Mat. Res. Soc. Symp. Proc. Vol. 792 2004 Materials Research Society R2.5.1 Kinetics of alkali ion echange of irradiated glasses Michael I. Ojovan, William E. Lee Immobilisation Science Laboratory, Department of Engineering Materials, University of Sheffield, Mappin Street, S1 3JD, UK. E-mail: M.Ojovan@sheffield.ac.uk Abstract The kinetics of alkali ion echange of irradiated glasses were investigated using the structural energy barrier model for ion echange of glasses. Derived rates of alkali ion echange depend both on irradiation dose D(Gy) and dose rate q(gy/s) illustrating that some effects cannot be simulated by eternal irradiation and require in-situ measurements. Higher D and q lead to increased ion echange rates. Significant changes occur in the activation energies demonstrating a 4-6 times decrease depending on glass composition. Radiation-induced changes are higher at relatively low temperatures and are diminished by increased glass temperature. Numerical estimations show that changes in alkali ion echange kinetics occur at D far below damaging doses. Introduction The kinetics of alkali leaching from silicate glasses are controlled by the etent and rate of reactions between aqueous solutions and the glass network. It is broadly accepted that the early stage of alkali leaching is controlled by alkali ion echange whereas the later stages are controlled by network hydrolysis and dissolution [1]. However, the duration of the initial stage may be very long, many hundreds of years depending on specific conditions where network hydrolysis and dissolution is suppressed, such as for silica saturated solutions or at low temperatures. Hence the ion echange may control the behaviour of alkali silicate glasses, this being the case for a number of glasses, including low radioactivity sodium silicate glasses [2-5]. The role and effect of irradiation (including self-irradiation inherent to nuclear waste glasses) on alkali ion echange has been the subject of few studies. A particular goal is a model description of ion echange accounting for radiation-induced effects at relatively small doses and dose rates characteristic of radioactive glasses. McGrail et. al. [4] developed a theoretical model describing quantitatively the kinetics of ion echange from alkali silicate glasses, e.g. the structural energy barrier model. This model gives for the alkali ion echange rate (mol/m 2 s): r = ω ep( E / RT), (1) where ω is the echange attempt frequency, E is the energy barrier to ion echange, R is the universal gas constant and T is temperature. The structural energy barrier model gives an ecellent description of the kinetics of alkali ion echange accounting for changes in composition of alkali silicate glasses and isotope effects [4]. The purpose of this paper is to demonstrate that the structural energy barrier model can be readily used to account for the effects of irradiation on the kinetics of alkali ion echange. It will be shown, that the rate of alkali ion echange of irradiated alkali silicate glasses can be determined by the equation: r = ω ep( E / RT) + ωf ( q, D) ep( E / RT), (2) ir
R2.5.2 with a much lower energy barrier to radiation-induced ion echange E ir <<E, and ƒ(q,d) as a simple function of the dose rate q and absorbed dose D of radiation. Equation (2) provides a description of leaching of irradiated glasses consistent with eperimental data. Structural energy barrier model The interdiffusion of H + (or H 3 O + ) and the alkali ion M + echange reaction: Si-O-M + H + (or H 3 O + ) Si-O-H + M + (and + H 2 O) (3) control the early stages of leaching of alkali silicate glasses in aqueous solutions leading to the well known t 1/2 time kinetics [6]. The ion echange interdiffusion of H + (or H 3 O + ) rather than diffusion of M + controls the kinetics [6, 7]. The structural energy barrier model considers the M + - H + echange as an ion hopping process. Since a hydrogen ion is much lighter than a sodium ion the number of attempts for a H ion to jump into a site containing a Na atom is proportional to the vibration frequency of the O-H bond. Assuming the non-bridging oygen (NBO) sites are the only sites susceptible to ion echange the echange attempt frequency ω is proportional to the vibration frequency of the O-H bonds ν H, the hydrogen ion concentration in adsorbed water molecules on the glass surface C H +, and the concentration of NBO sites C NBO : ω ν H C H +C NBO. The probability of successful jumps depends upon the structural energy barrier E through the Boltzmann factor, resulting in equation (1) for the rate of alkali echange. This equation provides a means to account for effects of structure of glass. In addition, since equation (1) comprises the term proportional to the vibration frequency of the O-H bond, it correctly describes the known isotope effect of a ~30% reduction of echange rate when replacing the hydrogen by deuterium. The temperature dependence of ion echange rate (equation (1)) is determined by the energy barrier E. The energy barrier for M + - H + echange is the sum of bond energy (enthalpy E b ) of the alkali M on the NBO site and elastic strain energy (enthalpy E s ) associated with the distortion of the glass network. E =E b +E s (4) The enthalpies E b and E s were calculated in [4] for a number of Na 2 O-SiO 2 -Al 2 O 3 glasses demonstrating that elastic strain energy accounts for only about 20% of the total echange enthalpy, e.g. E s << E. An energy barrier increase from 34 kj/mol for a Na 2 O-2SiO 2 glass to 49 kj/mol for a 38Na 2 O-47SiO 2-15Al 2 O 3 glass was determined correctly accounting for bonding of sodium to NBO sites and an increasing stiffness of glass network with increasing alumina content. Equation (1) has been used to calculate the rates of sodium ion echange for the leaching of Na 2 O-SiO 2 -Al 2 O 3 glasses in silica saturated and slightly alkaline solutions (ph=8) which resulted in sodium release rates from 20 to 80 times faster than rates of glass hydrolysis and dissolution. Ecellent agreement was demonstrated between the calculated and measured sodium ion echange rates along with the correct description of 30% slower rates in silica saturated D 2 O solutions [4]. Effect of irradiation on alkali ion echange We now consider the rates of alkali echange under conditions of irradiation, in which doses and dose rates are far below those leading to macroscopic changes. Moreover we are following the constraints of the original model of alkali ion echange [4] and considering the kinetics of alkali ion echange in conditions of both constant temperature and solution ph. Hence we consider the
