2553
The Oxidation of Water by Cobaltic Aquo Ions Michael Anbar and Israel Pecht Contribution from the Weizmann Institute of Science, Rehovoth, Israel. Received December 6,1966
Abstract: The oxidation of water to oxygen by C0(111)~~ has been investigated using 0 ’ 8 as tracer. It has been shown that the oxidation reaction, which proceeds by inner-sphere ligand t o metal electron transfer, involves the dimer by a Coaq3+ ion. The HOz radical, which is being formed in the rate-determining step, oxidation of a C0(111)~~ was shown t o contain substantial amounts of is therefore the immediate precursor of molecular oxygen. C0(111)~~ dimers even in strongly acid solutions. B.D.H.) were used. Triply distilled water was used throughout this obaltic aquo ions in aqueous solution a r e k n o w n study. to oxidize water to oxygen. The kinetics o f this C0(111)~~ solutions were prepared by electrolytic oxidation of 0.6 reaction have been studied b y several investigators ;1-3 M Co(ClO& in 6 M HCIOl at 0’. The electrolysis (current density however, the mechanism o f this reaction has not been of 300 ma/cm2) was carried out in a cell consisting of a platinum anode inside a nonglazed ceramic vessel surrounded by a copper foil hitherto unequivocally elucidated. Mechanisms procathode. The solution was electrolyzed for 2 hr, Le., considerable posed f o r this reaction have been based solely on kinetic time after the oxidation of Co(I1) was complete. As soon as the evidence, on which conflicting information is available. electrolytic oxidation was stopped, the Co(1IQaq solution was On thermodynamic g r o u n d s it m a y be inferred that the frozen in liquid nitrogen and kept as such until used. The Co(11) primary step of oxidation o f water to oxygen b y content of the Co(II1) solutions prepared by this method is negligible. C0(111)~~ does not involve a single electron transfer, CO(III)~~ labeled with 1 8 0 was prepared by the same method as the reaction C0(111)~~ HzO + C0(11)~~ OH HP0 (-70 atom 1 8 0 ) as solvent. H+ has a positive free energy of about 24 k ~ a l / m o l e . ~ - ~using Methods. One milliliter of the concentrated stock solution of An alternative primary product is hydrogen peroxide CO(III)~~ (0.6 M ) enriched in 1 8 0 was mixed under vacuum at room temperature (22-24”) with 20 ml of degassed, triply distilled water formed b y the concerted reduction o f t w o C0(111)~~ ions. of natural 1 6 0 content. Any oxygen dissolved, or formed in the The kinetics of reduction of C0(111)~~ b y water as interC0(111)~~ solution, was quantitatively removed before mixing by preted b y B a w n and White2 support a bimolecular pathcontinuous pumping from the two solutions until the moment way ; however, no experimental evidence f o r the formaof mixing. The oxygen formed after mixing was completely tion of H202as a primary product of oxidation o f water is pumped off by a Toepler pump at predetermined time intervals (3, 6, 10, 30, 120, and 360 min following time of mixing), and its available. A third possibility suggested b y Baxendale i n isotopic composition was mass-spectrometricallyanalyzed. Masses his interpretation of the kinetic data3is the formation o f 28 and 40 were also measured to check on air contamination. The HO2 a s primary product b y the oxidation o f a C0(111)~~ experiments were repeated using C O ( I I I ) of ~ ~natural 1 8 0 content dimer by Coaq3+. T h i s mechanism is thermodynammixed with water enriched in 1 8 0 . The experiments designed for detection of the intermediary ically m o r e favorable than the former one b y a AF formation of hydrogen peroxide were carried out as follows: 50 o f a b o u t 3 kcal/mole.5,6 p1 of 0.6 M C0(111)~~ stock solution were injected