Kinetics and Mechanism of the Thermal

Kinetics and Mechanism of the Thermal Decomposition of Hexaamminecobalt(lI1) and Aquopentaamminecobalt(II1)Ions in Acidic Aqueous Solution

Can. J. Chem. Downloaded from www.nrcresearchpress.com by Hunan Normal University on 06/05/13 For personal use only.

ANTHONY MARTINNEWTONA N D THOMAS WILSONSWADDLE' Depurtnlent nf Chernirtry, The U t z i ~ ~ r s iof t y Calgnry, C n l g c r r ~Alhertn , T2N I N 4 Received March 18, 1974

MARTIPI. NEWTONand THOMAS WILSOUSWADDLE. Can. J. Chern. 52,2751 (1974). ANTHONY + acidic aqueous solution is the The initial step in the thermal decomposition of C O ( N H , ) ? ~ in replacement of N H 3 by H 2 0 , which occurs by a hydrogen-]on independent path, first order in complex, with rate coefficient k , = 7.9 x lo-' s-' (140.4"), AH* = 36.6 kcal mol-', and AS* = 10.7 cal deg-' mol-' in 0.1 M HCIO,. For CO(NH,),OH,~+, there is a similar initial aquation path with k l = 12.6 x lo-' s-' (140.6'), AH* = 41.9 kcal mol-', and AS* = 24 cal deg-' mol-' and also a path first order in complex but inverse first order in [ H + ] with k 2 , = 6.2 x lo-' M s-' (140.6'), AH* = 43.5 kcal mol-', and AS* = 26.7 cal deg-I mol-I, in perchlorate media of ionic strength 1.0 M. The effects of electrolyte type and concentration on the rates of these reactions have been examined. Subsequent aquation steps are relatively rapid because of the predominance of inversely [Hf]-dependent pathways and are followed by redox to CO(H,O),~+,NH,+, N2, N 2 0 , and a minor amount of 0 2 . A mechanism involving O H and N H 2 radicals is proposed for the redox step. ANTHONY MARTINNEWTON et THOMAS WILSONSWADDLE. Can. J. Chem. 52,2751 (1974). L'etape initiale lors de la decon~positionthermique du CO(NH,),~+en solution aqueuse acidifite est le remplacement d'un NH, par une molecule d'eau; cette reaction se produit par une voie n'inpliquant pas d'ions hydroghe, elle est du premier ordre en complexe avec un coefficient de vitesse k , = 7.9 x lo-' s - ' (140.4"), AH* = 36.6 kcal mol-I et AS* = 10.7 cal deg-' mol-' dans HCIO, 0.1 M. Dans le cas du CO(NH,),OH,~+, il existe une equation similaire pour le chemin initiale et k, = 12.6 x s-' (140.6"), AH* = 41.9 kcal mol-' et AS* = 24 cal deg-' mol-'; cette reaction est aussi du premier ordre en complexe mais d'un ordre inverse du premier en [H+]avec k 2 ' = 6.2 x lo-' M s - I (140.6"), AH* = 43.5 kcal mol-' et AS* = 26.7 cal deg-I mol-' dans un milieu perchlorate ayant une force ionique de 1.0 M. On a aussi examine les effets de type tlectrolyte et de concentration sur les vitesses de ces reactions, les tquations des Ctapes subsequentes sont relativement rapides a cause de la predominance d'un chemin dependant d'une f a ~ o ninversement proportionnelle a la concentration [H+1, ces Btapes sont suivies par une reaction d'oxydo-reduction conduisant a C O ( H ~ O ) ~ ~ + , NH,+, N2, N 2 0 et des quantitks mineures de 0,: On propose un mecanisme impliquant des radicaux O H et NH, pour I'ttape d'oxydo-reduction. [Traduit par le journal]

