6490

Inner-Sphere and Outer-Sphere Mechanisms in the Reductions of

Oxalatocobalt (111) Complexes by Vanadium( 11)
Bernard Grossman and Albert Haim*
Contribution from the Department of Chemistry, State University of New York,
Stony Brook, New York 11790. Received March 30, 1971

Abstract: The reductions of Co(en)(Cz04),- and Co(C~04)~~- by vanadium(I1) have been studied in aqueous
solutions of ionic strength 1.0 M . For C0(en)(C,0~)~-, the second-order rate constant (25", [H+] = 0.56 M) is
(1.08 f 0.04) x 102 M-1 sec-1 with AH* = 9.3 i 0.3 kcal/mol and AS* = -18 f 1 eu. For C O ( C ~ O ~ ) ~ ~ - - ,
the second-order rate constant (25", [H+] = 1.0 M ) is (2.04 f 0.06) X lo4 M-l sec-l with AH* = 2.2 i 0.2
kcal/mol and A S * = -31.5 f 0.8 eu. Both reactions produce VCz04- and V3+ as primary products, and it is
concluded that parallel inner-sphere and outer-sphere mechanisms are operative. The reduction of C ~ ( e n ) ( C ~ o ~ ) ~ -
by Fez+ has also been studied. At 25O, [H+] = 0.60 M , the second-order rate constant is 3.15 X M-l sec-l
with AH* = 15.7 f 0.3 kcal/mol and AS* = -17.5 i 1.0 eu. Comparisons of vanadium(I1) and iron(I1) re-
ductions of various cobalt(II1) complexes are made and interpreted mechanistically.

0 xidation-reduction reactions of vanadium(II), in

common with those of chromium(I1) and iron-
(11), fall into the inner-sphere-outer-sphere mechanis-
tic classification. However, in contrast with the
inner-sphere reductions by chromium(I1) and iron(I1)
for which electron transfer within the precursor binu-
clear complex is rate-determining, inner-sphere reduc-
tions by vanadium(I1) have been understood on the
basis of a rate-determining substitution into the co-
ordination sphere of vanadium(II).2 It has been sug-
gested that redox reactions of vanadium(I1) that pro-
ceed with rate constants in the range 1-50 M-' sec-'
(25 ") and activation energies in the range 11-13 kcal/
mol are of the inner-sphere type., However, when
the rate constants exceed appreciably the upper limit
of 50 M-' sec-I and the activation energies fall below
10 kcal/mol, then the outer-sphere mechanism becomes
operative. Examples of reactions that conform to
+
this generalization are c i ~ - C o ( e n ) ~ ( N ~ ) ~Vz+,
+ Cr-
+
SCN2+ V2+,4C O ( N H ~ ) ~ C ~ OV2+,j + ~ + FeNCS2+
+
V2+,4and FeN32+ V2+.4 The first three reactions
are known to proceed by an inner-sphere mechanism
with rate constants 32, 8, and 45 M-' sec-I ( 2 5 O ) , re-
spectively. The last two reactions are known to be of
the outer-sphere type, and proceed with rate constants
6.6 X IO5 and 5.2 X IO5 M-' sec-I (25"), respectively.
Confirmation of the adequacy of the mechanistic as-
signment based on the magnitude of the second-order
rate constant comes from recent measurements of the
rate constant for water exchange between solvent and
the aquovanadium(I1) ion.6
In the present work, we have extended our previous
studies' of the Co(NH&C204+-V2+ reaction to the
complexes Co(en)(CzO& and C O ( C ~ O ~ ) ~Since ~-.
equilibrium between vanadium(II1) and oxalate is
achieved relatively slowly,5 these systems offer an op-
portunity to establish the mechanism by direct obser-
(1) This work was supported by Grant No. GP-9669 from the Na-
tional Science Foundation.
(2) N. Sutin, Accounts Chem. Res., 1,225 (1968).
vation of the products. Moreover, we were interested
in the possible transition from an inner-sphere to an
outer-sphere mechanism as the reactivity of the cobalt-
(111) center in C O ( N H ~ ) ~ C * O
is ~increased
+ by succes-
sive substitution of the amine ligands by oxalate ligands.
The trends in rate constants in systems of known mech-
anism are of particular importance because recent
work8 on the vanadium(I1) reductions of various com-
plexes of the class Co(en)2XC1"+ (X = NH3, py, C1-,
H 2 0 ) has led to the suggestion that all these reactions,
which proceed with rate constants in the range 1.9-460
M-I sec-I (25 "), are mechanistically of the outer-sphere
type.
Experimental Section
Materials. Iron(I1) perchlorate solutions were prepared by
electrolytic reduction of solutions of recrystallized iron(II1) per-
chlorate in perchloric acid. The iron(I1) and hydrogen ion con-
centrations were determined as described previous1y.Q Vanadium-
+ (11) perchlorate solutions were prepared by reduction of solutions
of vanadium(V) oxide in perchloric acid with amalgamated zinc
under a n atmosphere of argon, and were used immediately after
~tandardization.~The preparation of lithium perchlorate solutions
and the purification of the argon have been described previously. lo
Distilled water was passed through a Barnstead ion-exchange
demineralizer and then was distilled in a Corning Model AG-lb all-
glass distilling apparatus. Sodium bis(oxa1ato)ethylenediamine-
cobaltate(II1) monohydrate was prepared by Dr. Jay Worrell.
The observed maxima at 540 and 382 nm with extinction coefficients
109 and 184, respectively, agree reasonably well with literature
values.12 Potassium tris(oxalato)cobaltate(III) sesquihydrate was
prepared by a standard procedure.13
Kinetic Measurements. The reduction of Co(en)(Cz04)z- by
iron(I1) was studied by conventional spectrophotometric methods.
The cobalt(I1I) complex was placed in the spectrophotometric
cell and the desired amounts of water, perchloric acid, and lithium
perchlorate were added. The cell was placed in the thermostated
compartment of a Cary 14 recording spectrophotometer, and,
following temperature equilibration, the desired amount of iron(I1)
solution, previously equilibrated to the same temperature, was
added. The absorbance at 540 nm was then recorded as a function
of time. The iron(I1) was in large excess to ensure pseudo-first-
(8) P. R. Guenther and R. G. Linck, J . Amer. Chem. SOC.,91, 3769
11969)
,----I.

