Compositional effects on crystallization in reheated tephra through simple crystallization experiments

Abstract:

Microcrystalline inclusions within microlite-poor matrix are surprisingly common in low intensity eruptions around the world, yet their origin is poorly understood. Inclusions are commonly interpreted as evidence of crystallization along conduit margins. Alternatively, these clasts may be recycled from low level eruptions where they recrystallize by heating within the vent. We conducted a series of heating experiments using untreated basaltic andesite and basaltic lapilli, as well as basaltic lapilli encrusted with NaCl and in 3.5% and 10% NaCl solutions. All samples were heated from temperatures below the glass transition (~690°C) to ≥ 1100°C for 2 to >60 minutes. In basaltic andesite samples at 690°C < T < 800°C, crystallization is evident after heating for ~20 minutes; at T > 800°C, crystallization occurs in <5 minutes; at T ≥ 900°C, basaltic andesite samples recrystallize extensively in 2-10 minutes. Untreated, NaCl-encrusted basalt samples and those in 3.5% or 10% solutions, however, did not exhibit new crystallization until T ≥ 1000°C, at which extensive crystallization is observed after 10 minutes. Extensive crystallization in all samples included pyroxenes, Fe-oxides, and plagioclase.

Timescales of crystallization are similar for untreated, NaCl-encrusted samples, and samples in saline solutions. However, the extent of crystallization appears greater in samples with NaCl (both encrusted and in solutions), including localized areas of extensive crystallization; a texture not observed in untreated samples. These localized crystalline areas strongly resemble microcrystalline inclusions observed in lapilli from the submarine volcano NW Rota-1, Mariana arc, both of which have increased Cl concentrations within the matrix of the crystalline areas (determined through EDS element maps), which suggests Cl diffusion during heating and a potential signature of submarine recycling. Local extensive crystallization may be due to interaction with NaCl through encrusted salt or seawater, creating an increased availability of Na which allows more rapid plagioclase crystallization, leaving residual Cl within the matrix.

From our experiments, more evolved compositions begin to crystallize at shorter timescales and lower temperatures than basalt, while all samples crystallized extensively at T ≥ 1000°C. The compositional dependence of heating-induced crystallization may be due to small quantities of preserved H₂O, which can suppress crystallization (particularly of plagioclase). Once heated any residual volatiles will be further degassed which may induce additional crystallization.