1319
J. Am. Chem. SOC.1984, 106, 1319-1323
Inner- and Outer-Sphere Pathways in the Reaction between (Pyrazine)pentaammineruthenium( 111) and (Ethylenediaminetetraacetato)cobaltate( 11): Direct Proof That the Successor Binuclear Complex Represents a Dead End a Grant C. Seamadb and Albert Haim* Contribution from the Department of Chemistry, State University of New York, Stony Brook, New York 1 1 794. Received July 22, 1983
Abstract: The reaction between [RU(NH,)~(PZ)]~+ (pz = pyrazine) and [Co(EDTA)]" (EDTA' = ethylenediaminetetraacetate) proceeds via parallel outer-sphere and inner-sphere pathways. The outer-sphere pathway yields [ R U ( N H ~ ) ~ ( ~ and Z)]~+ [Co(EDTA)]- directly with a rate constant of 1.0 X lo3 M-' s-' at 25 OC and ionic strength 0.10 M. The inner-sphere pathway with a rate constant of 2.5 X lo3 M-' s-' at produces the dead-end binuclear intermediate [(NH,),RU~~(~Z)CO~~'(EDTA)]~ 25 O C and ionic strength 0.10 M. The intermediate disappears by intramolecular electron transfer (rate constant of 16.9 s-' at 25 OC and ionic strength 0.10 M) followed by dissociation to produce [ R U ( N H ~ ) ~ ( P Zand ) ] ~ [Co(EDTA)I2-, + which in turn react by the outer-sphere pathway described above. The rate constant for the outer-sphere reaction between [RU(NHJ~(PZ)]~+ and [Co(EDTA)]- has been measured as 3.2 M-I s-l at 25 " C and ionic strength 0.10 M. A detailed mechanistic discussion of the title reaction is presented, and comparisons are made with the analogous reactions where pyrazine is replaced by 4,4'-bipyridine or the cobalt center is replaced by ruthenium.
spectroscopically after reduction with ascorbic acid at pH 5 . The obThe detection of binuclear complexes represents a key obserserved maximum at 474 nm with a molar absorbance of 13 500 M-' cm-' vation in the elucidation of the mechanisms of oxidation reduction is in excellent agreement with the previously reported values 472 nm and reactions between transition-metal complexes.2 Often, the bi13 300 M-' cm-' for [ R U ( N H , ) ~ ( ~ Z ) ] ~ A+stock . ' ~ solution of cobalt(I1) nuclear complexes, whether precursor3 or successor,4 are transient chloride was prepared by dissolution of the reagent grade material in intermediates along the pathways that connect reactants to purified water. The concentration of cobalt in the solution was deterproducts. However, in some instances, e.g., the !Fe(CN),I3-mined spectrophotometrically as the pyridinedicarboxylate complex." [Co(EDTA)I2- ~ y s t e mthe , ~ binuclear complex detected as the Solutions of [Co(EDTA)I2-were prepared when needed by adding a 10% major product immediately after mixing the reactants6,' represents excess of Na2H2EDTAto the cobalt(I1) solution. Reagent grade lithium perchlorate was recrystallized from water. The sample of K[Co(EDTA)] a dead end. In other words, the binuclear species is not generated was from our previous work.5 The water used in all experiments was as an intermediate along the pathway that leads from mononuclear purified as described previously.s All other chemicals were of reagent reactants to mononuclear products but is formed in an unprograde and used as received. ductive side reaction. In previous work,8 we sought to characterize Kinetic Measurements. The reaction between [RU(NH~)~(PZ)]'+ and the binuclear complex that was anticipated to be produced as a [Co(EDTA)I2- was studied in a Durrum-110 stopped-flow apparatus. dead-end intermediate in the reaction between [ R ~ ( N H ~ ) ~ ( b p y ) 1 ~ +Absorbance vs. time data, displayed in an Omnigraphic 2000 recorder (bpy = 4,4'-bipyridine) and [Co(EDTA)I2- (EDTAe = ethyconnected to a Nicolet Explorer 111 A oscilloscope, were digitized with lenediaminetetraacetate). To our surprise,8 that reaction was found a Hewlett-Packard 9864A digitizer and then processed in a Hewlettto proceed exclusively (>95%) by an outer-sphere mechanism, Packard 9820 calculator. Values of kow, observed first-order rate constants, were calculated from a least-squares fitting of the equation In ( A , and, of course, no binuclear complex could be characterized. In - A,) = In (Ao- A,) - kobsdt,where A,, A,, and A , are absorbances at contrast, we have found that the reaction between [Ru(NH3j5time t, 0 (first point), and long times (8-10 half lives), respectively. The (pz)13+ (pz = pyrazine) and [Co(EDTA)I2- results in high initial reaction between [ R U ( N H ~ ) ~ ( ~and Z ) ][Co(EDTA)]~+ was studied in yields of the binuclear intermediate I, and we report herein our a Cary 17 spectrophotometer and the absorbance vs. time data were treated as outlined above in order to obtain values of kobsd.
I
detailed kinetic studies of the forward and reverse reactions in the [ R U ( N H , ) ~ ( ~3+/2+-[Co(EDTA)]-/2~)] system.
