3670
0.AMICHAI AND A. TREININ
in pure water solutions. After bubbling the gas mixture for 30 min and adding the catalyst, the flask was closed by means of a stopcock and shaken for successive 30-min periods. Gas chromatographic analysis was made of samples of the gas mixture withdrawn into an evacuated bulb; separation was on a column at -80". The ratio of N2 to Hz in the gas mixture progressively increased (Figure 1) with time in the presence of the Pt catalyst while insignificant changes occurred in the control experiment. In the presence of methanol, the rate of reduction of E20 was somewhat diminished but this probably arises because of competitive chemisorption of radicals (e.g., 3 C-OH) derived from the metha n 0 1 . ~ ~Catalytic hydrogenation was hence indicated unless the extreme and rather unlikely view is taken that catalytic hydrogenations themselves proceed by an eaq- intermediate, vix.
Pt
+ '/2Hz
Pt'H
+ H 2 0 Pt + H + + eaqeas- + oxidant +reduced product PtH
This seems rather improbable since, as in the electrode case, adsorbed H can be detected at the Pt surface, e.g., by infrared absorption, and hydrogenated surface complexes are known. At a cathode passing current, adsorbed atomic H is also present21)22and will undoubtedly have similar reducing properties. It is evidently difficult to conclude that N20is reduced specifically at Pt by eaq- or to interpret competitive experi-
ments with methanol when dissociative chemisorption occurs in a heterogeneous system.
6. Conclusions Difficulties are shown to arise in mechanisms of cathodic hydrogen evolution and for most, but not all, cases of metal dissolution if it is supposed that hydrated electrons are under most conditions the initally formed species at neutral or alkaline pH. Under certain circumstances, e.g., for dissolution of the basest metals in water, i e . , at sufficiently high negative potentials in S a amalgam dissolution or in photoassisted processes, a small steady-state concentration of eaq- could be established. It seems unlikely, however, that this can be the general mechanism for hydrogen evolution and other cathodic processes, particularly for the more catalytic metals which chemisorb H and at which H2evolution can proceed at high rates but at relatively low cathodic potentials near that of the reversible hydrogen electrode Acknowledgments. Grateful acknowledgment is made to the National Research Council, Canada, for support of this work and D. J. M.acknowledges the award of a National Research Council Graduate Scholarship. We thank Messrs. J. Esser and P. Schuchmann for performing the gas chromatographic analyses, and Professors G. Stein, U. Schindewolf, and P. Delahay for their comments on this paper prior to its publication. (53) V. S. Ragotskii and Y . B. Vasilev, Electrochim. Acta, 12, 1323 (1967); 11, 1439 (1966); see also M. W. Breiter, Discuss. Faraday Soc., 45, 79 (1968).
On the Oxybromine Radicals by 0. Amichai and A. Treinin Deparlment of Physical Chemistry, Hebrew Unbersity, Jerusalem, Israel
(Received March 91, 1970)
The pulse radiolysis of BrOs- in water and its photolysis in boric acid glass were investigated. Both techniques give rise to three intermediates with Amax at ca. 315, 350, and 475 nm. They were identified as Br03, BrO, and BrOz, respectively. BrO results from the decomposition of BrOa; this reaction appears to proceed by different mechanisms in water and in the rigid matrix.
The flash photolysis of BrOs- gives rise to transient absorption in the region 300-500 nm.' The band peaking at 475 nm was definitely assigned to Br02.2 The absorption below 400 nm is short-lived and overlapped by the spectrum of BrO-, which is a permanent is rather product Of the photolysis; therefore its difficult. Bridge and Matheson reported1 an absorpThe Journal of Phy/sical Chemistry, Vo2. 74, No. $Os 1970
tion with a double peak in the region 350-390 nm decaying by a first-order process which they assigned to Br08. An absorption centered around 360 nm was (1) N. K. Bridge and M. s. Matheson, J . Phys. Chem., 64, 1280 (1960). (2) 0 . Amichai, Czapski, and A. Treinin, Isr. J . Chem., 7, 351 (1969). (7.
