Is SiO2 mobile during the metamorphic history of the amphibolites (meaning that it can be observed in products of the previous reactions)? Is this SiO2 lost to a fluid phase during retrogression?
If metamorphic condition adequately medium grade amfibolite facies (or higer) SiO2 be mobile phases. Silica be formed viens, cracks and migmatites. Often on peak/regress stage formed metamorphic differentiations and metasomatic recycling (rock divide on basic and felsic layers).
For my PhD, I made some reflections about the contrasting behaviour of metapelites and mafic rocks during high-grade metamorphism. Maybe this could be interesting to you, even if it is not the exact anser to your question. I just add the chapter.
It was rather the question, where do metamorphic fluids stem from and are they preserved in a specific rock or do the escape from the rock? What differences do occur during differing PT-loops? Actually, PT-calculations are system conserving, i.e. nothing escapes from the system, but you can see what would escape (e.g. fluids). Calculations along a certain loop however where allowed to let escape a reasonable amount of fluid (if I remember well).
Of course, almost everything is soluble in metamorphic fluids, and as most veins tell us, especially quartz. But at high T, even in amphibolites, quartz and free fluid are involved in melting reactions and the fluid is sucked up in the melt until the latter is saturated. Probably only when saturation is reached a free fluid phase would coexist with the melt and would be able to dissolve quartz (in a fluid - not in a melt). However you would need quite a lot of fluid. At T below the wet solidus it would be easier to dissolve quartz in a fluid. Deformation would certainly increase the rate of quartz dissolution.
Probably field relations would help to decipher such questions. A closer look at the sites of quartz precipitation should provide important information.
And few add-ons: Fluid not always only SiO2, in amfibolites and high conditions be CO2 (more in granulite), Cl, Na etc. This is components was be included in scapolite, amfibole and carbonate. In articles Austerheim (Nordway) this components transform granulites in to eclogites.
The following papers may be helpful references in the discussed issue:
Heinrich C. A. (1982) Kyanite-eclogite to amphibolite fades evolution of hydrous mafic and pelitic rocks, Adula nappe, Central Alps. Contributions to Mineralogy and Petrology, vol. 81, 2, Issue 1, pp 30-38
Hauzenberger C. A., Mogessie A., Hoinkes G., Felfernig A., Bjerg E. A., Kostadinoff J., Delpino S. and Dimieri L. (2001) Metamorphic evolution of the Sierras de San Luis, Argentina: granulite facies metamorphism related to mafic intrusions. Mineralogy and Petrology, vol. 71, 1-2, pp 95-126