摘要:

1975 845Reduction of Tris(2,2'-bipyridyl) and Tris(1 ,lo-phenanthroline) Complexesof Iron(iii) and Osmium(iii) by Hydroxide IonBy Gwyneth Nord,* Chemistry Department I (Inorganic Chemistry), H. C. Brsted Institute, Universitetsparken 5,D K2100 Copen hagen 0, DenmarkOle Wernberg, Chemistry Department, University of Odense, D K5000 Odense, DenmarkReaction (i) (L = 1.1 0-phenanthroline, 2.2'-bipyridyl, and a series of methyl-substituted derivatives of these4 [M(LL)J3+ + 4 OH-+ 4 [M(LL)J2+ + 0 2 + 2 H2O ( 0ligands; M = Fe and 0 s ) has been studied in aqueous (1 M) NaCI-NaOH from 6.5 to 35 'C. The rate law through-out is that reported earlier for the unsubstituted iron complexes, i.e. d[Fe(LL)33+]/dr = k[Fe(LL)3S+] [OH-].Phenanthroline complexes react faster than bipyridyl complexes, Fen' complexes faster than the analogous Osmcomplexes. A linear free-energy relation between the rates and standard free-energy changes of the metal couples,analogous to that known for other reducing agents, holds only for each metal and each ligand type.It is con-firmed that the enthalpies of activation are less than those calculated for the standard enthalpy change of reaction(ii). From a consideration of the stopped-flow, potentiometric, and n.m.r. data, together with those from pulse-[M(LL),I3+ + OH- __t [M(LL),I2+ + *OH (ii)radiolysis and chemiluminescence studies for these and similar reactions, we conclude that the first-formed productis a highly reactive precursor complex and formulate this as [M(LL)2(R'-OH)]2+ where R'-OH is probably thepseudo-base of a radical formed by addition of an electron to a ligand which approaches the structure of a quater-nary-nitrogen cation.IN an earlier publication1 we showed that the rate ofreduction of the tris(2,2'-bipyridyl) and tris( 1,lO-phenanthroline) complexes of iron(Ir1) by hydroxide ionwas first order in complex and hydroxide-ion concen-trations. t We presented a reaction scheme whichaccorded with the suggestions of earlier workers2s3 inthat it also accounted for the final products (0, and[M(LL)JZ+}. We pointed out, however, that our datafor [Fe(bipy)J3+ over a range of temperature were accur-ate enough for us to show that the enthalpy of activationAH$ was less than that of AHe for the previously postu-lated rate-determining step k,, thus necessitating an[Fe(bipy)J3+ + OH- &= [Fe(bipy)J2+ + 'OH (i)intermediate [Fe(bipy),OHI2+ as the first reactionproduct.As rates of reduction of these reactant metalcomplexes (and of oxidation of the corresponding[M(LL),I2+ complexes (31 = Fe and Ru)}, together withvariations in rate with substitution in the ligands, havebeen much used to test theories of outer-sphere electron-transfer mechanisms, any evidence for intermediate-complex formation is of obvious i n t e r e ~ t . ~ We have nowextended our studies to complexes of FeIII and 0s1I1with a series of such substituted ligands and have alsoaccurately measured the temperature dependence of thereduction rate of [Fe(phen),I3+.The standard electrode potentials for the [M(LL)&3+-[M(LL)J2+ couples, used by earlier workers to constructlinear freeenergy-rate relations, have previously onlybeen measured in acid solutions and have been said to beaciddependent.We have measured these for M = FeIIIand LL = phen and bipy in alkali. We have also in-? The bidentate ligands 2,2'-bipyridine and 1,lO-phenanthrolineare referred t o below as bipy and phen, respectively, and collec-tively as LL.