OXYBROMINE RADICALS
3671
produced by the pulse radiolysis of Br03- and was similarly interpreted. However, no evidence was presented for this assignment. Our recent flash photolysis work2p4has shown that a t least two primary processes occur in excited Br03-
400 -
1
I
8rO-
I
1
350
400
300
+ 0- (or OH) Br03- +BrOz- + O(3P)
BrO3- +Br02
(1) (2) I
Under the conditions employed, the OH radicals react with Br03OH
+ BrOs-
---f
BrO3
+ OH-
-f; 200
w
(3)
Thus the photolysis is expected to yield Br03 radicals. 350 nm5) is However, the possibility that BrO (A, aiso produced should be carefully considered. The photolysis of IO3- yields IO3, 102, and IO; pulse radiolysis6 and matrix isolation' techniques were employed to reach this conclusion. Here we wish to report the results of similar experiments with Br03-.
100
250
300 X.nm
Experimental Section The pulse radiolysis setups and the preparation of boric acid glasses* were already described. The glasses were irradiated at 2537 8 and their absorption spectra were recorded with a 450 Perkin-Elmer spectrophotometer, using the intensity scale to measure optical 15%). densities down to (reproducible within A holder was designed to give good thermal contact with the glass; it was thermostated inside the cell compartment (flo). NaBrOB (Analar grade) was used without further purification. NaBrOz was supplied to us by the Soci6t6 d'Etudes Chimiques pour 1'Industrie et I'Agriculture, France.
*
Results The spectra of the oxybromine anions pertaining to our research are recorded in Figure 1. The absorption of BrO- is somewhat higher than that previously recorded; unsatisfactory correction for the dispropsrtionation reaction is responsible for the discrepancy. (1) Pulse Radiolysis. Figure 2 shows some transient spectra produced in air-free 0.1 M KBr03 at pH 4.6. Immediately after the pulse, the spectrum displayed a band peaking a t -315 nm (band A). Its fast decay ( n l 2= 5 psec, independent of the bromate concentration) proceeded in parallel with the growth of another band -350 nm; the latter decayed by a second(B) with A,, order process with 2k = (9.0 f 0.6) X 106€360M-' see-' (Figure 3). (€360 is the extinction coefficient at 360 nm.) In alkaline solution (pH 12) transient B was observed, but the spectrum of A was overlapped by that of some slowly decaying transient, with A, below 280 nm; a permanent absorption was finally left, similar to that of a solution containing BrO- and BrOz- in 1: 2 mole ratio (Figure 2). Below pH 5, BrOcould not be detected because of its rapid protonation to
Figure 1. The spectra of BrO-, BrOn-, and BrOs- in aqueous solution a t pH 13.6.
HBrO, which weakly absorbs above 300 nm;lo BrOz- is unstable under these conditions. 1' Therefore, band B appeared well defined only in acidic solution. Even in neutral solutions the transient absorption revealed a slowly decaying component which could be ascribed to BrO- reacting with H30+ (probably with some contribution from BrOz- below 330 nm). Above 400 nm the familiar spectrum of BrO2 peaking at 475 nm2m6was observed immediately after the pulse. A mixture of 0.44 M Br03- and 1.6 X M BrOz(air free) at pH 11.9 was pulsed. (Under these conditions all solvated electrons and most OH radicals should be scavenged by Br03-.2112) The decay of transient B was found to be much faster in this system. The effect of N20was tried with 4 X M KBr03 solution, so that most ea4- were effectively scavenged by N20. BrOz could hardly be detected but the absorption below 400 nm was enhanced by ~ 2 0 %compared with the same solution saturated with argon. On the other hand, ethanol did not affect the amount of BrzO (3) M. S. Natheson and L. M. Dorfman, J . Chem. Phys., 32, 1870 (1960).
(4) 0. Amichai and A. Treinin, Chem. Phys. Lett., 3 , 611 (1969). (5) G. V. Buxton and F. S. Dainton, Proc. R o y . Soc., Ser. A , 304, 427 (1968). (6) 0. Amichai and A. Treinin, J . Phys. Chem., 74, 830 (1970). (7) 0. Amichai and A. Treinin, J. Chem. Phys., 53, 444 (1970). (8) A. Gitter and A. Treinin, ibid., 42, 2019 (1965). (9) 0. Amichai and A. Treinin, to be submitted for publication. (10) L. Farkas and F. Klein, J . Chem. Phys., 16, 886 (1948). (11) J. Breiss, Ing. D. Thesis, University of Strasbourg, Strasbourg, France, 1959. (12) M. Anbar and P. Neta, Int. J . A p p l . Radiat. Isotop., 18, 493 (1967).
