2627
RADIOLYSIS OF CRYSTALLINE ALKALIMETALBROMATES
Radiolysis of the Crystalline Alkali Metal Bromates by Neutron Reactor Radiations'
by G. E.Boyd and Q. V. Larson Oak Ridge National Laboratory, Oak Ridge, Tennessee
(Received April 22, 1964)
The radiations in the active lattice of the Oak Ridge graphite reactor were employed to radiolyze bromate ion to varying degrees in each of the crystalline alkali metal bromates. The decomposition of BrOs- gave oxidizing fragments, bromide ion, and occluded oxygen gas in the crystals in amounts which depended on the compound and on the radiation dose it absorbed. Ceric sulfate dosimeter solutions and thermal neutron flux monitors were employed to measure the absorbed doses, and initial 100-e.v. yields a t 40" were derived for bromate decomposition, Ga(-BrOs-), oxidizing fragment formation, Go("Ox"), and bromide ion production, Ga(Br-). The yields for all the salts except LiBrOs were the same within experimental error as observed previously with GoB0y-rays. The production of high LET tritons and a-particles by Lis fission was responsible for the significantly larger yields of LiBr03. The nonlinear dependence of the formation of oxidizing fragments and bromide ion on the absorbed dose was explained by a radiolytic niechanisin in which the concentrations were governed by a sequence of first-order irreversible reactions.
Thls study continues our investigations on the radiation chemistry of molecular ions in crystals. The principal objective has been to determine the decompositions produced in the alkali metal bromates by the radiations in the active lattice of the Oak Ridge graphite reactor (ORGR) and to compare these with the decomposit,ions produced in the same compounds by Co6O y-rays2 to answer the question whether there were any important differences in the mechanism of the radiolysis by the two types of radiation. The measurement of the doses imparted by the nuclear reactor radiations to the crystalline solids was an important part of this research. The decomposition of several crystalline inorganic compounds by the radiations present at the center of a heavy water reactor has been determined3 previous to our investigations. Oxygen gas produced in the radiolysis of N d 0 3 , K S 0 3 , and KCIOs was measured ; and 100-e.v. yields, G(02), were derived based on numerical calculations of the dose absorbed. Recently, however, much smaller G(O2) values have been obtained4 in the radiolysis of KaN03 and KN03 b y * C ~ y-rays. This discrepancy might be explained if the estimates of the radiation doses by Hennig, Lees,
and Matheson were too small or if there were a significant dependence of G(OJ on the LET. There is some evidence that LET effects may occur in nuclear reactor irradiations. In the radiolysis of several liquid aromatic hydrocarbons in the core of the ORGR the yields of Hz,CH4, and other gases have been found5 to be larger than those obtained.in the electron bombardment radiolysis of the same compounds. The yields for bromate ion decomposition, bromide ion, and oxidizing fragment formation which we shall report below have been based on dose measurements made with aqueous ceric sulfate solutions calibrated calorimetrically, and with flux monitors to estimate the energy imparted to the crystals consequent to the capture of thermal neutrons in them. Agreement (1) Presented before the Division of Physical Chemistry, 146th National Meeting of the American Chemical Society, Denver, Colo., Jan. 19-24, 1964. (2) G. E. Boyd, E. W. Graham, and Q. V. Larson, J . Phys. Chem., 66, 300 (1962). (3) G. Hennig, R. Lees, and M. S. Matheson, J . Chem. Phys., 2'1, ~664~(1953). (4) T.-H. Chen and E. R. Johnson, J . Phys. Chem., 66,2249 (1962). (5) T. J. Sworski and M. Burton, J . A m . Chem. Soc., 73, 3750 (1951).
Volume 68,Number 9
September, i96.4
2628
with the Co60 y-ray yields was obtained to within the limits of error in the reactor dose measurements (Le., f1070), and there was no evidence of LET effects except with one conipound, LiBrO3, wherein energetic tritons and a-particles were produced by Li6 fission. Studies also were made on the details of the radiolysis of bromate ion, and these gave support to a general radiolytic mechanism for the production of ?oxidizing fraginents and bromide ion. Equations capable of describing the nonlinear dose dependence of the yields of the latter were derived based on this mechanism.
