Kinetics and mechanism of substitution reactions of

0 downloads 0 Views 574KB Size Report
Kinetics and mechanism of substitution reactions of ... Siniple anation and solvent exchange reactions of Co(NH3)jDMSO3+ in DMSO are acconi- panied by ...... F. BASOLO and R. G. PEARSON. Mechanisms of inorganic reactions. 2nd ed.
Kinetics and mechanism of substitution reactions of (dimetEnylsulfoxide)pentaamminecobalt(III)iorm in dimethylsulfoxide solvent: complications due to redox and conjugate base processes S. T. DANNYLO, EVELYN bf. OUDEMAN, JEANC. HANSEK~ AND THOMAS W. SWADDLE' Depurrr71er7iof Clrer77isr~~., The L'riicet..\i!, of Culgar!,, Calgurj.. Altu., Cur7adu T21\'

IN4

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

Received June 17, 1976 S. T. DAKNYLO, EVELYN M. OUDEMAN, JEANC. HANSEX, and THOMAS W. SWADDLE. Can. J. Chem. 54, 3685 (1976). Siniple anation and solvent exchange reactions of Co(NH3)jDMSO3+ in DMSO are acconipanied by reactions in which the conjugate base of this coniplex undergoes either internal redox to cobalt(11) or relatively rapid substitution. These conjugate base reactions are eliminated by addition of H+, although a minor redox pathuay persists Llithin ion pairs of C O ( N H ~ ) ~ D M S O ~ + with chloride or bromide. The latter pathway is evidently not niechanistically related to the solvent exchange or anation reactions. which proceed by dissociative interchange (Id) according to the acti\ation parameters (\olun~eof activation = +10 cm3 mol-1 for DMSO exchange; enthalpy of activation = 123. 121, and --I26 kJ mol-1 for DMSO exchange and bromide and chloride anation respectively). Enthaipies of activation for solkent exchange and for the Iiniiting anation rate are shown to be better criteria of rnechanianl than the corresponding rate coeficients. S. T. DANNYLO, EVELYN M. OUDEMAN, JEANC. HANSENet THOMAS W. SWADDLE. Can. J. Chem. 54, 3685 (1976). O ~ +le DMSO Les reactions d'anation simple et d'Cchange de solvant du C O ( N H ~ ) ~ D M S dans sont accon~pagn&spar des rCactions dans lesquelles la base conjuguee de ce complexe subit soit une reaction redox interne conduisant B du cobalt(l1) ou des reactions de substitution relativement rapide. On elitnine ces reactions des bases conjuguees par addition de H+; toutefois un chemin mineur de rCductio11 persiste St l'interieur des paires d'ions du C O ( N H ~ ) ~ D M S O ~ + avec le chlorure ou le bron~ure. I1 est evident que cette dernitre voie n'est pas reliee d'une fafon mCcanistique avec 1'Cchange de solvant ou les reactions d'anation qui precedent par un echange dissociatif (Id) si l'on se fie aux parametres d'activation (volume d'activation = +10 cn13 mol-1 pour 1'Cchange DMSO; l'enthalpie d'activation = 123, 121 et -126 kJ mol-1 pour les Cchanges avec le DMSO et d'anation du chlorure et du bromure). On montre que les enthalpies d'activation pour l'echange du solvant et pour la vitesse limite l'etape determinante de I'anation sont des meilleurs critkres pour Ctablir 1e micanisme que les coefficients correspondants de vitesse. [Traduit par le journal]

Introduction We recently summarized (1) the evidence which indicates that, when water is the solvent, simple substitution reactions of cobalt(lI1) ammines [I] CofNH3)5(sol\ent)3-

KIP + Yn-+ / C ~ ( N H ; ) ~ ( s o l \ e n t )Yn-) ;~,

ILi C O ( N H ~ ) ~ Y ( ~f - solvent ~)+

proceed by a dissociative interchange (Id) mechanism (2), that is, one in which Co-solvent bond breaking within the ion pair or encounter complex (C~(IaiH~)~(solvent)~+, Yn-) is the rate 1 Author

to whom correspondence should be addressed.

determining process. We also pointed out, however, that cobalt(I1I) complexes may be anomalous in this respect, since the evidence collected to date suggests that, in general, octahedral cationic complexes of other trivalent transition metals M(II1) undergo simple ligand substitution by an associative interchange (I,) mechanism, in which bond making by Ynwithin the encounter complex is involved in the rate determining process (1). We now report on our attempts to extend the data available for reaction 1 to include dimethylsulfoxide (DMSO) as the solvent. This solvent was chosen, partly because of the considerable recent interest (3-10) in the complex C O ( N H ~ ) ~ D M S 0 3 in ~ itself, but mainly because the lower dielectric constant of DMSO relative to water