R2.5.3 effect of irradiation on the glass not involving any ph change caused by radiolysis or other radiochemical effects. Under such conditions the main result of irradiation is formation of point defects. Classification of point defects in irradiated silicate glasses is based on those in silica glass. There are two types of point defects in irradiated silica glasses: oygen-deficiency and oygen-ecess defects [8]. Oygen deficiency defects are neutral oygen vacancies and paramagnetic E centres. These are schematically represented as Si-Si and Si respectively, where depicts an unpaired spin residing on silicon. The measured number of paramagnetic centres at high doses of irradiation (eceeding 10 10 Gy for silica) achieves saturation values 10-3 of silicon atoms whereas the number of neutral oygen vacancies is larger perhaps by an order of magnitude or more [9]. Note that this saturation occurs below the level at which changes in density of the glass occur [9]. Oygen ecess defects in silica glass are the NBO hole centre and the peroy radical. These are schematically represented as Si-O and Si-O-O. Oygen deficiency centres can be detected in the ultraviolet region around 460nm and 280nm, and oygen ecess centres in the low photon energy region [8, 10]. The most important point defects in alkali silicate glasses are oygen-ecess centres [9]. There are two types of NBO hole centres in silicate glasses, one is an analogue to a NBO hole centre in silica glass and the second involves two NBOs on the same silicon. The NBO hole centres are formed by radiation-chemical reaction: Si-O-M + irradiation Si-O + M + (5) It is assumed that a free hole resulting from the irradiation is first trapped at the Si-O-M site and then the alkali ion is free to diffuse away to the site of a trapped electron where it stabilises the uncompensated charge [9]. Mobilisation of alkalis in irradiated glasses has been confirmed by a large number of studies (see references in the overview [9]), being confirmed also by molecular dynamic simulations of alkali migration in silicate glasses [11, 12]. These demonstrate that the association with NBO confines the alkali to local motion, whereas the absence of a co-ordinating NBO allows the alkali ion to eplore more easily its environment and to undergo long-range migration [11]. Moreover there was recently obtained evidence for the formation of sodium species, resulting from electron trapping by sodium ions [13] which also may affect the behaviour of sodium during ion echange of irradiated glasses. Reaction (5) provides in addition to thermally released alkali ions new radiation-released alkali ions enabling their free diffusion from the NBO sites. The concentration of NBO hole centres created by irradiation C NBOHC is proportional to concentration of NBO sites, C NBO, and depends both on dose rate and cumulative dose: C NBOHC C NBO ƒ(q,d). Function ƒ(q,d) generally holds its values within an interval 0 ƒ(q,d)<1, moreover ƒ(q, D) is equal to zero at q=d=0, then it grows by an increase in D achieving its maimum value ƒ ma <1 at very high doses (but below doses at which changes in macroscopic properties occur) [9]. Assuming only irreversible reactions ƒ(q,d) can be epressed in terms of q and D by a fractional eponent law [14]. At relatively low doses when kd<<1, where k is the rate constant, this leads to: f(q,d) = KD b, where b is the fractional eponent 0<b 1. Both K and b can be determined eperimentally, for eample paramagnetic defects in silica glass hold b = 0.80 [14]. Thermally and radiation-induced annealing of a proportion of the defects with a characteristic time τ back into the network configuration during the process of defect creation, will add a term proportional to dose rate. More generally, accounting for the Kohrausch relaation W(qτ) B, where fractional eponent B holds its values within an interval 0<B 1:
R2.5.4 f(q,d) = KD b + W(qτ) B (6) Reaction (5) demonstrates that under irradiation additional channels are generated for successful hops of an alkali ion to jump from a NBO site, where it is bound by the Coulomb interaction. Even without irradiation such jumps occur due to thermal activation, which is included in the ion echange rate (1) by the Boltzmann factor ep(-e /RT). The Boltzmann factor consists of two parts: the first eponent ep(-e b /RT) accounts for the release of an alkali from the NBO site and the second eponent ep(-e s /RT) accounts for the successful motion of an alkali away from the NBO site. Note that at low doses, when there are no significant changes in the glass network, it is unlikely that the motion of an alkali away from an NBO site can be significantly assisted by irradiation. According to the structural energy barrier model the NBO sites are the only sites susceptible to ion echange. The number of alkali ions thermally liberated from the NBO sites (or unbound alkalis at these sites) is proportional to C NBO ep(-e b /RT). The number of new unbound alkali ions, e.g. released by irradiation from NBO sites, is proportional to C NBO.