Introduction A striking feature of cobalt(II1) chemistry is the great decrease in reactivity which occurs on replacing several of the aquo ligands in Co(HzO)63+by NH,, to the extent that the ammine ligands in CO(NH,),~+and CO(NH,),X(~-")' (X = halogen, H,O, oxyanions, etc.) are commonly regarded as substitution inert in acidic aqueous solution. Nevertheless, kinetic studies of the aquations of X n - from CO(NH,),X(~-")' begin to suffer from the complicating effects of NH, loss from either the parent complex or (more usually) from the product Co(NH,),OH,3+ at temperatures above 80" (1). The present study seeks to determine the importance 'To whom correspondence should be addressed.

of this complication, as well as to gain insight into the mechanism of decomposition of Co(NH,),,' and Co(NH3),0Hz3+ in solution. Since the completion of our kinetic studies, Garner and his co-workers (2) have published data on the decomposition of Co(NH3),OHZ3+ and ~~S-CO(NH,),(OH,)~~ + in acidic perchlorate media, in addition to their papers on the decomposition of the presumed facial isomer of CO(NH,),(OH,),~+ (3) and cis-Co(NH,),(4). It appears that the initial replacement of an ammine ligand by water is the ratedetermining step in the decompositions of the penta- and tetra-ammines (2) but that redox decomposition to cobalt(I1) predominates in the decomposition of the triammine (3) (however, up to 17% of the observed rate of disappearance +


2752

C A N . J . C H E M . V O L . 5 2 , 1974

of the tetraammine could also conceivablv be due to redox (2)). cis-Diamminetetraaquocobalt(111) decomposes by direct redox to cobalt(I1) (4), whereas amminepentaaquocobalt(I1I) undergoes an extraordinary disproportionation to cobalt(I1) and diamminetetraaquocobalt(II1) (5). We shall address ourselves primarily to two salient questions which remain. Firstly, there is no information on the kinetics of decomposition of the hexaammine, and, secondly, the nature of the oxidized reaction products has not been established. The latter has an important bearing on the mechanism of the redox steps. In addition, new information pertaining to electrolyte effects on the reaction rates will be presented, together with improved activation parameters for the decomposition of CO(NH,),OH,~+.

The an~pouleswere then sealed and immersed in the thermostat bath at 130.6' for several half-periods of the decomposition reaction. The samples were then frozen at liquid nitrogen temperatures and the gaseous contents of the tubes were released into a VarianIMAT CH-5 mass spectrometer for analysis.

Results Throughout this article, the molar concentrations cited refer to solutions as at 25". Uncertainty limits quoted represent standard errors; where these are not stated for an experimental measurement, a single determination is implied.

Decor?zposition of Aquoperztaamminecobalt(III) Ion The spectrum of the final reaction products 5 M-I (absorption maximum at 505 nm, E cm-l) identified the Co-containing species as being entirely hexaaquocobalt(I1) (10). The Experimental absorbance A , at time t changed in accordance iClaferial~ with first order kinetics as the reaction proSalts of lithium, CO(NH,),~'. and Co(NH3),0HZ3+ ceeded, except during a short initial "induction were made by standard methods (6, 7) and were checked for purity by chemical analysis and by examination of period" (Fig. I ) ; the latter phenomenon was their visible spectra; these and all other spectral measure- much less marked than that reported (1 1) for the ments were made using a Cary Model 15 spectrophotom- deconlposition of cobalt(II1) ammines in molten eter. Baker Analyzed perchloric acid (72%) and Fisher NH4HS04 and in 9 7 z sulfuric acid. Further"purified" sodium perchlorate monohydrate and sodium nitrate were used directly. Distilled water was either more, as Garner and co-workers have observed passed through Barnstead deionizer and organic removal also (2), the visible spectra of partially reacted cartridges or else redistilled carefully from alkaline per- solutions showed isosbestic points at the early manganate before use; the kinetic results were the same in stages of the reaction, but these were not maineither case. tained as the reaction proceeded. These facts Kinetic Studiies ind~catethe presence of a fairly long-lived reacAliquots of solutions of the appropriate complex in tion intermediate in the decomposition. aqueous NaC10,-HC10, or LiCI0,-HC10, of the reBecause CO(NH,),(OH,),~+ is known (3) to quired ionic strength I were sealed into Pyrex ampoules, and these were preheated to about 95 (to facilitate ther- decompose rapidly under the conditions of these mal equilibration and to dissolve any solid hexaamn~ine- experiments, the intermediate was presumably a cobalt(II1) perchlorate) before immersion in an oil-filled mixture of cis- and trans-CO(NH,),(OH,),~',