T h e Journal of Physical Chemistry, Vo1. 74, No. 20, 1970
3672
0. AMICHAI AND A. TREININ eaqc.
300
/--
E '.
350
325
A,
375
nm
Figure 2. Absorption spectra of the trrtnsients produced by pulse radiolysis of 0.1 M Br03- a t pH 4.6, 5 gsec (a), 10 psec (b), 30 psec (c), and 50 psec (d) after pulse. Curve e: the spectrum of BrO from pulse radiolysis of 2 X lod4M BrOat pH 11.6. Curve f : the permanent absorption from pulse radiolysis of 0.1 ilf BrOa- a t pH 12.6, 40 msec after pulse. BrOz- in 1 : 2 molar ratio. Curve g: the spectrum of BrO-
+
+ Br03- -%BrOz + 20H-
Figure 2 records for comparison the spectrum of BrO which was produced by pulse radiolysis of air-free BrOat p H 11.6. BrO- was prepared by mixing equivalent amounts of C10- and Br-; C1- does not interfere since its reaction with OH is much slower.12 The spectrum of BrO thus produced is very close to that previously reported.5 (2) Matrix Isolation. The solubility of SaBrOa in boric acid glass and its extinction coefficient at 2537 A are rather low; therefore long irradiation times were required to obtain detectable spectral changes. The results are shown in Figure 4. Three bands emerged on irradiation: bands A, B, and C peaking at ca. 320, 350, and 470 nm, respectively. Band A could be readily annihilated by moderately warming the glass, bands B and C growing in parallel with its decay. At 41" the decay was found to be first order with k = 7 X sec-l. At 85", bands B and C began to fade and a new band (D) grew up, peaking at -410 nm. Under this treatment the glass was losing its transparency, so complete annihilation of B and C could not be achieved.
I
1
I
I
I
I
450
500
550
D
I
350
400
A.
nm
Figure 4. The photolysis of Br08- in boric acid glass at 2537 A. Absorption curves: (a) after 4.5-hr irradiation; (b and c) after 2 and 19 hr, respectively, at 41"; (d) after 2 min a t 85".
c,ysec
t,ysec
Figure 3. Plots for the decay of BrO under various conditions: (a) second-order decay in 0.1 M at pH 4.6; (b) secondorder decay corrected for the formation of BrO- in 0.08 M Br03- at pH 13; (c) first-order decay corrected for the M BrOzformation of BrO- in 0.44 M Br031.6 x a t pH 11.9.
+
(its rate of decay slightly increased) but completely suppressed the absorption at shorter wavelengths. (This result was obtained with 0.1 M Br03- and 2 X M ethanol.) This shows that within the limit of error (&lo%) the sole source of BrOz is the reaction5 The Journal of Physical Chemistry, Vol. 7& No. $0, 1970
Discussion Both techniques lead to similar results: in acidic solutions three intermediates are produced by irradiation, but B appears to involve A as its precursor. Band C (Figure 4) is evidently due to Br02, with A,, nearly the same as in water. However, contrary to the aqueous system, BrOz also results from some thermal process in the glass, which may involve transient A. The effects of ethanol and NzO on the pulse radiolysis indicate that OH radicals are the precursors of transients A and B. Ethanol scavenges OH while NZO converts eaq- to OH. If all the hydroxyl radicals were reacting with Br03-, then the amounts of A and B should
OXYBROMINE RADICALS
3673
be doubled in the presence of N2O (since G, = GOH). However, reaction 3 is relatively slow2 and thus under the conditions employed ((OH)o= 10-5 M and (Br03-) = 4 X A I ; these conditions were chosen to get detectable yields of radicals and to let NzO effectively scavenge e,,-) only -25% of the OH radicals react with BrOa-. The following considerations led us to assign band B to BrO. (a) In aqueous solution it closely resembles the 350-nm band of BrO (Figure 2). (The discrepancy below 330 nm may be due to BrOz- undergoing decomposition.) (b) The decay of band B at pH 4.6 is somewhat faster than that reported6 for BrO in the pH range 1114. This may still be within the limit of error or a genuine pH effect. In order to determine the rate of decay in alkaline solution we had to correct for the absorption of the reaction products. Assuming that BrO decays by the reaction5
2Br0 +BrO-
+ BrOz-
(4)
