Dissolution of nepheline, jadeite and albite glasses: Toward better models for aluminosilicate dissolution

SLB acknowledges many educational and entertaining conversations with Hal Helgeson (ranging from kinetics to bent head morphologies) over the last 17 years. To investigate the effects of changing the Al/Si ratio on plagioclase dissolution without complications of varying Na/Ca content or exsolution, three glasses with varying Al/Si ratios (albite, jadeite, and nepheline glasses) were synthesized and dissolved. Many similarities in dissolution behavior between plagioclase crystals and this suite of glasses were observed: 1) Dissolution was slowest at near-neutral pH and increased under acid and basic conditions; 2) Dissolution rate at all pH values increased with increasing Al/Si ratio; 3) the pH dependence of dissolution was higher for the phase with Al/Si = 1 than the phase with Al/Si = 0.3; 4) after acid leaching, the extent of Al depletion of the altered surface increased with increasing bulk Al/Si ratio from Al/Si = 0.3 (albite glass) to 0.5 (jadeite glass), but then decreased in nepheline glass (Al/Si = 1.0), which dissolved stoichiometrically with respect to Al; and 5) little to no Al depletion of the surface of any glass occured at pH > 7. In contrast with some observations for plagioclase dissolution, however, log (rate) increased linearly with Al content, and n, the slope of the log (rate) - pH curve at low pH, varied smoothly from albite glass to jadeite glass to nepheline glass (n = -0.3, -0.6, and -1.0, respectively). These results, plus the observation that the slope calculated at high pH, m. did not differ between glasses (m = 0.4 ± 0.1), may be consistent with an identical mechanism controlling dissolution of albite, jadeite, and nepheline glasses, although no Si-rich layer can develop on nepheline because of the lack of SiOSi linkages. Such a conclusion is consistent with a transition state for these aluminosilicates at high pH consisting of a deprotonated Q^Si₃ hydroxyl group (where Q^x_v refers to an x atom in a tetrahedral site with v bridging oxygens) or a five-coordinate Si site after nucleophilic attack by OH^-. At low pH, bridging oxygens between Q^Si₄ and Q^Al₄ may be rate limiting if they are slower to hydrolyze than Q^Si_v Q^Si_w linkages (v, w ≤ 3). According to this mechanism, dissolution rate increases from albite to jadeite to nepheline glass because hydrolysis of AlOSi bonds become more energetically favorable as the number of Al atoms per Si tetrahedron increases, a phenomenon documented here by geometry optimizations by use of ab initio methods. A model wherein Q^Al₄ Q^Si₄ linkages are faster to hydrolyze than lower connectivity linkages between Si atoms (Q^Si_v Q^Si_w v,w ≤ 3) may also explain aspects of this data. Further computational and experimental measurements are needed to distinguish the models.