82 1
The spectra calculated for the trans-chloro(so1vent) trans product compare complexes assuming 17 favorably with the spectra of trans complexes with ligands of similar strength. Spectra calculated on product ratios greatly different from this yield unreasonable spectrL for the trans complex. It seems unlikely that the reaction of cis-[CoCl(DMA)(en)zI2+in DMF, which gives anomalous activation parameters, has an s N 2 mechanism as this would involve a large decrease in entropy in the formation of the transition state, and thus AS* would be negative. We feel sure it has a similar mechanism to the other reactions, but we cannot explain the parameters. By comparing parameters for the reactions of a series of similar complexes in one solvent, as was done in the discussion of the activation parameters for the aquation of the cis-solvent complexes, it is established that the strengths of the metal-ligand bond are in the order DMSO > D M F > DMA. Since activation parameters could not be determined for the reaction of the cischloroaquo complexes, the position of H?O in this series could not be established. Little can be gained from a comparison of the few results obtained for trans complexes, as these have nothing in common. The results for substitution into
tr~ns-[CoCl(H~O)(en)~]~+ are complicated by chloride entry in all solvents except DMSO. In view of the weakness of water as a ligand,'j coupled with the high rate of exchange of cis-[Co(NO2)(DMSO)(en),]2+ (see below), it seems possible that solvent interchange may also be significant in some recently published anation reactions15 where direct anation has been considered as the sole mechanism of reaction in solvents quite likely to coordinate. A comparison of the rates of exchange of DMSO as a function of the other monodentate ligand and the comparison of these results with those on aquation of the monochloro complexes of the type [CoClA(en)z]+ is rewarding.I6 It can be seen that for C1-, Br-, or DMSO as the unreplaced ligand (A) the rate of DMSO exchange is not greatly different, but with Nos- the rate of DMSO exchange increases more than tenfold. This is interpreted5 as a result of a change in mechanism, the electron-withdrawing nitro group facilitating an W2 mechanism and inhibiting an SNI mechanism. The rate of exchange of DMSO by the cis-[CoCl(DMSO)(en)#+ ion appears to be uninfluenced by ion association, as indicated by using the bromide salt. (15) M. N. Hughes dnd M. L. Tobe, J . Chenz. SOC.,1204 (1965). (16) S. Asperger and C. R. Ingold, ibid., 2862 (1956).
Octahedral Cobalt (111) Complexes in Dipolar Aprotic Solvents. X. The Isomerization of cis- and trans-Dichlorobis (ethylenediamine)cobalt ( 111) Ions, [CoC1, ( en) 2 ] -, in Anhydrous Sulfolane W. R. Fitzgerald and D. W. Watts
Contribution from the School of Chemistry, The Uniuersity of Western Australia, Nedlands, Western Australia. Received September 6, 1966 Abstract: The isomerization of cis- and trun~-[CoCl~(en)~] ions has been studied in sulfolane (tetramethylene sulfone). No evidence has been found for solvent-containing complexes. The mechanism of isomerization for both species is interpreted as Ssl,the rates being influenced by ion pairing. Activation parameters have been obtained for the isomerization reactions commencing from both cis and trans isomers, and in addition the equilibrium constants and the standard-state enthalpy and entropy changes for this isomerization and for the formation of the ion with chloride ion have been measured. An energy-profile diagram for the ion pair of the ci~-[CoCl,(en),]~ system is presented, and correlation of the thermodynamic and kinetic results IS discussed. +
his work is a continuation of the work of this T g r o u p l - g and Tobe and on reactions of octahedral cobalt(II1) complexes in dipolar aprotic solvents. (1) L. F. Chin, W. A. Millen, and D. W. Watts, Austruliun J . Chsm., 18, 453 (1965). (2) W. A. Millen and D. W. Watts, ibid., 19, 43 (1966). (3) W. A. Millen and D. W. Watts, ibid., 19, 51 (1966). (4) W. R. Fitzgerald and D. W. Watts, ibid., 19, 935 (1966). (5) I. R. Lantzke and D. W. Watts, ibid., 19, 949 (1966). ( 6 ) I. R. Lantzke and D. W. Watts, ibid., 19, 969 (1966). (7) W. R. Fitzgerald and D. W. Watts, ibid., 19, 1411 (1966). (8) I. R. Lantzke and D. W. Watts, submitted for publication. (9) I. R. Lantzke and D. W. Watts, submitted for publication. (10) M. L. Tobe and D. W. Watts, J . Chem. SOC.,4616 (1962). (11) M. L. Tobe and D. W. Watts, ibid., 2991 (1964).
