Diffusion In Olivine Research Articles

Olivine diffusion constrains months-scale magma transport within Kīlauea volcano’s summit reservoir system prior to the 2020 eruption

Olivine diffusion constrains months-scale magma transport within Kīlauea volcano’s summit reservoir system prior to the 2020 eruption

The presence of silicate melt may enhance rates of cation diffusion in olivine

Olivine is commonly used as a ‘crystal clock’ to extract timescales relevant to pre-eruptive perturbations within mafic magmatic systems. Diffusion chronometry applications require accurate calibrations for the rates at which Fe-Mg or other commonly measured elements like Ni, Mn, and Ca diffuse through the crystal lattice. In the past, these rates have been mainly characterized using solid-solid diffusion couple experiments involving olivine single crystals, thin films, or powder sources. Despite the presence of melt surrounding olivine in natural magmatic systems, very few experiments involving magma have been performed, largely because controlling interface reactions is difficult. For this study, we carried out olivine-melt diffusion experiments as a test of the diffusion chronometry method, and to determine whether the presence of melt influences the calculated timescales. To approximate a natural system, we incorporated small natural Kīlauea and San Carlos olivine seeds within a natural Kīlauea basalt and tracked diffusive re-equilibration through time. To better control interface reactions, after some equilibration period at an initial superliquidus temperature of 1290°C, the runs were rapidly cooled to form a rim and left to dwell at various final temperatures (1200, 1220, 1240, 1255°C) for 6–84 h. Concentration gradients for Fe-Mg, Mn, Ni, Ca were measured, and the step-wise nature of the core-rim transition was ascertained using slow diffusing elements like P or Al. When these gradients are modeled using published diffusivities, the timescales retrieved are typically 10 times longer than the actual experiment durations. Thus, measured diffusivities are an order of magnitude faster than those previously obtained in olivine-solid source experiments, but they are in excellent agreement with the only two other melt-olivine datasets. We explore reasons for why melt-bearing olivine diffusion experiments tend to yield faster rates. The possible effects of (1) growth during diffusion, (2) diffusion during any initial dissolution step, and (3) extended tube or planar defects at the interface on calculated diffusivities are all considered but found to be inconsequential. Instead, we argue that additional point defects (vacancies) are likely created at the interface by higher concentrations in elements like Al or H in the basalt melt compared to other solid couple diffusant sources. Future applications of diffusion chronometry in olivine may require a complete re-evaluation of published diffusivities using melt-bearing experimental configurations.

Ascent rate of the Udachnaya-East kimberlite melts from olivine diffusion chronometry

Kimberlite melts are low volume, volatile-rich melts formed in the asthenosphere at depth >150-250 km. Due to their low viscosity and the exsolution of CO2 triggered by their interaction with the sub-cratonic mantle, these melts are highly buoyant and rise rapidly through the lithosphere, ultimately reaching the surface and erupting at sonic to super-sonic speed. Studies based on fluid/melt dynamics, fracturing and disaggregation of the transported mantle-derived cargo, as well as on H diffusion in olivine and Ar diffusion in magmatic phlogopite showed that the ascent rate of kimberlite melts can vary between 0.1 and 30 m/s. However, these rates are extrapolated from processes starting at different depths, and thus may not be representative for the whole ascent path of kimberlite melts through the sub-cratonic lithosphere. To shed light on this issue, we applied diffusion chronometry to the deepest feature ascribable to the liquid line of descent of the melts, i.e. early liquidus olivine formed around mantle-derived xenocrystic cores from the Udachnaya-East pipe (Siberian craton). Our results showed that the time elapsed between the formation of the Internal Zone II (early olivine on the liquidus) or rims (typical magmatic olivine) around xenocrystic cores and final cooling is between 5 and 74 days. Considering travelled distances of 80-110 km in accordance with thermobarometric calculations, this implies overall average ascent rates of kimberlite melts of 0.02 to 0.23 m/s, which are comparable to those of typical alkali basalts, albeit starting at much greater depth. This indicates that kimberlite melts start with moderate ascent rates and only later increase their speed, likely due to massive CO2 exsolution triggered by decarbonation reactions during melt/fluid-rock interactions in the lithospheric mantle. The overall timescales of 5 to 74 days obtained in our study overlap with the lowest values reported in the literature from chemical zoning of mantle-derived minerals (25 days to >100 years), confirming that the time elapsed between kimberlite-related mantle metasomatism and the ascent of kimberlite melts was short. Importantly, our data show that Fe-Mg interdiffusion in olivine can be used as a trustable chronometer in kimberlite systems, opening new perspectives on the study of ascent dynamics in kimberlites worldwide.

