Infrared spectroscopy was used to measure structural changes, water surface concentrations and effective hydroxyl diffusion coefficients in silica glass during isothermal hydration heat-treatments at temperatures from 80 to 1150 degrees C in 0.467 atm of water vapor. The observed glass structural changes were determined to be identical to relaxation during annealing, and it was found that infrared spectrometry may be used to measure glass fictive temperatures. Minute amounts of water had a pronounced accelerating effect on structural relaxation, and relaxation, in turn, affected the water content of the glass in three ways: (1) slow relaxation at low temperatures hindered the glass-water reaction or caused a slow increase of the reaction equilibrium constant; (2) expansion of the glass during water entry allowed an increase of the molecular water solubility; and (3) healing of the glass during bulk relaxation caused a decrease of the hydroxyl solubility. These processes occurred at different rates causing a peculiar increase and then decrease with time of both the surface hydroxyl concentration in thick specimens and total hydroxyl uptake in thin specimens. This observation was used to demonstrate that hydroxyl solubility measurements taken below 850 degrees C by other researchers are not true equilibrium solubilities. Additionally, the kink at 550 degrees C in the Arrhenius plot of D-eff,D-OH as observed by Wakabayashi and Tomozawa in 1989 was found to be a time-dependent phenomenon which is explained in terms of slow glass-water reaction during relaxation. Diffusion of water into silica glass is therefore suggested to be bounded by two extremes: a high-temperature (> 850 degrees C)/long-time extreme where relaxation and reaction are faster than diffusion and water diffuses according to the Doremus model and a low-temperature/short-time extreme in which relaxation and reaction are slower than diffusion and water penetration is limited only by the diffusion coefficient of molecular water in the glass.
