RADICAL AND MOLECULAR YIELDSIN IRRADIATED WATER ionization events. It is also important to remember that the diffusion model used here and elsewhere does not take into account any initial correlation of the radicals; i.e., the hydrogen atom and hydroxyl radical formed from the dissociation of a water molecule are probably much closer to each other than to other
1937
radicals in the spur. A large fraction of these (more than half) may recombine before the correlation is lost through diffusion. This specific recombination of pairs would not be counted in the diffusion model calculation so that the true initial yield of hydrogen atoms could be several times 0.62.
Radical and Molecular Yields in Water Irradiated by
y
Rays and Heavy Ions*&
by A. Appleby and Harold A. Schwarzlb Chemistry Department, Brookhaven National Laboratory, Upton, N e w York
11973
(Received October 1, 1 9 6 8 )
Nitrous oxide solutions containing separately ethanol, sodium deuterioformate, or potassium nitrite were irradiated with Co60y rays, %MeV deuterons, and 32- and 12-MeV helium ions in order to obtain the variation of the hydrated electron yield, the hydrogen atom yield, and the molecular hydrogen yield with the linear energy transfer (LET) of the radiation. It was found that the hydrogen atom yield does not increase with LET as the molecular hydrogen does nor does it decrease as rapidly as the electron yield. The results are compared with the diffusion model of spurs and tracks and are in agreement with the previous conclusion that hydrogen atoms are produced both by direct action on the water and by reaction of the hydrated electrons with hydrogen ion in the tracks. The much smaller perhydroxyl radical yield was also measured for these particles in copper sulfate-ferrous sulfate solutions.
Introduction I n water radiolysis, free-radical yields decrease and molecular yields increase with increasing LET of the radiation.2-6 This effect is believed to be due to increasing combination of radicals in the spurs and tracks as the density of radicals along the tracks increases. Previous studies of LET effects have not measured the separate yields of hydrogen atoms and hydrated electrons. It was therefore thought of interest to investigate the effect of varying LET on the yields of each of these species. I n particular it was hoped that such a study would throw some light on the origin of the yield of hydrogen atoms. These might be formed directly from the water or alternatively could be produced by a reaction of electrons with hydrogen ions in the expanding spur. Recently, accurate determinations of the rate constants of the reactions expected to occur in the spur have been made.6 This allows a more complete description of the processes occurring in the spur than has hitherto been possible. It was therefore hoped that the present study would allow a more rigorous test of the diffusion model predictions,
Experimental Section The helium ion and deuteron beams were produced by the Brookhaven 60-in. cyclotron. The energy of
the beam was measured by the technique of Schuler and Allen,' and our results are normalized to theirs by Fricke dosimetry. y irradiations were carried out in a cylindrical GoBosource at a dose rate of about 2.1 X 1020 eV/l. min. Irradiation Cells. The cells for cyclotron irradiations consisted of glass syringes containing 10 ml of solution. Stirring was provided by a glass-enclosed magnetic stirrer. The syringe was closed a t the tip by a glass cap, into which a wire was sealed to allow the ion beam current to go to ground through a current integrator. Both ends of the hollow plunger were removed, and the lower end was replaced with a thin (20 mg/cm2) glass window. The beam passed down the axis of the plunger. y irradiations were carried out in unmodified syringes. Materials and Analytical Methods. Sodium deuterioformate was obtained from Volk Radiochemical (1) (a) Research performed under the auspices of the U. 9. Atomic Energy Commission. (b) Correspondence should be sent to this author. (2) A. 0. Allen, Radiation Res., 1, 85 (1954). (3) E. J. Hart, W. J. Ramler, and 9. R . Rocklin, {bid., 4, 378 (1956). (4) N. F. Barr and R. H. Schuler, J . Phys. Chem., 63, 808 (1959). (5) G. L. Kochanny, Jr., A. Timnick, C . J. Hochanadel, and 0. D. Goodman, Radiation Res., 19, 462 (1963). (6) L. M. Dorfman and M. S. Matheson, Progr. Reaction Kinetics, 3, 239 (1965).
(7) R . H. Schuler and A. 0. Allen, J . Amer. Chem. Soe., 79, 1565 (1957).