G. E. BOYDAND Q. V. LARSON
distance of 8 in. This periodicity correlated with the location of the uranium-filled channels in the graphite passing a t a right angle to hole 71. Auxiliary experiments were performed which showed that the radiolysis observed in the alkali metal bromates was caused entireIy by reactor neutrons and y-rays. The teniperature of the air a t the irradiation position was maintained a t 40-45” ; the samples probably were slightly warmer. Approximately 5-g. quantities of the purified crystalline salts were placed in cylindrical polyethylene capsules (1.2 cin. in diameter X 2.1 cm. high) and were Experimental irradiated together with (a) weighed 20-mil diameter Preparation of Anhydrous Compounds. The syncobalt metal wires (30-50 mg.) to monitor7the “thermal theses of the crystalline anhydrous alkali metal broneutron exposure dose,” C$t& and with (b) 10-ml. mates employed in this work have been In aliquots of aqueous ceric sulfate solution held in sealed addition to the general requirement of high purity, quartz ampoules to measure the reactor y-ray and fast it was essential that the lithium contents of these comneutron dose rates. The dimensions of the alkali pounds be as low as possible because of the large metal bromate samples represented a compromise reaction cross section of thermal neutrons with this between the requirement that they be sufficiently thick element. Careful flame spectrophotoinetric analyses to perinit the establishment of “electronic equilibrium” showed that the NaBrO3, KBr03, RbBr03, and CsBrO3 in the scattering of the reactor -prays8 and the need to preparations each contained less than 10 p.p.m. of minimize thermal neutron and low-energy y ray “selflithium. shielding” effects caused by the strong absorption of The LiBr03 preparations were derived from spectrothese radiations in the outermost layers of the salts. chemically analyzed coinmercial LiBrO3. H20 (City Auxiliary experiments, confirmed by theoretical coniChemical Corp.) which was purified by three crystalputations based on neutron diffusion theorylg showed lizations. The anhydrous compounds were sepathat all samples but LiBrO, were sufficiently small to rated from a hot (>52’) aqueous solution. The Li6 make self-shielding negligible. abundance in the LiBrO3-1 and LiBrOJ-2 preparations Radiation Dosimety. The absorption of radiant was lower than the accepted natural abundance of energy by the alkali metal bromates was assumed to 7.5Y0; mass spectrometric analyses gave 3.78 f 0.09 be directly proportional to the thermal neutron dose,1° and 3.70 f 0.09 atom % Li6, respectively. They also &ht, estimated from the bombardment time, t , and the contained 19 p.p.m. and < 5 p.p.m, of bromide ion, C060 y-ray activity induced in 20-mil pure cobalt respectively. wire monitors employed in each irradiation. RadioTwo preparations of Li7Br03 were made starting activity measurements were conducted with a 4 r with separate lots of spectrocheinically pure lithium geoinetry ionization chamber filled to a pressure of 40 metal containing 99.99 and 99.994 atom % Li7, reatni. with krypton gas. The chamber readings on the spectively. The metal was converted to Li7~S04 monitors were converted to disintegration rates per and treated with purified Ba(BrOJz.HzO to give Li7B r 0 3 . H z 0 , which after one recrystallization a t 60” (6) M. E. Ramsey and C. D. Cagle, Proc. Intern. Cor$. Peaceful Uses gave thP anhydrous salt in about 63% yield. An At. Energy, 2 , 281 (1956). (7) Although the reactor was operated a t a constant, nominal power analysis of thcse preparations gave 213 and 25 p.p.m. of 3400 kw., changes in the control rod positions, uranium loading, of broniide ion, respectively. and the insertion and removal of experimental equipment elsewhere in the active lattice produced alterations in the neutron flux of sufIwadiation of Samples. All irradiations were perficient magnitude that a monitor for every irradiation was essential formed in the Oak Ridge graphite reactor6 in vertical if the exposure dose were to be known to within 1-2’35. See also, J. Wright, Discussions Faraday SOC.,12, 60 (1952) hole 71 a t a point approximately 24 in. above the (8) J. Weiss, Arucleonics, [7] 10, 28 (1952). center plane of the active lattice. The thermal neu(9) P. F. Zweifel, ibid., [ l l ] 18, 174 (1960). tron flux, &h, a t this location was 7.4 X lo1’ cin.