HAROLD -2.SCHWARZ AND
5636
JANE
Vol. 80
M. HRITZ
[CONTRIBUTION F R O M rIiE CHEMISTRY DEPARTMENT, BROOKHAVEN x A T 1 O N A L LABORATORY ]
The Radiation Chemistry of Deaerated Ferrous Chloride Solutions1 BY HAROLD ,4. SCHWARZ AND JANE M. HRITZ~ RECEIVED APRIL 18, 1958 The oxidatioti of deaerated F e c i z solutions in 0.4 H C l is qualitatively similar to the oxidation in H2S04. However, the decrease in yield due to accuniulation of Fe(II1) as the reaction proceeds is much more marked in HC1 solutions. I n the presence of Fe(III), the yield approaches equivalence with the molecular Hz formed by the radiation. This property allows accurate measurement of the hydrogen yield in 0.4 M HC1 solutions and the variation of the yield with Fe(II1) concentration. T h e ratio of rate constants for the reactions H Fe(II1) -4 H + Fe(I1) and H H+ Fe(I1) -+ H Z Fe(II1) is found t o be 170, and the molecular Ht yield is found to vary according to the relation G R ~= 0.45 - 0.64(FeIII)'/s where GH*is the number of H Zmolecules produced per 100 e.v. adsorbed in the solution.
+
The radiation induced oxidation of FeS04 solutions in 0.4 M HzS04 has received considerable attenti~n.~ In particular, for C060 and high energy electrons, 3b the yield in air-saturated solution is 15.5 Fe(I1) ions oxidized per 100 e.v. absorbed in the solution. In deaerated solutions the yield is only 8 . 2 . 3 d , g In a study of the oxidation of Feel2 in air-saturated 0.4 Jf HC1, the initial yield was 15.8 ions oxidized per 100 e.v., in agreement with HzS04 solution^.^ However, the yield decreased a t high doses, showing that the Fe(II1) chloride complexes compete effectively with oxygen for H atoms. The ratio of the rate constants for the reactions H
+ Fe(II1) --+ 11 4-Fe(I1) H + 02 +HO:! -
was found to be k F e ( I I I ) l k o t = 0.21 in 0.4 iM HCl. This ratio was dependent on chloride concentration indicating that the various ferric complexes react a t somewhat different rates. Fe(II), however, is not appreciably cornplexed in either HCl or HzS04 solutions. To a first approximation, one might expect the k o Z I k F e ( I I ) rate constant ratio, 1200 in 0.4 1 11 H2S04,3d to be the same in HCl solutions. The product of these two rate constant ratios should give the ratio of rate constants for H atoms reacting with Fe(I1I) versus Fe(I1) in HCl solution, ~ F ~ ( I I I ) / ~ F ~_N ( I 0.21 I) X 1200 N 230. This ratio would apply in the kinetics of oxidation of Fe(I1) in deaerated HC1 solution. By analogy with HzS04 solutions,$f we might expect the mechanism H,O -+ I ~ II-TKI., ~ , 11, o I r Fe(II1) -+ €1'. Fe(I1I H HFe(I1) --+IT2 Fe(II1) OH Fe(I1) --+ O H Fe(II1) H20s 2Fe(II) +2OIT 2Fe(III) H
+
+ + + +
+ + + +
(1) (2) (3) (4)
I n HC1 solutions, reaction 3 probably proceeds via the intermediate production of a chlorine atom, but this does not matter in the present work, as the net result in either case is the oxidation of one Fe(I1). This mechanism and the equation of material balance, 2 G ~ , o , Go11 = ~ G H ,j- GH lead to the equation for the rate of production of
+
(1) Research periormed under the auspices of t h e U. S. Atomic Energy Commission. ( 2 ) Alfred University, Alfred, New York. (3) (a', C. J . Huchanadel and J . A . Ghormley, J . Chem. P h y s . , 21, 880 (19.731; (b) R . H . Schuler and .4.0. Allen, i b i d . , 2 4 , 56 (l95G); (c) T. Rigg, G . Stein and J . W e i s s , Pvoc. R o y . Soc. (Loizdon), A a l l , 375 (1952); (d) A . 0. A l l e n and n-. G. Rothschild, Radiation Reseavch, 7, 591 (1957); ( e ) A. 0. Allen, 1'. Hogan and W. G. Rothschild, ibid., I , 603 (19571, ( f ) W. G. Rothschild and A . 0. Allen, ibid., 8 , 101 (1958,; ( g ) N. F. Barr and C. G. King, THISJOIJRXAL, 78, 303 (1956). ( 4 ) H. A. Schwarz, ibid.. 79, 534 :10:7>