Volume 79, Number 6 June 1069
A. APPLEBYAND HAROLDA. SCHWARZ
1938
Go., and the isotopic purity was stated to be better than 987& A value of 98.5% was assumed. Mixtures of N 2 0 with argon and N 2 0 with helium were obtained from Matheson Co., passed through a Dry Ice-acetone trap and over solid MnO, and bubbled through alkaline pyrogallol solution, water, a solution equivalent to that to be irradiated, and then finally the solution to be irradiated. A bubbler system for filling the syringes was similar to that described by Seddon and Sutton.8 Gas analyses were performed using a Perkin-Elmer vapor fractometer (Model 154) equipped with a Linde molecular sieve column (Type 5A). For Hz analyses, the carrier gas was argon. For N2 and O2 analyses, helium was used. After replacing the glass syringe cap with a metal needle (kept in water continually flushed with carrier gas), irrbdiated samples were injected through a rubber serum cap directly into a bubbler on the vapor fractometer and the dissolved gas was subsequently flushed into the instrument by a stream of carrier gas.g Sodium deuterioformate solutions were irradiated in sealed glass cells similar to those described by Schuler and Allen.’ Gaseous products were removed on a vacuum line and analyzed for H2, HD, and N2 by mass spectrometry.
reactions are
cas- + N2O H
€ 2 0I
NZ
+ C~HBOH Hz
Nitrous Oxide-Potassium Nitrite Solutions. Hydrogen yields from NzO-KN02 solutions are given in Table I. The yields were independent of energy
Table I: Hydrogen Yields
.y
Coao y rays
18-MeV D+ 32-MeV He2+ 12-MeV He2+
0.43 0.68 0.96 1.11
io-
M NOZ(ref 12)
0.41
0.64 0.89 1.19
absorbed. Solvated electrons react rapidly with both NzO (k = 8.67 X lO9)lo and NO2- (k = 4.58 X 109).11 Hydroxyl radicals and hydrogen atoms also react with NO2-; hence, the yields in Table I represent the molecular yield of hydrogen, G H 2 (molecules per 100 eV absorbed), produced in water. Hydrogen yields have been determined a t the same particle energies by Schwarz, et a1.,I2 for nitrite solutions. Their results for l O - 3 M KNO2 should be comparable to ours and are also given in Table I. The hydrogen yields of Table I are repeated in Table V as the molecular hydrogen yields. Nitrous Oxide-Ethanol Solutions. Both the electron yield and the hydrogen atom yield can be measured in solutions of nitrous oxide and ethanol. The pertinent The Journal of Phusical Chemistry
(1)
(2) The hydroxyl radicals are also removed by the ethanol, and the resulting ethanol radical does not produce nitrogen or hydrogen. Hence, the yields of nitrogen and hydrogen are given by G(Nz) = Gena-
+
G(H2) = G H ~ GH (Yields of species in parentheses represent observed yields while species given in subscript represent the derived values of molecular and radical products.) Some difficulty was experienced as solvated electrons also react with two radiolysis products-acetaldehyde and hydrogen peroxide. Of these, the effect of the hydrogen peroxide reaction (eq 3) predominates ( ~ , , , + C H ~ C H O = 3.5 X lo’; k e . p + ~ 2 = ~ 21.23 X 101o).lo e,,
+ HzOz
3
OH
+ OH-
(3)
The variation of nitrogen yield with dose is therefore given by
+ (ks(H2Oz)/kl(N20))]
where Go(N2) is the initial yield. At low doses the ratio of hydrogen peroxide to nitrogen is approximately constant, and the above equation may be integrated to give (N2) C ( N Z )=~ Go(N2) X dose (1) where
+
C 10-4 &f NO^--^ x 10-4 IigO (this work)
OH
CZHAOH
4
d(Nd/d(dose) = Go(N2)/[1
Results
Particle
-- ++
=
k s G 20~ 2/2kiG (N2) ( NzO) =
0 . 7 1 G ~ ~ o ~ / G ( (NaO). Na)
This dose effect is most easily observable with %MeV D+ and the nitrogen production vs. dose plot at high doses is given in Figure 1 together with the calculated curve. All other experiments were carried out in lower dose regions, and the yields were corrected by eq 1. The largest correction was 4%. Hydrogen yields were found to be independent of dose. The molecular and radical yields are somewhat dependent on solute concentration.l* This effect is believed to be due to the reaction of the radicals with (8) W. A . Seddon and H. 0. Sutton, Trans. Faraday SOC.,59, 2323
(1963). (9) J. W. Swinnerton, V. J. Linnenbom, and C. H. Cheek, Anal. Chem., 34, 483 (1962). (10) 5. Gordon, E. J. Hart, M. 9. Matheson, J. Rabani, and J. K. Thomas, Discussions Faraday Soc., 36, 193 (1963). (11) J. K. Thomas, S. Gordon, and E. J. Hart, J. Phys. Chem., 68, 1524 (1964). (12) H.A. Schware, J. M. Caffrey, and G. Scholes, J. Arner. Chem. Soc., 81, 1801 (1959). (13) See, e.g., A. 0. Allen, “Tha Radiation Chemistry of water and Aqueous Solutions,” D. Van Nostrand 00.. Inc., Kew York, N. Y.,1961.