$ 1 ~ = 1 mol dm-3.k - lG. Nord and 0. Wernberg, J.C.S. Dalton, 1972, 866.N. Sutin, unpublished work quoted in A. A. Green, J. 0.Edwards, and P. Jones, Inovg. Chem., 1966,5, 1858.vestigated the lH n.m.r. spectra of the reactant [Fe-(phen)J3+ and product [Fe(phen),I2+ in D,O-NaODsolutions as another approach to the determination of themode of attack of OH- (OD-) on the reactant complex.EXPERIMENTALAll reagents were either specially purified or were ofAnalaR grade.Water doubly distilled from quartz glasswas used throughout. Osmium tetraoxide was obtainedfrom Drijf hout, Amsterdam, the ligands from G. F. Smith.and Co., U.S.A. The complexes were prepared by pub-lished methods. The absorption spectra of the 3+ and 2+complexes were in excellent agreement with those given inthe literature.Kinetic measurements were made with a Cary 14 record-ing spectrophotometer and with a stopped-flow apparatus.The latter was the apparatus used in ref. 1 b u t was auto-mated by attachment of a Biomation 802 TransientRecorder together with a Facit 4070 tape puncher.Allsolutions were 1 . 0 0 ~ in NaCl or NaNO, (see Table) andrates of reaction were all measured under pseudo-first-orderconditions, [OH-] [M(LL),3+].$N.m.r. measurements were made by Dr. Jonas Pedersenof Chemistry Department V of this institute. NaOD InD,O was added to a solution of [Fe(phen),13+ in D,O. Thereduction reaction occurred rapidly (see above) and , beforemeasurement of the product, Na,[S,O,] was added toremove 0,. This is necessary to stabilise the productsolution for the n.m.r. measurements, since in the presence4 [Fe(LL),I3+ + 8 OH- + 0, + 2 H,O ---+of 0, the overall reaction (ii) is known to take place.526 Alloperations were carried out in a glass apparatus under N,.Potentiometric titrations were carried out with a pH-stat(Radiometer) fitted with a gold electrode and a standardcalomel electrode.The apparatus was thermostatted(25 f 1 "C). Parallel experiments with solutions of theM. Anbar and I. Pecht, Trans. Faraday SOL, 1968, 4, 744.J . N. Braddock and T. J. Meyer, J . Amer. Chem. SOC., 1973,G. Nord and T. Pizzino, Chenz. Comm., 1970, 1633.G. Nord, Acta Chem. Scand., 1973, 27, 743.4 Fe(OH), + 1 2 L L (ii)95, 3168846 J.C.S. Daltonsame composition were made using an 0, electrode (Radio-meter type E 5064). The apparatus was filled with eitherN, or H, and solutions were l~ in NaCI.RESULTSRate Constants .-Pseudo-firs t-order rate constants werecalculated from the rate of increase of optical density due tothe [M(LL),I2+ product a t the wavelength of maximumabsorbance in the visible range.These all increased linearlywith (excess) [OH-] and were converted into second-orderrate constants which were independent of the initialsystems in the presence of added ligand.7 These are freefrom the competing product reaction (ii) (see above) andpossibly also of any dissociation of the reactant [M(LL)J3'.The potential of the gold electrode in the reaction mixturewas automatically recorded and from this curve the poten-tial at the time corresponding to the first half-life of thereaction was obtained. It follows from the stoicheiometrythat the concentration of [Fe(LL),I3+ at this point is 50%and that of [Fe(LL),I2+ is 40.5% of the initial [Fe(LL)J3+concentration. Thus the potential measured is E = E* +(RTIF) In (50/40.5) = E* + 5 mV.The E* values of theRate constants and activation parameters *AG: (26°C) - AH: A S In (k/dm3 mol-1 s-1)Ligand, LL kJ mol-l kJ mol-l J K-l mol-l at 25 "C(a) Reactants [Fe[LL)Js+ and OH--41.8 (2.8)5, 6-Me2phen 61.2 (0-3) 38.0 (0.6) - 78.0 (2.0)4,7-Me2phen 64-22 (0.08) 43.0 (0.0) -71.6 (2-0)- 17.0 (2.8)phen 67.92 (0.16) 46.6 (0.8)5-Mephen 60.00 (0.4)- 12 (2) 63 (8)62.4 (0.8)3,4,7,8-Meghen 66 (2)bipy a 66.24,4'-Me2bipy 71.8 (0.2) 54 (6) -60 (20)(b) Reactants [OS[LL)~J~+ and OH--336 (16)-80 (16)-13 (20)-46 (6)phen 60.6 (0-2) 50 (5)386,.!