JAMES DEEWHITEAND H. TATJBE
4142
Reduction of Cobalt(II1) in Cobaltammines Induced by the Decomposition of Persulfate Ion by James Dee White and H. Taube* Department of Chemistry, Stanford University, Stanford, California 94806
(Received May $6,1070)
The stoichiometric relations proposed by Thusius and Taube' for the production of Nz from cobaltammines Co(NHa)sOHza+ SzOa2- = Co2+ !/zNz 4NH4f induced by the decomposition of SzOaZ-: Hi. HzO 2sodZ- has been confirmed under conditions such that the production of COZ is virtually eliminated. The stoichiometric relation CoI'I:SzOaz-:Nz = 2:2: 1 has been extended to CO(?JH~)~(OH~)~*+ as substrate. Production of NzO is also observed, amounting to 5% of the gas in the case of the pentaammine and 15% when the tetraammine is the substrate. When C O ( N H ~ ) Z ( O His~substrate, ) ~ ~ + aside from adventitious COZ,0 2 is the only gaseous product. The stoichiometricrelation CoIII:SzOs2-: 02 was observed to be 2: 2: 3, but it is not certain whether this is invariant with concentration conditions. In the reaction producing Nz only that amount of SZOI- which would have decomposed in the absence of CoIII is implicated. For each of the ammines, it is assumed that S04- produced in the decomposition of 5 ~ 0 8 ~produces CoIV. Nitrogen tracer experiments show that the formation of the N-N bonds in NZis an intermolecular process, and it is suggested that NZarises from the production of NH in the coordination sphere of CoIV. Nitrogen and oxygen tracer experiments show that the formation of the N-O bond in NzO is intramolecular. The reaction CoIY --c CoZ+ NHzOH is suggested as the primary step leading to NzO production. CoIV is implicated as the agent which takes NH to Nz,and as producing isomerization in trans-nitrogen labeled Co(NHa)LOHza+.
+
+
+
+
+
+
+
Introduction Earlier studies' have shown that the decomposition of Sz0s2- in the presence of Co(NHa)sa+ or Co(NH3)SOHza+ induces the reduction of the cobaltammine to Co2+, with NH4+ and N2 appearing as the major nitrogen-containing products. At low Corrl concentration, the rate of production of Co2+ increases with [CoIII], but when this is in the millimolar range, the rate reaches a saturation value determined only by the rate of the reaction2,a SzOa2- =
2so*-
(1)
The main net change in the rate saturation region for the pentaammine as substrate is
H+
+ C O ( N H ~ ) ~ O H+~SzOs2~+ = CO'+ + '/&z + 4NH4+ + HzO + 2S042-
(2)
It should be noted that only that amount of SZOs2-is brought into reaction which in the absence of the cobaltammine would have decomposed to 804'- and 0 2 . The experiments by Thusius and Taube' indicated that NzO in minor amounts was also formed, but conclusions on the contribution by this reaction and on some other significant aspects of the chemistry of the system were obscured by the fact that COz, arising from the oxidation of adventitious organic matter, was present in substantial amounts among the gaseous products. We have returned to the study of these systems and have succeeded in reducing the amount of COZ so that it is only a minor component of the product gases, making it possible to get quantitative data on The Journal of Physical Chemistry, Val. 7.4, No. B,1.970
the reaction producing NzO. The work has been extended to encompass the tetraammine and diammine complexes of CoIII, and it includes nitrogen and oxygen tracer studies which help to clarify the mechanism of production of N2and N20.
Experimental Section Preparation of the Complexes. [Co(NH&OHa](C104)3. Carbonatopentaamminecobalt(II1) nitrate4 W perchloric acid, and the solid was treated with cold 2 i which formed was recrystallized three times from 0.1 M perchloric acid. Anal. Calcd for [Co(NHa)sOHz](ClO4)a: H, 3.7; N, 15.4; Clod, 64.8. Found: H, 3.7; N, 15.3; c104, 64.7. trans- [CO("a) 4( lsNHa)O H z ]((?lo4) 3. The compound [Co(NHs)&3Oa]zSOa was prepared according to the synthesis of Werner and Gruger.S A modification of Sargeson's procedureaJ was used for the synthesis of the 16N-labeledcomplex. [ C O ( N H ~ ) ~ S O was ~]~S dissolved ~~ in concentrated hydrochloric acid and immediately precipitated with
* To whom correspondence
should be addressed.