1939
RADICAL AND MOLECULAR YIELDSIN IRRADIATED WATER Table 111: Gas Yields for Different Radiations M Ethanol-7 X 10-4 M Nitrous Oxide) (Solution of
2.70 1.48 0.72 0.42
y rays 18-MeV D+ 32-MeV H& 12-MeV He2+ C060
1.05 1.32 1.38
Table IV: Product Yields from 10-2 M DCOONa-7 X 10-4 M NzO Particle
G(Hd
G ( H D )cor
GWa)
Figure 1. Nitrogen produced as function of energy input for 18-MeV D+. Dashed line is corrected results according to eq I. The energy input is given as eV/6.02 X lo1*so that the slope gives the yield directly as molecules per 100 eV.
Co’O y rays
0.43 1.23
0.55 0.23
2.81 0.49
the solute before the radicals recombine in the spurs and tracks. Consequently interpretation of the data is simpler a t low solute concentration where these effects are minimal. The effects of nitrous oxide and ethanol concentration in Co60y-irradiated solutions are given in Table 11. Our results a t higher nitrous oxide concentration agree well with those of Allan and Beck,14who studied solutions of 2-propanol containing 3 X 10” M to 0.02 M nitrous oxide. The decrease of G(H2) with increasing nitrous oxide concentration is a composite effect on G H z and G H and cannot be separated into its components in this work. Nitrogen and hydrogen yields from loV2M ethanol7 X 10-4 M nitrous oxide solutions determined for Co60y rays, 18-MeV D+, 32-MeV He2+,and the nitrogen yield for 12-MeV He2+ are shown in Table 111. The nitrogen yield is repeated as the solvated electron yield in Table V. The hydrogen atom yield given in Table V is the difference between G(H2) and the molecular yield, G H ~ . The hydrogen yield was not measured for 12-MeV He2+ as the contribution of GH to G(H2) was expected to be small while experimental errors a t 12 MeV are large. Instead, deuterioformate was substituted for ethanol. Sodium Deuterioformate-Nitrous Oxide Solutions. The pertinent reactions are reaction 1 and
Hydroxyl radicals are also removed by deuterioformate. Scholes and Simic16 pointed out that formate-nitrous oxide solutions give higher nitrogen yields than 2-propanol-nitrous oxide solutions, presumably due to a reaction of some organic radical with nitrous oxide. Hence, the nitrogen yield is not a reliable estimate of G,,,. The competition between reactions 4 and 5 was pointed out by Scholes and Simic.la Yields of H2,HD, and nitrogen from deuterioformatenitrous oxide solutions are given in Table IV. The HZ and H D yields were independent of dose and the nitrogen yields were corrected by eq I. The H2 and H D yields were corrected for the small amount of HCOO- present using the value 6.6 for ke/k416
H
+ D C O O - 4 H D + COOH + DCOOproduct
(4)
(5)
-j
Table 11: Gas Yields from Coao ?-Irradiated Ethanol Solutions Containing Dissolved NzO
13-MeV He2+
H
[NaO], M
G(Hd
G(Nz)
5 x 10-2 5 x 10-2 5 x 10-2 10-2
2.4 X 5.1 X 10-8 7.1 x 10-4 7.1 x 10-4
0.97 1.02 1.os 1.05
3.12 2.85 2.73 2.70
(6)
The yields of H2 should be equal to G H of~ Table V and are in good agreement with them. The data of Table V are somewhat more accurate. The H D yield produced by Coco y rays is smaller than G H determined in ethanol-nitrous oxide solutions in agreement with Scholes and Simic.*6 The difference is due to reaction 5 and leads to the conclusion that k6/(k4 k6) is 0.55/0.61 or 0.9. This factor was used to derive GH from G(HD) for 12-MeV He2+ (Table V). Ferrous Sulfate-Cupric Sulfate Solutions. Hart? suggested that a small yield of perhydroxyl radical is produced in water radiolysis in the absence of oxygen. He noted a small yield (0.02) of oxygen in ferrous sulfate-cupric sulfate solutions and attributed it to the reaction
+
H02
[EtOHI, M
+ HCOO- + Hz + COO-
+ Cu2+
4
H+
+ CU++
0 2
(7)
(14) J. T. Allan and C. M . Beck, J. Arne?. Chem. Soc., 8 6 , 1483 (1964). (16) G. Scholes and M . Simic, J. Phys. Chem.. 6 8 , 1731 (1964). (16) G.Scholes and M. Simic, { b i d . , 6 8 , 1738 (1964). (17) E.J. Hart, Radiation Res., 2, 33 (1955).