(1.6)6-Mephen 61-8 (0.2)bipy b 69.2 (0.3)4,4'-Me,bipy 76.9 (0.3) 64-0 (1.6)[OH-] = 0-006--0.076 and [M(LL),"] = 1 X W3-l x ~VM.From ref.1. Also in 1-00~-NaNO,. ~ H , o / ~ D , o = 2.26 at 26 "C.6.042 (0.06)5.26 (0.16)4-76 (0.14)3-56 (0.04)2-83 (1.0)2.760.60 (0.12)5.06 (0.10)4-63 (0.10)1.64 (0.12)- 1-14 ' (0.12)* Standard deviations (f) are given in parentheses.[M(LL)3a"J (see Table). The activation parameters are alsogiven in the Table. These were calculated from plots ofIn k / T against I/T using the expressions from reaction ratetheory, k = (kBT'/h)e-A*$/RT and AGS = Am - T A s .Under the experimental conditions used (see Table), sidereactions were negligible. This was not so for more con-centrated solutions of complex and more dilute solutions ofbase, cf. the potentiometric measurements below.N.M.R. Spectra.-The spectrum of the product [Fe-(phen),]2+ produced by treating OD' with [Fe(phen)d3+ inD20 was identical with that of a sampIe dissolved in D,Oand prepared from iron(I1) sulphate and ligand in H20, andagreed with the published spectrum.7 The relative peakheights were easily measured.No D-H exchange thusoccurs during the reduction of [Fe(phen),J3+ by OD-.Potentimetric Studies.-Potential changes during theautomatic addition of NaOH to maintain a constant con-centration (1 x 10-4-2.3 x 10-*M) in rather concentratedsolutions of [Fe(phen),],+ and [Fe(bipy),]3+ (1 x2 x l o - 3 ~ ) were used to determine Ee values for the2 f-3 f couples. The pH-stat curves were reproducibleand the overall stoicheiometry was as in (iii). The ligand[Fe(LL).J3+ + 1.38 OH- ---t 0.81 [Fe(LL),I2+ +and 0, concentrations are given in parentheses because thefree ligand was not measured and the oxygen concentrationmeasured with the oxygen electrode was only about one halfthat predicted.We hope to investigate whether or not thisdiscrepancy is real by studying the corresponding Osmt We showed in ref. 1 that addition of free ligand did not affectthe rate of reduction, but did not determine its effect on thestoicheiometry.$ We thank J. C. Sullivan of the Argonne National Laboratory,U S A . , for kindly bringing this reference to our notice.0.19 Fe(OH), + (0.57 LL + 0.20 0,) (iii)couples, veysus the standard hydrogen electrode but in1.00M-NaC1, were as follows:fFe(phen) J3+-[ Fe (p hen) J ,*E*/mV = 1008, 1 026, 993, 1 040, 1 030, 1016, and 1 026(average 1019 f 6)[Fe( bipy)J3+-[Fe (bipy),] ?+E*/mV = 976, 970, 987, 988, and 990(average 982 f 4)DISCUSSIONSince the publication of our first paper1 carefulcalculations of the energetics of reaction (i) have givenAH* = 102 -& 15 k J mol-l for bipy and 109 & 16 k J mol-lfor the corresponding reaction of phen.For the former,AHJ extrapolated (ionic strength&) from our data1is 71 &- 5 kJ mol-l giving AHe - AHo$ = 38 & 20 kJmol-l, which is probably rather less than our estimate(50 kJ mol-l) at I = IM. For the phen reaction (seeTable), using the same correction as that found for thebipy system, AHe - AHo$ = 58 & 20 k J mol-l. Alter-natively we note that the E* values for the two metalcomplex couples are in agreement with literature valuesfor acid solutions 9~10 and differ only by ca.0.4 kJ mol-l.Both couples are isentropic l1, $ so that neither the differ-ences in AH$ nor in AS$, 15.9 & 1-5 kJ mol-l and 24.8 &7 R. H. Prince and R. D. Miller, J. Chem. SOC., 1965, 470;J . Chem. SOC. ( A ) , 1966, 3185. * P. B. Monkhouse, Chemistry