+
+
+
+
Fe(II1) (including any H + concentration dependence of reaction 2 in the rate constant k z )3d
As kl/kz is of the order of 250 and GH? and GH are approximately 0.4 and 3.7, respectively, then a t (FeIII)/(FeII) ratios as small as 0.05, the rate of oxidation of Fe(I1) should be governed principally by the magnitude of G H ~ . This suggests that a study of deaerated FeClz solutions might lead to accurate measurements of G H ~in acid solutions. The following work was performed to check the above mechanism and to study G H ~and its possible dependence on Fe(II1) concentration. Experimental Standard solutions of reagent grade FeS04.7H20, Fez(S04)z.XHzO and HC1 were prepared. rlliquots of the stock solutions were added to the triply distilled water.6 The HC1 concentration was maintained near 0.4 Af. I n the worst case, about 5% of the Fe(II1) was in the forin of FeS04+. Kormally it was less than 1%. The solutions mere poured into tubes 15 cm. long and 10 mm. i.d. and deaerated by N2 bubbling.6 They were stoppered and irradiated in a cylindrical Co60 s o ~ r c e : .T~h e dose rate was approximately 2.5 X l0l9e.v./l. min. With this procedure, the scatter in the data was slightly greater than was expected (amounting t o 1 3 p M Fe(111)) and all of the results appeared high by about 7 p M Fe(II1). The effect was still present but smaller when 400 p M of Fe(II1) mas added. If 0.4 N HC1 was irradiated, TT-ith neither Fe(I1) nor Fe(II1) added, a n optical density increase of about 0.010 was noted at 3.50 mp. T h e origin of this was not determined, b u t the results were found to be more reproducible if the 0.4 N HC1 was prepared and preirradiated with about IO5 roentgen before adding the stock iron solutions. This procedure was followed in all of the runs with added Fe(II1). A blank correction, 7 H M of F e ( I I I ) , was subtracted from all of the data collected previous to this on solutions without added iron (Fig. 1). The Fe(II1) was estimated by measuring its absorption a t 335 mp in a Beckman DU ultraviolet spectrophotometer. (The molar extinction coefficient E F ~ ( I I I )= 1340 in 0.408 N HCl.)4 \!.hen Fe(II1) was added initially, the irradiated solution \Tas compared directly to the unirradiated solution in separate cuvettes by setting the instrument a t 100:; transmission when the unirradiated solution was in the light beam. n'ith this method, small changes in the optical density could be measured even though the total optical density was greater than unity. The 3 X and 10- 3 I Fe(II1) solutions were diluted 1 : l and 1 : 4 with 0.4 N HC1 before comparing to avoid stray light effects a t very high optical density.
Results and Discussion The oxidation of deaerated Fe(I1) in 0.4 fl and lo--' ilf Fe(I1) HC1 is shown in Fig. 1 for (.5) E. R. Johnson and A . 0. Allen, ibtd., 7 4 , 4147 (1952) (6) H. A. Schwarz and A. J. Salzman, Radioliorz Resenvch, in press. 1.7) H. .4.Schwarz and A . 0. Allen, Sucleonics, 1 2 , KO.2 , 5 8 (1954).
RADIATION CHEMISTRY OF DEAERATED FERROUS CHLORIDE
Nov. 5. 1958 1
9
1
-
-
1
-
I
4
lo
0
I
2 DOSE.ev/C
3
4
I-
5637
i
o/";;)
5
XIO-~'
Fig. 1.-Oxidation of Fe(I1) in deaerated 0.40 M HC1: M Fe(I1); 0 , lo-* M (Fe"). A blankcorrectionof 0.5 X 1019molecules/l. has been subtracted from the data (see text). Curves are calculated from eq. 2.
