Diffusion of water in silicate melts

To better understand diffusion in silicate melts, we first briefly review silicate melt structure. In natural silicate melts from basalt (about 50% Si02) to rhyolite 

As can be seen from the interconversion reaction between two hydrous species H20m and OH (Reaction 3-79), the coefficient for OH is 2. This seemingly trivial point turns out to be the key in understanding H2O diffusion. Because of this factor of 2, the equilibrium constant K involves the square of OH concentration but only the first power of H20 i concentration (Equation 3-80b). That is, the species concentrations of OH and H20 i are not proportional to each other, and neither concentration is proportional to total H2O concentration (H20t) in the entire concentration range. Solving Equations 3-80a,b for the two unknowns of H20m and OH yields 

Substituting Equations 3-83a and 3-83b into Equation 3-78 leads to a complicated nonlinear diffusion equation for [H20t], which must be solved numer- 

Because the derivatives are complicated functions of H20t, HhzO, is a complicated function of H20f 

Concentration profiles of H20m and OH after diffusion experiments can be measured by infrared spectroscopy. For the species profiles to reflect the concentrations at the experimental temperature and not altered during quench, the experimental temperature needs to be intermediate (usually 673 to 873 K) so that species reaction during quench is negligible. From such experiments, assuming Don and are independent of H20t, it is possible to fit all concentration 

---

Wasserburg G.J. (1988) Diffusion of water in silicate melts. /. Geol. 96, 363-367. 

---