(1) D. D. Thusius and H. Taube, J . Phys. Chem., 71, 3845 (1967). (2) I. M. Kolthoff and J. K. Miller, J . Amer. Chem. SOC.,7 3 , 3055 (1951). (3) W. K. Wilmarth and A. Haim in "Peroxide Reaction Mechanisms," J. 0. Edwards, Ed., Wiley-Interscience, New York, N. Y., 1962. (4) A. Lamb and K. Mysels, J . Amer. Chem. SOC., 67, 468 (1945). (5) A. Werner and H. Gruger, Z . Anorg. Chem., 16, 406 (1898). (6) D.A. Buckingham, I. I. Olsen, and A. M. Sargeson, J . Amer. Chem. SOC.,89, 5129 (1967). (7) A. M.Sargeson, private communication.
REDUCTION OF COBALT(III) ethanol. The precipitate was recrystallized from a solution of concentrated hydrochloric acid and lithium chloride, then from an ammoniacal solution with ethanol, washed with ether, and treated with 3% aqueous LiOH for 5 min. trans- [Co(NH&(OH)S03] crystallized upon the addition of ethanol. Freshly prepared [Co(NH3)4(0H)SO3] (0.63 g) was added to a solution containing 0.5 g 15NH4C1 (99.570 15N) and 0.26 g of LiOH in 50 ml of H2O. trans-[Co(NHs)4(16NHs)S03]C1 was crystallized from solution. After 10 min ethanol was added to complete the crystallization. After drying, the trans- [Co(NH3)4(l6NHa)S03]C1was dissolved in a solution of LiCl in concentrated HCl. The solution was heated on a water bath for approximately 10 min. After cooling in an ice bath, ethanol was added to precipitate trans- [Co(NHa)4(16NHa)C1]C12.This material was dissolved in 0.1 M HC1O4 and treated with a solution of mercuric perchlorate. The resulting solution was ion-exchanged and trans- [CO(NH~)~('~NH~)OH~](C~O~)~ was precipitated with 7270 HC104. The 16N-labeled complex was recrystallized twice and stored under vacuum in a desiccator a t ca. 0". The precaution of keeping the temperature low was taken in order to minimize isomerization of the ~ o m p l e x . ~ Proton nrnr was used to determine the purity of the complex using Sargeson's peak assignmentsa6 No cis-15NHa protons were observed. I n addition, the extinction coefficients a t 490 and 345 mp were determined.* The value of e found at X 490 and 345 nm were 47.4 and 44.8, to be compared to the literature values of 47.5 and 44.8. [CO(NH~)~(OH&](CZO~)~. This salt was prepared by Schlessinger's p r o ~ e d u r e . ~Carbonatotetraammine (30 g) was dissolved in 600 ml of cold deionized water and 40 ml of 72% HC104 was then added slowly. Precipitation of the salt was induced by the addition of ethanol; the solid was collected and was further purified by recrystallizing from dilute perchloric acid solution. Anal. Calcd: H , 3.47; N, 12.14; clod,64.7. Found: H , 3.51; N, 12.15; clod, 64.7. Equilibrationlo" between cis and trans forms of the tetraammine has been reported to be quite rapid. Experiments done since this work was completed and other auspiceslob suggest that the tetraammine exists in the cis form. Attempts to prepare the trans form have proved to be unsuccessful. [CO(NH~)~(OH~)~](C~O~)~. The blue variety of K [Co(NH3)z(COa)z] was prepared according to Mori's procedure." K[Co(NH&(CO&] (10 g ) was dissolved in 100 ml of cold 1 M HC104. The resulting solution was filtered and charged onto a thermostated column of Dowex 50x2 100-200 mesh cation exchange resin. Cold 1 M HC104 was used to elute Coli. A violet species was eluted with 2 1%' HC1o4. On the basis of its spectrum and ion-exchange behavior we believe this