Experimental Section Materials. [ R U ( N H ~ ) ~ ( ~ Zwas ) H ]prepared B ~ ~ from Ru(NH,)+l, as described previ~usly.~The purity of the product was ascertained (1) (a) This work was supported by Grant CHE-8203887 from the National Science Foundation. (b) Abstracted in part from the B.S. Thesis of G.C.S., State University of New York at Stony Brook, May 1981. (2) Haim, A. Prog. Inorg. Chem. 1983, 30, 273. (3) Gaswick, D. G.; Haim, A. J . Am. Chem. SOC.1974, 96, 7845. (4) Melvin, W. S.; Haim, A. Inorg. Chem. 1977, 16, 2016. (5) Rosenheim, L.; Speiser, D.; Haim, A. Inorg. Chem. 1974, 13, 1571. (6) Adamson, A. W.; Gonick, E. Inorg. Chem. 1963, 2, 129. (7) Huchital, D. H.; Wilkins, R. G. Inorg. Chem. 1967, 6, 1022. (8) Phillips, J.; Haim, A. Inorg. Chem. 1980, 19, 76.
0002-7863/84/1506-1319$01.50/0
Results Stoichiometry. When solutions of [ R U ( N H ~ ) ~ ( P Z )(] ~ ' M) and [Co(EDTA)12M) are mixed at room temperature, a purple color develops immediately and then fades, within a few seconds, to the yellow color characteristic of [ R U ( N H ~ ) ~ ( P Z ) ] ~ + . Spectrophotometric examination of the product solution in the 600-400-nm range showed no additional absorbance changes for at least several minutes. The stoichiometry of the overall reaction (reactants to yellow product) is described by eq 1. With [Ru[ R U ( N H ~ ) ~ ( P Z )+] ~[Co(EDTA)]'+ e [Ru("3)5(pz)I2'
+ [Co(EDTA)I- (1)
(NH3)s(pz)]3+(1.03 X loe5M) and [Co(EDTA)I2- (7.26 X M), the forward reaction in eq 1 proceeds quantitatively as seen (9) Yeh, A.; Haim, A,; Tanner, M.; Ludi, A. Inorg. Chim. Acta 1979,33, 51. (10) Creutz, C.; Taube, H. J . Am. Chem. Soc. 1973, 95, 1086. (11) Hartkamp, H. Z. Anal. Chem. 1961, 251.
0 1984 American Chemical Society
1320 J. Am. Chem. Soc., Vol. 106, No. 5, 1984
Seaman and Haim
Table I. Kinetics of the Reaction between Ru(NH3),pz3+ and CoCDTA2- at 25.0 "C, pII 4.40 t 0.05, Ionic Strength 0.10 Ma isosbestic wave-
3.06 0.222 17.7 532 6.12 0.451, 0.424 518 12.2 0.713, 0.730 0.867, 0.889 20.7, 21.9 1.18, 1.41 23.4, 26.7 509 24.5 1.07, 1.03 1.21 26.6 509 24.5 1.05, 1.00 1.61, 1.28 24.4, 25.4 509 24.5b 1.05, 1.09 1.57. 1.31 27.5. 27.1 509 24.5' 1.06, 1.11 1.27' 25.6 509 24.5d 1.04 1.51, 1.49 29.4, 30.2 505 36.7 1.21, 1.30 36.7 1.33 1.37 1.48, 1.63 34.8, 32.7 504 38.4 1.28, 1.47 1.66 35.6 504 48.4 1.53 62.9 1.46, 1.53 1.70, 1.85 39.2,40.6 502 62.9 1.53 1.76 1.83, 1.96 42.0, 40.9 501 72.6 1.63, 1.48 a With [ R U ( N H , ) , ( ~ Z )=~ (0.9-1.1) +] X 10.' M. Each entry is the average of four replicate measurements with the same pair of solutions. [ R U ( N H , ) , ( ~ Z ) ~=+ ]2.41 X and 2.58 X M for the first and second entry, respectively. ' [ R u ( N H , ) , ( ~ z ) ~=+ ] M for the first and second entry, reand 7.43 X 4.91 X spectively. d [ R U ( N H , ) , ( ~ ~ ) ~=* 1.05 ] X M. by comparing the calculated absorbances at 474 (maximum for [ R U ( N H , ) ~ ( ~ Z )with ] ~ +molar absorbance 1.35 X lo4 M-' cm-'; molar absorbance of [Co(EDTA)]- 81 M-' cm-I) and 538 nm (maximum for [Co(EDTA)]- with molar absorbance 301 M-l cm-I; molar absorbance of [ R U ( N H , ) ~ ( ~ Z ) 2.09 ] ~ ' X lo3 M-' cm-') with the observed values. The comparison, observed absorbances (range in instrument 0.2-1-cm cell) 0.690 and 0.125 at 472 and 538 nm, respectively, vs. calculated values, 0.695 and 0.123, shows that the reaction proceeds to completion (>99%) as expected from the equilibrium constant K2 = 343 calculated from the E" values at 25 "C and ionic strength 0.2 MI2 or 0.1 MI3 for the [ R U ( N H , ) ~ ( ~ Z ) ] ~(0.52 + / ~ +V) and [CO(EDTA)]-/~(0.37 VI4) couples. Kinetics. Mixing of [ R U ( N H , ) ~ ( ~ Z )with ] ~ ' [Co(EDTA)I2in the stopped-flow apparatus clearly showed that the overall reaction represented by eq 1 proceeds in two stages. Observations were made at three wavelengths: 474 nm (the maximum for [ R U ( N H , ) ~ ( P Z ) ] ~530 + , nm (the predicted maximum for I; see below), and an intermediate wavelength. At 474 nm the absorbance increases in a biphasic mode. For example, with [Co(EDTA)]" (7.26 X M), there is a rapid absorbance increase that is complete in about 0.1 s and is followed by a slower increase that ends in about 2 s. Under the same conditions, at 530 nm the absorbance first increases rapidly (-0.1 s) and then decreases slowly (-2 s). The time at which the absorbance reaches a maximum and the magnitude of the absorbance at that time vary with changing [Co(EDTA)]'- concentration. The maximum absorbance increases with increasing [Co(EDTA)I2- concentration whereas the maximum time decreases with increasing concentration. At an intermediate wavelength, a single exponential increase in absorbance with the characteristic time of the rapid phase is observed. The wavelength at which only a monophasic reaction is observed decreases with increasing [Co(EDTA)]'concentration. Rate constants for the slow phase were measured at 474 and 530 nm. Good first-order plots (- 3 half-lives, correlation coefficient 0.999) were obtained at both wavelengths as long as the initial absorbances were taken after 8-10 half lives of the fast phase had elapsed. Values of kobsd474 and kobsdS3' for the measurements at 474 and 530 nm are listed in columns 2 and 3 of (12) Creutz, C.; Kroger, P.; Matsubara, T.; Netzel, T. L.; Sutin, N . J. Am. Chem. SOC.1979, 101, 5442. (13) Yeh, A. Ph.D. Thesis, State University of New York at Stony Brook, 1978 _.