through a seroA n o t h e r open question is whether hydrogen peroxide logical cap into 250 ml of 10-4 M H202solution, thoroughly deor H 0 2 , if formed, a r e produced b y an inner-sphere gassed, and vigorously stirred. The isotopic composition of the electron transfer f r o m the or b y an outersolvent water was different from that of the coordination sphere of C0(111)~~; the HtOz was of natural isotopic composition. The sphere oxidation of the solvent molecules. By the use acidity of the reaction mixture was preadjusted by NaOH to give a of lSO as tracer, w e were able to show that H 0 2 is final pH of -4. After the reaction was complete the molecular formed as primary product in the oxidation o f water oxygen formed was pumped off for isotopic analysis. The solution b y C 0 ( 1 1 1 ) ~and ~ that this reaction proceeds b y an innerwas then thoroughly degassed, and the residual hydrogen peroxide sphere oxidation o f water bound to t h e metal. was decomposed by injecting either a suspension of platinum black in water or a solution of ceric ions. The Oz produced from the Experimental Section residual H202was then analyzed for its isotopic composition. Materials. Oxygen-18 enriched water was supplied by the Results and Discussion isotope separation plant of the Weizmann Institute of Science. The working hypothesis of this study w a s that the It was distilled several times over alkaline permanganate in an alloxidation of water b y C0(111)~~ proceeds b y an innerglass still. Co(ClO& (reagent grade, G. F. Smith Chemical Co., Columbus, Ohio) and hydrogen peroxide 50 (Analar grade, sphere electron transfer f r o m the coordinated ligand water to the central ion. This hypothesis could be (1) A. A. Noyes and T. J. Dahl, J . Am. Chem. Soc., 59,1337 (1937). proven b y demonstrating that the molecular oxygen (2) C. E. H. Bawn and A. G. White, J . Chem. Soc., 331 (1951). formed originates f r o m water molecules b o u n d to the (3) J. H. Baxendale and C. F. Wells, Trans. Faraday Soc., 53, 800 metal i o n in the inner coordination sphere. Cobaltic (1957). (4) H. Taube, J . Gen. Physiol., 49, 29 (1965). ions are k n o w n to f o r m substitution inert complexes. IO (5) W. M. Latimer, “The Oxidation States of the Elements and Their The C O ( I I I ) ~ ~ - H ~isotopic O exchange has been estiPotentials in Aqueous Solutions,” Prentice-Hall, Inc., New York, N. Y., mated to be a relatively slow processlo unless C0(11)~~ 1952, p 21 1 . (6) P. George in “Oxidases and Related Redox Systems,” Vol. I, T. ions are present; these may catalyze the water exchange E. King, H. S. Mason, and M. Morrison, Ed., John Wiley and Sons, I ) ~ ~ exchange. It seemed b y C O ( I I I ) ~ ~ - C O ( Ielectron Inc., New York, N. Y., 1965, p 5. (7) M. Anbar in “Mechanisms of Inorganic Reactions,” Advances in plausible therefore that when the aquo complex of Chemical Series, No. 49, American Chemical Society, Washington, Co(II1) labeled with reacts with H2160, the rate o f D. C., 1966, p 126.
C
+
+
+
(8) F. P. Gortsema and J. W. Cobble, J. Am. Chem. Soc., 81, 5516 (1959). (9) D. D. Thusius and H. Taube, ibid., 88, 850 (1966).
(10) H. Taube, Chem. Rev., 50, 69 (1952). (11) H. L. Friedman, H. Taube, and J. P. Hunt, J . Am. Chem. Soc., 73, 4028 (1951).
Anbar, Pecht / Oxidation of Water by Cobaltic Aquo Ions
2554 Table I. Calculated and Experimental Values of the Isotopic Distribution of Oxygen in 0 2 Evolved from Co(III)~,and Water
% of Co(II1) decomA t o m z 180 B C posed A 0-100 0-20 20-39 39-59 59-89 0-9 9-31 31-79 0-16 16-40 40-72 0-10 10-22 22-78 0-8 8-13 13-63 0-17 17-29 29-42 42-77 0-8 8-14
65 63 63 63 63 63 63 63 63 63 63 65 65 65 33 33 33 33 33 33 33 33 33
2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.6 2.0 2.0 2.0 1.4 1.4 1.4 1.4 1.4 1.4