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

3686

C A N . J. CHEM. VOL. 53, 1976

was expected to allow ion pairing t o proceed Y n - is varied. This is because the enthalpic part essentially t o co~npletiona t practical concentra- of the free energy of activation is alwiys that tions of Y7'-and so permit the limiting first order required to break the same metal-solvent bond, rate coefficients kito be measured directly, this within a given Id series; clearly, no such conbeing inlpossible in water for rnononega- stancy of enthalpies of activation would be tive Y-. Moreover, conlplexes of the type expected for an I, process, and indeed the latter I v I ( N H ~ ) ~ Y ( ~ - are " ) + known for a wide variety behavior characterizes 1, substitution at of M(III), and it was intended that the present chromium(1II) centers (12). We present here evidence that the mechanism study would form the basis of a n extended of reaction I does remain Id on going from water n comparative study of ligand s ~ ~ b s t i t u t i omechanisms in nonaqueous solvents in relation t o the to DMSO as solvent, and that refinement of the nature of M. In addition, there was the possi- Langford-Muir criterion in terms of enthalpies bility that the predominant mechanism of reac- of activation does appear to be necessary for DMSO. Unfortunately, our experiments were tion 1 might change on changing the solvent. Finally, the relatively large size and bulky complicated by unexpected conjugate base reacshape of the DMSO molecule raise the possi- tion pathways and redox phenomena, but these bility that the need for a refinement of the are of interest in thenlselves and are of importLangford-Muir criterion of an Id mechanisni ance in considering past and future studies in (1 I) might become apparent for reaction 1 when this field. DMSO is the solvent (footnote 3 of ref. 12). Experimental Where the solvent molecules (e.g. water) are small, movements t o and from the reaction site .?.Ii~ierials are relatively free, and Co-solvent bond breaking Tetraethylammonium ciiloride anci broniide (Baker will generally lead either to net exchange with Cheni. Co.) ucre recrystallized from hot tlimethylthe ubiquitous solvent molecules or to unseiec- formamide ( D M F ) by adding ether, anci were dried ~ ~ n d e r Iiigh \ acuuni. Anal) tical reagent grade p-tolue~iesulfonic tive scavenging of Y n - within the encounter acid hydrate (Fisher Scientific Company) was used complex. One may therefore expect, with Lang- without further recry~tallization.Dirnetli) lsulfoxide (ACS ford and Muir (1 l), that ki will be about the grade, Fisher Chemical Corp.) and DbISO-c16 (Merck, same for all Y n - in a given solvent, and less than Sharp and Dohme Canada Ltd.) \\ere puritied by a (or, at most, equal to) the first order rate coeffi- methoci analogo~tit o the purification of D M F as described elsewliere (13). cient k,, for solvent exchange in free CO(NI-T~)~- [Co(NH3)5D.LISO](C104)3. 2 H Z 0 was prepared anti re(~olvent)~', in the absence of some general cry5tallized by the method of Mac-Coll ant1 Beyer (10). O , ) ~prelabilization of the complex on forming a n ion [Co(NH3)jCI](C104)2and [ C O ( N H ~ ) ~ B ~ ] ( C Iwere pair. These expectations may not, however, be pared as deicrihed previously (14). The purity of these realized if the bulkiness of the solvent molecules coniplexei wa5 clieclted by micioa~ialysis. rZ1~~1~1it~rnle~ifs resulted in a high probability of recombination Ki/i~~iic The progress of the exchange reaction between the cowith the same solvent molecule, rather than a O ] ~D.\lSO+ ordinated D hfSO-i16 in [ C O ( W H ~ ) ~ D ~ I Swith dijferent one, after cobalt-solvent bond fission, ri6 solbent u a s follo~vedby remo\ing samplcs at appros o that the measured value of k,, would be con- priate intervals from thermostatted solutions of the siderably lower than the true bond breaking rate perchlorate salt. freezing then? ~111tilrequired, and recoefficient. Similarly, access to the reaction site cording the pnir spectra of the thawed sample5 with a would be easier for small than for large Yn-. Varian A-60 spectrometer at roorii temperature. The height of the coordinated DMSO-/i6 peak (7 = 7.27 ppm) Thus, ki could vary somewhat with the nature was nieasured relative t o that of the NH3 resonance of Yn- and could exceed k,,, especially when (7 = 6.07). The pressure dependence of the reaction rate Y n - is small, even though the mechanism re- was similarLy measured by removing samples from a premained strictly Id. These complications are seen pressurized syringe, as described previously (14). The progress of the anation reaction^ \$as f o l l o ~ e d t o be essentially statistical, i.e., entropic, in spectropiiotometr-icaliy using a Cary Model 17H spectroorigin, and serve to emphasize that the ger~eral photometer as described elsewhere (12). The relevant kinetic charucteristic of cin Id nzechanism is the molar absorptivity (t) data are recorded in Table I . conslaizcj of rhe enfha!pj of actirjation clssociutecl Cohiilr(1l) A1ia1j.sc.s with the rate cocficietlrs k, ~ n c lk,, throughout The concentration of cobait([l)in the reaction mixtures a series of reactions such as reaction I in which was determined ipectroscopicaily a i either CoCI4'-