(14) Tanaka, N.; Ogino, H. Bull. Chem. SOC.Jpn. 1965, 38, 1054
Table 11. Kinetics of the Reaction between [Ru(NH3),(pz)12t and [Co(EDTA)]- at 25.0 "C, pH 4.40 r 0.05, Ionic Strength 0.10 Ma
[ R u ( N H 3 ) , p z 2 + ] ,hi X 10'
9.93, 9.83 19.7, 19.9 40.0, 38.9 79.1, 79.2 a With [CoEDTA-] = 1.7 X 10-3 M.
kobsd384,S-'
3.19, 3.13 6.72, 7.05 13.3, 13.2 25.2. 25.8
M and [ascorbic acid] = 1.8 X
Table I. It will be seen that the values of kobsd474 or k o b ~ d 'are ~~ independent of the [ R U ( N H ~ ) ~ ( P Z )concentration ]~' and depend upon the concentration of [Co(EDTA)I2- according to eq 2. a[(Co(EDTA))'-] kobd474
Or
kobdS30
=
1
+ ~[(CO(EDTA))~-]
(2)
Nonlinear least-squares calculations yielded values of a and b (25 "C, ionic strength 0.10 M) equal to (8.8 f 0.3) X lo2 M-I s-] and 423 f 35 M-' at 474 nm and (1.09 f 0.05) X lo3 M-I s-l and 457 f 58 M-' at 530 nm. The wavelengths at which only the rapid, monophasic absorbance changes occurred (herein after named isosbestic wavelength) were determined experimentally at each [Co(EDTA)]'- concentration. Measurements were made near the anticipated wavelength. At wavelengths longer than the isosbestic, the absorbance was clearly seen to reach a maximum and then to decrease. At wavelengths shorter than the isosbestic, a biphasic absorbance increase was easily discerned. In this manner, the isosbestic wavelengths were determined within 1 nm. Plots of In ( A , - A,) vs. time at these wavelengths were linear for at least 3 half-lives, and the derived values of koMi" and the corresponding isosbestic wavelengths are listed in columns 4 and 5, respectively, of Table I. koMw varies with the concentration of [Co(EDTA)I2- according to eq 3, where (least-squares calculation) c = 16.9 f 0.6 s-I and kobdiSo
=c
+ ~[(CO(EDTA))~-]
(3)
d = (3.53 f 0.14) X lo3 M-' s-' at 25 "C and ionic strength 0.10 M. With ordinary concentrations of [RU(NH,)~(PZ)]~' and [Co(EDTA)]- the reverse reaction in eq 1 does not proceed to completion. Therefore, the reaction was studied in the presence of excess ascorbic acid. The latter is known to react slowly' with [Co(EDTA)]- and rapidly', with [ R U ( N H ~ ) ~ ( ~ Z )Therefore, ]~'. as rapidly as [ R u ( N H , ) , ( ~ z ) ] ~and + [Co(EDTA)I2-are formed by reaction between [ R U ( N H , ) ~ ( ~ Z )and ] ~ +[Co(EDTA)]-, the [ R U ( N H ~ ) ~ ( ~ isZ reduced ) ] ~ + back to [ R U ( N H , ) ~ ( ~ Z )by] ~the + ascorbic acid. Under these circumstances, the system under study reduction of [Coconsists of the [R~(NH,)~(pz)]*+-catalyzed (EDTA)]- by ascorbic acid, the rate-determining step being the reverse of eq 1. The disappearance of [Co(EDTA)]- was monitored at 384 nm, a maximum for [Co(EDTA)]- with molar absorbance 235 M-I cm-I, and a near minimum for [Ru(NH3)5(pz)]2' with molar absorbance -800 M-I cm-I. Plots of In ( A , - A,) vs. time were linear for at least 3 half-lives, and the corresponding values of kobsd'84, listed in column 2 of Table 11, were found to vary with the concentration of [Ru(NH3)j(pz)I2+ according to eq 4, where e = (3.5 f 2.0) X lO-'s-' a n d f = 3.20 f 0.04 M-' at 25 "C and ionic strength 0.10 M.