Project Report, Queen MaryCollege, University of London, 1973.0 P. George, G. Hanania, and D, H. Irvine, J . Chem. SOC.,1969, 2648.l o M. H. Ford-Smith and N. Sutin, J . Amer. Chem. SOC., 1961,88, 1830.l1 J. Lin and W. G. Brack, Canad. J. Chem., 1965, 48, 7661975 8475.6 J K-1 mol-l, respectively, can be attributed to changesin AHe or AS* for the overall reaction (i).Another approach is to assume that the thermo-chemical data for the reaction OH =+ *OH + e- arein error * and that the rate-determining step is in fact areversible electron transfer with the product *OHrapidly reacting further.In this context, pulse-radiolysisresults l2 for the back reaction, [Fe(phen),I2+ + 'OH 4Products, are relevant. It has been found that aproduct is formed at a rate close to the diffusion-con-trolled limit (k = 9 x loB dm3 mol-l s-l). This, unlikethe intensely absorbing radical product of the free ligandwith *OH, which was also measured, is stable in neutralsolution (during the time scale of the experiments) butboth decays in alkaline solution and has an intensespectrum very different from that of [Fe(phen),]3+. Theauthors tentatively proposed a reaction involving attackon a ligand which remains bonded to the complex[equation (iv)! and which would decay as in equationskFt(phen),I2 - I *OH + [Fe(phen),OHI2 (iv)(v) arid (vi).There is thus additional evidence for tlie[Fe(phen)J3+ + OH- (v): Fe(phen),R-]27 + HzO (vi)[ Fe ( phen),OHI2 + <=non-occurrence of a simple reversible electron-transfermechanism.Figure 1 illustrates that the dependence of rateon ligandtype persists with complexes of the substituted ligandsand also that it is larger than that found at 25 "C for the[Fe(LL),]3+-Fe11 reaction which has been used to test theRlarcus theory of outer-sphere electron tran~fer.~ Inboth cases changes in rate from methyl substitution ineither phen or bipy parallel changes in E* for themetalcouples. It is of interest that different activationenergies, which are small but mgative, have recentlybeen reported for tlie reduction of [Ru(phen),]3+ andiRu(bipy),]3+ by [Fe(OHz),J2+ and that the similarity ofthe activation parameters to those for [Fe(phen)A3+ with[Fe(OH,),]2+ is considered by the authors as evidence fora common mechanism.In both series it was necessaryto invoke reaction by a ' non-Marcus ' path.On replacing the inetal centre FeIII by OsIII rates ofreduction consistently decreased (see Figure 2). ForRuIII complexes the rates are qualitatively known to bevery fast l3 so that there is an overall increase in rate withincrease in E* for the three metal couples. Of particularinterest here are the results of Lytle and Hercules l4 whofound that addition of a concentrated solution of a strongbase to concentrated solutions of [Ru(bipy),I3+, formedin sit% by PbO, oxidation of [Ru(bipy)J2+, gave as anintermediate the same excited form of [Ru(bipy),]2+ as* Using AG* €or reaction (i) to obtain K , ( = k , / k - , ) , substitutionof the measured k ( = 2 k,) would of course give an impossiblylarge value for k-,.12 P.Pagsberg and K. E. Siekierska, Progress Report on theRedox Kinetics of Tris(o-phenanthro1ine)iron Ions Studied byPulse Radiolysis, Symposium of Nordisk Forening for StrAlings-forsag og StrHlingsindustri, Otaniemi, Finland, 1973.that produced photochemically in [Ru(bipy)J2' solu-tions. These authors follow Fleischauer and Fleisch-2.5 -2.0 -I I I4 5 6log(kFe1~/drn3 mol-'s-l)FIGURE 1 Plot of log,, (k/dm3 mol-l s-l) for the reactions(Fe(LL)J3+ + OH- against log,, (k/dm3 mol-l s-l) for thereactions [Fe(LL)J3+ + Fe2+ a t 25 "C (ref.12 and this work inO.