and ignoring the contribution of BrOz- (see Figure l), we derived the expression
l - 0 = -1+ - 2k4t A - PA0 A O EBrOi
(5)
2 is the optical path; A and A0 are the optical densities at the given time and zero time, respectively (the latter was found by extrapolation; the short formation period of BrO was ignored); 0 = €BrO-/2EBrO, where EBrO- and eBrO are the extinction coefficients of BrO- and BrO, respectively. The validity of eq 5 was checked at 360 nm, setting E B ~ O = 9005 and B B ~ Q -= 200 A I - l cm-l. A straight line was obtained (Figure 3) from which 2k4 = (4.8 f 0.7) X lo9 M-l sec-l was derived in agreement with previous data.5 (c) BrO is known5 to react with Br02- (in alkaline solution) as in
BrO
+ Br02-
---)
BrO-
+ Br02
(6)
This explains the effect of BrOz- on the decay of B. By correcting for the absorption of BrO- produced, the expression In ( A - 2PA0)
=
In ( A o- 2pA0) - ka(BrOz-)t (7)
was obtained with all the symbols as previously defined. The plot for 360 nm is shown in Figure 3, from which k6 = (4.0 f 0.5) X los nd-lsec-'masobtained. This value agrees (within the limit of error) with previous data. (d) Above 0.1 ill BrOa-, reaction 3 should be effectively completed within the pulse duration2 Therefore, transient B which still grows after the pulse cannot be due to BrOa, as previoulsy supposed.'S3 Of Bra' In Transient A appears to be the boric acid glass the conversion of A to BrO is first order,
occurring at relatively low temperatures. This leads us to assign band A to BrOa and postulate the reaction BrOI +BrO
+ O2
(8)
as responsible for the generation of BrO. An analogous reaction occurs with 1 0 3 in boric acid glassa7 In general, boric acid glasses doped with 1 0 8 - and Br03- show close resemblance. The X03radicals most readily decompose while XOz and XO must diffuse for annihilation. All the oxyiodine radicals finally yield iodine.7 The oxybromine radicals are expected to behave similarly; indeed, band D (Figure 4) is most 410 nm in the gas probably due to Brz which has A, phase. 13 There is one major difference between the two systems: with IOa--doped glass IOzis produced only by primary photodissociation; there is no thermal generation analogous to that of BrOz. This thermal reaction does not occur in aqueous solution, where Br03 is much less stable than in the rigid system. There is no evidence that reaction 8 occurs in one stage and that Br03undergoes the same reaction in both fluid and rigid phases. Thus in boric acid glass the decomposition may occur in two stages with BrO2 being the intermediate. Little is known on the decay of in aqueous solution; the short lifetime and weak absorption made the analysis of its kinetics unreliable. The lack of (Br03-) effect rules out its reaction with bromate. Still, the profound phase effect suggests some basic difference in mechanism, more than a simple cage effect. (In aqueous solution 1 0 3 undergoes a second-order decaya but this process does not generate IO.) The assignment of band A to Br03 is also in keeping with the following regularities? (a) the transition energies of the oxyhalogen radicals increase in the order X03 > XO > X02; (b) hulllaxof XO, increases almost linearly with the electronegativity of the halogen atom. This strongly supports our analysis of the transient spectra. By extrapolating the yield of BrO to zero time (ignoring its short period of formation) and taking G B r O = GOH, we obtained Empx(BrO) SO0 M-I cm-I in fair agreement with previous data.6 E,,,,~(BTO~) -1000 M-' cm-1 was estimated from the absorption at 315 nm immediately after the pulse. According to the mechanism proposed, nearly equivalent amounts of BrO and Br02 are generated by the GoH). BrO- is produced only from pulse (since G, the decay of BrO (reaction 4)) while BrOz- owes its formation to reaction 4 and to the decay of BrOZ5
-
2BrO2 +Br02-
+ Br03-
(9) 2 , obtained
This explains the ratio (BrOz-)/(BrO-)= in alkaline solution (Figure 2 ) . From the amount of radicals produced in the glass (13) A . G. Briggs and R. G. W. Norrish, Proc. R o g . Soc., Ser. A , 276, 51 (1963).