kobsd384 =e
+ A(Ru("3)dpz))"l
(4)
Inner- and Outer-Sphere Pathways
J . Am. Chem. SOC.,Vol. 106, No. 5. 1984 1321
Discussion
concentrations of all species were calculated as a function of time. Input parameters, chosen to be near the values estimated from The bulk of the observations are consistent with mechanistic the experimental measurements, were k l = 3 X lo3 M-' s-I, k-, Scheme I. According to this scheme, the reaction between = 15 s-], k, = 1 X IO3 M-I s-I, k-2 = 3 d, [ [ R U ( N H ~ ) ~ ( ~ Z ) ] ~ + ] ~ [ R U ( N H ~ ) ~ ( P Z )and ] ~ ' [Co(EDTA)],- is viewed as proceeding = 1X M and [[Co(EDTA)]*-] = (1-7) X IO-) M. For each via parallel outer- and inner-sphere pathways. The outer-sphere pathway k, produces directly the final products [ R U ( N H ~ ) ~ ( P Z ) ] ~ ' concentration of [Co(EDTA)12-,values of A474,,4530, and Ai,,, the absorbance of the system at 474 and 530 nm and the isosbestic and [Co(EDTA)]- and accounts for the observed rapid increase wavelength, were computed as a function of time (molar absorin absorbance at 474 nm. The inner-sphere pathway k l (also bances of [ R U ( N H ~ ) ~ ( ~ and Z ) ] I~were + taken as 1.4 X lo4 and rapid) yields the binuclear complex I. The slow absorbance 5.0 X lo3 at 474 nm and 3.0 X lo3 and 2.0 X lo4 at 530 nm). increase at 474 nm is then accounted for on the basis of back kobsd530, and koMiSowere obtained by fitting the Values of kobsd474, electron transfer in I (k-l path) followed by the outer-sphere absorbances to (A, -A,) = (A, - A,) exp(-kobdt) (for 530 and reaction k,. The rapid absorbance increase at 530 nm followed 474 nm, A, values were taken after 8-10 half-lives of the rapid by a slower absorbance decrease is also accounted for on the basis phase had elapsed). The resulting values of kobsd474 or kobdS3Oand of Scheme I. By analogy with the similar binuclear complexes kobsdiso were fitted to eq 2 and 3, respectively, to obtain values of [(NH3)5Ru11(pz)Co111(CN)s] (maximum at 524 nm, molar aba, b, c, and d . At 474 nm, a = 1.05 X IO3 M-I s-l a nd b = 2.57 sorbance 1.86 X lo4 M-l an-'), [(NH3)SRu11(pz)Ru111(EDTA)]+ X 10, M-I. At 530 nm, a = 1.03 X lo3 M-I s-l and b = 2.65 X (maximum at 520 nm, molar absorbmce 1.72 X lo4 M-' cm-I),l2 10, M-I SKI.At the isosbestic wavelengths, c = 14.5 s-I and d [(NH3)5Ru11(pz)Rh111(NH3)s]s+ (maximum at 528 nm, molar = 3.73 X lo3 M-I s-l. Now we consider the significance of the absorbance 1.8 X lo4 M-l cm-l),lo and [(NH3)5Ru11(pz)Rh111kobsd474or kobsd530and kobsdisoon the basis O f Scheme I. The (EDTA)]+ (maximum at 529 nm, molar absorbance 1.9 X lo4 M-l cm-I),l2 the binuclear complex [ ( N H 3 ) 5 R ~ 1 1 ( p ~ ) C ~ 1 1 1 -measurements of the rapid phase at the isobestic wavelength are expected to be a measure of all pathways for the disappearance (EDTA)]+ is expected to have an absorption maximum near 530 of [ R U ( N H ~ ) ~ ( ~ and Z ) ]its ~ +equilibration with I, and therefore nm with molar absorbance ca. 2 X lo4 M-I cm-I.l5 Since [Ru(NH3)s(pz)]3+does not absorb at 530 nm and [ R U ( N H ~ ) ~ ( ~ Z ) ] ~k+o b s p = k-I (k, k2)[[Co(EDTA)I2-]. The measurements of the slow phase at 530 or 474 are given by the expression kobsdS3O and [Co(EDTA)]- have relatively small molar absorbances at this = ~,[[CO(EDTA)]~-]/(I(~,/~_,)[[CO(EDTA)]~-]) or kobsd474 wavelength (3.0 X lo3 and 2.9 X lo2 M-' cm-I, respectively), the ifthe rapid phase has reached equilibrium before the onset of the rapid formation of I via pathway k l is accompanied by a rapid slow phase. It will be seen that according to the above interincrease in absorbance. The subsequent disappearance of I via pretation and the input parameters, kobsiSo= 15 + (4 x Io3)the sequence k-I followed by k2 accounts for the slow absorbance [[CO(EDTA)]~-]and kobsd474 or kobsd530= 1 X 103/(1 200decrease. The increase in the maximum absorbance at 530 nm [[Co(EDTA)I2-]), to be compared with the results of the kinetic with increasing [Co(EDTA)],- concentration is also accounted = 14.5 (3.73 X 103)[[Co(EDTA)]2-],kobsd474 modeling kobsdiso for by Scheme I. Since the disappearance of I is relatively slow = 1.05 X 103/(1 (2.57 X 102)[[Co(EDTA)]2-]),and kobsd530 compared to its formation, when the maximum absorbance is reached, I is in near equilibrium with [ R U ( N H ~ ) ~ ( ~ Zand ) ] ~ + = 1.03 X 103/(1 (2.65 X 1O2)[[Co(EDTA)I2-]). Evidently, within 6% accuracy, a, c, and d can be identified with k2, k-l, and [Co(EDTA)I2-, and thus the concentration of I increases with kl + k2, respectively. However, b (2.6 X 10,) is somewhat larger increasing concentration of [Co(EDTA)I2-. Finally, the occurthan kl/k-, (2 X lo2) and almost equal to d / c (257), which is rence of an isosbestic wavelength that decreases with increasing identified as (k, k2)/k-,. It is apparent that the interpretation [(CO(EDTA))~-]is nicely accounted for on the basis of Scheme offered above of the significance of kobsdiSo and kobd474or kobsd530 I. At the end of the rapid phase, near equilibrium between the represents an acceptable approximation to estimate the constants reactants and I has been achieved. The slow phase corresponds in Scheme I, except perhaps kI/k-,.l7 The later value can be to the disappearance of an equilibrium mixture of [Ru(NH,),estimated as (d - a)/c. and (pz)] [Co(EDTA)]*-,and [(NH3)5Ru11(pz)Co"1(EDTA)]+ The measurements of the slow phase conform to the equations the appearance of [ R U ( N H ~ ) ~ ( ~ ZThe ) ] ~ effective +. molar abk0M474= 8.8 X 102/(1 + (4.2 X 102)[(Co(EDTA))b])and kobd530 sorbance of the equilibrium mixture, aeff,is given by eq 5 where = 1.1 X 103/(1 (4.6 X 102)[(Co(EDTA))Z-]),respectively. aB[ll Therefore, k, = (1.0 f 0.1) X lo3 M-I a nd k1/k-, = (4.4 & aeff = (5) 0.2) X lo2 M-I. The measurements of the rapid phase conform [I1 + [[Ru("3)s(Pz)13+l to kobdiso= 16.9 (3.53 X 103)[(Co(EDTA))2-].Therefore, kl UB is the molar absorbance of the binuclear complex, and the + k2 = 3.5 X lo3 M-' & a nd k-, = 16.9 s-'. Finally, the meaabsorbances of [ R U ( N H ~ ) ~ ( P Z )and ] ~ ' [Co(EDTA)12- are nesurements of the reverse reaction conform to the equation kobsd384 glected. When equilibrium between the reactants and I obtains, = 3.5 X 3 . 2 [ ( R ~ ( N H ~ ) ~ ( p z ) ) ~ Therefore, ']. the rate eq 5 becomes aeff = aB[[Co(EDTA)]-]K1/(l + K,[[Coconstant for direct reduction of [Co(EDTA)]- by ascorbic acid (EDTA)],-]), where KI = k,/k-l. At a wavelength where the M-' s-l and k-, = 3.2 M-' s-I. It will be seen that value of a,, is equal to the molar absorbance of [ R U ( N H ~ ) ~ ( ~ Z ) ] ~ +is, -2 X the experimental results conform reasonably well to those of the then the transformation of the equilibrium mixture into [Rukinetic modeling with two exceptions. First, the kinetic modeling (NH3)5(pz)]2+proceeds without an absorbance change, Le., an shows that the rate constants for the slow phase should be isosbestic wavelength, dependent upon [(CO(EDTA))~-],obtains. wavelength independent, whereas we observe that the values of Although Scheme I accounts nicely for most of our general k,m474(and the derived value of k,) are 15-20% smaller than observations, there are some problems with the detailed interthe values of kobsd530.Second, the kinetic modeling shows that pretation of some of the data. However, before analyzing our data and ( k , + k 2 ) / k I derived kl/k-l derived from kobd474or kobsd530 on the basis of Scheme I, it is important to determine the relafrom kobsdiSoare almost equal, whereas the experimental results tionship between the measured constants k0M474,koM5m,and kMh show a discrepancy of almost a factor of 2.2 between these two and the rate constants kl, kl,k2, and k-,. This was accomplished quantities. We do not have an entirely satisfactory explanation by performing the following kinetic modeling. Scheme I was for the discrepancies, but kinetic complications in systems of this integrated by using the Runge-Kutta approximation,I6 and the type have been observed previously7J8and have been ascribed to
+
+
+
+
+ +
+
+
'+,
+
+
+
-
(15) Since the 474-nm band of [ R U ( N H ~ ) ~ ( P Z )is] ~a +metal to ligand charge-transfer band, addition of electropositive substituents to the remote N of pyrazine is expected to produce a bathochromic shift in the MLCT band. See,for example: Jwo, J. J.; Gaus, P. L.; Haim, A. J. Am. Chem. SOC.1979, 101, 6189. (16) Calculations were performed on a Digital Equipment Corporation 11V03 computer using the program RLINKLIT: Johnson, K. J. "Numerical Methods in Chemistry"; Marcel Dekker: New York, 1980; p 351.
(17) An exact treatment of the data does not seem feasible at the present time. Neglecting the back reaction k.2 (a good approximation since the equilibrium in eq 1 is more than 99% complete at all concentrations of [Co(EDTA)I2-). the differential equations derived from Scheme I can be integrated. However, not enough points were collected during the initial rapid phases at 530 and 474 nm. (18) Isied, S . S.; Taube, H. J . Am. Chem. SOC.1973, 28, 8198.