-2 (10) A proof of the validity of this assumption for a reactor similar sec.-I with the reactor operating a t 3400 k-cv.; the to the ORGR has been reported by A. R. Anderson and R. J. ’I?’aite. “Calorimetric Measurement of Energy Absorbed from Reactor variation of 4 t h with distance was about 0.70j0/in. Radiation in BEPO,” AERE-C/R 2253, Harwell, Berkshire, England, 0.1 X 1O1O tin.? The fast neutron flux, + f 1 was 1.0 1960; see also, A. R. Anderson and J. K. Linacre, “Selected Topics in Radiation Dosimetry,” IAEA, Vienna, 1961. set.-' and varied periodically by about 12% over a The Journal of Physical Chemistry
2629
RADIOLYSIS OF CRYSTALLINE ALKALIMETALBROMATES
gram of cobalt by conzparisonwith CoB0standards whose d.p.m. had been determined by coincidence rate measurements. Decay corrections were made with the accepted Co60half-life of 5.27 zt 0.01 years. Neutron selfshielding by the 20-mil wires was determined as 17.5y0 by comparisons with the specific activity of cobalt induced in an “infinitely dilute” alloy of cobalt (0.151, wt. Yo) in pure aluminum. A value of 41.7 barns Waf4 derived for the effective thermal neutron activation cross section for C Ofrom ~ ~the measured cadmium ratio for the alloy (9 8 =k 0.6) together with the value of 37.0 barns for the 2200 m. sec.-l cross section and the cobalt resonance integral, lo= 75 barns, following published recommendations. l1 The cobalt-cadmium ratio measured with 40-mil cadmium and dilute cobalt-aluminum alloy was almost independent of the position in hole 71. The Maxwellian temperature of the neutrons was taken as 500°K. (0.043 e.v.) and the cadmium cut-off as E , = 0.52 e.v. for 40-mil cadmium. The Neutron Capture Dose. The total energy deposited in the samples consequent to neutron capture may be considered as made up from three contributions : (a) the energy absorbed from the energetic y-rays emitted a t the moment of neutron capture; (b) the energy deposited by the nucleus which recoils into the crystal on emitting capture y-rays; (c) the radioactive decay of the productt nucleus and the absorption of its p- and y-rays in the sample. The energy absorbed per unit thermal neutron exposure dose from the instantaneous neutron capture y-rays generated within a given compound may be estimated with the relation
Dn,, (e.v. mole-1 per neutron cm?)
=
aerfN~&@, (1) where (Teff is the iLeffectivel’capture cross-section1?; N A , the Avogadro number; &, the energy released as y-radiation in e.v. per neutron captured13; and 8,, the fraction of the internally generated y-ray energy absorbed. This fraction is given by 8, = 1 - exP[-(PLa/P)hJl
(2)
where L is the average distance traversed in the sample by the y-ray, p is the density, and ( p L . / p )is the energy absorption mass attenuation coefficient. The distance L for right cylinders is equal to two-thirds the mean chord length, 1, given by Cauchy’s relation: 1 = 4V/S where V is the sample volume and S its surface area. For the ca. 5-g. samples used in this research, 1 =; 0.933 cm.; hence, L = 0.622 cm. The density of the powdered samples was approximately 2.0. The appropriate energy-dependent ( p a l p ) value for eq. 2 is determined by the character of the neutron capture
y-ray spectrum which is usually complex and unfortunately not well known for energies of less than 1 Mev.14 However, an average energy, E,, of sufficient accuracy may be estimated from the ratio of Q to the average number of quanta emitt,ed per neutron capture, V. A multiplicity of v = 3 appears to be typical for many nuclei.l5 Published value^^^'^' of (Teff for the alkali metals were employed when available, or were estimated from the 2200 m. set.