Lauda NS-HT thermostat bath (-t0.lC).Light levels in the bath were negligible. Timing of the reactions was begun 2 min after inlnlersion of the samples, and ampoules were withdrawn periodically and chilled to room temperature. The samples were then analyzed spectrophotometrically at 427 nm (for the hexaammine) or 344 or 490 nm (for the aquopentaammine). Alternatively, the cobalt(I1) content of the samples were determined by making r. ml of the aliquot and (6 - c ) ml 0.1 M HCIO, up to 25 ml with concentrated HCI, and measuring the optical absorbance of the resulting blue solution at 690 nm ( E 471 M - ' cm-' according to a calibration curve established using solutions of pure cobalt(11) nitrate) (8). The cobalt(1I) concentrations so measured were also checked in some cases by Kitsen's method (9), with essentially identical results. Mass Spectra o f t h e Gaseorts Decotnposition Prodrict~ Solution aliquots, made up as for the kinetic experiments, were placed in breakseal ampoules and thoroughly degassed on a vacuum line by repeated freeze-thaw cycles.

-

MINUTES

FIG. 1. T ~ m edependence of the opt~calabsorbance A , (490 nm, 10 mm opt~calpath, 25 ) o f a solut~on~nrtially 0.0106 M aquopentaamn~inecobalt(lIl) perchlorate In HC10,-NaC104 (I = 1 .O M) at 140.6'.

NEWTON AND SWADDLE: COBALT(I1I) AMMINES

TABLE 1. First-order rate coefficients k,,, for the decomposition of C O ( N H ~ ) ~ O in H 0.1 ~ ~ M HCIO," +

Temperature


rc)

102[NaC104] (M)

l o 2[NaN03] (M)

l~~k,,, (s - I)

,,,

"[CO] = 0.0106 M. bComplex present as its nitrate salt; elsewhere, as the perchlorate.

TABLE 2. Hydrogen Ion concentration dependence of the first-order rate coefficient k,,, for the decon~pos~t~on of C O ( N H , ) ~ O H , ~" + --

Temperature ("C)

=

IO4k0b, wl)

1.06 M , adjusted with NaCIO,. Initial

with the former predominating, as Garner and CO-workers suggested (2). These decompose more rapidly than their progenitor (2), and so the slopes of the linear portions of the plots of log (A, - A,) against t gave the first-order rate coefficient kobsfor the rate-controlling reaction H+ [I] C O ( N H ~ ) ~ O + 3 H ~C ~O ( N H ~ ) , ( O H ~ ) ~ ~N H 4 + +

--

-

[HCIO,] (MI

+

The values of kob,collected in Tables 1 and 2 confirm that parallel acid-independent (rate coefficient k,) and inversely acid-dependent (k,) paths operate (2), as per eqs. 2 and 3, and that high concentrations of perchlorate ion have no drastic effect on the reaction rate.

-

lo5kl (ssl)

105kz (M s-')

[Co(NH3),0H13-I = 0.0106 M. K,

[2]

C0(NH3)50Hz3+

* C0(NH3)50H2+ + H +

+ H.0 H+ CO(NH~),(OH~+ ) ~ NH3 ~ -+ NH4+ k2/

+

[3] kobs= k1

+ k , ~ , [ H + l - ' = kl +

k,,[H+]-'

In order to facilitate comparison with the CO(NH,),~+system (in which the low solubility of the perchlorate salt imposes some restrictions on the kinetic experiments), the effect of nitrate ion on the rate was also examined, and this anion was found to cause a modest acceleration