1322 J . A m . Chem. SOC.,Vol. 106, No. 5, I984
Seaman and Haim Table 111. Comparison of 1:xperimcntal and Calculntcd Values of A n4'4 103-
1.22 2.45 3.67
2.1, 1.8 1.7, 1.9. 2.1, 2.2 2.4, 2.9, 3 . 0 3.1, 3.2, 3.9 3.2, 3.6 4 . 0 , 4.4
0.94 1.84
0.4 1 0.84
2.60
1.23
4.84 3.20 1.56 oxidation of the binuclear complexes [ (NC)5Fe11(CN)Cor116.29 3.80 1.92 (EDTA)I5' or [(H20)(NH3)5Ru1rNCSH4-4-C02C011r(NH3)5]4+ 7.25 4.12 2.13 by [Fe(CN)J3- or [RU(NH~)~(OH~NC~H,-~-CO~]~+. Similarly, we postulate the oxidation of I by [ R U ( N H ~ ) ~ ( ~ Z eq ) ] ~6.+ , All values normalized to [ R U ( N H ~ ) ~ ~ Z= ~1 +. 0]X, lo-' M , 1-cm path length. Obtained by extrapolation to zero time of [ ( N H 3 ) 5 R ~ 1pz)Co"'(EDTA)]+ r( [Ru( NH3),(pz)]3+ F= the At474vs. timc plots. Calculated from eq 6 on thc basis of Scheme I with k = 2.5 X l o 3 hIY s - l , k , = 1.0 X l o 3 &I-' sC1 [(NH3)5Ru11'(pz)Co"1(EDTA)]2+ + [ R U ( N H ~ ) ~ ( P Z )(6) ]~+ k., = 16.9 s-'. a Calculated from eq 6 on the basis of Schcn;c I1 withk,=3.5X1O3X' s~',k.,=16.9s~',k,=4.8~~'. Adding eq 6 to Scheme I accounts for some of the discrepancies noted above.I9 contributions of Scheme I and I1 is contained in the absorbance On the basis of the above considerations, we believe the values vs. time data at 474 or 530 nm. The calculations are not very of k2 (1.0 X lo3 M-I s-l), k, k2 (3.5 X lo3 M-ls-I ), k-1 (16.9 sensitive at 530 nm, and therefore only the 474-nm data was s-I), kl/k-l ((3.5 X 103-l.0 X 103)/16.9 = 148 M-I), and k-2 (3.2 processed. By integration of the mechanism represented by M-' sd) to be rather good approximations of the true values. An Schemes I and 11, it can be shown that the absorbance of the slow independent estimate of the value of k2 can be obtained by comphase extrapolated to time zero is given by eq 7, where aRuis the bining the measured value of k-2 (3.2 M-' s-l) with the value of the equilibrium constant K2 ( = k 2 / k 2 ) calculated from the reaBT' aRuT'(L2- 1) duction potentials of the r u t h e n i ~ m ' ~and - ' ~ cobalt14couples. With 4,474 = (7) (L2 - &)MI K2 = 343, we obtain k2 = 1.1 X lo3 M-' s-l, in remarkable agreement with k2 derived from the measurements of the slow molar absorbance of [ R U ( N H ~ ) ~ ( P Z ) ] ~T'' , = (k, + k2 phase. Additional confirmation of the value of k2 comes from ~lkl)[(R~("3)s(Pz))3+l,, LI = W ( k - 1 + k3 - MI), L2 = calculations using the Marcus cross relationship (ion-pair apk-,/(k-l k3 - M2). M1,2 = [(kl k2 + k-1 + k3)/2] f [0.25(kl proach20). With rate constants for self-exchange 4.5 X lo5 M-' k2 - k-, - k3)2 klk-l]1/2. It will be seen that the value of s-I and 3.5 X M-l s-I a nd radii 3.5 X and 4.0 X is sensitive to the mechanism. If only Scheme I1 is operative, cm for the ruthenium2, and cobalt* couples, respectively, the only I contributes to If Scheme I is also operative, then calculated value of k2 is 5.9 X lo2 M-' s-l, in excellent agreement will be larger because of the contribution to A,,"74of the with the measured value. [ R U ( N H ~ ) ~ ( P Z produced )]~' in the rapid phase. Experimental So far, back electron transfer in I followed by outer-sphere values of are listed in column 2 of Table 111. In columns reaction between [ R U ( N H ~ ) ~ ( ~ Z and ) ] ~[Co(EDTA)I2+ has been 3 and 4 we list values of A0474calculated under the assumption taken to be the only route from I to the final products [Ruthat either Scheme I or Scheme I1 is operative.25 It will be seen (NH3)s(pz)]2+and [Co(EDTA)]-. An alternative or additional that, except for the lowest concentration of [Co(EDTA)]'-, the )]~' route involves direct dissociation of I to [ R U ( N H ~ ) ~ ( P Z and agreement between columns 2 and 3 is excellent, and we conclude [Co(EDTA)]- (Scheme 11). The mechanistic patterns embodied that Scheme I constitutes the major reaction pathway and estimate in Schemes I and I1 were considered previously5~*for the [Fethat Scheme I1 contributes less than 10% to the overall reaction. (CN),l3--[Co(EDTA)l2- and [Fe(CN),LI2- (L = pyridine or It is instructive to compare the relative outer- and inner-sphere 4,4'-bipyridine)-[Co(EDTA)I2systems. With the information reactivities of [ R U ( N H ~ ) ~ ( P Z )and ] ~ ' [ R ~ ( N H ~ ) ~ ( b p yvs. )l~+ available for the iron systems, Schemes I and I1 are kinetically [Co(EDTA)I2-. The experimental ratio of the outer-sphere rate indistinguishable. However, from the indirect arguments based constants kpzOS/kbpyM is 13, in good agreement with the value 33 on comparisons of substitution rates of Co"'-EDTA complexes, calculated from the square root of the difference in reduction I1 is not operative; it was a r g ~ e dc~o.n~~ i n c i n g l ythat ~ ~ , Scheme ~~ potentials between [ R U ( N H ~ ) ~ ( P Z ) (0.52 ] ~ ' V) and [Rui.e., the successor binuclear complex I represents a dead end. (NH,)S(bpy)]3+(0.34 V) expressed