6.65 16.5 6.49 3.26 2.77 32.0 5.50 2.70 22.8 3.85 2.47 31.4 5.51 2.52 17.8 11.0 2.20 11.54 5.80 3.12 1.51 16.2 11.5
P
X
Y
2
0.0677 0.229 0.0640 0.0109 0.0027 0.487 0.048 0.001 0.336 0.0239 0.0012 0.463 0.0497 0.0019 0.509 0.291 0.0071 0.321 0.139 0.0535 0.0035 0.469 0.321
0.829 3.14 0.88 0.149 0.033 6.66 0.657 0.013 4.60 0.327 0.015 5.67 0.608 0.023 22.83 13.08 0.318 14.40 6.28 2.40 0,156 21.07 14.40
3.08 10.69 3.00 0.508 0.123 22.69 2.24 0.045 15.68 1.11 0.053 21.08 2.26 0.086 22.49 12.88 0.312 14.19 6.18 2.36 0.153 20.75 14.19
2.86 9.10 2.56 0.433 0.114 19.3 1.90 0.039 13.35 0.95 0.045 19.57 2.098 0.080 5.54 3.17 0.077 3.49 1.52 0.582 0.037 5.11 3.49
zk mole fraction I m 88.81 73.10 88.76 93.83 94.61 48.70 90.30 94.77 62.96 92.60 94.76 51.13 90.52 95.08 47.18 68.10 95.36 66.02 83.62 92.00 96.88 51.59 66.02
4.37 3.90 4.74 5.01 5.05 2.60 4.82 5.06 3.10 4.56 4.54 2.51 4.45 4.68 1.93 2.78 3.89 1.87 2.37 2.61 2.75 1.46 1.87
>
0.053 0.051 0.062 0.066 0.067 0.034 0.063 0.066 0.037 0.055 0.056 0.030 0.054 0.057 0.019 0.028 0.039 0.013 0.016 0.017 0.019 0.011 0.013
a
f
P
g
12.03 5.22 11.60 17.03 18.30 2.19 12.88 18.56 3.59 16.39 20.60 2.40 13.58 19.95 2.87 5.18 22.77 5.01 10.51 18.90 22.67 3.27 5.00
11.27 5.14 10.96 16.32 18.00 1.95 11.98 18.60 3.37 16.04 20.22 1.98 19.41 19.55 2.72 5.00 24.65 5.12 9.86 18.20 22.67 2.95 8.11
0.39 0.63 0.33 0.091 0.035 0.76 0.25 0.021 0.71 0.18 0.022 0.83 0.32 0.0003 0.22 0.20 0.028 0.21 0.18 0.12 0.10 0.23 0.21
0.34 0.60 0.29 0.065 0.015 0.61 0.23 0,022 0.54 0.16 0.024 0.53 0.053 0,018 0.20 0.18 0.072 0.16 0.13 0.092 0.075 0.050 0.14
ligand oxidation might compete with that of ligand and 1 6 , 1 8 0 2 and found it to b e far from the statistical: exchange resulting in the formation of molecular oxygen [16,1802]2/[18~1802][16,1602] = 4.12 If the O2 formed ions labeled with l 8 0 , originated from two C0(111)~~ partly labeled with l80. To check on these assumptions the oxygen should contain 18,1802. However, in view we have treated l8O labeled C0(111)~, with H2160and of the extensive isotopic exchange, it seemed improbable determined the l80 content of the oxygen formed. that 18,180z would be formed in any appreciable amount As there are no reliable data available on the rate of if the oxygen originated from the reaction between two the Co(III),,-H20 isotopic exchange, it was impossible separate C O ( H ~ O ) ions. ~ ~ + On the other hand, if the to predict the relative rates of isotopic exchange us. oxidation of water a priori. l81‘SO2 would originate from H2l80 labeled C0(111)~~ in a chemical form which does not readily undergo It has been found that (6.65 - 2.6)/65 = 6.2% of the isotopic exchange with the solvent, one could calculate oxygen evolved originates from the inner coordination the isotopic distribution of l 6 , l 6 0 2 , 1 6 v 1 a O ~ , and 1 8 , 1 8 0 2 sphere of C0(111)~~.This result could be interpreted assuming that the evolved 0 2 originates from two in two ways: either that 6.2 % of the reaction proceeds sources: O2 labeled with I 8 0 formed from the labeled by an inner-sphere oxidation mechanism and the rest Co(III),, exclusively and the rest of the O2 from the by an outer-sphere process, or that all the oxygen formed water exclusively. Following this assumption the originates from the ligand water but that an over-all mole fractions of 1 6 , 1 6 0 2 (x), 16,180~ ( y ) , and 1 * , 1 8 0 z average of 93.8% of the ligand Hz180 has exchanged ( z ) originating from the C0(111)~~ containing an atom with the solvent before oxidation. The second iny z ) of are in isofraction A = ( 0 . 