Hide and Seek—Trace Element Incorporation and Diffusion in Olivine

Olivine, once overlooked as a host of trace elements, is becoming increasingly important for our understanding of the kinetic and equilibrium behaviour of these elements. Much of our understanding of trace element substitution and diffusion in geological materials comes as a result of experimental and petrological studies of olivine. Here, we consider trace element concentrations and incorporation mechanisms, and how these relate to diffusive behaviour. If we understand trace element behaviour in olivine, we have a powerful tool kit that can be directly applied to address many problems in petrology and volcanology. Perhaps more importantly, what we have learned from olivine can be applied to other minerals and aid us in addressing other far-reaching questions from across the Earth sciences.

Petrology of the 2020–21 effusive to explosive eruption of La Soufrière Volcano, St Vincent: insights into plumbing system architecture and magma assembly mechanism

Abstract The 2020–21 eruption of La Soufrière, St Vincent began with extrusion of a viscous lava dome, which was destroyed upon transition to a major explosive phase. Here we present petrological data to reconstruct the processes leading up to these events. Bulk-rock SiO 2 contents range from 52.8 to 55.4 wt%, classifying the lava and the subsequent scoria as basaltic andesite, the latter being slightly more mafic. Macrocrystal chemistry and modes (plag–cpx–opx–tmt–ol) and crystallinity (45–50 vol%) are largely identical for both phases of the eruption. Pyroxenes are homogenous and precipitated mostly from andesitic melts. Conversely, plagioclase shows strong normal zonation resulting from magma ascent and stalling at multiple crustal levels. Clinopyroxene thermobarometry reveals that crystallization predominantly took place between 8 and 13 km depth at temperatures of 997 − 35 + 18 ∘ C . A lack of evidence for mafic recharge and changes in volatile content and the omnipresence of xenoliths, suggests pre-eruptive destabilization of an andesitic–dacitic melt pocket that disrupted and entrained antecedent mush. Olivine diffusion profiles show that this interaction preceded the onset of eruption. Low dissolved sulfur contents (≤270 ppm S) place constraints on the total SO 2 gas release. Melt–mush disruption appears to be a dominant driver of eruptions at La Soufrière.

Anisotropic copper diffusion in olivine: Implications for Cu preservation in olivine and olivine-hosted melt inclusions

The rate of copper diffusion in olivine governs the preservation of copper in olivine and olivine-hosted melt inclusions, but Cu diffusivity in olivine has not been well determined especially regarding its anisotropy. In this study, Cu diffusion in single crystal olivine (Fo91Fa09) cuboid, prepared to comply with the three principle crystallographic axes, was investigated at 1273–1673 K and 1–2.5 GPa in a piston cylinder apparatus, using a diffusion-in approach with Cu or Cu2O as the source. Copper concentration profiles were analyzed by LA-ICP-MS scanning on the recovered samples along multiple directions. We found Cu diffusion in olivine to be appreciably anisotropic, with Cu diffusivity along the c axis being greater than that along the a axis by nearly an order of magnitude (Dc ≈ 3Db ≈ 7.5 Da). The activation energy for Cu diffusion is ∼210 kJ/mol, with little dependence on crystallographic orientation. Cu diffusivity at 2.5 GPa is only slightly lower than that at 1 GPa, yielding a small activation volume of 2.3 cm3/mol. While not being as rapid as the migration of H and interstitial Li, Cu diffusion, especially along the c axis, is significantly faster than the major elements of olivine (Fe-Mg, Si and O). Diffusion modeling indicates that the timescale for re-equilibration of Cu between olivine-hosted melt inclusion and external melt is relatively short (<60 yr at 1000 °C and <3000 yr at 800 °C) compared to the typical lifetime of magma chambers. Cu concentrations in olivine-hosted melt inclusions may therefore have been significantly modified from the pristine values. For Cu profiles preserved in olivine phenocryst, our determined Cu diffusivity can be used to constrain its residence time in magma chamber, and the diffusion anisotropy can provide a cross-check.