-' cross section18 and resonance activation integral. l 9 Redeterminations of uefffor the production of 4.4-hr. BrSom, 18-min. BrE0g,and 35.9-hr. Br82were made during the course of this work because of the importance of these quantities in the calculation of the neutron capture dose. Values of 3.25, 10.2, and 4.1 barns, respectively, were found; details on these measurements will be published elsewheren20 Values for ( p * / p ) for the elements have been tabulatedz1; the required values for each of the alkali metal bromates were derived as the average over the coefficients for the constituent elements weighted in proportion to the abundance of each element by weight in the compound. The Recoil Atom Dose. Neutron capture also produces energetic heavy charged particles (Le., recoils) whose motion through the crystal lattice causes appreciable decomposition. The energy deposited, D,, is given by a relation identical 1vit.heq. 1 except that Q is replaced by E , (e.v.) = 537EY2/Mwhere M is the mass of the nucleus emitting a capture y-ray of energy, E , (MeV.). All of the energy of the recoil atoms will (11) R. W. Stoughton and J. Halperin, Nucl. Sei. Eng.,‘ 6, 100 (1959). (12) The effective cross section, serf,is defined by the relation: R = $ t h s e f f , where R is the reaction rate per atom for a substance at high dilution or in very small amount, including thermal and epithermal (but not fast) neutron capture reactions, and $)th is the thermal flux. See ref. 11. (13) D. J. Hughes and J. a.Harvey, “a4merican Institute of Physics Handbook,” McGraw-Hill Book Co., New York, N. Y., 1957, Section 8h, Table 8H-1. (14) E. Troubetakay and H. Goldstein, Nucleonics, [ l l ] 18, 1’71 (1960). (15) C. 0. Muehlhause. Phys. Rev., 79, 277 (1950). (16) W. S. Lyon, Nucl. Sei. Eng., 8 , 378 (1960). (17) A. P. Berg and R. M. Bartholomew, Can. J . Chem., 38, 2528 (1960). (18) D. J. Hughes and R. B. Schwartz, “Neutron Cross Sections,!! Report BNL-325, Brookhaven National Laboratory, Upton, K. Y., 1958. (19) R. L. NIacklin and H. S. Pomerance, “Progress in Nuclear Energy,” Series I, McGraw-Hill Book Co., Sew York, K , Y . , 1956 Chapter 6. (20) G. E. Boyd, J. S. Eldridge, and Q. V. Larson, J . Inorg. 2Vucl. Chem., to be published. We have also confirmed that a hitherto unobserved 6.2-min. BrEZm is produced by thermal neutron capture in Br81. (21) E. P. Blizard, “Xuclear Engineering Handbook,” H. Etherington, Ed., McGraw-Hill Book Co., New York, N. Y., 1958, Chapter 7-3, Table 2.
Volume 68, Number 9
September, 1904
2630
G. E. BOYDAXD Q. V. LARSOK
be deposited in the solid (0, = 1) because of their short ranges. Because of the relative smallriess of E,, the contribution of D , to the total dose was almost negligible. Neutron capture in compounds containing Li6 produces energetic triton (2.736 Mev.j and a-particle (2.051 Mev.) recoils. The entire energy of these fission fragments is deposited in the crystal, and large D, values are found because of the large ueff for Li6 fission and the large energy per event (4.787 MeV.). Neutron capture by Li7produced 0.8-sec. Lis which decays to Be8 which promptIy splits into two a-particles with the release of 2.99 MeV. of energy.22 I n this case, however, ueff is so small that the energy deposited is almost negligible. The Decay Radiation Dose. @-rays from neutroninduced activities radiolyze the alkali nietal bromates because of their efficient absorption by these coiiipounds. The dose, D,, is given by
D, = aerfNAEpOpS‘
(3)
where ueff and NAhave the same meaning as in eq. 1, E , is the average p-ray energy, 0, is the fraction of prays absorbed, and S‘ is the integrated saturation factor. Values of E, are found by integration over the continuous @-ray spectrum; for “allowed” spectra & has been tabulated as a function of the atomic number of the daughter nucleus and the maximum @-ray energy.23 The fraction, 8,, may be estimated from the sample geometry and the range, R, of the P-rays in the compound.24 p-Ray ranges were estimated from a suitable empirical equation relating R to the density of the powdered sample, p , and to E,.25 The factor, X’, is included in eq. 3 to correct for the fact that the dose does not increase linearly with irradiation time, t b , because of the time dependence of the induced radioactivity
X’
=
1 - (1 -
e-Xtb))/Xtb
(4)
where X (sec.-1) is the decay constant for the induced activity. When AtB is large, as with long irradiations, S’ becomes independent of time. The energy absorbed from decay y-rays was estimated froin an equation identical with eq. 3 except that was replaced by
