. Price, J . Chem. SOC. ( A ) , 1971, 3108.l3 J. Burgess, R. D. Peacock, and A. M. Petric, J.C.S. Dalton,1973, 902.l4 M.J. Blandamer, J. Burgess, S. P. Edwards, and G. P. Gisby,unpublished work.15 M. J. Blandamer, J. Burgess, and R. I. Haines, J.C.S. Dalton,1976, 385.16 F. Franks, in ‘ Hydrogen-bonded Solvent Systems,’ eds. A.Covington and P. Jones, Taylor and Francis, London, 1968, p. 31.l7 F. Franks in ‘ Water-A Comprehensive Treatise,’ ed. F.Franks, Plenum Press, New York, 1973, ch. 1.M. J. Blandamer and J. Burgess, Chem. SOG. Rev., 1975,Q. 551976 1297interpolated l5 using the Guggenheim-Scatchard 19*20equation for expressing the dependence of GE on molefraction of the aqueous mixtures. Thus 8,AGt re-presents the change in AGt on going from a solution inwater to a mixture where the mole fraction of co-solventis x,, i.e. 6,AGX = AGS(x,) - AGt (x, = 0).The functionGE expresses the extent to which the molar Gibbsfunction of the mixture differs from that of the corre-sponding ideal mixture.There is a satisfactory correlation between the kineticand thermodynamic parameters. In the present case,unlike that of aquation of [Fe(5-N0,phen)3]2f, but likethat of t-butyl chloride solvolysis, results in aqueousglycol can be accommodated within the general 6,AGt-GE correlation. The close similarity of the 8,AGt-GEpatterns for these three dissociative substitutions isgratifying.Relative Stabilities of [Fe(CN),LI3- Com$lexes.-Thevalues of K in Table 4 represent the relative stabilities of[ Fe( CN),(3,5-Me2py)]3- and [Fe( CN),(py)I3- anions inthe respective solvent mixtures.Values for the corre-sponding differences in standard-state Gibbs freeenergies (AGe = --RTlnK) are plotted against therespective GE values for the solvent mixtures in Figure 5.I GEfJ mol-’”\ \ a- 4 000FIGURE 5 Relation between AG*, the Gibbs free energy for theequilibrium between the pyridine and 3,5-Me2py complexes[Fe(CN),LI3- [equation (1) in text], and GE in binary aqueousrmxtures. Symbols as in Figure 4As for the analogous kinetic plot (Figure a), there is agood correlation of AG* with GE for t.a. mixtures.But the AGe values for the t.n.a.n. solvent glycerol donot fall on the correlation line for the t.a. mixtures. Aswith similar plots for aquation of [Fe(5-N02phen),]2+and for solvolysis of t-butyl chloride, a consistentpattern is obtained for t.a.mixtures, but a complexpattern emerges for t.n.a.n. mixtures.The above kinetic and thermodynamic aspects ofreactions of [Fe(CN),LI3- anions are related, cf. equations(1)-(3). Trends in stability constants should bereflected in the discriminatory tendencies of the transient[Fe(CN),I3- intermediate, in other words in the curvatureof plots of kobs. against the concentration of the incomingnucleophile of the type illustrated in Figures l(a) and 2.In practice, the curvature of such plots for reactions of[Fe(CN),LI3- anions with cyanide, pyrazine, or N-lS G. Scatchard, Chem. Rev., 1949, 44, 7.2o E. A. Guggenheim, Trans. Faradav SOC.. 1937. 33. 151.methylpyrazinium is too sharp for accurate inter-comparisons between different solvent mixtures. Atr 12- Q& 0-the other extreme, reaction with thiocyanate is un-suitable as this ligand does not compete sufficientlystrongly with substituted pyridines for reaction toproceed far enough towards completion.Solvatochromism of the [Fe(CN),(mpz)]2- A Tiion.-Thecharge-transfer (c.t.) spectrum of the product of thereaction of [Fe(CN),(X-py)j3- anions with the N-methyl-pyrazinium cation, anion (I), varies markedly withsolvent composition. The anion (I) bears a resemblanceto the betaine (11), whose solvent-sensitive intra-molecular c.t.band is the basis for Reichardt’s scale ofsolvent ET values,21 in that intramolecular charge transferI 5ot / RI t 814 000 14 530 15 000V/crn-IFIGURE 6 Dependence of the frequency of maximum absorptjon(v) for the charge-transfer band of the [Fe(CN),(mpz)l2- anionon solvent ET values in binary aqueous mixtures.Co-solvents:(A), MeOH; (0)’ EtOH; (0)’ Pr’OH; (0)’ acetone; (a)’dioxan; (A)’ wateris possible across an aromatic ring to a quaternarynitrogen. The c.t. spectrum of (I) might therefore be21 C. Reichardt, ‘ Losungsmitteleffekte in der OrganischenChemie.’ Verlag Chemie. Weinheim. 19681298useful for providing a comparable scale of empiricalsolvent-polarity values of direct relevance to inorganicsolutes .The sodium salt of anion (I) is soluble in many binaryaqueous mixtures, but insoluble in all the organicsolvents tried. The frequencies of maximum absorptionfor the c.t. band of the anion are plotted against therespective solvent Er values in Figure 6. The slightcurvature of the isopropanol and dioxan plots is con-sistent with preferential solvation by water in aqueousmixtures containing these co-solvents {cf. [Fe(CN),-(phen)J, ref. 22). The tetraphenylarsonium salt of (I)is soluble in many organic solvents. The wavelength ofmaximum absorption varies between 676 nm in aceto-nitriIe and 721 nm in tetrahydrofuran (655 nm inJ.C.S. Daltonwater). There is no correlation with solvent Er values,which is consistent with the very different solvationcharacteristics for (I) and (11), but contrasts with thegood correlation reported for such complexes as [Fe-We are grateful to Mi-s. Christine Stokes for carrying outpreliminary equilibrium studies, to Dr. E. F. G. Heringtonfor helpful discussions, and to the Royal Society for agrant-in-aid for the purchase of the spectrophotometer.[5/1215 Received, 23rd June, 1976322 J. Burgess and S. F. N. Morton, J.C.S. Dalton, 1972, 1712.23 J. Bjerrum, A. W. Adamson, and 0. Bostrup, Acta Chem.Scand., 1956, 10, 329; J . Burgess, Spectp.oclzinz. Acta, 1970, A%,1309, 1957.24 J. Burgess, J . Ovganometallic Chem., 1969, 19, 218.(CN),(biPY)J 23 and cw(co),(biPY)l.2