as an equilibrium constant. Similar indirect arguments can be advanced for the present The agreement between the experimental value of the ratio and system.24 However, it is not necessary to do so: the data obtained the value calculated on the basis of the thermodynamic barrier give direct proof that Scheme I provides the bulk of the reaction only implies that the rate constants for self-exchange in the pathway, with Scheme I1 providing at the most a minor contri[ R U ( N H ~ ) ~ ( ~ Z ) ]and ~ + [/ R ~ '~ ( N H ~ ) , ( b p y ) l ~ systems + / ~ + are bution. The information necessary to determine the relative substantially the same, as observed for other complexes of pentaammineruthenium with pyridine derivatives.2' With regard to the inner-sphere reactivity kpzis/kbp?, it is noteworthy that the (19) A kinetic modeling as described above but incorporating eq 6 to inner-sphere pathway was not detected in the [ R u ( N H ~ ) ~ Scheme I yields the following values (input values in Scheme I are the same (bpy)]3'-[Co(EDTA)]2- system* whereas a rate constant of 2.5 as above; rate constants for the forward and reverse reactions in eq 6 are taken X lo3 M-I s-l is found in the corresponding pyrazine system. as 1 X IO' M-' s-', the self-exchange rate constant for [ R U ( N H ~ ) ~3 cL/ 2 c Assuming that the reactivity differences between the pyrazine and complexes where L is a pyridine derivative): k2(474 nm) = 1.2 X 103 M-I SKI; kl/k-l(474 nm) = 304 M-I; k2(530 nm) = 1.7 X IO3 M-l s-l. , k i/k-1(530
+
+
+
+
+
nm) = 429 M-I; k-,(isosbestic) = 15.3 s-l; kl k2(isosbestic) = 3.67 X 10' M-l s-l. (20) Oliveira, L. A. A,; Haim, A. J . Am. Chem. Soc. 1982, 104, 3363. (21) Brown, G. M.; Krentzien, H . J.; Abe, M.; Taube, H . Inorg. Chem. 1919. - , 18. 3374. (22) Taube, H . Pure Appl. Chem. 1975, 44, 25. (23) Reagor, B. T.; Huchital, D. H . Inorg. Chem. 1982, 22, 703. (24) On the basis of Scheme I1 and the measured constants, k3 is calculated to be a c / ( d - a ) = 6.7 s-l and represents the loss of [ R U ( N H , ) ~ ( ~ Z ) ] ~ ' from the coordination sphere of a pentadentate Co(II1)-EDTA complex. Similar reactions, such as loss of CI- from [CO(EDTA)CI]~-,occur with rate constants -lod SKI. Therefore, k3 = 6.7 SKI is considered an unreasonably high rate for a substitution reactions at the Co"'(EDTA) center.
--.--
+
+
+
(25) The data from the kinetic modeling of Scheme I were also treated, and it was found that the values of A0474obtained by extrapolation to f = 0 of the A,"4 data are ([Co(EDTA)I2- concentration in parentheses) 0.0101 (1.0 X 0.0189 (2.0 X 0.0317 (4.0 X and 0.0438 (7.0 X Values calculated from eq 6 are 0,0101, 0.0191, 0.0324, and 0.0441. A kinetic modeling of Scheme I1 was also carried out with k , = 3 X 10' M-' s-' k -1 = 15 s-', and k3 = 5 S K I , Values of A0474obtained by extrapolation are 0.00268 0.0112 (4 X loT3),and 0.0196 (7 X (1 X 0.00555 (2 X Values calculated from eq 6 are 0.00258, 0.00554, 0.01 14, and 0.0197. The calculations show that the comparison between extrapolated values of A o and calculated values on the basis of either Scheme I or Scheme I1 serves to make a distinction between the two mechanisms. 3
~
~
~
J . Am. Chem. SOC.1984, 106, 1323-1332 bipyridine complexes arise solely from thermodynamic differences, the ratio kpzis/kbpyisshould be 33, the same value as for the corresponding outer-sphere reaction. However, the distance between the metal centers is presumably different for inner-sphere and outer-sphere transition states. Calculations' suggest that outersphere electron transfer corresponds to a distance of closest approach between metal centers of 7 X lo-* cm in [Ru(NH3),L]3+-[Co(EDTA)]z-, independent of the nature of L when L = pyridine or 4,4'-bipyridine. In contrast, the metal to metal distances are very different in the [ ( N H , ) , R u ( ~ z ) C O ( E D T A ) ] ~ (6.9 X lo-' cm) and [(NH,),Ru(bpy)Co(EDTA)]+ (11.1 X lo-' cm) inner-sphere transition states. In previous work2 we found that the distance dependence of rate constants for intramolecular electron transfer in [(NC),Fe(L)Co(NH,),] and [(EDTA)Ru(L)Co(NH,),]+ (L = pyrazine or a bipyridine) is given by eq 8,
the outer-sphere reorganization energy,26where a , and a2 are the radii of the two reactants, r is the distance between the metal centers in the transition state, and Dopand D, are the optical and static dielectric constant of the medium, respectively. On the basis of eq 8, the expected ratio of inner-sphere reactivities for pyrazine and 4,4'-bipyridine is 65 (to be compared with experimental values of 21 and 30 for [(NC),F~(L)CO(NH,),]~'and [(EDTA)Ru(L)CO(NH,),]+~*, respectively). On the basis of the thermody(26) Callahan, R. W.; Keene, F. R.; Meyer, T. J.; Salmon, D. J. J . Am. Chem. SOC.1977, 99, 1064. (27) Szecsy, A. P.; Haim, A. J. Am. Chem. SOC.1981, Z03, 1679.
