Post-glacial explosive eruptive history of Mentolat volcano of the Andean Southern Volcanic Zone of Chile is investigated using the petrology and geochemistry of six explosively derived tephra. Each eruption contains two texturally and compositionally distinct glasses: a microlite-poor rhyolitic glass (~74 wt% SiO₂) and a microlite-rich andesitic to dacitic glass (60-69 wt% SiO₂). The tephras are crystal-rich containing plagioclase, amphibole, orthopyroxene, clinopyroxene, olivine, ilmenite, magnetite, and apatite. The high crystallinity and the compositional similarity of minerals within crystal clots and the matrix glasses indicate the presence of a mush zone within the reservoirs feeding Mentolats explosive eruptions. Plagioclase phenocrysts have bimodal core compositions with a prominent mode between Anorthite (An) 60-40 and a less significant mode at An 85-70. Plagioclase with high An (~80) core compositions are strongly resorbed and mantled with lower An (60-50) oscillatory zoning. Temperatures and pressures estimated using several thermobarometers (two-pyroxenes, magnetite-ilmenite, and amphibole-plagioclase) produce generally similar results within each eruption and indicate that there have been minor variations in the intensive parameters of the magmatic systems throughout the last 18,000 years. Pre-eruptive temperature estimates range between 800-1000 °C, an oxygen fugacity of NNO + 0 to 1.5, and a dominant pressure range between 5-8 kbar. Some magnetite and ilmenite pairs produce higher temperatures at the grain boundaries compared to the cores indicating late-stage temperature increase in four of the six eruptions. These results indicate that explosive eruptive activity at Mentolat commonly involves the mixing and mingling of two distinct melts prior to the eruption which disrupts and entrain crystal-rich material from a mush zone. These mixing processes cause minor reheating of an evolved shallow reservoir which may be an important mechanism triggering explosive activity at Mentolat.

Melt inclusions can retain the chemical composition and pre-eruptive volatile content of their parental magmas, and they are most frequently investigated toward these aims. In volcanic systems, entrapped melts are susceptible to post-entrapment modification due to changes in thermobarometric conditions and chemical potentials affecting the system as an unavoidable consequence of eruption. Whilst mechanisms of chemical exchange can affect the fidelity of melt inclusions as accurate recorders of pre-eruptive compositions, these processes are fundamentally diffusion-controlled, and thus potentially recorded as fossil chemical gradients, especially near inclusion margins, to which diffusion chronometry could be applied. Due to the rapidity of volcanic eruptions, eruption-linked chemical gradients might be expressed over only very short distances from inclusion walls within a mineral host. Meanwhile, owing to rapid diffusivity and resultant homogenization of chemical components in the melt phase, chemical gradients preserved in glass can be equally abrupt and likewise only observable in immediate proximity to inclusion walls. The extraordinary spatial resolution of modern nanoanalytical techniques (TEM-EDS, NanoSIMS, Atom Probe Tomography) can readily access these gradients even for rapid explosive eruptions. The accuracy of these techniques can be augmented by conjunctive EPMA analysis, normally performed prior to sample preparation for nanoanalysis.We have investigated melt inclusion wall interfaces from the 2007 caldera-forming eruption sequence of Piton de la Fournaise volcano. Long wavelength diffusion gradients (tens of microns) at these inclusion margins have been investigated previously to estimate timescales of magmatic residence in the weeks and months before and after caldera collapse [Albert et al. 2019 EPSL 515:187-199]. We expand on this study by examining near-interface (micrometer and submicrometer) gradients formed during the hours and minutes of the eruption using both TEM-EDS and nanoSIMS.