4143 material to be a dihydroxy bridged dimer of the diammine; it will be reported on in greater detail in a separate publication. To obtain monomeric diammine, a solution of the dimer was condensed using a rotary evaporator with no external heating of the solution. After reducing the solution to a small volume, it was further concentrated on a vacuum line until solid material formed. The solid is very hygroscopic and difficult to weigh; it was immediately put into solution for later use. Because of the difficulty in manipulating the solid, analyses were performed on solutions. The ratios of COT:Corrl:NH3 in two separate preparations proved to be: 1.00:0.99:1.98, 1.00:1.02:1.98. COT represents total cobalt and CoIII represents that present in the 3f oxidation state. The methods of determining these quantities will be dealt with presently. Stock Reagents. Water. Deionized water was distilled three times in consecutive glass stills, the first distillation being from alkaline permanganate. The water was further treated by adding KzSzOs, 3 g/l. of water, refluxing for 8 hr, and then distilling the water. The liquid so obtained was used for the persulfate reactions and for recrystallizing the complex compounds used in this research. Cerium(1V). Primary standard grade ammonium hexanitrato cerate, obtained from the G. Frederick Smith Chemical Co., in HzS04 served as the source of CerV. Iron(1Z). Mallinckrodt analytical grade Fe(NH4)2(S04)z.6HzO dissolved in 2 N H 8 0 4 was used as a reducing agent. Dichromate. Primary Standard potassium dichromate (Mallinckrodt) was used in CorIr determinations. Bufer System. I n most of the experiments, a phosphate buffer was used, made up using AR sodium monobasic phosphate and sodium dibasic phosphate. The total level of phosphate was 0.054 144'. Potassium Hydrogen Phthalate. Matheson Coleman and Bell's alkimetric standard K.H.P. was used to standardize NaOH solutions. All other chemicals were readily available ACS reagent grade and were used without further purification.
Andy ses CobaZt(l1). Cobalt (11) was determined spectrophotometrically as the thiocyanate complex in acetone according to Kitson's method.12 CobaZt(ZZ1). Excess Fe(NH&(SO& in 2 N H2SO4 (8) A. M. Zwickel, Ph.D. Dissertation, University of Chicago, 1954. (9) G.Schlessinger, Znorg. Syn., 6, 173 (1960). (10) (a) R. G. Yalman and T. Kuwana, J. Phys. Chem., 59,298 (1955); (b) 3. D. White and T. W.Newton, to be published. (11) M. Mori, M.Shibato, E. Kyono, and T. Adachi, Bull. Chem. soc. Jap., 29, 883 (1956). (12) R. E.Kitson, Anal. Chem., 22, 664 (1950).
The Journal of Physical Chemistry, Vol. 7.4, No. 23,1970
JAMESDEEWHITE AND H. TAUBE
4144 was added to the Corrl solution and back-titrated with Crz0,2- using sodium diphenylamine sulfonate as the indicator. l a Total Cobalt (COT). This was determined by the Kitson12 method. Persulfate. Excess Fe(NH4)z(S04)2 in 2 N HzS04 was added to the persulfate solution and back-titrated potentiometrically with CeIV. Cerium(1B). The CeIV concentration was determined spectrophotometrically using extinction coefficients measured by Robson14and Medalia and Byrne. l5 Dichromate. A solution of potassium dichromate was made alkaline with NaOH to convert the dichromate to chromate and the concentration was spectrophotometrically determined at 373 mp.lS Iron(I1). Ferrous ion was determined potentiometrically with either CeIV or Crz0T2- using diphenylammine sulfonate as an indicator. Ammonia. A known volume of aquoamminecobalt(II1) solution was reduced with FeS04. Sodium hydroxide pellets were added and the liberated ammonia was distilled into a standardized perchloric solution. The acid was then titrated with a sodium hydroxide solution with methyl red serving as the indicator. Chloride. Chloride was determined gravimetrically as silver chloride. l7 HzlsO. The Coz equilibration method described by Huntla was used to determine the l80content of enriched water.
