1:45 PM - 2:15 PM
[MIS18-01] Thermodynamic analysis of water dissolution of silicate glass
★Invited Papers
Radioactive waste generated from the reprocessing of spent nuclear fuel is vitrified and then disposed of in geological repositories. The glass used for vitrification is required to have high chemical resistance to water.
The leaching of glass components is caused by the gradient of chemical potential, and it is considered that the process approaches a virtual chemical equilibrium for the reaction between glass and water:
Glass + xH2O + yH+ = aHSiO3- + bB(OH)3 + cAlO2- + dCa2+ + eNa+ (1)
where x,y, a, b, c, d and e are stoichiometric coefficients.
The equilibrium constant K for the reaction at temperature T is related to the Gibbs energy of glass dissolution (ΔdissG) by
lnK=-ΔdissG/RT (2)
The ΔdissG can be calculated using the Gibbs energy of formation of glass at 298K (G(glass, 298)), the sum of the Gibbs energies of oxide component i (Σx(i)G(ox, i, 298)), and the sum of Gibbs energy of dissolution of i (Σx(i)ΔdissG(ox, i, 298)) by the equation,
ΔdissG = - G(glass, 298)+Σx(i)G(ox, i, 298)+Σx(i)ΔdissG(ox, i, 298) (3)
Linard et al. (2001) measured the G(glass, 298) by a solutoin calorimetry, and determined the equilibrium constant from the Eqs. (2) and (3).
In this study, we investiageted the difference between equilibrium constant for the reaction (1) calculated using thermodynamic data and that determined by dissolution experiments.
We calculated the equilibrium constant (lnK(calc)) using thermodynamic database for silicate melts optimized by CALPHAD method, which has been developed in recent years, and the heat capacity data of glass for the calculation of G(glass, 298). The equilibrium constant (lnK(calc)) was determined from the Eqs. (2) and (3).
Dissolution experiments was carried out by the MCC-3 method for SiO2-B2O3-Al2O3-CaO-Na2O glassses with different amout of SiO2 and Al2O3, which are components that improve water resistance, and the amount of Na2O, which is a component that reduces water resistance. Glass powder (40-100 μm) of a 0.235 g was leached into 1 liter of ultrapure water (pH = 9) at 90°C for 48 hours, and 4 ml of the solution was collected at a given time. The concentrations of Si, B, Al, Ca, and Na were measured by ICP-AES.
The activity of compnents in the soltion was determined using the measured concentration C and the activity coefficients by extended Debye-Huckel equation. The equilibrium constant (lnK(exp)) was calculated using those activity values.
The lnK(exp) decreases with increasing the network-forming oxide (NF)/network-modifying oxide (NM) ratio of the glass, and increased with decreasing Al2O3 content at constant NF/NM ratio.
RTlnK(exp) varied in the range of -65 to +15 kJ/mol and were in agreement with RTlnK(calc) with an error of about ±20 kJ/mol. This indicates that the thermodynamic data of the CALPHAD method may be applied to predict the water resistance of glass. For glasses with high water resistance, the lnK(exp) tended to be lower than the lnK(calc). This is thought to be due to the formation of a silica gel layer on the glass surface, which hinders the reaction.
The leaching of glass components is caused by the gradient of chemical potential, and it is considered that the process approaches a virtual chemical equilibrium for the reaction between glass and water:
Glass + xH2O + yH+ = aHSiO3- + bB(OH)3 + cAlO2- + dCa2+ + eNa+ (1)
where x,y, a, b, c, d and e are stoichiometric coefficients.
The equilibrium constant K for the reaction at temperature T is related to the Gibbs energy of glass dissolution (ΔdissG) by
lnK=-ΔdissG/RT (2)
The ΔdissG can be calculated using the Gibbs energy of formation of glass at 298K (G(glass, 298)), the sum of the Gibbs energies of oxide component i (Σx(i)G(ox, i, 298)), and the sum of Gibbs energy of dissolution of i (Σx(i)ΔdissG(ox, i, 298)) by the equation,
ΔdissG = - G(glass, 298)+Σx(i)G(ox, i, 298)+Σx(i)ΔdissG(ox, i, 298) (3)
Linard et al. (2001) measured the G(glass, 298) by a solutoin calorimetry, and determined the equilibrium constant from the Eqs. (2) and (3).
In this study, we investiageted the difference between equilibrium constant for the reaction (1) calculated using thermodynamic data and that determined by dissolution experiments.
We calculated the equilibrium constant (lnK(calc)) using thermodynamic database for silicate melts optimized by CALPHAD method, which has been developed in recent years, and the heat capacity data of glass for the calculation of G(glass, 298). The equilibrium constant (lnK(calc)) was determined from the Eqs. (2) and (3).
Dissolution experiments was carried out by the MCC-3 method for SiO2-B2O3-Al2O3-CaO-Na2O glassses with different amout of SiO2 and Al2O3, which are components that improve water resistance, and the amount of Na2O, which is a component that reduces water resistance. Glass powder (40-100 μm) of a 0.235 g was leached into 1 liter of ultrapure water (pH = 9) at 90°C for 48 hours, and 4 ml of the solution was collected at a given time. The concentrations of Si, B, Al, Ca, and Na were measured by ICP-AES.
The activity of compnents in the soltion was determined using the measured concentration C and the activity coefficients by extended Debye-Huckel equation. The equilibrium constant (lnK(exp)) was calculated using those activity values.
The lnK(exp) decreases with increasing the network-forming oxide (NF)/network-modifying oxide (NM) ratio of the glass, and increased with decreasing Al2O3 content at constant NF/NM ratio.
RTlnK(exp) varied in the range of -65 to +15 kJ/mol and were in agreement with RTlnK(calc) with an error of about ±20 kJ/mol. This indicates that the thermodynamic data of the CALPHAD method may be applied to predict the water resistance of glass. For glasses with high water resistance, the lnK(exp) tended to be lower than the lnK(calc). This is thought to be due to the formation of a silica gel layer on the glass surface, which hinders the reaction.