5 ~ z)/(x terpretation seems to be much more plausible in view of the ‘ 8 0 content of 0 2 collected in successive fractions topic equilibrium, i.e., y2/xz = 4. Analogously the with water (Table I). (m) during the reaction of C0(111)~~ molar fractions l 6 , l 6 0 2 (k), 1 6 , 1 8 0 2 (E), and 18,1802 of oxygen originating from solvent contain an atom It may be seen that, whereas the first fraction collected contains up to half of the oxygen originating from the fraction B of l80( B = (0.51 m)/(k 1 m ) and inner sphere, the last fractions collected have an isotopic P]mk = 4). The oxygen formed is a mixture of molecomposition identical with that of the solvent. These cules originating from the two sources and contains an m)/ atom fraction C of l80:C = ( 0 . 5 ~ 0.51 z findings show that the relatively low yield of I 8 0 in the molecular oxygen formed is due to H20solvent-H~0~igand (x+ y+z+k+I+m),and,asx+ y + z + k I m = 1, C = 0.5(y I) z m. If one deisotopic exchange, and that it is very likely that all the fines the mole fraction of 0 2 originating from the oxygen evolved originates from the inner sphere of hydration. It is experimentally hard to sample oxygen C0(111)~~ as P = x y z, then C = A P B(l at the very early stages of C0(111)~~ decomposition, - P ) or P = (C - B)/(A - B). From these equations one may derive that x = P ( l - A ) 2 ; y = 2 A P . which could allow the isotopic composition of the evolved oxygen at time zero to be obtained. Thus a ( 1 - A ) ; z = P A ’ ; k = ( 1 - P)(1 - B)’; 1 = 2B(1 - B)(1 - P ) ; m = ( 1 - P ) B 2 . Thus the values of the small contribution ( < 5 %) of an outer-sphere mechanism ratios CY = 1 6 , 1 6 0 2 / 1 6 , 1 8 0 2 = (x k)/(y I ) and cannot be completely excluded. p = 1 * , 1 8 0 2 / 1 6 , 1 8 0 z = (z m)/(y 1) could be calculated In view of the competition between the ligand exon basis of the measured values of A , B, and C and change and ligand oxidation it was of interest to examine the exact isotopic distribution of in the (12) F. Z. Roginsky, “Theoretical Principles of Isotope Methods oxygen molecules produced. We have examined therefor Investigating Chemical Reactions,” Academy of Sciences, U.S.S.R. 0 ~Moscow, 1956. fore the isotopic distribution of l80 between 1 8 ~ 1 8Press,
+ +
+
+
+ +
+ + + + + + +
Journal of the American Chemical Society j 89:11 1 May 24, 1967
+ +
+ +
+ + +
+
2555
ative mechanisms are possible. (a) One is the intracompared with the ratios between the masses 16,160~ molecular conversion of the dimeric complex of Co(II1) l a 0 2 (36)experimentally deterinto a binuclear peroxy complex of Co(I1) mined by the mass spectrometric analysis of the evolved gas. f = and g = The results presented in Table I show a satisfactory agreement between the calculated values of CY and p and the experimental values off and g, respectively. Thus the assumption that the oxygen formed is a mixture of two types of oxygen molecules with practically no isotopic equifollowed by hydrolysis of the cobalto peroxide to H202 librium between them has been verified. As the value 2Co(II). The following steps would then be H202 of % was much smaller than % and 3, the accuracy of C O ( H ~ O ) --t ~ ~C+O ( H Z O ) ~ ~ +H+ HO2 and HO2 g is smaller than that off. Consequently the agreement C O ( H ~ O ) + ~ ~ C + O ( H ~ O ) ~ ~ 0 + 2 H+. (b) The between CY and f is evidently better than between fl and other is an attack of C O ( H ~ O ) on ~ ~ the + dimeric aquo g. cobaltic complex The only explanation of these results we are able to offer is the following. Oxygen is formed from binuclear [ ( H Z O ) ~ C O O H+ ] ~ ~C+o ( H z 0 ) ~ ~+ + H 61complexes of C0(111)~~and these oxygen molecules 0 originate exclusively from their inner hydration sphere. The C0(111)~~ solution (0.6 M in 6 it4 HC104) contains (HZO)~CO:: \:;CO(HZO)~] Co(HIO)sz+ a certain percentage (
