Comparing the reactivity of glasses with their crystalline equivalents: the case study of plagioclase feldspar

Abstract

To evaluate the impact of atomic short- and long-range orders on silicate dissolution kinetics, the dissolution of amorphous and crystalline oligoclase was investigated at pH 1.5 and 10 at 90 °C. Experiments in solution saturated with respect to SiO2 am were additionally performed to constrain the effect of Si-rich surface layer formation on dissolution rates. The face-specific dissolution rates of the crystalline oligoclase and of the oligoclase glass were determined from element budget in solution and surface retreat measured by vertical scanning interferometry. The results show that atomic ordering primarily impacts solid reactivity, irrespective to the pH of the solution. A strong relation between the crystal surface orientation, the evolution of its topography and its dissolution rate was observed. The (001), (010) and (101¯) flat faces containing the strongest bonds dissolved the most slowly and their dissolution rates remained constant throughout the experiments. In contrast, the stepped (11¯1) face was characterized by the highest initial dissolution rate, but progressively decreased, suggesting that the preferential dissolution of stepped sites expose afterwards more stable planes. The differences in terms of etch pit density from one surface to another also explained the difference in dissolution rates for the (001) and (010) faces. The fluid chemistry suggested the formation of very thin (100–200 nm) Si-rich surface layers in acidic conditions, which weakly affected the dissolution rate of the pristine crystal. At pH 1.5, oligoclase glass dissolves at a rate similar to that of the fastest studied faces of the crystal, suggesting the absence of structural effect on oligoclase dissolution. Whereas Si-rich surface layers likely formed by interfacial dissolution-reprecipitation for oligoclase crystal, molecular dynamic calculations suggest that the slightly more open structure of the glass could also allow ion-exchange following water diffusion into the solid. This mechanism could explain why the surface layer of the glass is characterized by a different chemical composition. Results at pH 10 are strikingly different, as the oligoclase glass dissolves up to 20 times faster than its crystalline equivalent. This non-linear response of the material upon pH was linked to the density of critical bonds in oligoclase that is indeed pH-dependent. In acidic pH, the preferential dissolution of Al leaves a highly polymerized and relaxed Si-rich surface, whereas in basic pH the preferential dissolution of Si leads to a complete destructuration of the network because of the lack of Si[sbnd]O[sbnd]Al bonds.