(3) J. H. Espenson, J . Amer. Chem. Soc., 89,1276 (1967).
(4) B. R. Baker, M. Orhanovic, and N. Sutin, ibid., 89, 722 (1967);
M. Orhanovic, H. N. Po, and N. Sutin, ibid., 90,7224 (1968).
(5) H. J. Price and H. Taube, Inorg. Chem., 7, 1 (1968).
(6) M. V. Olson, Y . Kanazawa, and H. Taube, J . Chem. Phys., 51,
289 (1969).
(7) C. Hwang and A. Haim, Inorg. Chem., 9,500 (1970).
(9) P. Benson and A. Haim, ibid., 87,3826 (1965).
(10) J. R. Ward and A. Haim, ibid., 92,475 (1970).
(11) F. P. Dwyer, I. I<. Reid, and F. L. Garvan, ibid., 83, 1285 (1961).
(12) B. E. Douglas, R. A. Haines, and J. G. Brushmiller, Inorg.
Chem., 2,1194(1963).
(13) W. G. Palmer, "Experimental Inorganic Chemistry," University
Press, Cambridge, England, 1954, p 550.

Journal of the American Chemical Society J 93:24 1 December I , 1971

 6491
order conditions. Pseudo-first-order rate coefficients, k,, were inclined to favor the latter interpretation. The in-
evaluated from the relation A t - A , = (A0 - A,)e+pt, where A t , crease in rate with [H+] is quite modest and falls well
Ao, and A , are the absorbances at times t,,0, and a,respectively.
The values of A were fitted t o the above equation by means of a within the range observed for medium effects in other
nonlinear least-squares program,14 with At and t being the de- redox reactions." In particular, because of the high
pendent and independent variables, respectively, and Ao, A,, and negative charge on the cobalt(II1) complex, the present
k , were treated as adjustable parameters. Second-order rate con- system may be quite susceptible to the nature and con-
stants were calculated from the expression k = k,/[Fez+]. The
reductions of Co(en)(CnO&- and C O ( C ~ O ~ by ) ~ ~vanadium(I1)
-
centration of the cations in solution,18 and therefore
were studied in the rapid-flow apparatus.I6 The procedure and the ca. 15% change in rate observed on replacing 0.90
treatment of the data have been described previously.lo The M H+ by 0.9 M Li+ can reasonably be ascribed to a
wavelengths used were 540 nm for Co(en)(C,O,),- and 290 and 280 medium effect. Included in Table I are the measured
nm for Co(CzO&3- second-order rate constants at [H+] = 1.0 M and 15
and 35". Using these values and the values at 25"
Results
(average 2.04 X IO4 M-' sec-I), the activation param-
The results of the kinetic measurements of the Co- eters computed by using the nonlinear least-squares
(C204)33--V2+ reaction are summarized in Table I. program are: A H * = 2.2 f 0.2 kcal/mol and A S +
= -31.5 f 0.8 eu.
Table I. Kinetics of the C O ( C ~ O ~ ) ~ ~ -Reactiono
-V~+ Since the rate of the C O ( C ~ O ~ ) ~ ~ -redox
- V ~ +reaction
is quite rapid, whereas the rate of equilibration for the
10~[Co(III)], vanadium(II1) oxalate system (eq 2) is relatively slow,j
M 104[V(II)], M [H+], M 10-4k, M-l sec-l
this system offers an opportunity to establish whether
2.19 2.54 1 .OO 1.92, 1.92 an inner-sphere or an outer-sphere mechanism obtains
10.9 12.8 1 .OO 1.88, 1 . 9 5
5.47 13.0 1.00 2.03, 2.16 V3+ + H~C204I_ VCzO4+ + 2H+ (2)
10.9 25.3 1.00 2.19, 2.09
4.37 4.92 0.85 1.90, 1.96 by determining the nature of the primary vanadium(II1)
4.37 4.92 0.70 1.87, 1.85 product. Since much of the necessary information
4.37 4.92 0.70 2.00, 2.03 about reaction 2 is not available, the measurements
4.37 5.08 0.50 1.82, 1.83 were made on a comparative basis. First, the reaction
4.37 5.08 0.50 1.77, 1.75
of 4.38 X M vanadium(II1) with 1.32 X M
4.37 7.58 0.30 1.82, 1 . 8 0
4.37 7.58 0.30 1.92, 1.97 H2C204in the presence of 0.0158 M vanadium(I1) at
4.37 5.05 0.10 1.75, 1.71 [H+] = 0.92 M and 25" was examined at 290 nm. The
2.19 2.50 1 .oo 1.68, 1 . 72c absorbance was found to increase, with a half-life of
4.37 5.05 1 .OO 1.70, 1.8gC 0.36 sec. The value of AA/[V3+I01, where AA is the
4.37 5.05 1.00 1.74, 1 . 8 6 ~
4.37 5.05 1.00 1.56, 1. 5OC change in absorbance upon reaction, [V3+]ois the initial
4.37 5.01 1.00 2.38, 2. 32d vanadium(II1) concentration, and 1 is the path length
4.37 5.01 1 .oo 2.40, 2.27d of the observation tube of the rapid-flow apparatus,
a Ionic strength 1.0 M maintained with HC1O4-LiC1Oa. Mea-
was 261 M-' cm-I. Next, the reaction of 4.38 X
surements at 25" unless specified otherwise. Each entry represents M C O ( C ~ O J , ~with
- 0.0201 M vanadium(I1) at [H+] =
the average of two to four replicate measurements with the same 0.92 A4 and 25" was examined at the same wavelength.
pair of solutions. First and second entries from 280- and 290-nm Under these conditions, first a rapid decrease in ab-
measurements, respectively. At 15". At 35".
sorbance was observed (disappearance of C O ( C ~ O3-,~ ) ~
E 8 X IO3, and formation of V3+, E lo), and this was