TABLEI . Ylolar absorpti~ities(in din; mol-1 cni-1) of perchiorate sa!ts of pentaan~n~inecobalr(lll) coniplexes in DI\ISO hIolar absorptivity at wa\elengli~inrn)

-

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

Coniplex

312

"[cornpiex]= [p-toluci~csulfon~c acid] ii[con~plexl= [p-tolucnc~ulfonicacid]

419"

= =

50ja

544"

( I to 2) X 10F3 n?ol d,n-3: [CI-1 mol dm-" [[Br-] (1 to 2) X

(ah\orption maximum in DMSO at 6 8 3 nni u'ith t = 30-1 dm; i~lol-1 cm -1 j or CoBr4'- (n:axini~ii??ii: DXlSO at 703 nil? ~ ~ i ttl l= 108.9 dm3 moi-1 cin-1). In some cases, rile cob;ilt(il) content was determined gravirnetricall~a s clit!lioc) anatotetra~~yridinecobait(11 j (I 5).

54Ab

2

----

j80b

~[cuniplex]. ijcrn:?>;,icr]

redox decompositinn p h e n o n ~ e ~ The ~ a . additin11 of hjdrogen icin !as ilie poorly con-~plexirig p-toluenesulfonic or trifiuoror?~ethrrnes!~lfnnic acids), however, eliniil?attcl cnbal!i!l) formatior;, except where chloride or bromide ions uere Results present, in which crises reduction of the cobalt In preliminary experiments, the anat;1011reac- center \,bas still greatly diminiilred. Tile relative tion (reaction 1 , with Yn- = Cl-, Br-: NCS-; amount of cobalt(il) present in tiles.- latter cases solvent = IIMSO) was follo\a;ed spectrophoto- at tile end of the subsiitilrioll reactions incrcascd metrically (at 340, 580, and 3 12 nm, respectii,eiy) with increasing halide ion conch-ntrntisn, hut at 45 ' C , ~lsingonly [CO(NI-I~)~DMSO](CIO~)~ and excess (C2Hj)&C1, (C2615)~NBr,or KNCS as solutes. In no case was the expected first order depmckiice on [ C O ( N H ~ ) ~ D M S O ~observed '] beyond a few percent reaction; rather, the reaction appeared to be autocatalytic (see, for example, Fig. l), most notably when Y n - was NCS--, and also was slightly accelerated by exposure t o the room lights at ieast when - f v L was Br-. Nevertheless, the initial rates of the reactions in the dark suggested apgroxinlate ~ a l ~ of ~ ethe s limiting first order rate coeficient ki of (1.6, i .0, and 1.2) S 10-4 s-I at 45 "C for Yn- = C1-, Br-, and NCS-. Closer examination of the prodilc?~of reaction 1 arter completion at 45 'C, however, revealed the presence of 97; and 7% cobalt(i1) along with C O ( N H ~ ) ~ Y ( ~ -for " ) -Y1+= Cl- and Br- respectively. while, over the same tirile but in the absence of other solutes, CO(NH;)~L)MSO~+ and C O ( N H ~ ) ~ E perchlorates ;~~+ yielded 1 I yc and 4 7 cobalt(I1) respectively in DMSO at 45 "C. Great care was taken to free the starting materials of cobalt(1B) by recrystallization, and I i. ____0.6Lit was established that the reaction rates were 0 50 900 independent of the mode of purification of the T i m e ",min l solvent and of the presence or absence of added FIG 1. Reaction of Co(Nh';jjDMSO3- (2.0 r~imoi I O ~ ) neither ~. water or [ C O ( D I ~ ~ S ~ ) ~ ] ( CThus, dm-3) with E r (15.0 rn111oi cim-3) ill EXZSO ar 45.0 - C cobalt(i1) zlorle nor trace contaminants were without added acid. Optical zbiorliar?ce; A, n~easuredat responsible for the autocatalytic substi"itio11 and 580 nm.