0,
The results agree qualitatively with the mechanism given in the introduction and expressed in eq. 1, ;.e., the rate of oxidation is not linear with dose and is greater in : I IFe(I1) than in M Fe(I1). I n order to obtain a quantitative comparison, eq. 1 is integrated with the assumption that (FeII) remains constant, This assumption does not alter the form of the equation and any change in the constants (which amounts to a few per cent.) is taken up in k l l k z . However the meaning of the equation is more obvious in this form.
A(FeII1) is the increase in (FeIII) concentration and (FeII1),/(FeII)o is the initial ratio of concentrations. When no Fe(II1) was added initially, this ratio amounted to 2.5 X because of the Fe(II1) contained in the reagent FeS04.7H20. GH was taken as 3.703d4and G H ~was taken as 0.41. The only constant left, k l / k 2 , was varied to give the best fit with the data, yielding k l / k 2 = 170. The curves drawn in Fig. 1 are calculated from eq. 2 with these constants and are in excellent agreement with the data. Furthermore, the value of kl,lk2, 170, is in reasonable agreement with the value predicted in the introduction. 250, considering the experimental errors involved and possible effects of the two different media. Figure 2 illustrates the effect of adding Fe(II1) initially. The rate of Fe(I1) oxidation is approximately linear in these solutions, as is predicted by eq. 2. Applying eq. 2 to the data in Fig. 2, we can obtain precise values of G H ~ . The logarithmic term in the equation is small compared to A(FeII1) and was computed assuming a value of 0.40 for G H ~ . This simplified the calculation of this term and can introduce only a negligible error, 0..j70 in the worst case. The best values of G H ~obtained are given in Table I. G H ~decreases as Fe(II1) increases, which is to be expected as analogous to the behavior of GH, in solutions of other solutess capable of reacting with hydrogen atoms. ( 8 ) (a) H A Schwarz, THISJ O U R N A L , 77, 4980 (1955); (b) J A. Ghormley and C J Hochanadel, Radzntzon Research, 3, 227 (1955);
1 0
IO
1 30
20
I
I
1
1
I
40
50
60
70
80
1 1 90
DOSE, ev/< x 16': Fig. 2.-Oxidation of Fe(I1) in deaerated 0.40 M HCI containing added Fe(II1): (A), (Fe(II1)) = 0.996 X lo-* M; M; (C), (Fe(II1)) = 0.940 X (B), (Fe(II1)) = 3.80 X M ; (D), (Fe(II1)) = 3.96 X M ; (E), (Fe(II1)) = 1.24 X lo-' M . For curve A, (Fe(I1)) = 3 X M ; for all others, (Fe(I1)) = 1 X iM. Note the displacement of the curves on the ordinate. Curves D and E have been raised 10 and 1 units from their respective origins.
I n the radical diffusion mechanism proposed to explain the appearance of HZand H2Oz among the products of the radiolysis of aqueous ~ 0 1 u t i o n s , * ~ ~ ~ ~ ~ 0 H atoms and OH radicals are assumed to be the only products formed initially. These are formed in small spurs, each containing an average of about TABLE I CALCULATED GHZI N SOLUTIONS CONTAINING Fe(II1) (FROM EQUATION 2) (Fer111 a v , A4
GH2
x
10-4 10-4 10-3 10-3
0.420 .405 ,375
1.05 X lo-*
.315
1.94 4.55 1.09 3.08
x x
x
.360
three dissociated water molecules, and widely separated in the case of y-rays. The radicals diffuse out of the spurs, occasionally encountering one another to form Hz, H202 or HzO. If they escape this fate, they can react with a solute present. Thus, in the present case, the combination reaction of two H atoms is in competition with diffusion H+H+H* A. Mahlman and J. W.Boyle, J . Chem. P h y s . . 27, 1434 (1957), (d) R.G.Snowden, THISJOURNAL, 79, 1263 (1957). (9) A. 0. Allen, Disc.Faraday SOL, 12, 79 (1952). (10) A. K. Ganguly and J . L. Magee, J . Chem. Phys., 26, 129 (1955). ( c ) H.