Previous studies of equilibria of the type ci~-[CoX~(en)~]+ = trun~-[CoX~(en)~]+
where "en" represents ethylenediamine and "X," a halide ion, in various solvents have been complicated by a number of factors. In dimethyl sulfoxide" and dimethylf~rmamide~~b the solvent (SOL) containing species [ C O X ( S O L ) ( ~ ~ ) ~is] *an + appreciable part of the equilibrium mixture. In the case of the [CoCh(en)s]+ species in dimethylacetamide, l 2 where no significant amounts of solvent-containing species are formed, (12) I. K. Lmtzke, unpublished results.
Fitzgrruld, Wutts / Isomeriiution of cis- cind trans- Dictilurobis(etk~lenediamine)cobalt(III)Ions
822
subsequent reduction reactions make equilibrium measurements inaccurate.10'12 Equilibria involving the [ C ~ C l ~ ( e n > ~species ]+ in methanol13 are not complicated by the presence of methanol-containing complexes. However, in methanol, the ci~-[CoCl~(en)~]+ is not a significant part of the system at equilibrium, and thus equilibrium studies are not practicable. In dipolar aprotic solvents, the cis ion is stabilized by ion-pair formation and by solvation of the dipolar cis isomer by the dipolar solvent, In the protic solvent methanol, anions are more strongly solvated owing to hydrogen bonding,14 so that ion pairing is less. [E.g., the ion association constant for the . . C1- ion pair in methanol (dielectric ci~-[CoCl~(en)~]+ constant = 33.214) is 13513 at 25" and in dimethylformamide (dielectric constant = 37.614) is 5500 at 25 ".? The present results in sulfolane (dielectric constant = 4414) extrapolate to a value of 53,000 at 25O.1 In the [ C ~ C l ~ ( e n ) system ~ ] + in sulfolane, no detectable amounts of solvent-containing complex are formed, and the ratio in concentration of the cis and trans ions is always measurable and varies considerably with chloride ion concentration. There is no evidence of subsequent reactions, such as reduction, in the time required to establish the isomerization equilibrium. For these reasons this system is an ideal model for the study of isomerization kinetics and equilibria.
were added by weight. The density of sulfolane (1.2623l5) was then used to find volume concentrations. Spectrophotometric measurements were made using a Unicam SP 500 spectrophotometer fitted with a thermostated cell compartment. Measurements were made at 3150 and 3400 A. The cis[CoCln(en)n]+ion is convenient for this study in sulfolane because the spectra of the cis-[CoC12(en)2]+ion and the cis-[CoC12(en)2]+.. . CI- ion pair have an isosbestic point at 3400 A, and the spectrum of the cis-[CoClp(en)?]+ion is flat around 3150 A. The extinction coefficients of the free ion and the ion pair at 3150 and 3400 A at the three temperatures used (30.2, 40.2, and 49.8") are recorded in Table I.
Table I. Ion-Pair Association Constants for the cis-[CoClz(en)2]A Ion and Chloride Ion in Sulfolane Temp, "C 30.2 40.2 49.8
CCIP
ECIP =
€0
€0
(315 mp) (315 mp) (340 mp) 1282 1316 1348
787 795 804
616 623 628
K2,1. mole-' 42,000 A 800 29,000 f 600 21 ,ooo=t400
Results and Discussion Reproducible equilibria are found in the reactions of both cis- and frans-[C~Cl~(en>~]+ in sulfolane, the positions of equilibrium being independent of the starting material. Because the trans isomer was studied over a greater range of chloride ion concentration, the equilibrium results presented are those using frans[ C ~ C l ~ ( e n )as ~ ]starting + material. Experimental Section Figure 1 shows the variation with chloride ion con(a) Preparation of Compounds and Purification of Solvents. centration of the position of this equilibrium at 70". The cis- and tr~ns-[CoCI,(en)~]Clwere from batches used preThe results are similar to those obtained for other viouslyj and were recrystallized as the perchlorates. systems, the cis isomer being stabilized by ion&id. Calcd for cis-[CoCl~(en)$2104: complexed C1, 20.29. Found : complexed CI, 20.23. Calcd for trans-[CoCl~(en)~]CIO~: pair formation. complexed C1: 20.29. Found: complexed C1,20.31. The system can be represented as previously1° by the Tetraethylammonium chloride was recrystallized from hot following equations dimcthylacetamide. washed with acetone and ether, and dried under 4358
vacuum for 6 hr at room temperature. Sulfolane was purified and dried by shaking with 4-A molecular sieves and allowing the mixture to stand with the sieves overnight at 30". The solvent was then distilled under reduced pressure through a fractionating column, the middle 7 0 x being collected. With very impure batches a second distillation was required. Scveral attempts were made to prepare sulfolane-containing complexes by methods previously used for the preparation of the analogous campounds containing dipolar aprotic solvents,' but as qct no compounds have been isolated. (b) Kinetic Technique. The rates of isomerization were studied using the wavelength 5420 A. At this wavelength, the molar e\tinction coefficient of the cis-[CoCL(en)?]' ion is 102.5 and of the ri.cii~,~-[CoCI~(en)~]+ ion is 7.3. T o check if a third component was presznt at any stage, measurements were also made at the isosbestic ions at point in the spectra of the cis- and the tr~ns-[CoC12(en)~]+ 4700 A at which the extinction coefficient is 29.0. For r u n s starting with cis-[CoCL(en)s]', samples were withdra\%n at known times, frozen rapidly, and, when convenient, thewed: their spectra were measured on an Optica 220 double-beam automatic spectrophotometer thermostated at 30". During sampling. the reaction vessel was flushed with solvent-saturated dry nitrogen in order to exclude moisture. With reactions of tr.nns-[C~Cl~(en)~]+ the complex precipitated as the chloride salt when samples were cooled. Thus these reactions were studied in an absorption cell thermostated in the spectrophotometer. Even at 80", the temperature fluctuation was no more than 1 0 . 1s . (c) Technique for Obtaining Ion-Pair Association Constants.
