Si-Al-rich alkaline glass inclusions in mantle xenoliths have been linked to metasomatic processes in the lithospheric mantle. Despite their importance, there is little agreement on their origin. It has been shown that glass of similar composition to that found in mantle xenoliths can be produced by reaction between orthopyroxene and a variety of Si-undersaturated alkaline melts at 1 atmosphere pressure (see Annual Report 1997). In the present study, the dissolution behaviour of orthopyroxene in a basanitic melt under anhydrous, hydrous and CO2-bearing conditions at pressures between 0.4 and 2 GPa has been investigated. The goal of these experiments was to determine if Si-Al-alkaline glass inclusions can be produced by melt - orthopyroxene reaction in the mantle as has been suggested in the literature.
Because melts that infiltrate the mantle will likely equilibrate with the surrounding wallrock, a series of experiments were performed to examine the composition of the basanitic melt after equilibration with orthopyroxene. The secondary melts formed in these experiments are only slightly silica-enriched and are alkali-poor relative to natural glass inclusions. This result indicates that orthopyroxene - melt reactions that reach equilibrium are unlikely to be responsible for the formation of Si-Al-rich alkaline glass.
Disequilibrium, time-series dissolution experiments using orthopyroxene spheres allowed measurements of dissolution rates and examination of textures and secondary melt compositions. Dissolution rates are rapid ranging from 2.5 µm/min at 0.4 GPa with an anhydrous solvent to 15 µm/min at 2 GPa with a hydrous solvent. Addition of water to the solvent melt results in a large increase in the dissolution rate whereas addition of CO2 decreases the dissolution rate relative to that measured under anhydrous conditions. Anhydrous dissolution rates are time dependent. In experiments of less than 10 minutes duration the reaction zone consists of olivine and Si-enriched glass. In experiments longer than 10 minutes clinopyroxene coexists with olivine indicating that Ca-diffusion caused saturation in the clinopyroxene component. This time dependence of clinopyroxene crystallisation is the reason for the observed time-dependent dissolution rates. The volatile content of the solvent melt also affects the textural development of the reaction zone. Addition of CO2 suppresses olivine crystallisation and results in a larger proportion of clinopyroxene. The same is true to a lesser extent for pressure: the amount of clinopyroxene in the reaction zone increases steadily with pressure.
Glasses in the reaction zones of anhydrous and hydrous experiments at 0.4 and 1 GPa show a close chemical correspondence to those found in mantle xenoliths. At higher pressures the glasses are only slightly enriched in silica and alkalis and do not compare favourably with natural glasses. Addition of CO2 to the solvent results in only minor silica and alkali-enrichment at all pressures studied.
The mechanism of orthopyroxene dissolution is complex involving incongruent breakdown of the crystal, which results in a reaction zone that contains up to 60 modal % olivine. Inward diffusion of Ca results in clinopyroxene saturation and uphill diffusion of K, and to a lesser extent Na, from the Si-poor solvent to the Si-rich reaction zone melt account for the strongly alkaline character of the secondary melt. Olivine, and where present clinopyroxene, are transiently stable and are continuously precipitated in the reaction zone and then redissolved in the solvent as the orthopyroxene is progressively dissolved.
Although there are compelling reasons for supporting the hypothesis that Si-Al-rich alkaline melts are produced by orthopyroxene dissolution in the mantle there are several complications particularly regarding quenching in of disequilibrium reaction zone compositions and the mobility of highly polymerised melts in the upper mantle. It is likely that the formation of Si-Al-rich alkaline melts by reaction is a common process during the ascent of xenoliths. However, static, in-situ reactions within the mantle will likely lead to equilibrium compositions. Thus, secondary melts will be only moderately siliceous and relatively alkali-poor.