~M-H,SO,) : LL = phen (a), 6-Mephen ( b ) , 6,6-Me2phen (c).4,7-Me,phen ( d ) , 3,4,7,8-Me4phen ( e ) , bipy (n, and 4,4'-Me,bipy(g)7%T 5560 65 70 75 80AG+IodLLl,33+/ kJ rnol''FIGURE 2 Plot of AGt for the reactions [Fe(LL)J3+ + OH-For keyauer l5 and suggest that an electron added to the x-systemof the ligands is eventually transmitted to the metal-ion88,4745.199.against AG: for rhe reactions [0s(LL),l3+ + OH-.see Figure 1la J . D. Miller, Ph.D. Thesis, Cambridge University, 1966.l4 F. E.Lytle and D. M. Hercules, J . Amer. Chem. Soc., 1966,l5 P. D. Fleischauer and P. Fleischauer, Chem. Rev., 1970, 70848 J.C.S. Daltonation of the free electron in the radical cation. Similarly,if the Ee sequence for the metal complex couples,Ru > Fe > Os, reflects decrease in the enthalpy of form-ation of [M(LL)J3+ then this may be correlated with thedegree of double bonding in both reactant and inter-mediate. The depiction of step (1) as rate determiningtogether with the overall stoicheiometry gives the rateconstant for this step as half the observed rate con-stants given in the Table, while the postulated pseudo-base formation could explain both the rate law and theactivation parameters. Finally, step (2), the rapid reac-tion with further OH-, can explain the observed effectof OH- on the product of the [Fe(phen)J2+ + ’OHreaction l2 and also finds analogy with other reductionsof [M(LL)J3+ complexes.Thus the rate law for reduc-tion of [Fe(bipy)J3+ (ref. 18) and of [Os(bipy)J3+ (ref. 19)by X- (X = SCN and I) is as in equation (ix). Weinterpret the second term as due to formation of X2- andhave therefore tentatively included H,02- as a possibleproduct in our postulated reaction scheme. The fastcentre, the observed chemiluminescence corresponding tothe intramolecular electron transfer 3d-x*(t22x*)+ld-d(tW6) + hv [cf. stage (3a) in the reaction sequencegiven below]. It has previously been pointed out thatconceptually the chemiluminescent reactions of[Ru(LL),I2+ are very similar to some oxidation-reductionreactions of CoIII complexes where ligand-reducedradical-anion intermediates have been postulated. How-ever, for the reduction of [M(LL),I3+ complexes by[Fe (OH,) J ,+ the exot hermicity associated with theformation of a precursor complex remains to be ex-~ l a i n e d .~We derived our scheme for the reaction sequenceduring OH- reduction by first considering two typicalreactions of quaternary cations formed from the ligandsand similar molecules which we use as models * for thecomplexes [M(LL)J3+. The reactions are reversible up-take of one electron, A2+ + e- + A+, e.g. a s in equation(vii),16 and reversible pseudo-base formation (B+ +OH- # BOH, BOH + H+ === B+ + H20) which inthe example chosen is in the stopped-flow time range[equation (viii)] .179 t We incorporated the two types ofreaction in the following sequence using as illustrationsone example of the resonance structures which wouldcontribute to the electronic structures of the reactantcomplex, the postulated intermediate, and the productcomplex.The depiction of the intermediate as a radical canexplain, in terms of increased conjugation, why reactionsof the phen complexes are faster than those of theanalogous bipy species, In accordance with this,changes in redox potentials for the couples A2+-At withchanges in ligand structure have been correlated l6 withthe degree of resonance stabilisation due to the delocalis-* We are grateful t o Professor Gerold Schwarzenbach ofZurich for suggesting this approach.t We thank Professor Martin Ettlinger of Chemistry Depart-ment I1 of this institute for providing this reference.IS E. Steckhan and T. Kuwana, Ber. Bunsengesellschaft. Chem.,1974, 78, 263.l7 J. W. Bunting and W. G. Meathrel, Canad. J . Chem., 1973,51, 1965.l8 W. K. Wilmarth, J. E. Byrd, and Jia-yong Chen, PPOC. 15thInternat. Conf. Co-ordination Chem., Moscow, 1973, p. 46.G. Nord and Britta Pedersen, unpublished work.