Volume 75,Number 6 June 1969
A. APPLEBYAND HAROLDA. SCHWARZ
1940 Table V: Experimental and Calculated Yields for Various Particles Particle
Cow y rays
18-MeV D+ 32-MeV
He2+
Z,spurs/cm X 10-6 Exptl
Calcd Exptl 1.83 (eq 111) I . 97 (eq 11) Exptl
5.69 (eq 111) 9.74 (eq 111) 12-MeV He2+
Exptl
9.30 (eq 111) 17.2 (eq IT) a
Reference 12; air-saturated 10-8 M KBr solutions.
b
GH 2
GH
Gear
GHaO*a
0.43 0.42 0.68 0.72 0.73 0.96 0.86 0.90 1.11 0.90 0.93
0.61 0.63 0.64 0.62 0.61 0.42 0.43 0.33 0.27 0.33 0.23
2.70 2.86 1.48 1.20 1.14 0.72 0.53 0.33 0.42 0.34 0.19
0.61‘~ 0.58 0.91 0.84 0.84 1.00 0.88 0.89 1.08 0.89 0.90
A . 0. Allen
Donaldson and Miller1*irradiated these solutions with PoZl0y particles and found a much larger yield of oxygen [G(02) = 0.251. It was obviously desirable to measure this H 0 2 yield as a function of particle energy. Unfortunately, no system has been developed for measuring HO2 (or 0 2 - ) yields in neutral solution, but it was expected that any acid effect would be small between M H2S04 and neutral solution. Oxygen yields were determined for helium ion M FeS04-10-2 M CuS04 irradiated solutions of adjusted to pH 2 with sulfuric acid and are given in Table V as CEO2. The oxygen yields were assumed to be independent of dose (as observed by Hart”) but the scatter in our results (=t50/,) could mask some curvature.
Discussion The hydrogen peroxide yields of Schwarz, et a1.,12 measured in air-saturated low3M potassium bromide solutions have been included in Table V. The molecular hydrogen and hydrogen peroxide yields increase while the yields of hydrated electrons and hydroxyl radicals decrease with increasing LET as was known previously. This behavior is in agreement with the diffusion model in which hydrogen and hydrogen peroxide are formed by combination of the radicals as they diffuse out of the spurs. As the LET increases, the spurs are formed closer together and radical combination is increased. The most interesting feature of Table V is the behavior of GH which is essentially independent of radiation quality between Coaoy rays and 18-MeV deuterons. I n contrast, the electron and hydroxyl radical yields for 18-MeV D+ are about half those for GoBoy rays. The hydrogen atom yield for 32-MeV He2+ is 3101, smaller than for Co60y rays, but the corresponding decrease for the solvated electron yield is %yo. This qualitative difference in the behavior of GH compared to G,,, and G O H argues against direct production of hydrogen atoms by action of the radiation on water as the sole source. This mechanism would predict that the The Journal of Physical Chemistry
and R.
GHO~
0.026
0.05 0.07
A. Holroyd, J. Amer. Chem. Xoc., 77, 5852 (1955).
variation of GH with LET would be similar to the variation of G,,,. An alternative mechanism is production of some of the hydrogen atoms by reaction of hydrated electrons with hydrogen ions as the spur expands eaq-
+ H+ -+ H
(8) Hydrogen ions are assumed to be present in the spur to balance the negatively charged electrons. The results of experiments in which strong bases were added to react with the hydrogen ionigsuggested that about half of the observed hydrogen atom yield came from reaction 8 and half from hydrogen atoms produced directly from water.20 The yield of reaction 8 would increase with LET, but the hydrogen atom would react with other radicals so the net result would be that the observed yield would tend to be independent of LET. This mechanism has been tested quantitatively using the spur model as described by Schwarz by which all of the molecular and radical yields can be predicted for the deuteron and helium ion experiments.20 The LET of the radiation is introduced through a parameter 2 which is the energy-averaged spur density along the particle track, i.e.