At constant hydrogen ion concentration, the second-
order rate constants defined by eq 1 are seen to be
independent of the concentrations of the two reactants l6
and of the wavelength used for the measurements.
However, the second-order rate constants decrease
with decreasing hydrogen ion concentration. The
value of k is approximately 15 % higher at [H+] = 1.O
M as compared to 0.10 M . The significance of this
observation is not clear. Two alternate interpretations
are plausible. On the one hand, this may be viewed
as a chemical effect caused by the operation of two
parallel paths, one acid independent (ko)and one acid
dependent (kl) with k = ka +
kl[H+]. On the other
hand, the increase in rate with increasing acidity may
be viewed as the manifestation of medium effects as
lithium ions are replaced by hydrogen ions. We are
(14) Los Alamos Report No. LA 2367, March 4, 1960, and Addenda,
Jan 14,1963.
(15) G . Dulz and N. Sutin, Znorg. Chem., 2,917 (1963).
(16) We were particularly interested in the possibility of outer-sphere
complex formation between the reactants. This type of interaction
would lead to a less than first-order dependence on [Vz+]at higher values
of [Vz+l. Unfortunately, it is not possible to increase [V?+] above the
values given in Table I because the reaction becomes too rapid.
followed by a slower increase in absorbance (equili-
bration in reaction 2, effective extinction coefficient of
the equilibrium mixture 271). The minimum in the
absorbance us. time curves occurred at 0.022 sec. The
increase in absorbance was associated with a half-life
of 0.36 sec and a value of AA/[Co(III)]ol of 219 M-'
cm-l. In this case AA was measured as A , - A,,
where A , was the absorbance at long times and A , the
absorbance at the minimum. A comparison of the
observations for the V3+-HzCz04 reaction and the
C O ( C ~ O ~ ) ~ ~ - reaction
- V ~ + suggests that the latter re-
action proceeds oia the sequence of steps given by eq
3 and 2. On the basis of this sequence, the values of
tmin, the time for minimum absorbance,lg and of A A /
+
C O ( C Z O ~ ) ~ ~vz+
-
H'
COZf + 3HpCzO4 + v3+ (3)
(17) D. E. Pennington and A. Haim, Inorg. Chem., 6,2138 (1967).
(18) The effect of ionic strength on the analogous system c0(c?04)33-
+ Fez+ has been studied in detail: J. Barrett and J. H. Baxendale,
Trans. Faraday Soc., 52,210(1956).
(19) tm was calculated from the expression [l/(ki+ +
k2)I In [1 (ki -
k?)(el - € 2 - 63)/kZ(e4 - ea)], where kl is the pseudo-first-order constant
for reaction 3, k? is the pseudo-first-order rate constant for approach to
equilibrium in reaction 2, €1, €2, and € 3 are the molar absorbances of
Co(C?O4)33- (8.0 X lO3), Co?+ (0.01), and V3+ (lo), respectively, and
e4 is like the effective molar absorbance of the equilibrium mixture of
VCzOa+ and V3+ under the prevailing condition (e 271).