where f f is the number of y-rays of energy, E,,, per disintegration, and 0, is the fraction of E , absorbed (eq. 2). The required numerical values for the half-lives, branching ratios, and energies for the decay of the The Journal of Physical Chemistry
alkali metal and bromine neutron capture induced activities were taken from current tabulations. 26 The Reactor y-Ray Dose. The energy absorbed by the crystalline bromates from the reactor y-rays was computed from the measured dose rates made with aqueous 0.0316 h/i ceric sulfate solutions 0.4 M in H2S04. The production of Ce(I1I) in these solutions was caused solely by y-rays and energetic neutrons: the contribution from thermal neutron capture by hydrogen was negligible. Ceric ion was reduced by the ionization and excitation of the water by y-rays and by the fragments produced by 0.1-1-hlev. energy recoiI protons resulting from the scattering of fast and epithernial neutrons. The yields ( i e . , G-values) for Ce(1V) reduction bj7 y-rays and neutrons differ slightly because of differences in the LET for these radiations.*’ Values of G = 2.40 and G = 2.85 molecules of Ce(II1) per 100 e.v. absorbed were assumed for each component a t 40°,28 respectively; and an effective yield, Geff = 2.70, was coniputed based on the observationzgthat energetic neutrons contribute two-thirds of the total energy absorbed by water exposed to the radiations in the active lattice of the ORGR. The average of many determinations with the ceric sulfate dosimeter gave a dose of 3.9 h’0.2 X lo4e.v. g.-l/neutron cm.?. This average was in agreement with an independent value of 4.1 X 104 obtained from calorimetric measurements of the rate of heating of mTater.2gb The y-ray dose per unit thermal neutron exposure dose, D,R,o = 1.4 X lo4 e.v. g.-l/neutron cm-2, was employed in the conversion of the dose in water to those absorbed by the solid salts. The y-ray doses absorbed by the crystalline alkali metal bromates, DY%IBr03, were computed with the relation
where N(E) is the differential energy flux of y-rays (22) T. A. Griffy and L. C. Biedenharn, AVucZ.P h y s . , 15, 636 (1960). (23) L. Slack and K. Way, “Radiations from Radioactive Atoms,” U.S.A.E.C., 1959, p. 73 ff. (24) P. I. Richards and B. A. Rubin, YucZeonics, [6] 7 , 42 (1950). (25) K. 2. Morgan, “Radiation Hygiene Handbook,’’ H. Blot%, Ed.. McGraw-Hill Book Co., New York, N. Y . , 1959. (26) K . Way, ’ ’ Kuclear Data Sheets,” National Academy of Sciences, National Research Council, Washington, D. C., 1964. (27) C. J . Hochanadel, “Comparative Effects of Radiation,” John Wiley and Sons, Inc., New York, S . Y., 1960, Chapter 8, pp. 151184.
(28) C. J. Hochanadel and J. A. Ghormley, Radiation Res.. 16, 653 (1962). (29) (a) C. D. Bopp and R. L. Towns, iVuc1. Sei. Eng., 13, 245 (1962) ; (b) D. M. Richardson, A. 0. Allen, and J. W. Boyle, Proc. Intern. Conf. Peaceful Uses At. Energu, 14, 209 (1956): see also ref. 10.
RADIOLYSIS O F CRYSTaLLINE ALKALIl / I E T A L BROMATES
(photons :\Iev.--l) cni.?, ( p a l p ) is the energy-dependent, energy absorption mass attenuation coefficient21 for the conipound irradiated, and E is the y-ray energy in MeV. The ratio of integrals (taken over the range E = 0.1 to E = a ) appearing in eq. 5 arises from the fact that a y-ray spectrum exists in the reactor. I n the numerical evaluations of this ratio use was made of an empirical equation for N ( E ) derived for the unperturbed spectral distribution a t the irradiation position we
+
N ( E ) = C ~ i ( 0 . 0 2 0 e - ' . ~ ~0.048e-2.5E1
2631
irradiated bromates showed that the reduction of hypobromite by arsenite was quite rapid but that broinite or other as yet unidentified fragments reacted slowly. A negative test for the presence of elemental bromine was obtained. One-gram fractions of extensively irradiated salt were dissolved in water and extracted with pure CCl,, or were pulverized under CCl,; and the ultraviolet. absorption spectrum of the organic phase waso measured. KO absorption a t the wave length (4150 A.) characteristic of Brz was found.