where Eo is the particle energy, B is the energy required to form a spur, and - dE/dx is the rate of energy loss of the particle to the medium. This description of the track is inaccurate because the spurs do not lie precisely along a line but rather each spur is somewhat displaced from the track center due to the uncertainty in position of the energy lossz1 Ar =
1/(2E/m)(h/r)
(18) D. M. Donaldson and N . Miller, Trans. Faraday Soc., 52, 652 (1956). (19) A. Appleby in “The Chemistry of Ionization and Excitation,” G. R . A. Johnson and G . Scholes, Ed., Taylor and Francis, Ltd., London, 1967, p 269. (20) H.A. Schwarz, submitted for publication.
A PULSERADIOLYSIS STUDY OF SOMECo(II1) COMPLEXES where E and fiz are the energy and the mass of the particle. The value of Ar is 4.6 X 10-8E1/2for helium ions and 6.5 X 10-8E1/2for deuterons where E is in MeV. Since neighboring spurs are most important in determining the extent of reaction, it is suggested that a better approximation of the interspur distance, d, would be d = [{e/(-dE/dx))'+ (A"']''' and a better Z given by Z =
-1 1
Eo
d-1dE (111) Eo a The calculations were performed for lo-' M ethanol7X M nitrous oxide solutions using the values of the constants and parameters given by Schwarz and using e = 62.5 eV per spur. This is not the average spur energy but the single spur energy giving results closest to the calculated yields for y-ray irradiations which were averaged over all spur sizes. The results using both eq I11 (lower LET) and eq I1 are given in Table V and it may be seen that the hydrogen atom yield is computed quite precisely by this mechanism. It is interesting that the observed molecular yields and hydrated electron yields tend to be higher than the
1941
calculated yields. This effect was noted earlier and attributed to correlation between the initial positions of the radical pairs.12 There would be a tendency for the hydrogen atom and hydroxyl radical formed from one water molecule to be closer together than those from different water molecules so that more dissociation actually takes place than is accounted for by the diffusion model for y-ray reactions. This correlation would tend to fade into the background for particle tracks where the over-all radical density is higher thus leading to an apparent increase in total decomposition yield for particle tracks. Perhydroxyl Radical Yields. The yield of perhydroxyl radicals also increases with LET (Table V) . This effect is not accounted for in our diffusion model as the reaction
OH
+ HzOz
+
HOz HzO (9) is not fast enough to produce the Kupperman has suggested that the radical is produced by a small yield of oxygen atoms reacting with OH.23 4
(21) J. L. Magee, Ann. Rev. Phys. Chem., 12, 389 (1961). (22) H. A. Schwara, J. Phys. Chern., 66, 255 (1962). (23) A . Kuppermann in "Radiation Research 1966," North-Holland Publishing Do., Amsterdam, 1967,p 212.
A Pulse Radiolysis Study of Some Cobalt(II1) Complexes. Occurrence of Electronically Excited Cobalt(I1) Products. by William L. WaltzIband Ralph G. Pearson Department of Chemistry, Northwestern University, Evanston, Illinois 60601, and the Chemistry Division, Argonne National Laboratory, Argonne, Illinois 60439 (Received October 3, 1968)
The rate constants for the reaction of the hydrated electron with some cobalt(II1) complexes in aqueous solutions have been measured and are in the range of 101o-lO1lM-l sec-'. In conjunction with results of others' work, it is suggested that the initial products of these reactions are electronically excited cobalt(I1) complexes of low-spin electronic configuration. Direct evidence for the presence of low-spin Co(bipy)++formed by the reaction of the hydrated electron and Co(bipy)g3+in aqueous, alcoholic media is given. The absorption spectrum of low-spin C0(bipy)3~+ between 400 and 615 mp has been recorded and the kinetics of the disappearance of this species has been studied. The rate constant, kobsd, associated with the disappearance of low-spin Co(bipy)Pat 578 mp is expressed as kobsd = ko 4-kl[Co(bip~)3~+1 -f k~[&],where ko 5 X lo3sec-', k~ = 8 X lo8 M-' sec-l, and k~ = 106-1010M-' sec1 for quenching by various compounds such as oxygen. Possible quenching mechanisms are discussed. Less extensive evidence is given for the formation of low-spin Co(bipy)32+ by the reaction of hydrogen atom with Co(bipy)g3+.