Methods Reactions with Persulfate. Water distilled from persulfate was stored in a glass-stoppered bottle with a cover over the stopper to prevent dust from contaminating the neck of the bottle. Glassware which was used for handling reagents was cleaned with a warm solution of X'azCr20, in concentrated HzS04, rinsed profusely with water distilled from persulfate, and dried in an oven at ca. 100". Cobalt complexes were recrystallized from perchloric acid solutions, which were made from water distilled from persulfate. A trace of sodium persulfate was added to the cobalt solutions before heating to oxidize organic impurities. Care was taken at all stages to minimize contamination by particulate matter. Kinetics of Reactions with Persulfate. For the reactions with persulfate, the desired amounts of sodium persulfate, disodium phosphate, and monosodium phosphate were added to redistilled (from persulfate) water and the mixture was heated in a constant temperature bath at 50" for 4 to 6 hr to oxidize organic impurities. It should be noted that only 2% or so of the 5202decomposes in this interval of time. After cooling the solution to room temperature, the pH was adjusted to 3.5 by dropwise additions of perchloric acid and sodium hydroxide. The solution was heated for 1 hr a t 50". The Journal of Phyaical Chemistry, Vol. 74, No. 93, 1970
The cobalt complex was weighed from a weighing bottle and added to a known aliquot of the persulfate solution, which had been allowed to reach ambient temperature. Upon dissolution of the complex, the solution was placed in a constant temperature bath at 50" to initiate the reaction. The initial rate of CoII production was followed by taking aliquots from the solution, chilling, and determining the amount of CoII using the Kitson method.12 Optical densities were determined with a Cary 14 recording spectrophotometer. I n all runs, the sample optical density was corrected by using a ColIr ammine blank of appropriate concentration. The initial rate of the reaction producing reduced cobalt was determined from the slope of the plot Corl vs. time. Rate constants for the process are defined as
k =
d [Co'I] dt [S208*-]
Analysis of the Gaseous Products of the Reactions with Persul$ate. The reaction vessel consisted of a roundbottomed flask with a constricted neck ending in a standard top joint, and a side arm topped with breakseal. The reaction vessel was cleaned with a warm solution of NazCrz07 in concentrated HzS04 and rinsed profusely with quadruply distilled water. The vessel was then filled with a persulfate solution and heated for 6 hr at 50". The vessel was emptied, a known amount of cobalt complex was added directly from a weighing bottle, and then the persulfate solution of known concentration. The solution was degassed by vacuum techniques, whereupon the vessel was sealed off a t the constriction with a torch. The reaction was initiated by placing the vessel in a bath at 50" and was allowed to proceed for a time varying from 15 to 60 min. The reaction was quenched by freezing the solution in a Dry Ice-acetone bath. The vessel was attached to a vacuum line, the break-seal was opened, and the gas was transferred by means of a Toeppler pump through a trap cooled in liquid nitrogen to a gas buret where the amount of gas was measured. The noncondensable gases were transferred to a Urey tube. To collect the condensable gases, Dry Ice-acetone was substituted for liquid nitrogen in the trap. The gases were analyzed either by means of a mass spectrometer or a gas chromatograph. Results Co(NH3)60Hza+. Some experiments were performed to check on the rate law (13) D. A. Skoog and D. M. West, "FundamentrJs of Analytical Chemistry," Holt, Rinehart and Winston, New York, N. Y., 1983, p 457.
(14) R. Robson and H. Taube, J . Amer. Chem. Soc., 89,8487 (1967). (15) A. I. Medalia and B. Byrne, Anal. Chem., 2 3 , 453 (1951). (16) E.A. Deutsoh, 1'h.D. Dissertation, Stanford University, 1967. (17) See ref 13,p 202. (18) H.R. Hunt, Ph.D. Dissertation, University of Chicago, 1957.