(6)
The Energetic N e u t r o n Dose. The energy deposited by epithernial and fast neutrons was estimated assuming that (a) only elastic neutron scattering occurred ; (b) the neutron flux, $ ( E ) ,was uniform throughout the sample; and (c) the neutron mean free path was large coinpared with the dimensions of the sample. If, then, scattering at all energies is spherically symmetrical, the dose rate will be given by
0'
E
15.00
v)
Po
(7)
+
where a = [ ( A - 1)/(A 1)12,with A being the target nucleus iiiass number, 5, is the average macroscopic neutron scattering cross section over the energy range indicated, and E is the neutron energy. If DnHrO,the energetic neutron dose per unit thermal neutron dose for water, is known, values for DnM~rOs will be given by
Average values for the microscopic neutron scattering cross sections for water, t+HzO, and the alkali metal bromates, ~ M B ~ Ow~r , e estimated froin the data for the elements tabulated by Hughes and Schwartz. l 8 The calorinietric value for the dose in water, DnIl10 = was employed. 2.7 X lo4e.v. g.-'/neutron Analytical Methods for Radiolytic Products. A meaningful study of primary radiolytic processes requires a very low product conversion and, thus, the eniployinent of sensitive analytical techniques is mandatory. Measurements of the total bromate radiolyzed, the bromide ion formed, and the oxidizing fragments produced were made following previously described microtitration procedures. The thoroughly mixed bombarded crystals were dissolved in a 0.2 M NaHC03 solution containing an excess of sodium arsenite, and a riiinimuni period of 30 min. was taken for the complete reduction of the oxidizing fragments to bromide.30 I\'Ieasureinents of the visible and ultraviolet absorption spectra on aqueous solutions of the
8 10.00
n
5.00
E=xZ3:00
d o
5.b0
EXPOSURE DOSE
&bo
%bo
8.00
9.w
.J
x
Figure 1. Decomposition of the alkali metal bromates by nuclear reactor radiations. (Symbols with flags indicate differing crystal preparations.)
45.00
40.00
EXPOSURE DOSE
( a tx
10l6)
Figure 2. Production of bromide in the decomposition of the alkali metal bromates by nuclear reactor radiations.
(30) Bromate in neutral and alkaline solutions is not reduced by arsenite. It was assumed that all radiolysis products, except B r ion, were reduced by AsOs- in 0.2 M NaHCOl solution
Volume 68, Xumber R
September, 2 9 C q
2632
G. E. BOYDAND Q. V. LARSON
Table I : Radiolytic Yields for Nuclear Reactor Radiations"
Salt
LiTBr03
NaBrOl KBr03 RbBrOa CsBrOl LiBrOl a
T-Ray dose
-----Neutron Capture y-rEYB
2.25 2.46 2.72 3.46 5.37 2.25
capture dose-----
1.41
0.223 0.002 0.002 0.002 0.002 8 3 . 4ZC
1.47 1.76 1.43 7.77. 1,146
Dose values in MeV. mole-' per neutron cm.-*.
-
2.0
4.0
'
6.0
EXPOSURE DOSE
Decay radiations
Recoils
(Olh
8.0
1.72 1.66 1.65 1.67 1.54c 1.39" Ref. 2.
(0.c
I X Nl'6)
Figure 3. Production of oxidizing fragments in the decomposition of alkali metal bromates by nuclear reactor radiations.
Results and Discussion Radiolytic Yields. The estimated doses absorbed by the crystalline alkali metal bromates from the reactor radiations are summarized in Table I, and the data m i their radiolysis are plotted as a function of the thermtl neutron dose, &t, in Fig. 1-3. Several observ,itions of possible interest may be made. (a) The n )st significant result (Table I) is that more than one-li,, f of the total dose is generated within the crystals themselves as a consequence of thermal and epithernial neutron capture. The samples become radiation The Journal of Physical Ch,emistry
Energetic neutron scattering dose
0.18 0.15 0.13 0.15 0.17 0.18
Total dose
5.78 5.74 6.26 6.78 14.85 88.38
Decomposition, Gross mmoles mole-' per reactor neutron cm.-z yield, X 1018 Ca(-BrOa-)
0.437 1.367 1.279 2,694 7.570 21.198
0.47 1.4 1.3 2.4 3.1 1.4
coao ?-ray yield,b Qo(-BrOa-)
0.31 1.42 1.32 2.33 3.4 0.31
Includes correction for neutron self-shielding.