2754

C A N . .I.C H E M . V O L . 5 2 , 1974

of the decomposition reaction (Table 1). We attribute this to the fact that nitrate ion has appreciable Br~nstedbasicity in aqueous solutions at high temperatures (the pK, of aqueous nitric acid rises from - 2 at 0" to + 2 at 300" (12)) and will therefore decrease the free hydrogen ion concentration and so increase the contribution of the inversely hydrogen-ion dependent pathway. The reaction product Co2+ has been reported to catalyze the radiolytic decomposition of hexaammine- and aquopentaammine-cobalt(111) (13), and to inhibit the reduction of Co(NH,),OHZ3' by S,0a2-/Ag+ (14). The effect of added cobalt(I1) perchlorate on the rate of the spontaneous decon~positionof CO(NH,),OH,~' in 0.1 M HC10, was therefore studied with reference to control experiments in which Zn2+ was added in place of Co2+, as these ions have closely similar ionic radii and therefore similar medium effects. At 130.6", with an initial aquopentaamminecobalt(I11) perchlorate concentration of 0.0106 M , and [M(CIO,),] = 0.0109 M, k,,, was (3.52 0.13) x lo-, and (3.49 f 0.06) x lo-" s-' for M = Zn and Co, respectively; the corresponding data for [M(CIO,),] = 0.109 M were (2.57 i 0.02) x lo-" and (2.61 k 0.03) x lo-" s-I, as against (4.01 0.01) x lo-" in the absence of added M(I1) perchlorates. Thus, divalent metal perchlorates exert a small retarding effect on the reaction (as does NaC10,). but there is no specific effect attributable to cobalt(11). Nevertheless, at high concentrations of added cobalt(II), an unidentified black precipitate formed as the reaction proceeded; this was not observed when Co(I1) perchlorate alone was heated in 0.1 M perchlor~cacid at 130" for several days, which suggests that Co(I1) in high concentrat~onscan react with one of the decomposition products of CO(NH,),OH,~+ formed after the rate-controlling steps.

+

0 440

480

520

560

WAVELENGTH nm

FIG.2 . Spectrum (25', optical path 10 mm) of a solution originally 0.0106 M CO(NH,),~+in 0.1 M HCIO,, ( a ) initially, and after (b) 100, (c) 360, and ( d ) 830 min at 130.3".

+

Decomposition of Hexaamniinecobalt (111) Ion For this ion, as for the pentaammine, the final cobalt-containing product was CO(H,O),~+ but the spectral changes occurring during the decomposition reaction (Fig. 2) make it clear that an intermediate Co(111) complex was again involved, and this is verified by the appearance of the semilog kinetic plots (Fig. 3, cf. Fig. 1). Ion exchange chromatography of a solution of partially-decomposed CO(NH,),~ on Dowex 50W-X4 resin with 1.0 M HC10, produced a +

MINUTES

FIG.3. Decomposition of Co(NH,),,+ in HCI0,NaC10, ( I = 1.0 M ) at 149.6-, followed by spectrophotometric determination of cobalt(I1) by Kitsen's method (9).

rapidly-moving red band (hexaaquocobalt(I1)) and a thin, slow-moving pink band closely followed by an extensive yellow band of Co(NH3),3t; the pink band was so small as to render its isolation impracticable, but it almost certainly consisted of Co(NH,) ,OH2, +,as the more highly aquated ammines would not have survived in detectable quantities under the experimental conditions (2, 3). The rate coefficients k,,, (Table 3) for the decomposition of hexaamminecobalt(III) were obtained from the linear portions of semilog

NEWTON A N D SWADDLE: COBALT(II1) AMMINES

TABLE 3. First-order rate coefficients k,,, for the decomposition of Co(NH,),j+ in acidic aqueous solution


Temperature ("C)

[HCI04] (MI

Ionic strength (MI

0.033

1. O

0.100

1 .O

1 .oo

1 .o

Supporting electrolyte

1O4kObs (S-

NaClO, NaCIO, LiCIO, NaClO, NaCIO, LiCIO,

.'[Co'+] measured by HCI method. b[CoZ+] measured by Kitsen's method (9).