Grossman, Haim 1 Reductions of OxaIatocobaIt(IIi)Complexes by Vanadium(I0

6492
[ C O ( I I I ) ] are
~ ~ ~0.020
~ sec and 245 M-I cm-l, respec- activation parameters calculated from the measure-
tively. The reasonable agreement between the mea- ments at [H+]= 0.56 M are: AH* = 9.3 f 0.3
sured and the calculated values is noteworthy, and it kcal/mol and AS* = - 18 f 1 eu.
is concluded that the proposed scheme (eq 3 and 2) is In addition to providing the rate constant for the
basically correct. However, the measured value of Co(en)(Cz04)2--V2+ reaction, the fifth and sixth ex-
AA/[CO(III)]~Zis somewhat smaller than the calculated periments in Table I1 provided information about the
value. If it is assumed that the discrepancy is caused nature and yields of the primary vanadium(II1) prod-
by a parallel inner-sphere reaction (eq 4) that produces ucts. For the fifth experiment, transmittance us. time
VCz04+ as the primary vanadium(II1) product, then measurements were also performed at 290 nm. Plots
Hf of log ( A , - A,) us. time were linear for at least three
+~ - +CO"
C O ( C ~ H ~ ) ~V2+ + 2HzCzO4 + VCz04' (4) half-lives and yielded a second-order rate constant of
it can be shown that the value of AA/[CO(III)]~Iis given 124 .f 3. Calculations of log ( A , - A , ) us. time at
290 nm were carried out under the following assump-
by tions: (a) that VC204+was produced in 100% yield
and then dissociated to produce the equilibrium mix-
ture of VC2O4+and V3+appropriate to the experimental
conditions used; (b) that V3+ was produced and then
associated with HzCz04 to produce the equilibrium
mixture of V3+ and VCZO4+. Under either of these
(€4 - 4 [kz exp( - klt,) - ki exp( - kztm)] assumptions curved log ( A , - A , ) us. time plots were
(ki - kz) obtained, and the initial slopes yielded second-order
rate constants in the range of 80-100 M-' sec-I. Since
wherefIs is the fraction of the C O ( C ~ O ~ ) ~ ~ - -reaction
-V~+ the experimental log ( A I - A , ) us. time plots at 290 nm
that proceeds via an inner-sphere mechanism and kf were linear and yielded second-order rate constants in
is the pseudo-first-order rate coefficient for the forward excellent agreement with those measured at 540 nm,
reaction in eq 2. Using the values AA/[CO(III)]OZ= it is concluded that both VCz04+and V3+ are produced
219, kl = 320 sec--I, kz = 1.93 sec-I, t1 8.0 X IO3, in the redox reaction in quantities approximately equal
c2 0.01, t3 10, e4 271, and tm 0.022 sec, it is calculated to the equilibrium values appropriate to the experi-
that fIs/kf = 0.054. Using the equilibrium quotient mental conditions used. Using the equilibrium quo-
of -looz1 for reaction 2 and the measured value of tient of -looz1 for reaction 2 and the concentrations
kz (1.93 sec-I) under the experimental conditions, kr = of V(III), HzCz04,and H+ used in the experiment, it is
1.1 sec-I, and therefore frs = 0.06. It is concluded calculated that the ratio [VCz04+]/[V3+]at equilibrium
that the overall Co(C20J33--V2+ reaction proceeds is 0.57, and therefore it is concluded that the Co(en)-
via parallel outer- (94%) and inner-sphere (6%) paths. (czo4)2--v2' reaction yields approximately 36 % VCz-
