A Problem For the Mush Model: Mismatch Between Trace Element Patterns Of Rhyolites And Dacite Matrix Glass
Abstract
Extraction of liquid from granodiorite mush—the "mush model" of crystal-liquid separation—is a widely accepted method for production of high-silica rhyolite. An obvious test of this hypothesis is to analyze glasses in dacite lavas and tuffs and see if they match the geochemical characteristics of rhyolites. We have done so for glass matrices of 12 dacites from 11 Pliocene-Quaternary volcanoes in the central Andes and find that their glasses, although broadly rhyolitic, fail to match the compositions of rhyolites in several key elements. The mush model thus fails this test. The sample set, analyzed by LA-ICPMS and FE-microprobe, comprises dacites and rhyodacites (avg whole-rock SiO2 67 wt%) with glassy or pumiceous textures and tens of % phenocrysts (dominantly hbl+bio+plag±qtz±san); reference dataset is 3400 silicic volcanic rocks (72-80 wt% SiO2) from circum-Pacific convergent margins. SiO2 in the glasses, calculated volatile-free, ranges up to 79 wt%, with several above the nominal igneous limit of 77.4 wt% and beyond nearly all rhyolites. Although elements that should be held by the crystal fraction (e.g., Ba, Sr, Zr, etc.) overlap with the reference dataset, Y and middle-heavy REE are uniformly lower in the glasses than in rhyolites, and REE patterns are uniformly more scoop-shaped than rhyolites. Modal titanite has only been found in 5 of the dacites; MREE depletion may reflect high modal hornblende. U and Th are significantly higher in the glasses than in rhyolites. It is unlikely that the 12 samples analyzed are all outliers, and we conclude that they contradict the hypothesis that convergent-margin rhyolites form by extraction of liquid from crystal mushes that congeal into granodiorite plutons. Our data are consistent with other failed tests of the mush hypothesis (e.g., comparisons of volcanic and plutonic rocks from the same setting; compositions of aplites) and physical reasoning (e.g., difficulty of separating sub-mm crystals from highly viscous liquid). We suggest that the mush hypothesis for the generation of rhyolites should be abandoned.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFM.V44A..07G
- Keywords:
-
- 1211 Non-tectonic deformation;
- GEODESY AND GRAVITYDE: 7280 Volcano seismology;
- SEISMOLOGYDE: 8145 Physics of magma and magma bodies;
- TECTONOPHYSICSDE: 8439 Physics and chemistry of magma bodies;
- VOLCANOLOGY