The Journal of Phgsical Chemistry, Vol. 74, N o . 10,1970
NOTES
3674 the quantum yield of reaction 1 was estimated as ca. 0.1, which is close to that in aqueous solution. Thus for both IO3- and Br03- the reaction is not hampered in the rigid matrix by cage recombination. The OH radicals produced by reaction 1 (the matrix is highly acidic*) are expected to diffuse fast and react withX03-. The 0 atoms produced by reaction 2 (there is no ando-
gous reaction with IO3- 4, may behave similarly, but their reaction with Br03- should yield BrOz- or Br04- l4 which hardly absorb above 300 nm. The nature of the slowly decaying transient observed in alkaline solution is still not clear. (14) E. H. Appelman, Inorg. Chem., 8 , 2 2 3 (1969).
NOTES
Activity Coefficients in Equimolal Mixtures of Some Divalent Metal Perchlorates
final concentrations were found from the change in weight. Details of the isopiestic experiments were as described in ref 1.
by Zofia LibuA and Teresa Sadowska
Results and Discussion
Department of Physical Chemistry of the Institute of Chemical Engineering and Measuring Techniques, Technical University of Gdahsk, G d a h k , Poland (Received February 5 , 1970)
The concentrations of the isopiestic C O ( C ~ O ~ ) ~ X(Cl04)2 and Zn(C104),-Ni(C104)2 mixed solutions, along with the corresponding water activities as well as the values of A, as defined below, are given in Tables I and 11. The water activities were determined by reference to the data concerning the corresponding single salt solutions. Inspection of the tables shows that in each case the total concentration of the isopiestic solutions shows but a very small variation with the mole fraction of one of the two metal perchlorates. The variation does not exceed 0.5% in the case of the C O ( C ~ O ~ ) ~ - T U ’ ~mixtures, ( C ~ O ~ ) ~and 1.0% in the case of the Zn(C104)2-Ni(C104)2mixtures. In order to check whether the relation between the total concentration of isopiestic mixtures, on one hand, and the mole fraction of the component salts, on the other, may be approximated by a straight line, the deviation from linearity has been calculated defined as A = na - ml@xl- m2@x2 for each experimental point. In this expression m denotes the total concentration of the mixed solution, ml$ and mZddenote the concentrations of the corresponding isopiestic pure solutions, and x1 and x2 are the mole fractions of component salts in the mixture defined as: x1 = ml/m and x2 = m2/m,ml and mi!being the concentrations of salts 1 and 2, respectively. As is seen, the values of A are, in general, smaller than 0.2% of m. On this basis we assume that the dependences under consideration are linear, within the experimental error, for each series of the isopiestic solutions studied in this work. Further conclusions arising from the above results
Previous work has shown that aqueous solutions of manganese(II), cobalt(II), nicltel(II), and zinc(I1) perchlorates display the same, in approximation, concentration dependences of their osmotic and activity coefficients in the whole concentration range studied, ie., up to approximately 3 m.l In the present paper we report the results obtained for mixtures of some of the divalent metal perchlorates completing, in this way, our study on the osmotic and activity coefficients of the divalent metal perchlorates in aqueous solution.
Experimental Section Stock solutions of C0(Cl04)~and Ni(ClO& were obtained and analyzed as described in ref 1. Zinc perchlorate, obtained by dissolving reagent grade zinc oxide in aqueous analytical grade perchloric acid, was purified by several recrystallizations, the last operation being performed using conductivity water. The stock solution of Zn(C10e)2was analyzed for zinc by standard EDTA titration as well as gravimetrically by the pyrophosphate method. The results obtained by the two methods were consistent to within 0.01%. Equimolal solutions of Co(C104)2, Ni(C10&, and Zn(C104)2were prepared from the corresponding concentrate stock solutions and were further used in the preparation of mixed solutions having constant total molalities, but different contents of the component salts. Every six equimolal solutions thus obtained (four mixtures and two single salt solutions) were equilibrated in the isopiestic apparatus at 25’ and their T h e Journal of Physical Chemistry, Vol. 7 4 , N o . $0,1970
(1) 2.Lib& and T.Sadowska, J . Phys. Chem., 73, 3229 (1969).