HAROLD A. SCHWARZ AND
5638
away from each other and reaction with Fe(III), reaction 1. Addition of Fe(II1) to the solution will tend to repress the combination reaction, as is characteristic of competing xeactions, lowering the H2 yield. Sworski has noted the empirical relation that the molecular yields tend to decrease linearly with the cube root of the solute concentration.l’ This relation breaks down with large changes in G H , but ~~ should be valid in the range covered in this work. A cube root plot of the data of Table I is given in Fig. 3. The linearity is in agreement with the
\ I
0.30
0
~
3.05
0.10
015
0.20
8:
__ 0.25
( Fern)
Fig. 3.-The variation of the hydrogen yield with Fe(II1) concentration; G H ~is calculated from eq. 2, and Fe(II1) is the average concentration present during the irradiation. The dotted line is the curve predicted by a radical diffusion mechanism.
other systems. Since this cube root relation is empirical, a comparison is also given in Fig. 3 with a one parameter curve representing an approximate solution of the diffusion-combination mechanism outlined above.8a The agreement with the prediction of this mechanism is considered good in view of the approximations that were made in calculating the curve. This is the first demonstration of the validity of the cube root relation for hydrogen yields in highly acid solutions, indicating that the possible reaction3c does not interfere with the solute dependH
+ He+
Hz’
( 5)
ence. Rothschild and Allen have suggested that this reaction is very rapid, competing effectively with O2for H If a new reaction scheme is (11) T. J . Sworski, THISJOURNAL, 76, 4687 (1954).
JANE
M. HRITZ
Vol. 80
proposed for our system in which the H atom assumes another form, such as H2+, the kinetics will remain unchanged, except that the rate constant ratio determined by us will apply t o the new species rather than the H atom. Another point of interest in Fig. 3 is the extrapolated value of G=* at “zero” Fe(II1) concentration. This value, G H ~= 0.45, is the same as is found in neutral solutions, GH, = 0.45. Mahlman and Boyle have obtained the same yield in dilute KBr solutions in 0.4 N H~S04.l~Previously i t was believed that the H2 yield was approximately 1570 lower in the acid s o l ~ t i o n s . ~ JThe ~ measurements leading to this conclusion were made principally in solutions saturated with 0 2 or containing iodine or ceric ion. These solutes react efficiently with H atoms and would be expected to lower the H Zyield somewhat. As was mentioned in the Introduction, the ratio of rate constants for H atoms reacting with 0 2 versus Fe(II1) is 4.8. A solution containing 1.3 X M 02 would be expected to correspond to a Fe(II1) solution 4.8 times this concentration, or 6.2 X M Fe(II1). From Fig. 3, the value of GH, predicted for this concentration is 0.34, in excellent agreement with the observed value of Ghormley and Hochanadel of 0.35.8b This appears ~ to be an adequate explanation for the previously observed low yields. The value of GH, used in eq. I, 0.41, is in agreement with Fig. 3 since the average Fe(II1) concentration in Fig. 1 is about M. The Hz yield in neutral 02-saturated solution is somewhat higher, 0.40.8b Since the yields in the absence of 02 are the same, 0 2 is apparently more effective in reducing in acid solution. The yields probably vary with the cube root of the 0 2 concentration, and the ratio of the slope of the curve in acid solution to the slope in neutral solution is (0.45 - 0.35)/(0.45 - 0.40) = 2. The ratio of the effectiveness in the two cases is the cube of this, hence 0 2 is 8 times more effective in acid solution. From the radical-diffusion model, this factor is proportional to (k0,r&/D,8a where kopis the rate constant for the reaction of H atoms with 0 2 , ro is the average radius of a spur of radicals when i t is formed, and D is the diffusion coefficient of the hydrogen atoms. However, any increase in D is likely to be reflected in an increase in KO,. Thus either the rate constant or radius of the spur or both could be increasing in going to acid solution. UPTON,Kmv YORK (12) H. A. Mahlman a n d J. W. Boyle, ibid., 8 0 , 773 (1968). (13) H. A. Schwarz, J. P. Losee and A . 0. Allen, ibid., 7 6 , 4 6 9 3
(1954).