sources and are self-radiolyzed. With LiBrOa the decomposition is produced almost entirely by the energetic triton and a-recoil particles which are formed in large yield because of the large neutron capture cross section of Lie and are totally absorbed within the compound. (b) The relatively large contribution of capture y-rays to the neutron capture dose is noteworthy. The influence of the large cross section of cesium is evident; the other alkali metals possess small cross sections, and the capture y-ray dose in their salts is contributed predominantly by bromine. (c) The dose imparted by the decay radiations was nearly independent of the salt as a consequence of the fact that the main contributor was the short-lived BrsoS activity which was formed with a large cross section and decayed with energetic 6- and y-rays. (d) The dose absorbed from the reactor y-rays increased with the atomic number of the cation in the salt. This reflects the strong dependence of the energy absorption mass attenuation coefficient on atomic number, especially for low-energy y-rays which are abundant in the reactor spectrum (cf. eq. 6). (e) An almost negligible fraction of the dose comes from the energetic neutrons in the reactor. The total doses for each salt are listed in column 7 of Table I. These values were employed to convert the thermal neutron doses to energy absorbed in e.v. mole-I. The existence of Iarge differences in the radiolytic stability of bromate ion between the various compounds was found (Fig. 1). A similar dependence on the cationic constituent was found in the decompositions of the same compounds by Co60 y-rays2 and has been observed in the radiolysis of the alkali metal nitrates. These differences have been related to the "free space" within the crysta12~3~31~32 with varying degrees of success. (31) J. Cunningham and H. G. Heal, Trans. Faraday SOC.,54, 1355 (1968). (32) J. Cunningham, J . Phys. Chem., 65, 628 (1961).
2633
RADIOLYSIS OF CRYSTALLINE ALKALIMETALBROMATES
The data for the radiolysis of KBr03 in Fig. 1 are in good agreement, ‘with the earlier measurements reported for this compound33when it was irradiated in a different position in the active lattice of the ORGR. This concordance may be taken as a further confirmation of the primary assumption made in this work that the absorption of radiant energy by the salts was directly proportional to (btht. The increase in bromate decomposition with dose (Fig. 1) was nonlinear above (btht = 10l6neutrons cm.-2 for all the salts except Li7Br03. Accordingly, the decompositions below this dose (see column 8, Table I) were employed to derive initial 100-e.v. yields, Go ( -BrOs-), for comparisons with the radiolysis by Co6O y-rays. As may be observed from Table I, the reactor yields were the same as with Co60-prays within the estimated error ( * 10%) in the former values for all but the lithium salts. This result, although not entirely unexpected, is of interest in that it indicates that high yield radiolytic processes were largely absent in these reactor bombardments. It may be concluded that either the LET of reactor radiations are about the same as for Co60 y-rays, or that the initial yield for bromate decomposition is only slightly dependent on the LET below 0.1 e.v./W. The much larger Go(- Br03-) value observed with the Li6Br03 crystals does indicate, however, that an appreciable dependence on LET exists for higher values of this parameter.34 In our next paper a detailed study on the radiolysis of Li13rOa will be reported, and the hypothesis that the decomposition is caused by “thermal spikes” will be discussed. The dependence of the production of bromide ion on $tht (Fig. 2) may be contrasted with the bromate decomposition data shown in Fig. 1. After an initial increase, the rale of flormation of bromide appeared to become and remain approximately constant, suggesting that as the bromate radiolysis progressed increasing amounts of the bromide ion found jn the irradiated crystals were formed via intermediates, Initial 100-e.v. yields for bromide, Go(Er-), for the salts are given in the second row of Table I1 where it may be noted that their values differ much less from one another than do the yields for bromate decomposition. Evidence that bromine species in addition to bromide ion are formed in the radiolysis of the alkali metal bromates by Co60 ?-rays has been reported.2 The reactor irradiated salts also possessed ozidizing properties, and these were talcen as a measure of the production of radiolytic intermediates. The data in Fig, 3 show the dose dependence of this “oxidizing power”; it may be noted that the concentrations of these fraginents increased nonlinearly and seemed to approach
Table I1 : Yields for Radiolysis Products of the Crystalline Alkali Metal Bromates
Go(-BrOa -)
GdBr-1 Go( “OX”) 20
Go(BrO-) GdBr02-1
LiBrOs
NaBrOs
KBrOa
RbBrOa
CsBrOs
1.4 0.58 0.76 2.2 0.26 0.47
1.4 (0.59) 0.81 2.5 0.22
1.3 0.63 0.74 2.5 0.20 0.54
2.4 0.96 1.54 2.7 0.23 0.31
3.4 0.61 2.86 3.4
0.59
... ...