REDUCTION OF COBALT(III)
4145
Table I : Analytical Results on the Products of the Reaction of Co( NH&OHz*+ with S20s2[SzOs~-I, M
Expt'
3 4 5 6 =
Time, min
Nmol of
4.61 25.4 5.76 4.97 6.00 4.87
60.0 60.0 16.0 15.5 15.0 15.0
5.40 10.90 7.21 6.75 9.18 9.24
0.15 0.30 0.73 0.73 1.03 1.03
1 2
aT
LCOI"1, 108 M
a t concentrations of [s20S2-]higher than those used by Thusius and Taube.l I n a series a t 50" and pH 3.5 with [CoIII] = 4 X M (note that the rate is within 5% of saturation1 at 1 X M CoIII) and with [s208'-] = 0.15, 0.15, 0.53, and 1.03 M , k (sec-l X 106) was found to be 0.99, 1.00, 1.01, and 1.03, respectively. As concluded in the earlier work, any contribution to the reaction in the "saturation" region by a second-order process is small indeed. An experiment was performed to determine whether ~ - place. significant complexation of CoI'I by S ~ O Stakes A solution containing C O ( X H ~ ) ~ O H(5~ X ~+ IM) and KnS208(0.1 M ) was heated a t 50" for 2 hr and scans of the absorption were made approximately every 15 min. The spectrum did not alter significantly with time, except for the 10-15% decrease resulting from the reduction of Co'II. This is not a very sensitive test of complex formation, but the experiment does indicate atleast that there is no major complexation of CoIII with SZOS2-. The results of determinations of stoichiometry are shown in Table I, and an analysis of the stoichiometric data appears in Table 11.
Table I1 : Stoichiometric Relations for Co(NHs)bOH2'+ $- S20s2Equiv coz+obSd/
Expt
(NIO) (NzO N?)
1 2 3 4 5 6
0.051 0.044 0.041 0,042 0.061 0.061
0.94 0.94 1.02 0.98 1.00 1.04
+
C02+oalod
Equiv
oxidQ agents consumed,
x
106
16.2 32.5 21.2 20.8 27.7 27.6
(Nz
+ Nz0)
formedb x 106
16.7 33.5 20.6 20.1 26.7 26.0
+
Z [ S Z O ~ ~ -decomposed ] = equiv [ C O I ~ ~and ] [ S Z O ~ ~decomposed -] was calculated using the -] k = 1.0 X lob6 expression d[SzOsz-]/dt = k [ S ~ 0 8 ~ where sec-1, the value measured in the absence of CoIII. [pmol of N2 X 6 pm of NzO X 81 = equiv of (NzO Nz) formed. [Coz+]
*
pmol of Nz
pmol of NzO
pmol of
2.59 5.27 3.25 3.17 4.07 3.98
0.14 0.24 0.14 0.14 0.26 0.26
0.02 0.04 0.05 0.02 0.26 0.26
CO?
pH 3.5; the reaction volume was 10 ml in all cases.
5 O ;
d[Co2+] _ _ _--k[SzOsZ-] dt
a
cot
[S2082-] used.
+
+
The second column of Table I1 shows that the NzO content is rather invariant to concentration conditions. In the third column are shown values of the ratio C02+obsd/COz+cplcdwhere Co2fcalcdis obtained on the basis of assuming that 2 mol of Co2+is produced for each mole of N2 (cf. eq 2) and 4 mol for each mol of NzO (the results are not sensitive t o the stoichiometry assumed for this reaction). The entries in column 4 are obtained by calculating the number of equivalents of SZOs2-which would have decomposed in the absence of CoIII and adding to them the number of equivalents of CoIII consumed, these together making up the number of equivalents of oxidizing agent consumed. The last column shows the number of equivalents needed to produce the N2 and NLOwhich formed, 6 equiv being required for each mole of the former and 8 each for each mole of the latter. An experiment was done by letting [Co(NH3)50H2I3+ (10-2 M ) react with S2082-(1 M) in Hz180 (1.52 atom% enriched) for 15 min at 53" and pH 3.5. The cobalt complex, persulfate ion, and phosphate were all used at normal levels of enrichment. If there were no induced exchange, the residual aquo complex would have proceeded 18% of the way to isotopic equilibrium with the solvent. The N20 was found by mass spectrometric measurements to contain 0.72 atom% l80. Making allow-an~e~~ for the incorporation of l80into the complex by spontaneous exchange, the result shows that N 2 0 derived no more than 17% of its oxygen from the solvent. Though there is a possibility that oxygen from S2082-or phosphate appears in the NzO, it seems a remote one and it will be assumed that the N2O derives its complement of oxygen of normal isotopic composition from the coordination sphere of the complex. Two experiments were performed by heating transC O ( ' ~ N H ~ ) ( N H ~ ) ~(5 O HX~ lov3 ~ + M) in SZOs2- (0.2 M ) a t 50" and 53" for 60 and 100 min, respectively (pH 3.5). I n the first experiment the ratios I4N2, 14N16N,I6Nzwere determined and were found to be 16:s: 1, precisely the value expected for random formation of nitrogen from one part of 15Nand four parts of (19) H. R. Hunt and H. Taube, J. Amer. Chem. Soc., 80, 2642 (1958).