The results obtained in the study of the Co(en)(Cz- Ob+and 64% V3+. A more accurate determination of
04)z--V2+ reaction are summarized in Table 11. It was the yields of VCZO4+and V3+ was carried out by mea-
suring the changes in transmittance as a function of
time at 290 nm under conditions where the VCz04+-V3+
Table 11. Kinetics of the Co(en)(CzOa)z--V2 Reaction"
+
equilibrium favors dissociation of VC2O4+. This is
104[CO(III)], the sixth experiment listed in Table I1 for which it is
M 102[V(II)],M Temp, "C k , M-l sec-l calculated that the ratio [VC204+]/[V3+]at equilibrium
2.45 2.01 15 ! 4
68 == is 0.08, Under the conditions used, the rate of the
2.53 2.01 15 61 =!= 3 Co(1II)-V(I1) reaction is faster than the rate of dis-
2.50 1.01 25 104 & 4 sociation of VC204+,and therefore the log ( A , - A,)
2.48 2.01 25 112 i 4 us. time plots consisted of two exponential decays, the
22.0 1.00 25 123 3~ @
4.03 2.02 25 116 i 3d second (and slower) decay corresponding to the dis-
2.53 2.01 35 187 i 7 sociation of the VC2O4+ formed in the first reaction.
5.00 2.01 35 194 i 3 By extrapolating the second exponential to time 0, it
a Unless specified otherwise, ionic strength = 1.0 M maintained
can be shownz2that f&, the fraction of reaction that
with HC1OrLiC1O4; [H+] = 0.56 M. Each entry is the average proceeds via the inner-sphere mechanism and produces
of four replicate measurements with the same pair of solutions. VCZO4+,is given by
c In the presence of 2.20 X Madded HZC204; [H+] = 1.0 M .
At [H+] = 1.0 M. IS = [m/(es - d 1 [ 1 - (k4/kd1 + kf'/k,
where m is the intercept obtained by extrapolating the
assumed, rather than proved, that the reaction is first second exponential in the log { ( A , - A,)/[Co(III)]d)
order in cobalt(II1) and first order in vanadium(II), us. time plots to time 0.2z Using the values m = 285,
and the calculated second-order rate constants are 10, e& 5 X lo2, k3 = 2.34 sec-I, k4 = 0.65 sec- l , and
listed in column 4 of Table 11. As was the case for the (22) The dependence of ( A t - A,)/[Co(III)]ol is given by ( A t -
C O ( C ~ O3--V2+
~ ) ~ reaction, there is a modest increase +
A,)/[co(III)]ol = ( € 5 € 6 - €2 - €3) exp(-k3t) + (e7 - ~3)[kf'/(k3
-
in rate with increasing [H+] (compare entries 3 and 4 ka) exp(-kat) + k3fIS/(ka - ka)[exp(-kd) - exp(-kat)l - kskr'/
ka(k3 - kr) exp(- kd)]. where € 5 , 8 6 , and e, are the molar absorbances Of
with 5 and 6). Again the increase in rate is ascribed to Co(en)(C20a)z-, V2+, and VCzOa+, respectively, k3 is the pseudo-first-
a medium effect rather than a chemical effect. The order rate coefficient for reaction 5, ka is the pseudo-first-or? rate
coefficient for approach to equilibrium in reaction 2, and kr is the
(20) Am/[Co(III)]ol was calculated from the expression (a4
(€1 - €2 - €3) exp(-klr,) +
( € 4 - €3)/(kl - kz)[kz exp(-klt,)
+ - kl.+
€2) pseudo-first-order rate coefficient for the forward reaction in eq 2.
At long times, ( A t - A,)/[Co(III)]ol = (e7 - 63)(k3k4f18 - k3kr')/
exp(-k&)l. (k3 - ka)ka exp( - kat). Extrapolation to time 0 yields the desired rela-
(21) C. Hwang, unpublished observations. tion.