The method used was similar to that described by Millen and W a t t s 2 However, as sulfolane cannot be accurately dispensed volumetrically at room temperature, the stock solutions of anion and complex (13) R . G. Pearson, P. M. Henry, and F. Basolo, J . A m . Chem. SOC., 79, 5382 (1957). (14) A. J. Parker, Quart. Reu. (London), 16, 163 (1962).
Journal of the Americun Chemical Society
/ 89:4 / February
1: + (1) tr~ns-[CoCL(en)~]+ + kt + c
cis-[CoCI2(en)2]+
+
Cl. .
cis-[CoCl2(en)y]*. .C1-
Cl-
tu~ns-[CoCl~(en)~]+. . 'C1-
where K1, Kz, and K3 are equilibrium constants for reactions 1, 2, and 3, and k, and k , are initial first-order rate constants for the removal of cis-[CoCl~(en)~]+ and trans[ C ~ C l ~ ( e n ) ~respectively ]+, (Kl = k,/k,). Under conditions of chloride ion concentration in which the cis isomer exists almost exclusively as the ion pair cisC1-, the forward rate constant is sym[ C ~ C l ~ ( e n )* ~. ]' + bolized kcIp. These subscripts are maintained throughout the discussion and apply to all relevant thermodynamic and activation parameters. (e.g., AHOZ is the standard enthalpy of the ion-pair reaction 2 and AH°CIP* the standard enthalpy of activation for the removal of the cis isomer under conditions of chloride concentration in which formation of the ion pair is comple te) . Values of Kl, K,, and K 3 have been calculated as previouslylo from a consideration of the cis and trans equilibrium results and their chloride concentration dependence and are 36 + 5, 4800 ==I 700 1. mole-', and 4 i 1 1. mole-' at 70". (15) R. Fernandez-Prini and J. E. Prue, Trans. Faraday Soc., 62, 1257 (1966).
15, 1967
823 l o
I
bI
-
Fraction of t o t a l complex
9 Rate cons! "?)
0
a
mu-
05 .
0
50 103
50
0 103
[cq
(mol0
t 3
100
Figure 1. Equilibria of cis- and rruns-[CoCl2(en)~]+in sulfolane at 70": total complex concentration 4 X M ; 0 , fraction of fraction of rruns-[CoC12(en)2]+. cis-[CoCly(en)g]+; 0,
The value of KBis almost certainly too low because of difficulties associated with the ion pairing of chloride with the tetraethylammonium cation. However, it is clear as in other ~ y s t e m s , ~ ~ that * ~ ~the ~ ~ cis 0 ~ ion ~ ' is appreciably more ion paired than the trans ion. No account is taken of the activity coefficients of the various species in the evaluation of Kz. For the ion pair which has no net charge, the activity coefficient will be close to unity. However, for the cis-[CoClz(en)z]+ and the C1- ions, the activity coefficients will be about 0.75 at the ionic strengths used (15 X le3 M) according to the Debye-Hiickel limiting law. (This value of 0.75 has been obtained using the dielectric constant for sulfolane at 30O.) Therefore, the value of the ion-pair constant (Kz)expressed in the form Kz
=
fion pair] [free ion][Cl-]fih
where f l and f2 are the activity coefficients for the free complex ion and the chloride ion, will be approximately 9000 1. mole-' at 70". This value based on the displacement of the isomerization equilibrium agrees reasonably well with the more precise value obtained from direct measurements on the ion-pair equilibrium discussed below. From this a value of 11,000 f 300 was obtained for Kz at 70". The initial rates of isomerization of the cis and the trans isomers (k, and k,) at various chloride concentrations are shown in Figure 2. The rates show firstorder characteristics and are affected by ion-pair formation. It is expected that the rate of isomerization of the trans isomer would eventually flatten off as does the cis rate if the chloride concentration could be raised sufficiently to bring about complete ion pairing. A feature of the reactions studied in sulfolane is that they are very much slower than the corresponding reactions in other solvents, particularly hydroxylic solvents. This correlates with expected poorer solvation of the dissociating anion by sulfolane in the formation of a trigonal bipyramidal intermediate. For example, the half-life for isomerization of cis-[CoClz(en)z]+ at zero Fitzgerald, Watts
100
[cil ( m d o i'1
Figure 2. Dependence of k, and kt on chloride concentration: total complex concentration 4 X M ; 0 , k , at 80"; 0 , kt at 70".