ƒ(q,D). Hence under irradiation the total number of successful jumps of alkali ions from the alkali ions NBO sites (or unbound alkalis at these sites) will become proportional to C NBO ep(-e b /RT)+ C NBO ƒ(q,d). This results in equation (2) of the rate of alkali ion echange of irradiated glasses with activation energy of radiation-induced ion echange E ir : E ir =E s, (7) where the elastic strain enthalpy E s is associated with the distortion of the glass network. E ir is always much lower than the activation energy of thermally induced ion echange E. For eample, Na 2 O-2SiO 2 glass has E =33.6 kj/mol, and a much lower E ir =E s =7.87 kj/mol [4], whereas 38Na 2 O-47SiO 2-15Al 2 O 3 glass has E = 48.5 kj/mol and correspondingly E ir =E s =8.58 kj/mol [4]. Since the radiation-induced ion echange has a much lower activation energy its contribution can be considerable even at low doses. Hence irradiation provides an important input to ion echange due to releasing alkali ions through radiochemical reaction (5), which makes alkali ions much more mobile and ready for ion echange. Equation (2) gives the rate of alkali ion echange of irradiated non-radioactive glasses, for which D>0 but q=0, and actual radioactive glasses or glasses being irradiated, when both D>0 and q>0. Consider equation (2) for the rate of alkali ion echange. Assume q=d=0, this returns us to the standard structural energy barrier model of ion echange and (2) coincides with (1) since ƒ(0,0)=0. However, under irradiation the rate of alkali ion echange is always higher compared to non-irradiated glasses since ƒ(q,d) is nonnegative. The higher the dose rate and the absorbed dose the higher the contribution of irradiation to the rate of ion echange. The maimum input of irradiation in the rate of alkali ion echange is achieved at saturation density of NBO hole centres, e.g. at high doses when ƒ(q,d) ƒ ma. Equation (2) is supported by the observed behaviour of the rate of alkali leaching from irradiated glasses. The eperimental data indicate a three-four times faster leaching rate for irradiated glasses the effect being characteristic to short term leaching (see [9] and corresponding references). Comprehensive studies of leaching from actual radioactive borosilicate glasses demonstrated that the initial (lasting for a few tens days) leaching rates were about 5 times higher for radioactive glasses compared to non-radioactive simulants [15]. Since the initial leaching is controlled by ion echange this result corresponds to conclusions drawn from equation (2).
R2.5.5 Discussion Equation (2) highlights that the rate of alkali ion echange retains activation character in irradiated glasses. However, the activation energies of the two terms in equation (2) are very different. The activation energy for thermally induced ion echange E is at the level of tens of kj/mol. The additional term accounting for irradiation has the activation energy E ir =E s, which is much less than E. This means the contribution from the irradiation is always more significant at low temperatures and may vanish at higher temperatures against the background of thermallyactivated processes of ion echange. To compare alkali ion echange in irradiated glasses to that in un-irradiated glasses the relative ion echange rate rr is defined by dividing the rate of alkali ion echange of an irradiated glass to that of an un-irradiated one, e.g. dividing equation (2) by equation (1): rr =1 + ƒ(q,d)ep(e b /RT) (8) Note that the relative rate is dimensionless and always rr 1, meaning that irradiation can only increase the echange rate assuming that doses and dose rates are low enough to avoid radiation damage. Since the activation barrier E b is rather high (e.g. of the order of tens of kj/mol) it follows that rr has a very strong dependence upon temperature, rapidly decreasing when temperature increases. We may consider that the increase of ion echange rate is significant in eperiments for eample at rr 2, and not important when rr <2. Hence we neglect any increase, that is less than two times. From equation (8) it follows that there is a certain (critical) temperature T* below which we can observe the increase in alkali ion echange and above which this change is not important: Eb T = (9) Rln 1 f ( q, D) The critical temperature achieves its maimum at high doses of irradiation, when ƒ(q,d) ƒ ma. Equation (9) shows that at low doses of irradiation when D 0, the critical temperature also tends to zero, making observation of radiation-induced effects impossible. The lower the bond strength the lower the critical temperature. Since the minimum reasonable T* is ~0 o C, e.g. T* min ~273 K, there is a critical dose D* below which the radiation-induced increase of ion echange may be not taken into account: D * =(1/K 1/b )[ep(-e b /RT min )-W(qτ) B ] 1/b (10) If the irradiation dose D<D* the change in the rate of ion echange is insignificant. Considerable increase in the ion echange rate can be observed only by eceeding D*. Note that D* is always lower if q>0, e.g. for actual radioactive glasses meaning an earlier detection of radiation induced effects. The critical dose is lower at high rates of formation of point defects, and at higher bond strength of alkali ions to NBO. Increased ionic bond strength thus causes a significant acceleration of alkali ion echange. This is quite surprising, demonstrating that a higher strength of alkali binding to the glass network, which obviously results in lower initial leaching rates for non-irradiated glasses, may however result in higher increase of leaching rates under irradiation. This conclusion however must account for possible changes in the rate of formation of defects under irradiation, which is given by the rate constant k and has different values for different glasses.