1323
namic barrier and the distance effect, the calculated value of kp2/kbpyisis 33 x 65 or 2.1 x 10, and kbpyis= 2.5 x 103/(2.1 x lo3) = 1.2 M-' s-'. Evidently, such small contribution of the inner-sphere path to the overall [ R u ( N H ~ ) , ( ~ ~ ~ ) ] ~ + - [ C O (EDTA)I2- reaction (kbp? = 77 M-' s-') could not have been detected.' Finally, it is of interest to compare the value of the rate constant 16.9 s-l for intramolecular electron transfer in [(NH,),Ru"(pz)Co"'(EDTA)]' measured in the present work with the value 8 X lo9 s-] reported previously12 for intramolecular electron transfer in the very similar system [(NH3)sRu11(pz)Ru111(EDTA)]'. Since the thermodynamic factors are probably not very different for the Ru"/Co"' and Ru"/Ru"' systems,29the difference in intramolecular electron-transfer rates reflects mostly the difference in rate constants for self-exchange in [Ru(EDTA)]2-/- vs. [CO(EDTA)]~-/-.U n d ~ u b t e d l ya, ~major ~ portion of the difference in the rate constants for exchange is associated with the much smaller inner-sphere configuration changes in [Ru(EDTA)-/" as compared to [CO(EDTA)]-/~-.Whether orbital symmetry considerations are also important in determining the difference in the rates of exchange remains to be seen.31 Registry No. I, 88440-69-1; [ R U ( N H ~ ) ~ ( ~ Z ) ]38139-16-1; -'+, [Co(EDTA)I2-, 1493 1-83-0. (28) Oliveira, L. A. A. de; Haim, A. "Abstracts of Papers", 183rd Meeting of the American Chemical Society, Las Vegas, 1982; American Chemical Society: Washington, D.C., 1982; INOR 235. (29) E" for ( N H , ) J R ~ " ' ( p ~ ) R ~ " l E D T A+ + e(NH,)5Ru"'(pz)Ru"EDTA has been estimated as 0.37 V,I2 identical with the Eo for CoEDTA-I2-. (30) Sutin, N. Prog. Inorg. Chem. 1983, 30, 441. (31) Brunschwig, B. S.; Creutz, C.; Macartney, D. H.; Sham, T.-K.; Sutin, N . Faraday Discuss. Chem. SOC.1982, 74, 113.
-
Reactivity of 2-( Dipheny1phosphino)pyridine toward Complexes Containing the Quadruply Bonded Re26+Core: Ortho Metalation and Redox Chemistry Timothy J. Barder,lPF. Albert Cotton,*Ib Gregory L. Powell,Ib Stephen M. Tetrick,la and Richard A. Walton*la Contribution from the Department of Chemistry, Purdue University, West Lafayette, Indiana 47907, and the Department of Chemistry and Laboratory for Molecular Structure and Bonding, Texas A&M University, College Station, Texas 77843. Received July 28, 1983 Abstract: The quadruply bonded dirhenium(II1) complexes (n-Bu4N),Re2C18and Re2C16(PR3),(R = Et or n-Bu) react in methanol with 2-(diphenylphosphino)pyridine (Ph2Ppy)to afford complexes that are derivatives of the triply bonded dirhenium(I1) core, Re2,+. The complex Re,CI4(Ph2Ppy), (I) is formed initially and is found to eliminate HCI to give the ortho-metalated complex Re2C13(Ph2Ppy)2[(C6H5)(c6H4)Ppy] (II), the first example of a reaction of this type occurring at a metal-metal multiple bond of the M2L8 type. Both chloride and hexafluorophosphatesalts of the [Re2CI2(Ph2Ppy),lztdication (111) have been isolated (Cl-, IIIa; PF6-, IIIb); 111 constitutes a rare example of a multiply bonded dimetal unit complexed by four neutral bridging ligands. [Re2Cl2(Ph2Ppy),]CI2can be converted to either I or I1 under certain conditions. When acetone is used as the solvent, then Re$&(PR3)2 reacts with Ph2Ppy to give Re2CI4(Ph2Ppy),(PR3)(R = Et, IVa; R = n-Bu, IVb), whereas with acetonitrile as the solvent, PhzPpy converts (n-Bu4N),Re2C1, to the dirhenium(II1) complex Re2(p-C1)2(p-Ph2Ppy)2C14; Le., substitution occurs but without concomitant reduction. The complexes 11,1114 and IVa have been structurally characterized by X-ray crystallography. Compound I1 crystallizes in space group P2Jn with a = 12.160 (6) A, b = 21.418 (7) A, c = 18.464 (2) A, p = 99.04 (3)'. and Z = 4. Complex IIIb forms crystals in space group P 2 , / c with a = 24.316 (3) A, b = 12.101 (2) A, c = 27.070 (3) A, p = 105.670 (9)O, and Z = 4. Two acetone molecules per dimer unit are also present in the lattice. Compound IVa crystallizes in space group Pi with a = 10.683 (2) A, b = 20.033 (5) A, c = 10.544 (3) A, (Y = 102.29 (3)O, @ = 107.69 (2)O, y = 94.19 (2)O, and Z = 2.
Monodentate tertiary phosphines form a variety of well-defined complexes with the halides of Nb, Ta, Mo, W, and Re that contain dimetal units with metal-metal bonds of order two ( N b and Ta), three (Re), or four (Mo, W, and Re).2 In the case of N b and (1) (a) Purdue University. (b) Texas A & M University.
0002-7863/84/1506-1323$01.50/0
Ta, these complexes are chlorine-bridged molecules of stoichiometry M2X6(PR3),,whereas the Mo, W, or Re species are or M,X,(PR,), ( M = Mo, W, or Re) and either Rf&(PR3)2 possess eclipsed M2L8-typerotational geometries.2 However, with (2) Cotton, F. A.; Walton, R. A. 'Multiple Bonds between Metal Atoms"; Wiley: New York, 1982 and references therein.
0 1984 American Chemical Society