3688

.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

y TL -

C A N . 1. CHEM. VOL. 54, 1976

TABLE2 Relati\e aniounts of Co(l1). Co(NH3),DMS03'. and C O ( N H ~ ) ~ nYj f' ~pre5ent at the completion of reaction 1 in DMSO solbent at 50 7 C ln~tial Initial Initial Final Final Final [ C O ( N I H ~ ) ~ D ~ ~ S[CH3C6H4S03H] O~-]~ [(C2H5)4NY] [Co'-]b [ C O ( N H ~ ) ~ D M S O ? - ][ C O ( N H ~ ) ~ Y ' + ] (nimol dm-3) (nirno! dm-?) (mrnol dnir3) (7,) (%,) (77,)

"As perchlorate snit. hDetermined spi-ctrophotometrically a s CoY,I-; uncertainty, k 10%. CUsing [Ca(h'I-li)jYl(C!O~):a s starting m a t e r ~ a ! ,1.0 mmol dm-'. ~ [ C o ( r ; H ~ ) j C l ] C 1which 2, h a s piecipitdtcd o n completion o f the reaction, was removed by filtration before determination of Co(l1). eAt 80.0 "C,

was independent of [HA] so long as this was on the order of the initial [Co(lll)], and did not increase further after the end of the substitution reaction; it was shown separately that Co(NH3)CI2- and C ~ ( T z d i d ~ ) ~ B did r ~not + undergo reduction ur;der these experimental conditions. These observations, ~ h i c hwere quantitatively reproducible, are summarized in Table 2: and indicate that reduction of cobalt(1ll) takes place primarily through a conjugate base species, presumabi) Co(N%-f3)4jNHz)DMSOi-, H hich can be expected also to undergo substitution relatively rapidly (1, 2, 46).

Reduction to cobalt(l1) results in labilization of the cobalt center leading to ammonia release [3] C O ~ N H ~ ) ~ ( N H ~ ) D M S>- O ? '

Co2' $ 5NM3

+ oxidized DblSO

which forces reaciion 2 to the right and so is autocatalytic. The persistence of a small degree of cobalt(1l) folrnation under acidic conditions when Ci- or Br- were present b a s significant orily at the lower reaction temperatures, and it was shown that the presence of small amounts

of cobalt(I1) did not affect the rate of reaction 1. In the light of these findings, all further rate measurenlents were carried out with the addition of acid. For the anation reactions (reaction 1, y n - - Cl- or Br-, solvent = DMSO), plots of In (A, - A,) us. time t , where A, is the optical absorbance at time 1, were then linear over the first 505; reaction for 'in-= Br- (later, A , changed slightly 1c.s.~rapidly than the first order rate law required), and over the first 6 5 4 0 % for y n - - Cl-. Exposure to the room lighting did not affect the reaction rates. The anation reaction proceeded effectively to completion for Yn- = Cl-, so that the observed pseudo-first order rate coefficient kobs equalled the anation rate coeficient k,,,, but for Yn- = Br-, anation was not quite complete (cf: Table I), and kobs was the sum of k,,, (which was directly estimable from the initial reaction rate, on the assumption of eventual complete reaction) and the reverse (solvolysis) rate coeficient, k,,,. These data are collected in Tables 3 and 4. The values of k,,, indicate limiting values ki at 45 "C of approximately 1.4 X s-I for Yn- = CI-and, by interpolation, 4.2 X s-I for Yn- = Br-. The latter value is less than half that observed in the absence of added acid, which

3689

LO ET AL.