Journal of the American Chemical Society 93:24 December 1, I971

 6493
kr'/kc = 0.08, the calculation yielded the valuefIs = sumed that an ion pair or outer-sphere complex is
0.50, and it is concluded that the Co(en)(Cz04)2--V2+ formed rapidly from the cobalt(II1) complex and vana-
react: 3n proceeds uia parallel inner- and outer-sphere dium(I1) (eq 6). The ion pair can proceed forward
paths as shown in the scheme of eq 5 . according to one of two possible reactions: (a) a slow
outer-sphere electron-transfer process (eq 7) producing
Co2+ + VC2O4+inner sphere V3+ (eq 9) as the primary vanadium(II1) mononuclear
Co(enXC,OI); + V2+ Co2+ + V 3 + outer sphere (5 1 product, and (b) a slow inner-sphere substitution into
60%
the coordination sphere of vanadium(I1) (eq 8) to pro-
The results obtained in the study of the Co(en)- duce the precursor binuclear complex COL(CZO~)~-
(C20&---Fe2+ reaction are summarized in Table 111. V(2-n)+, This slow complex formation is followed by
rapid electron transfer within the binuclear complex
Table 111. Kinetics of the C O ( ~ ~ X C ~ O ~ ) ~Reaction"
--F~~
+ and dissociation to cobalt(I1) and VC204+ (eq 10).
104[Co(III)],M 102[Fe(II)], M Temp, "C 103k, M-' sec-l According to this mechanism, the observed second-
order rate coefficients are given by eq 11. Under the
5.00 1 .oo 35 6.96
5.00
6.62
9.26
9.26
35
35
7.26
7.40
+
k = ( k ~ s kos)Q/<l +
Q[V2+I> (1 1)
10.1 9.26 35 7.46 experimental conditions employed, Q[Vz+] << 1.
5.33 3.71 35 6.33b Therefore, the observed second-order coefficients are
5.00
6.24
5.56
9.26
35
25
6.80c
3.15 +
(kIs kos)Q, and separate values for Q and for k ~ s +
6.66 9.26 25 3.15 kos could not be obtained from the kinetic measure-
5.48 9.26 45 17.2 ments. However, according to the proposed mech-
5.53 9.26 45 16.7 anism, kIs/kos = [VC204+]/[V3+],the ratio of the pri-
~