R2.5.6 Conclusions The structural energy barrier model for ion echange of glasses developed by McGrail et al [4] was used to calculate rates of alkali ion echange of irradiated glasses. There is an increase of rates of alkali echange due to irradiation within this model. The increase of ion echange rates occurs with an activation energy that is much less than the thermal activation energy. Changes in rates of ion echange are higher at low temperatures diminishing with increased temperature. For every dose of irradiation there is a critical temperature above which the radiation increase of ion echange is negligible. For every temperature there is a critical dose of irradiation below which the increase of ion echange is negligible. Below the critical temperature and above the critical dose the increase of alkali ion echange rate is significant. This may be important for performance assessment and modelling behaviour of nuclear waste glasses and for processes involving ion echange, e.g. etraction by leaching. Literature 1. P. Van Iseghem ed. Glass in its disposal environment. Proc. Int. Topical Workshop on Glass in its Disposal Environment. J. Nucl. Mat., 298 (1, 2) (2001). 2. Y. Chen, B.P. McGrail, D.W. Engel. Mat. Res. Soc. Symp. Proc., 465, 1051-1057 (1997). 3. J.W. Sheng, S.G. Luo, B.L. Tang. Nuclear Technology, 123 (3) 296 (1998). 4. B.P. McGrail, J.P. Isenhover, D.K. Shuh, P. Liu, J.G. Darab, D.R. Baer, S. Thevuthasen, V. Shutthanandan, M.H. Engelhard, C.H. Booth and P. Nachimuthu. J. Non-Cryst. Solids, 296, 10-26 (2001). 5. M.I. Ojovan, R. Burcl. Proc. 2003 EPRI Int. Conf. in Conj. with IAEA, July 16-18, 2003, New Orleans, LA, USA, CD ROM, EPRI-S08-P7.pdf (2003). 6. R.H. Doremus. J. Non-Cryst. Solids, 19, 137-144 (1975). 7. F.M. Ernsberger. J. Amer. Chem. Soc., 66 (11), 747-750 (1983). 8. L.Skuja. J. Non-Cryst. Solids, 239 (1-3) 16-48 (1998) 9. W.J. Weber, R.C. Ewing, C.A. Angel, G.W. Arnold, A.N. Cormack, J.M. Delaye, D.L. Griscom, L.W. Hobbs, A. Navrotsky, D.L. Price, A.M. Stoneham, M.C. Weinberg. J. Mater. Res., 12 (8) 1946-1978 (1997). 10. K. Moritani, I. Tagagi, H. Moryama. J. Nucl. Mat., 312, 97-102 (2003). 11. A.N. Cormack, J. Du, T.R. Zeitler. J. Non-Cryst. Solids, 323, 147-154 (2003). 12. J.M. Delaye, D. Ghaleb. Phys. Rev. B. 61 (21) 14481-14493 (2000). 13. B. Boisot, G. Petite, D. Ghaleb, N. Pellerin, F. Fayon, B. Reynard, G. Calas. Nucl. Instr. and Meth. in Phys. Res. B 166-167, 500-504 (2000). 14. V.A. Mashkov, W.R. Austin, L. Zhang, R.G. Leisure. Phys. Rev. Lett., 76 (16), 2926-2929 (1996). 15. T. Advocat, P. Jollivet, J.L. Crovisier, M.del Nero. J. Nucl. Mat., 298 (1, 2) 55-62 (2001).