In this study, we analyzed δ³⁰Si in-situ, in euhedral olivine and their hosted melt inclusions (OHMIs) from 3 basalts from the Mid-Atlantic Ridge. To date, only bulk δ³⁰Si have been reported for oceanic crust samples and mantle peridotites with a range from -0.19‰ to -0.44‰ [1]. Silicon isotopes were first measured in OHMIs with a Cameca IMS1280HR. A calibration using optical basicity was used to correct for instrumental mass fractionation [2]. Average δ³⁰Si in OHMIs for the 3 samples range from -0.20‰ ± 1.27 (2sd) to 0.06‰ ± 0.70 (2sd). Importantly, δ³⁰Si variability could be up to 2‰ within a sample, which is 4 times larger than in Mid-Ocean Ridge Basalt (MORB) [<0.5‰; 1], and 5 times larger than reproducibility measured in BHVO (0.4‰, 2sd). In order to verify this variability, δ³⁰Si were also measured in the same OHMIs, as well their host olivine crystals using laser-ablation-MC-ICP-MS following the protocol published in [3]. OHMIs analyzed by laser ablation display a variability of 0.54‰, by far smaller compare to SIMS measurements and similar to the reproducibility of the standards measured by LA-MC-ICP-MS (0.23 ‰, 2sd). The average δ³⁰Si of all OHMIs measured by LA-MC-ICP-MS is -0.47 ± 0.17‰ (2sd), slightly lower than average MORB glasses from the Atlantic (-0.27 ± 0.06‰ (2sd); [4]).The different δ³⁰Si variability measured in OHMIs by two different in situ methods suggest that despite careful calibrations (R² > 0.99), an analytical artefact resulting in larger variability and a small bias toward higher values still needs to be identified for SIMS δ³⁰Si measurements in glasses. REFERENCES [1] Poitrasson, 2017, RIMG 82[2] Tissandier et Rollion-Bard, 2017, RCMS 31[3] Guitreau et al., 2020, JAAS[4] Savage et al., 2010, EPSL 295

In situ chemical analytical techniques have long played an important role in geoscience. SEM-based x-ray spectroscopic techniques, EPMA in particular, have been widely used for high-precision chemical analyses applied to a range of geological problems. Analysis of natural samples is more challenging than analysis of synthetic materials, owing to greater chemical and structural complexity and variability at every spatial scale. This spatial variation is, in fact, the essence of why in situ chemical analytical techniques are often preferred for geoscience applications. As our ability for nanoscale characterization improves, we must do more than simply change the scale bar. We are forced to re-evaluate the way in which we use data and reframe our understanding of natural systems to these new spatial scales. Important considerations include, firstly, whether we are able to attain comparable accuracy and precision (to microanalysis) so that the same classic methodologies can be used to evaluate phase equilibria in the context of geothermobarometry and other applications. Secondly, because of the greater time commitment required for these techniques, we are forced to reduce the number of observations, thereby potentially reducing their statistical significance. Thirdly, we must consider that nanoscale observations are fundamentally different from microscale observations in that as we begin to approach atomic and molecular spatial scales, we also reach the limits of “bulk” behaviour and material properties.  In this work, we survey three nanoanalytical techniques now increasingly applied in geosciences and, using example datasets, evaluate and compare their suitability and capability in common geological applications. For example and comparison, we have elected to characterize pyroxene because it is a common phase in igneous and metamorphic systems and also because major and trace element systematics of the pyroxene group are commonly applied to a variety of interpretive applications such as thermobarometry, diffusion chronometry, and volcanic petrogenesis. 

Iron isotopic signatures in pyrites are widely used as a proxy for tracking past microbial activity, paleoenvironmental and local redox conditions. Although bulk rock analyses help to understand iron cycling at global scale, few studies focused on the investigation of intra and inter-grain isotopic variabilities. Indeed, the measurement of micro-pyrites less than 20μm in size has been restricted so far by analytical techniques. A gain in spatial resolution is achieved through the development of a new radio-frequency plasma ion source (Hyperion-II) which increases the beam density 10 times compared to the previously used Duoplasmatron source. We present here a procedure of high spatial resolution measurements of iron isotopes in micropyrites using a 3nA primary intensity focused in a 3µm beam. This method has been cross-calibrated on two Cameca IMS 1280-HR2 ion microprobes at CRPG-IPNT (France) and at SwissSIMS (Switzerland). We obtained a reproducibility for repetitive δ⁵⁶Fe measurements on Balmat pyrite standard better than ±0.25‰ (2SD, standard deviation) and the internal error of each analyses was ~0.10‰ (2SE, standard error). None of the possible analytical effects, such as topography or crystal orientation seem to affect Fe isotope analyses. This method offers a new analytical way for searching micrometer scale Fe isotopic variations in natural samples in order to evaluate the part of primary, biological or abiotic, processes and secondary overprint on micropyrites.