aIonic strength 1.0M maintained with HC104-LiC104; [H+] =
0.60 M unless specified otherwise. b[H+] = 0.189 M . c[H+] =
0.284M .
In the limited concentration range studied, the reaction
obeys mixed second-order kinetics. The second-order
coefficient is seen to be dependent on hydrogen ion
concentration (cf. entries 1-6 of Table 11), but, as was
the case in the C O ( C Z O ~ ) ~ ~ - - reaction,
V~+ the change
in rate with [H+] is likely to be a medium effect. The
activation parameters calculated from the measure-
ments at 0.60 M hydrogen ion concentration are A H S
= 15.7 i= 0.3 kcal/mol and AS* = -17.5 1.0 eu. * eters for the vanadium(I1) reductions of cobalt(I1)
Discussion
The results obtained in the present work conform,
as far as the major features are concerned, to the gen-
eralizations about vanadium(I1) reductions mentioned
in the introduction. The very rapid reduction of
C O ( C ~ O ~ ) k~ ~=- , 2.1 X l o 4 M-' sec-I at 25" with
AH* = 2.2 kcal/mol, proceeds predominantly by an
outer-sphere mechanism. For the slower reduction of
C0(en)(C~0~)~-, k = 1.1 X IO2 M-' sec-' at 25" with
AH* = 9.3 kcal/mol, an inner-sphere path is impor-
tant in bringing about the redox reaction. However,
for both systems, the detailed examination of the pri-
mary vanadium(II1) products shows that parallel inner-
sphere and outer-sphere mechanisms are operative,
the distribution between paths varying considerably
for Co(en)(C204),- and CO(C~O&~-.An entirely sim-
ilar situation has been previously demonstrated for the
VO*+-V2+ reaction.23 The proposed kinetic scheme
is described by eq 6-10 (L = en or C2042-). It is as-
+
CoL(C2O4)zn- V2+ =
COL(C204)2"-* v2++COL(C204)2("
C O L ( C ~ O ~ ) ~ ~ - rapid,
+ 1)- .v3+
.V~+Q
slow, kos (7)
V2+--f C O L ( C Z O ~ ) Z V ( ~slow,
COL(CZO~)Z"-* - ~ ) + k ~ s (8)
COL(GOa)2("+')-. v3 +
Cos+ L + + 2C2O42- + V3+ rapid
CoL(C204)2V(2-=I+ --f
Goa+ + L + C20n2- + VCz04+ rapid
(23) T. W. Newton and F. B. Baker, Inorg. Cbem., 3,569 (1964).
mary mononuclear vanadium(II1) products. There-
fore, from the stoichiometric measurements it is pos-
sible to calculate kIs/kos values of 1.0 and 0.064 for
Co(en)(C204),- and C O ( C ~ O ~ ) ~ respectively.
~-, 24 Ex-
pressing the inner-sphere and outer-sphere paths in
terms of the second-order rate coefficients Q k ~ sand
Qkos, respectively, we calculate (25 O), for Co(en)-
(C2OJ2-, QkIs = 55 M - l sec-' and Qkos = 55 M-'
sec-' and for C O ( C ~ O ~ ) ~Q~k-~, s= 1.3 X lo3 M-'
sec-' and Qkos = 2.0 X lo4 M-* sec-l.
A summary of rate constants (corrected for statis-
tical factors where appropriate) and activation param-
oxalate complexes is given in Table IV. It is seen that
the statistically corrected value of Q k ~ sfor Co(en)-
(C204)2-falls nicely in the range of rates of the other
vanadium(I1) inner-sphere redox reactions. The en-
thalpy of activation for the Co(en)(Cz04)z- reaction
(9.3 kcal/mol) is somewhat smaller than the value
characteristic of inner-sphere reactions (12 1 kcal/
mol). However, since the measured enthalpy of acti-
vation includes the contribution of the outer-sphere
path, the lower value is rea~onable.~SIn this context,
it is noteworthy that the enthalpy of activation for the
C O ( C ~ O ~ ) ~ ~ - -reaction,
V~+ which proceeds predomi-
nantly by an outer-sphere path, is only 2.2 kcal/mol.
The statistically corrected value of QkIs (420 M-'
sec-l) for C O ( C , O ~ ) ~is~ -somewhat higher than the
values for the other inner-sphere reactions. However,
values in the range 112-278 M-I sec-' ( 2 5 O , ionic
strength 1.0 M ) have been reported for other trineg-
ative ions,26 and when QkIs for C O ( C ~ O ~ is ) ~cor-
~-
(6)
rected for an additional statistical factor of 2,27then
the resulting figure of 210 M-I sec-1 falls well within
the range for the trinegative ions.
(24) The corresponding value in the VO2+-V2+ reaction23 is 1.85.
(25) R. S . Taylor, R. N. F. Thorneley, and A . G. Sykes, J . Cbem.
SOC.A , 856 (1970).
(26) K. M. Davies and J. H. Espenson, J . Amer. Chem. Soc., 91,
(9) 3093 (1969).
(27) The statistical corrections in Table I V yield rate constants per
oxalate ligand. However, in comparing Co(CN)sX3- with Co(C2-
(10) O4)a3-, the rate constant per oxalate ligand must be divided by 2 in
order to obtain the rate constant per donor oxygen atom. Note that
X- and not CN- is the entering ligand in the Co(CN)6X3--V2* reac-
tions.26

Grossman, Haim Reductions of Oxalatocobalt(IIl) Complexes by Vanadium(II)