TABLE3. Pseudo-first order rate coeffic~entsLobs, Lan, and k,,, the format~onof C O ( N H ~ ) ~ B ~ " from Co(NH,)5D.MSO?+ and Br- In DMSO [ C O ( N H ~ ) ~ D ~ ~ S[CH3C6H4S03H] O~*]~ [ ( C Z H ~ ) ~ N Temperature B~] (mmol dm-') (nimol dnir3) (rnmol dm-?) (cC)

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

1.0l5

1.053

117.9

55.7 60.8 40.0 50.7

85.9 57.6 29.7 19.0 9.2

105h,byh

I 05(X,,

(5-I1

19.0 + 0 . 2 18.5 5 0 . 2 37.5 5 0 . 7 1 ,9510.03 10 51.0 10.1 rk0.2 9.01 10.2 7.5 ~ 0 . 2 5.6 k0.1 4.310.08

1O5kSold

(5-1 )

11.2=0.2 10.310.2 9.350.2 7.410.2 5.4~0.i

(sSi)

1.1 1.3 1.8 1.8 I .I

aAs the perchlorate salt. bFrom initial reaction rate, assuming 100% co(NH3)1Br'- at infinite time. CFrom plots of in ( A - A t ) us. t , using experimental A m ; represents sum of forward (ksn)and back (h,,i) reaction rate coefficients. dBy difference.

TABLE 4. Pseudo-first order rate coeficlents Lobs for the forniatlon of C O ( N H ~ ) ~ Cfrom I ~ - C O ( N H ~ ) ~ D M S and O ~ -Cl- In DbISO [C~(NH~)SD~~SO [CH3C6H4S0,H] ~-]~ Temperature [(C2H5)4NC1] (rnniol dm-3) (rnmol dm-)) ( C) (rnn~oldm-')

105X,hsb (>-I)

'JAs the perchlorate salt. bExperimenta1 uncertainty 1 2 % .

indicates that a major portion of the bromide anation reaction proceeds via the conjugate base species of reaction 2 ~vhenno acid is added. The value of ki for chloride anation is also slightly less than in the absence of added acid. The ki data for Yn- Br- gave a linear Arrhenius plot with activation energy E, 123 kJ mol-I (corresponding to AHi* = 121 2 3 kJ mol-I, ASi* = +51 i- 9 J K-I mol-I); the corresponding data forYn- = Cl- are not available because of the poor solubility of [Co(NH3)5CI]C12at high [Cl-1, but the temperature dependence of kOb, over a limited (10 " C ) range at [Cl-] = 0.02 mol dm-3 indicates AHi* 126 kJ mol-I.

-

-

-

The solvent exchange reaction (Yn- = solvent DMSO in reaction 1) proceeded according to first order kinetics over 60-80% reaction, in the presence of added p-tol~~enesulfonic acid, and no cobalt(I1) could be detected even after 10 half-lives of the exchange reaction. The values of the rate coefficients k,, listed in Table 5 were obtained from the first order rate equation with a standard deviation of about 47;, on the average, but reproducibility was about i9$,. Plots of In k,, against 1; T and against pressure P were linear within the experimental uncertainty; the relevant parameters and their standard errors are AHe,* = 123 F 2 kJ mol-l, ASe,* = 61 _+ 6 =

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

3690

CAN. J. CHEM.

VOL. 54,

1976

TABLE5. Pseudo-first order rate coefficients for the mechanism (12). It would also see111 that the exchange of DhfSO-ci6 between Co(NH,)5DMSO-~i637a expressed in the ~ ~ , t ~ and ~ and DlbiSO-h6 solvent in the presence of elsewhere (121, regarding the relative valucs of p-toluenesulfontc acidD k,, and the various kiin an I, mechanism in a _ _ _ _ 'Temperature Freisare los~,,d solvent such as DMSO, are justified, since, a1 ("4 (MI'a)c (s-1) 45 OC, ki for chloride anation exceeds k,, for solvent exchange by a factor of 2.2 and the 55.7 (3. I interpolated k, for the larger anion, bromide 29" 14.8 50.7 (4.2 x 10-"-I), by a factor of 3.3. Similar 6,35 45 .O studies of reaction 1 with D M F as solvent, 1 . 35 .O

45.0

50.7 --

5.14

101.3

4.38

202.7

2.Y3

-

aConcentration 0.39 i 0.01 mol dm-'. bConcentration 0.41 5 0.01 mol dm-'. c0.1 M P R = 1 bar = 3.9869 a t m . dStandard deviations typically lt 4 % ; reproducibility, i9 %.