The Journal of Physical Chemistry, Vola74, No. $3, 1970
JAMES DEEWHITEAND H. TAUBE
4146
14N. The COz content of the gas was too high to lead Table IV: Analytical Results on the Products of the to good values of the isotopic ratios in NzO. Accord3+ with SzOs2Reaction of [Co(NH3)4(OH~),] ingly, the gas obtained in the second experiment was [Ca(NHs)rdistilled onto solid NaOH to absorb COZ. The ratio (OHz)zl**, 14N14K0,14N15N0and 15N14N0,15N15N0was found to [SzOs2-], x 102, #mol smolb Nrnolb pmalb be 40: 11:1. The I6Xcontent is much below that exExpt'" M M cor NZ Nz0 COz pected for random distribution of the isotopes. 1 0.15 0.45 5.44 2.20 0.57 0.02 The unreacted complex from the second experiment 2 0.24 5.07 8.60 2.88 0.85 0.03 3 0.24 11.9 8.62 2.63 0.80 0.02 was precipitated, dissolved in DMSO, and the proton 4 0.24 2.45 8.68 2.70 0.90 0.04 nmr spectrum was taken.6 This showed that the 5 1.03 2.4 36.5 12.63 2.27 0.10 W H 3 was randomly distributed in the complex. A 4.96 6 1.03 24.0 8.28 1.53 0.90 blank experiment was done in which t~ans-Co(~~NH3)a T = 50°, p H 3.5. I n all cases the volume is 10 ml and (NH3)40Hz3+was heated for 110 min at 53" in a soluthe time is 60 min, except in number 6 where the time is 40 tion at pH 3.5 in the absence of S2OsZ-. The proton min. nmr of the cobalt salt showed that the 15NH3label was not scrambled in the experiment, and we thus conclude that scrambling is induced by the decomposition of Table V : Stoichiometric Relations for the Reaction SzOs2-. The blank, incidentally, confirms for rather of Co( NHa)4(OHz)zs+with Sz012extreme conditions the conclusion reached by Buckingham, et al.,0 to wit, that substitution in the pentaammine Eauiv Eauiv complex in acidic solution takes place without isomerioxid-agent ("2 NzO) Nz0 COz+abed consumed formed zation. Expt Nz0 4- Nz C z x 106 x 106 C O ( N H ~ ) ~ ( O H ~The ) ~ ~ rate + . of production of Co2+ 1 0.21 0.82 16.2 17.8 in the reaction of C O ( N H ~ ) ~ ( O Hwith ~ ) ~SzO.sZ~ + follows 2 0.23 0.94 25.9 24.1 the pattern found with Co(KHa)50Hz3+. At low con3 0.23 0.96 25.9 22.2 centrations of CoIII, the rate is sensitive to this varia4 0.25 0.97 25.9 24.9 ble, but becomes independent of it when [CoIII] ex5 0.15 1.06 110.5 93.9 6 0.16 1.05 73.4 61.9 ceeds approximately 4 X M . (cf. Table 111.) +