6494
Table IV. Second-Order Rate Constants and Activation Parameters for VZ+ and Fez+Reductions of Cobalt(II1) Complexese
V2+ reductant
Complex k AH*
a Rate constants in M-1 sec-l at 25” and ionic strength 1.0 M . Activation energies and entropies in kcal/mol and eu, respectively.
sphere path. e Reference 5 . J. Espenson, Inorg. Chem., 4, 121 (1965). e Reference 7.
work. h Outer-sphere path. Corrected for statistical factor of 3. j A. Haim and N. Sutin, J. Amer. Chem. SOC.,88,5343 (1966).
It is instructive to compare the vanadium(I1) reduc-
tions of oxalatocobalt(II1) complexes with the corre-
sponding reductions by iron(I1). These results are
summarized in Table IV and in Figure 1, where log
kv is plotted against log kFeaZ8It will be seen that in
4
. the rate constants barely change. In contrast, the
Y 2
. .3
-4 -3 -2 -I 0 I
l o g kFe
Figure 1. Plot of log kv us. log k~~ for the reductions of oxalato-
cobalt(II1) complexes : 1, Co(NH&C?O4’-; 2, Co(en)(Cz04)~-;
3, Co(NH3)sCzOa+; 4, C O ( C Z O ~ ) ~ ~ - .
going through the series C O ( N H ~ ) ~ C ~ O Co(en)(C2-
~+,
04)2-, and Co(NH&Cz04+ the rate constants for re-
action with V 2+ are practically invariant, whereas the
rate constants for reaction with Fe2+ cover a range of
ca. lo3. In going from C O ( N H ~ ) ~ C to ~ OC~O+( C ~ O ~ ) ~ ~of
toe rate of vanadium(I1) reduction increases by a fac-
tor of ca. 4 X lo2, whereas the increase is ca. 25 for
iron(I1). The differences in sensitivity to changes in
the coordination sphere of cobalt(II1) displayed by
vanadium(I1) and iron(I1) reductions can be readily
(28) The values used in the plots are the measured second-order rate
constants divided by the number of oxalate ligands in the cobalt(II1)
complex.
Journal of the American Chemical Society j 93:24 j December I , 1971
-Fez+ reductant----
AS * k AH* AS * Ref
Inner-
’
Corrcected for statistical factor of 2. # This
understood on the basis of the mechanisms. The re-
actions of C O ( N H ~ ) ~ C ~ OC ~O+( ,N H ~ ) ~ C ~ O
and
~ +Co-
,
(en)(Cz04)z-with vanadium(I1) are predominantly sub-
stitution-controlled, inner-sphere reactions and, con-
sequently, are quite insensitive to nonbridging ligand
effects. 29 Therefore, in the series under consideration
corresponding reactions with iron(I1) are electron-trans-
fer-controlled, inner-sphere reactions,’! 30 with rate
constants that vary considerably with the nature of the
nonbridging ligand^.^ This accounts for the horizon-
tal portion of the log kv us. log kFe plot. Now, in going
from C O ( N H ~ ) ~ C ~toOC~O+ ( C ~ O ~ )the
~ ~ vanadium(I1)
-,
reaction undergoes a change in mechanism, from inner
sphere to outer sphere, and an abrupt increase in rate
obtains. In contrast, the iron(I1) reduction of Co-
(Cz04)33-is still inner sphere, and the rate continues to
increase gradually. This explains why point 4 in Fig-
ure 1 is off the horizontal line.
It is noteworthy that the present finding of a break
in the log kv us. log kFe plot when the vanadium(I1)
reactions change from an inner-sphere to an outer-
sphere mechanism contrasts with the finding of a linear
free energy relation for the vanadium(I1) and iron(I1)
reductions of a series of chlorocobalt(II1) complexes.8
On the basis of such linear relation, and after recog-
nizing that some of the vanadium(I1) reactions were
sufficiently fast to ensure the operation of an outer-
sphere mechanism, Guenther and Linck suggested that
all of the chlorocobalt(II1) complexes that they studied
reacted via an outer-sphere mechanism, but cautioned
that their interpretation was dependent on the finding
- , a system which displayed a change in sensitivity to
nonbridging ligands as the mechanism changed from
inner sphere to outer sphere. The present results prove
that such a change in sensitivity occurs, and conse-
quently they reinforce the conclusion* that the vana-
dium(I1) reductions of chlorocobalt(II1) complexes pro-
ceed by an outer-sphere mechanism.
(29) T. J. Przystas and A. Haim, manuscript in preparation.
(30) A. Haim and N. Sutin,J. Amer. Chem. Soc., 88,5343 (1966).