+

J M-I mol-l, and AY,,* = 10.0 rt_: 1.2 cm3 nol-l. The value of k,, a t 45 OC and 0.1 MPa is less than one-fifth of that reported by Reynolds et al. (8) for the same conditions but without added acid. This discrepancy is attributable t o the incursion of the conjugate base pathway in Reynolds9 experiments. The data of Table 2 ind~catethat it is Co(NH3)5DMS03-, and not C O ( N I - I ~ ) ~ C ~or~ - Co(IUH3)5BrzL, which is susceptible t o redox deco~nposition; accordingly, 'autocatalysis' of solvent exchange can be expected t o be even more marked than that of anatiorr, and it may be significant that Reynolds et al. (8) quote an unusuallji large experimental uncertainty (>20%) in k,,. The strongly positive value of Aye,* leaves little doubt that Co-DMSO bond breaking is the rate determining process in the solvent exchange reaction (I), and hence consideration of the generalized reaction 1 in the light of the principle of microscopic reversibility, together with experience with other systems (121, leads t o the expectation that the anation reactions will also proceed via an Id mechanism in DMSO just as in water in the absence of interfering effects (see below). This inference receives strong support from the fact that A H i * for anation by bromide and (with less certainty) chloride, and A&,* for the solvent exchange reaction, are the same within the experimental error, whereas they differ markedly for the analogous reactions of CI'(DMP;')~" in D M F solvent, a syste1-n in which substitution evidently proceeds by an I,

which we wil! report later, confirm that equality of AHi* and AH,,", rather than a close interrelationship of ki and k,,, is characteristic of an Id prccess. The foregoing conclusions refer to reaction 1 in DMSO solvent to which acid has been added, and appear to answer the questions originally posed. Unfcrtunately, the fact that conjugate base and redox reactions can be important, while of interest in itself, obliges us to reexamine those sinlplistic conclusions. Diniethylsuifoxide is somewhat more basic towards PI- than is water, but much less acidic (17, 181, and accordingly its a~~toprotolysis con(18-21).2 It is stant is about to difficult to generalize about the strengths of electrically neutral acid ~nolecules in DMSO, but it appears that the intrinsic basicity of the solvent, favoring ionization, is offset by the relatively low dielectric constant, so that most inorganic acids and organic oxyacids are sornewhat less dissociated in DMSO than in water (20-24). Protonic cations such as Co(NH3)5DMS03+, however, should be more acidic in DMSO than in water, since the lower dielectric constant should f l ~ a o r proton release on the basis of electrostatics. In retrospect, therefore, it is not surprising that cobalt(1II) ammine conjugate base formation (reaction 2) makes significant contributions to the kinetics of the anation and solvent exchange reactions in 'neutral' DMSO. These effects are absent from the Cr(DMSO)63+, DMSO, Vn- system (25,26), and, while recognizing that cobalt(ll1) is more suscqtible to conjugate base effects than is chromiunl(III), one may reasonably suppose that the conjugate base of C O ( N H ~ ) ~ D M S Oforms ~ + by deprotonation of an N H 3 ligand, rather than the DMSO, but an ambiguity does exist. >The value of 5 X 10-18, giken by KolthofF anti Reddy (17), seems to be anomaious.

d ~ ~ t i

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

LO ET AL.