(vii)c-cH2 H2further stages involving peroxide radicals, (3b), assuggested in our first publication,l now account for the13+( 1 1 1 + O H - (0,7Reactant complex 1 I n t e r m e d i a t e( d 5 )( 2 )0 2 3.-1-2 +Q-Q‘M’Stable product complex ( d 6 )(a), Rate-determining step A,; ( b ) , + OH- (kp); ( c ) , further re-action with reactant complex according to ref.1 ; (4,t,5X* -> t,1975final product being oxygen. We were unable to detectradicals in the reacting solutions under our experimentalconditions, but unidentified radicals have been detectedby Gusenius e' under undefined mixing conditions.= klcx- + K,(cx--)~ (ix) -d[M(LL):+] 1dt C[M( LL) JS+Reversible pH-dependent spectral changes have beenreported for other bipy complexes. Murray and Nord 21found that addition of sodium hydroxide to a red solutionof [Cr(bipy)JZ+ produced a blue solution, the spectrumof which corresponded to that reported for [Cr(bipy)J+.This fast colour change is reversed on immediate additionof acid without overall reduction of [Cr(bipy),I2+.Inlow-valent complexes such as those of CrI the presenceof the ligand as an anion seems likely. Also, Gillard 22has suggested that the reversible changes with pH in theU.V. spectrum of [Pt(LL)d2+ are due to OH addition at acarbon atom of the ligand. We also considered thepossibility that a seven-co-ordinate intermediate isformed in the present reactions but note first that suchcomplexes have only been reported for spin-free Fernand secondly that the increased rate of reduction of thephen complexes over those of bipy then becomes in-explicable.Summarising, we find that the rate law and activationenthalpies for the reduction of [M(LL)J3+ with OH-2o E. M. Gusenius, Diss. A h . , 1963, 24, 1386.21 R. Murray and G. Nord (Waind), Pvoc. 7th Internat. Conf,Co-mdinufdon Chem., Stockholm, 1962, p. 309.accord with the general equation (x). The rather largechanges in AS% are still unexplained. We find that[M(LL)J3+ + OH-kik-1 C X[M(LL),0Hl2+ --% Products (x)k*succeeding fast steps involving peroxide radicals may beapplicable as suggested earlier,l although we failed topoint out previously that if the products include thehydroxyl radical then the equilibrium 'OH + OH- +0- + H,O must be considered as the pK for *OH is11.9.% For the decomposition of the intermediate, ifX = OH- as in the reaction sequence given above, theobserved rate constants become equal to 2k,, while inthe other limiting case of X = H,O and a rapid pre-equilibrium, k = 2K1k2 where Kl = k,/k-,. The D-Hand 0, isotope 3 effects and also the salt effect accordwith either of these limiting cases.We thank the Carlsberg Foundation for making availablethe stopped-flow apparatus, Statens naturvidenskabeligeForskningsrLd for the award of a maintenance grant (to0. W.), Dr. R. H. Prince for kindly sending us photocopiedsections of ref. 13, since such Ph.D. theses, although accept-able as references in J.C.S., are not directly availableoutside the United Kingdom.[4/1079 Received, 3rd June, 1974322 R. D. Gillard and J. R. Lyons, J.C.S. Chem. Comm., 1973,*3 J. Rabani and S. Nielsen, J . Phys. Chem., 1964, 68, 1169.686

ISSN:1477-9226    

DOI:10.1039/DT9750000845

出版商:RSC    

年代:1975

数据来源: RSC