saturation (or ‘*steady-state”) values. The number of moles of bromine associated with the oxidizing f r a g ments is given by the differences between the data in Fig. 1 and 2 for each salt, respectively. The ratio of the measured “oxidizing power” (equiv. inole-’) of the crystals (Fig. 3) to the moles of partially reduced Br(V) per mole of bromate gives a value of (2 1) for the fragments, where Z is their average oxidation number. Initial 100-e.v. yields for the oxidizing fragments, Go(“Ox”),may be derived using the values of (2, 1) thus found for the average oxidation number, Zo, for small decomposition. As would be expected, the sum of Go(Br-) and GO(“Ox”) niust equal Go(-BrOs-). It is, however, of interest to note that a satisfactory agreement between Go(Br-) and GO(“0x”) for C060 y-rays and reactor radiations exists, for all the salts except LiBrOa. If it is assumed that the oxidizing fragments in the crystals are bromjte and hypobromite solely, values for their yields, Go(BrOz-) and Go(BrO-), may be derived from GO(“Ox”)and Zo (see Table 11). Clearly, bromine-containing fragments are formed in CsBrOs of higher average Z than bromite ion; the assumption that only BrOn- and BrO- are created may be incorrect, therefore, even for the other salts. The values of Table I1 also reveal that bromate ion breaks down much more completely in the lighter than in the heavier alkali metal bromates. This result is somewhat surprising if the crystal “free space” were actually the determining factor governing the radiolysis. The average oxidation number of the incompletely radiolyzed species was found to be nearly a constant for each salt or to approach constancy with increasing bromate decomposition. The tendency to a constant,, nonintegral Z has suggested that several partially rcduced bromine species probably were formed during radiolysis and that these were either further radiolyzed
+
+
(33) G . E. Boyd and J. W Cobhle, J.Phys. Chem., 63,919 (1959). (34) The yield for Li’BrOn in Table I may be reduced from 0.47 to 0.37 when n correction is made to its decomposition for the radiolysis caused by ca. 100 p.p.m. of Lie it contained.
Volume 68, Number 9 September, 1964
2634
G. E. BOYDAND Q. V. LARSON
or thermally decomposed so that ultimately a "steadystate" concentration of each was established. Mechanism of Radiolysis. The following general mechanism can be written
Br-
+
30
where n is an integer with values 0, 1, or 2 and Y 2 0. The species Br03--nY-1will represent the oxidizing fragments which may be either free radicals and/or bromite plus hypobromite ions and/or positively charged bromine-containing ions. If the oxidizing fragment concentration, N z , is governed' by a sequence of first-order irreversible reactions, an exponential relation for their concentration will hold
where Nz" is the "steady-state" concentration and is the thermal neutron dose. The data (Fig. 3) for the two salts, KBr03 and CsBr03, on which a sufficient number of measurements were made fitted eq. 9 quite satisfactorily. The respective values for N t m and kz were 13.2 and 88.7 mequiv. mole-' and 0.18 X and 0.30 X 10-16cm.2. According to the above mechanism, the concentration of bromide ion formed directly from bromate and indirectly by the decomposition of the oxidizing fragments will be governed by eq. 10
&ht
This relation describes the data of Fig. 2 within the limits of their errors. An additional elucidation of the radiolytic inechanisms in the alkali metal bromates will depend upon further knowledge about the chemical nature of the oxidizing fragments. The identification of these intermediates might be accomplished by direct observational methods, or their nature might be inferred indirectly. For example, information on oxygen gas yields, G(O,), might serve to test the hypothesis that BrOz- and BrO- ions are the species present in the crystals a t room temperature. Conceivably, perbromate, Br04-, is formed and stabilized in the crystal; if such a reaction did occur, an oxygen deficit might be observed. The production of substantial amounts of perchlorate in the reactor bombardment of SaClO3 has been reported.35 It is not altogether clear whether the oxidizing fragments decompose thermally or are radiolyzed. If the latter process occurs to an appreciable extent, a dose rate dependence of the bromide ion yield should be observed. Finally, we will remark that the larger G(O2) values found by Hennig, Lees, and Matheson3 in the reactor radiolysis of crystalline KaN03 and K1-03 probably should be interpreted as indicating the existence of an LET effect. The oxygen gas yields they reported (2.5 and 2.6 molecules/100 e.v.) for the decomposition of KC103 have recently been confirmed by radiolysis measurements with 50 KVP X-rays,36as would have been predicted from our work because LET effects should not be present with this compound. It seems likely that the W4(n,p)C14 reaction may contribute importantly to the radiolysis of nitrates in nuclear reactors, and this point deserves investigation. (35) L. J. Sharman and K. J. .McCallum, J . Chem. Phvs.. 23, 597 (1955). (36) H. G. Heal, Can. J . Chem., 37, 979 (1959).
The Journal of Physical Chemistry