3691

It is also clear from the data of Table 2 cobalt(I1) intermediates in the base hydrolysis that the conjugate base species provides the of C O ( N H ~ ) ~ C(3~ 1). ~ ' Thus, although redox main redox pathway, although reduction of the catalysis by halide ions of simple cobalt(II1) cobalt(l1I) center occurs to a minor extent in the substitutions has been suggested previously, presence of excess bromide or chloride ions. there is no direct evidence that it actually occurs, Neither C O ( M F I ~ ) ~ nor C ~ ~Co(NH3)5Br2+ + yielded and, conversely, where some reduction of cobalt(II), ho~vever;this shows that the reduc- cobalt(1iI) accompanies substitution, as in the tion of the cobalt(iI1) center is effected by present case, it does not follou that the substitucooi~clit~atecl DMSO, although it u a s not possible tion mechanism involves redox processes in any t o determine the ultimate fate of the DMSO way. Indeed, the fact that ki for Y n - = C1- is ligand. Furthermore, the minor degrees of reduc- more than three times that for the stronger tion of C O ( N H ~ ) ~ D M S observed O~~ in the reductant Br- in reaction 1 in DMSO is clear presence of free C1- or Br- are not attributable evidence against redox caealjrsis of substitution t o direct reduction of cobalt(1il) by these halide in these systems. The observations that conjugate base formaanions. This latter possibility may also be discounted on the grounds that chloride is thermo- tion (reaction 2) provides the most favorable dynamically a poorer reductant than bromide, pathway for redox (reaction 3), and that a yet is at least as effective in promoting reduction numerical connection seems to exist between of cobalt(ll1) amnlines in DIvISO.~ redox rates relative to anation within the differThe reduction of cobalt(il1) ammines by ent ion pairs,3 suggest that halide ion catalysis DMSO in the presence of CI- or Br- is reminis- of redox may operate through an internal cent of the catalysis by mildly reducing anions conjugate base process cvitliin the ion pairs, of topological changes in cobalt(ll1) complexes since neither HBr nor HCl is completely discontaining quadridentate or bidentate arsine sociated in DMSO (22, 23). This process, a ligands reported by Bosnich et ul. (27, 28), who proton transfer from Co(NI-J3)5DMS03-- to the attribute this catalysis t o the formation of a halide ion, would be similar in esser,ce to that labile, transient, but undetected coSalt(I1) inter- believed to occur in reactions of anions of weak mediate. Similarly, catalysis of the aquation of acids with complex ions in water (I). It could cobalt(I11) ammines through reduction to co- also explain (16) why kifor { C O ( N F I ~ ) ~ D M S O ~ - - , balt(I1) by hydroxide ion has been invoked by C1-} is more than twice k,, for the free complex Gillard (29) as an alternative to the now generally- ion, and more than three times ki for the accepted conjugate base mechanism (16). We bromide analog (HCL is a t least 10 times weaker have argued previously (30) against Giliard's than HBr in DMSO (22)). The close similarity proposal, partly on the basis of the complete between AHi* for C1- and Br- anation and lack of correlation of the catalytic power of AH,,* for solvent exchange, however, would various anions with their thermodynamic reduc- seem to rule out this attractive alternative to the ing power (as above), and indeed Gillard ef al. sirnple Id mechanism, since AH* values associhave recently reported failure to detect any ated with internal conjugate base processes are much higher than for simple interchange proe3At the same stoicI?iometric halide concentratic;ns, esses, a t least for water as solvent (1). Thus, the chloride produces more cobalt(l1) than does bromide (Table 2 ) , but the yields of cobalt(1l) are equal for the function of C1- and Br- in promoting tile redox reaction may involve nucleophilic attack of same extents of ion pairing of ColNH3),DMS03-- by C1and Br-. The data of Tables 3 and 4 relate t o various halide on the coordinated (and therefore polarionic strengtiis. so tliat constant KIP values (eq. I ) cannot ized) DMSO molecule, much as the attack of be obtained from them. but we can say that ion pairing is iodide on doubly-protonated DMSO in acidic (for example) 50"; complete when kohs = 0.50ki. This occurs (at 50.7 " C ) when [CI-] 5 mmol dm-j, or when aqueous DMSO initiates redox processes (32). In summary, although conjugate base phe[Br-] 17 mnlol dm-3, and Table 2 s h o ~ that s the yields of cobalt(l1) a t these concentrations are the same (0.7- nomena and redox processes complicate the 0.8';;) within the experimental uncertainty. Anation observation and interpretation of the kinetics of within the ion pairs, however, is 3.3 times faster for Clthan Br-, so that redox must be correspondingly faster in reaction 1, where the solvent is DMSO and '2'"the chloride ion pair, since it ceases o n completion of the is Cl-, Br-, NCS-, or DMSO itself, the substituanation reaction. tion processes observed in the presence of

-

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 121.52.144.245 on 05/28/13 For personal use only.

3692

CAN.