chloride concentration in methanolI6 is 8.3 min and in dimethylacetamidel2 is 350 min at 60", compared with 1310 min for the present work in sulfolane. Table I1 shows activation parameters for isomerization of both the trans and the cis isomers at various chloride ion concentrations. It can be seen that the activation energy for isomerization of the cis-[CoClz(en)z]+ ion is unaffected by ion-pair formation. The amount of isomerization of the trans-[CoCl~(en)~]+ ion at zero chloride concentration is too small to enable a rate constant to be obtained, and it is thus not possible to obtain an activation energy under these conditions. The values of AGOl for the isomerization of cis[ C ~ C l ~ ( e n ) at ~ ] zero + chloride concentration at various temperatures are listed in Table 111. These values of AGOl were calculated from spectrophotometric measurements on equilibrium mixtures obtained from the reaction of cis-[CoClz(en)z]+at various temperatures. N o significant variation of AGOl with temperature is observed so that AGO' = AH"' = -2.5 f 0.1 kcalmole-' and AS", = 0 f 0.5 cal deg-' mole-'. The energy-profile diagram for the isomerization at zero chloride concentration is shown in Figure 3. From the values AH",* = 25.0 f 0.2 kcal mole-' and AHOt = -2.5 f 0.1 kcal mole-' for cis-[C~Cl~(en)~]+ isomerization at zero chloride concentration, the value AHot* = 27.5 f 0.3 kcal mole-' is obtained for trans[ C ~ C l ~ ( e n ) ~isomerization ]+ at zero chloride. The activation energy for tr~ns-[CoCl~(en>~]+ measured at 4.3 X le3 M chloride is AH",* = 20.3 A 0.2 kcal mole-'. This discrepancy can be explained only by stabilization of the transition state by ion association. The values of KZ at various temperatures are listed in Table I. These values were obtained from spectrophotometric results as previously described2 and yield AHoz= -7.1 f 0.1 kcal mole-'and ASoz = -2.5 f 0.3 cal deg-' mole-'. That the activation energy for cis isomerization is not affected by ion-pair formation means the transition state is stabilized in ion association by the same energy as the reacting ~is-[CoC12(en)~]+ ion. Thus the energy of the transition state is lowered by 7.1 kcal mole-' (16) D. D. Brown and R.S.Nyholm,J. Chem. SOC.,2696 (1953).
1 Isomerization of cis- and trans- Dichlorobis(ethylenediamine)cobalt(ZII Ions
824 Table 11. Measured Activation Parameters in Sulfolane 10yc1-],
M
AH",+-
0 4 3 30 0
25 O f 0 2
-AHo+, kcal mole-'AH'cir*
AS=c*
AH"tF
ASc*, cal deg-1 mole-'---------LS=c*r*
-6 4 f 0 5
25 O f 0 2 25 O f 0 2
-4 7 1 0 5 -3 1 1 0 5
20 2 f 0 2
Table 111. Values of AGC1at Various Temperatures for the Isomerization of ~is-[CoCl~(en)~]+ at [CI -1 70 ' AGO, kcal mole-'
-2.4+0.1
80" -2.5k0.1
=
-18 4 f O 5
0 in Sulfolane
90
100"
116"
-2.4-f 0.1
- 2 . 5 i0 . 1
-2.6 & 0.1
when it is associated with chloride ion. At fairly low concentrations of tetraethylammonium chloride (-5 X 10-3 M ) , the ion pairing of the trans isomer will be small (
