Age and origin of quartz xenocrysts in a basaltic lava flow in Martinique 

Abstract

In south-western Martinique (Lesser Antilles), the Pointe Burgos Basaltic monogenetic strombolian cone and lava flow cut through a porphyritic dacitic lava dome (Morne Champagne) dated at 617 ± 52 ka (Germa et al., 2011). An exceptional feature of the basaltic lava flow is the occurrence of about 4% of large (up to 2 cm) quartz crystals. Previous studies suggested that quartz xenocrysts had been added to the basaltic magma upon mechanical magma mixing with the cooled shallow dacitic reservoir, with a 9:1 ratio. Indeed, plagioclase phenocrysts (>1 cm) present resorbed surface and reaction rims, a well-known evidence of crystal remobilization. However, no other textural evidence of magma mixing is visible in the basaltic edifice. Moreover, the quartz crystals present unusual habits, are heavily cracked, and appear as filling voids in the basalt. This led us to investigate the age of the basaltic eruption and of the quartz crystals to propose a scenario for the xenocrysts’ origin. The groundmass of the basaltic lava flow was K-Ar dated at 379 ± 25 ka, therefore ~240 ka after the eruption of the dacitic dome that it cuts through. Such a long time difference suggests that magma in the shallow reservoir was completely solidified when the basaltic magma ascended through it. Therefore, we would have expected to see also enclaves of dacite included in the basalt. Thermoluminescence (TL) dating allows for estimating the time since mineral formation or the last heating of a mineral above ~350 °C, thus representing an ideal tool to test whether the quartz xenocrysts formed synchronously with lava flow, or if they were formed later as substitutional minerals or void fillings. Here, we used red TL in combination with several protocols for dose determination, yielding consistent results. For assessing the dose rate, we took into account the quartz xenocrysts’ size distribution, the radioelement concentration of the basaltic matrix and the sites’ erosive evolution. The latter is particularly important because the basaltic matrix is comparatively poor in radioactivity so that the share of the cosmic dose rate becomes important. We obtain preliminary TL ages of 345 ka (with a total erosion of 20 m) and 377 ka (with an erosion of 50 m), each with a ~7% uncertainty. This result emphasizes the importance of reconstructing the lava flow geometry for answering the research question. As we believe that scenario (1) is more plausible, our results indicate the xenocrysts were likely formed post-eruption, challenging the initially suggested model of magma mixing and favouring secondary precipitation or mineral substitution. In case of TL resetting through geothermal activity, the age would represent the cooling following its cessation. However, given the xenocrysts’ pseudomorphic habit, it appears more likely that they represent secondary precipitates from hydrothermal fluids in voids left by strong geothermal alteration of the basaltic host rock.