1. CHEM. VOL. 54, 1976

12. S. T . D . L o and T. W. SWADDLE.Inorg. - Chem. 15, 1881 (1976). and D. W. WATTS.Aust. J. Chem. 19. 13. I. R. LANTZKE 949 (1966). 14. W. E. JONES:L. R. CAREY and T. W. SWADDLE.Can. J. Chem. 50, 2739 (1972). 15. A. I. VOGEL.A textbook of quantitative inorganic analysis. 3rd. ed. Longmans, London. 1961. 16. F. BASOLOand R. G . PEARSON.Mechanisms of inorganic reactions. 2nd ed. J. Wiley and Sons, New York, N.Y. 1967. pp. 177-193. 17. I. M . KOLTHOFF and T. B. REDDY.Inorg. Chem. 1, 189 (1962). 18. W. S. MACGREGOR. Ann. N. Y. Acad. Sci. 141, 3 (1967). and S. I. PETROV.Zh. Anal. Khim. 27, 19. L. N. BYKOVA 1076 (1972). and M. LE DEMEZET. C. R. Acad. 20. J. COURTOT-COUPEZ Sci. Ser. C , 266, 1438 (1968). 21. R. STEWART and J. R. JONES.J. Am. Chem. Soc. 89, 5069 (1967). 22. C . MCCALLUM and A. D. PETHYBRIDGE. Electrochim. Acta, 20, 815 (1975). and C. BUISSON.Electrochim. Acta, 18, 23. R . L. BENOIT 105 (1973). 24. C. D. RITCHIEand R . E. USCHOLD.J. Am. Chem. SOC.89, 2752 (1967). 25. D. L. CARLEand T. W. SWADDLE.Can. J. Chem. 51, 3795 (1973). Acknowledgment 26. S. T. D . Lo and D. W. WATTS. Aust. J. Chem. 51, 491,501 (1975). We thank the National Research Council of 27. B. BOSNICH,W. G . JACKSON, and S. B. WILD. Inorg. Canada for financial support. Chem. 13, 1121 (1974). 28. B. BOSNICH, W. G . JACKSON,and J. W. MCLAREN. I . T . W. SWADDLE. Coord. Chem. Rev. 14, 217 (1974). Inorg. Chem. 13, 1133 (1974). 2. C . H. LANGFORD and H. 5 . G R A YLigand . substitution 29. R . D. GILLARD. J. Chem. Soc. A, 917 (1967). processes. M.'. A. Benjamin, New York, N.Y. 1966. 30. W. E. JONES,R. B. JORDAN,and T. W. SWADDLE. A , E. TOMA,and E. GIESBRECHT. 3. L. A. D E O L I V E I RH. Inorg. Chem. 8, 2504 (1969). Inorg. Nucl. Chern. Lett. 12, 195 (1976). 31. 1. G. BROWNING, R. D. GILLARD,J. R. LYONS,and 4. W. L. REYNOLDS. Inorg. Chem. 14, 680 (1975). P. R. MITCHELL.J. Chem. Soc. Dalton Trans. 373 5. J. IMACB.HAKROWFIELD, A . M. SARGWON, B. SINGH, (1974). and J. C . SULLIVAN. Inorg. Chern. 14, 2864 (1975). 32. J. H. KRUEGER. Inorg. Chem. 5, 132 (1966). 6. N . RODRIGUEZ, E. K R E V E RC. , R . P I R I ZMAC-COLL, 33. I. R. LAKTZKE and D. W. W ~ n sJ.. Am. Chem. Soc. and L. BEYER.Z . Anorg. Allg. Chem. 412, 59 (1975). 89, 815 (1967); Aust. J. Chem. 20, 35 (1967); 20, 2623 7. D. A. BUCKINGHAM, W. MARTY,and A. M. SARGE(1967). SON.Inorg. Chem. 13, 2165 (1974). 34. L. F. CHIN,W. A. ,MILLEN,and D . W. W a n s . Aust. E. S. BARBER, and R. CRANDELL. 8. W. L. REYNOLDS, J. Chem. 18, 453 (1965). Int. J. Chem. Kinet. 6, 51 (1974). and D. W. W a n s . Aust. J. Chem. 35. W. R . FITZGERALD 9. W. L. REYNOLDS, M. BIRUS.and S. ASPERGER.J. 19, 1411 (1966); J . Am. Chem. Soc. 90, 1734 (1968). Chem. Soc. Daiton Trans. 716 (1974); J. Chem. Soc. 36. W. R. FITZGERALD, A. J. PARKER, and D. W. WATTS. Chem. Commun. 822 (1973). J. Am. Chem. Soc. 90, 5744 (1968). 10. C . R . PKRIZMAC-COLLand L. BEYER.Inorg. Chern. 37. M. L. TOBEand D. W. WATTS. J. Chem. Soc. 4614 12,7 (1973). (1962). 11. C . H. LANGFORD and W. R. M U I RJ. . Am. Chem. Soc. 89, 3141 (1967); Inorg. Chem. 7, 1032 (1968).

p-toluenesulfonic acid are evidently of the simple Id type. The strongly positive AV,," value, and with AHi*, supthe apparent equality of AH,.x* port this. Finally, in view of the large effect of conjugate base pathwaqs on the rates of reaction 1 in DMSO, a check was made on the effect of added acid on the DIMSO exchange rates reported by Lantzke and Watts (33) for cisC O ( ~ ~ ) ~ C ~ ( D M S It O ) was ' ~ . found that the exchange rates were virtually unaffected by the addition of acid in that case: the reasons for this contrast with solvent exchange between CO(NH~)~DMSO and ~ + DMSO may lie simply in a markedly lower acidity of the dipositive chelate complex relative to the tripositive ammine. In any event, it appears that our present findings d o not detract from the usefulness of the extensive contributions of Watts' school (13, 33-37) and others (e.g., 11) in this field, but future work should include routine checks for con.jugate base and redox pathways.