Years of magma intrusion primed Kīlauea Volcano (Hawai'i) for the 2018 eruption: evidence from olivine diffusion chronometry and monitoring data

Years of magma intrusion primed Kīlauea Volcano (Hawai'i) for the 2018 eruption: evidence from olivine diffusion chronometry and monitoring data

A possible sea-level fall trigger for the youngest rejuvenated volcanism in Hawaiʻi

Many intraplate oceanic islands undergo “rejuvenated” volcanism following the main edifice-building stage. Honolulu features Hawaiʻi’s most recent rejuvenated volcanism. K-Ar dating of Honolulu volcanism suggests that it started at ca. 750 ka and ended at &amp;lt;100 ka. Here, we present new 40Ar/39Ar ages and olivine diffusion modeling from Koko Rift lavas to resolve when the most recent Honolulu eruptions occurred and to evaluate possible mechanisms of rejuvenated volcanism and volcanic hazards. Diffusion modeling of olivine zoning profiles in Koko Rift basalts suggests that magmas were stored in the crust for many months prior to eruption. Six new 40Ar/39Ar ages cluster at 67 ± 2 ka (2σ), which demonstrates that Koko Rift is Hawaiʻi’s youngest known area of rejuvenated volcanism. The timing of Koko Rift eruptions coincides with the pronounced drop in global sea level (∼100 m) during Marine Isotope Stage 4. This major sea-level fall may have triggered the eruptions of Koko Rift magmas that were stored in the crust for months to years at &amp;lt; 15 km depth. The proposed mechanism is similar to that at other volcanic islands, which suggests that changes in global sea level may have significant control on the magnitude and frequency of eruptions at ocean island volcanoes.

Pre-eruptive timescales from the historical Hapaimamo eruption at Mauna Loa, Hawai‘i

The 240 yr B.P. Hapaimamo eruption occurred on the southwest rift zone (SWRZ) of Mauna Loa. A large tephra deposit and an extensive lava flow field were produced that extend from 1880 m above sea level down to the southwest coast. Detailed petrological study of olivine crystals from both deposit types are used to determine the timescales of crystal disaggregation from a cumulate mush zone. Over 70% of studied olivines in both, lava flow and tephra samples, exhibit normal zoning. Tephra and flow olivine core compositions are between Fo89–84 and Fo89–81. Olivine rims in the tephra have compositions of Fo86–81, whereas rims in the flow samples extend to lower values within the range Fo83–71. The remaining tephra olivines are unzoned and fall within the same core compositional range. Using FeMg interdiffusion chronometry, we calculate tephra olivine diffusion timescales from 13 to 1600 days (days to years), with over 70% of timescales <100 days, and lava flow olivine diffusion timescales from 25 to 1110 days with 65% of timescales >100 days. Zoning widths were defined as narrow (type I) or broad (type II) and together with the diversity of timescales indicate an ongoing process of crystal mush disaggregation into the erupted host melt.Comparison of tephra and flow diffusion timescales show consistently longer timescales across the lava flow population, as expected due to continued diffusive re-equilibration. Our results suggest that where textural evidence allows, and tephra timescales can be compared, olivines from lava flows can be used to provide temporal information for magmatic processes within the subsurface and the lava flow. Comparison of published elemental diffusion timescales between Mauna Loa and neighbouring Kīlauea suggest that magma mobilisation and transfer processes occur over similar timescales within the plumbing systems of both Mauna Loa and Kīlauea.

Coupled Li-P Zoning and Trace Elements of Olivine from Magmatic Ni-Cu Deposits: Implications for Postcumulus Re-Equilibration in Olivine

Abstract Olivine is an important mineral in mafic–ultramafic rocks and records various crustal and mantle processes in basaltic magma. Generally, phosphorus (P) is immobile in olivine and its zoning tends to record magmatic processes. In contrast, lithium (Li) is one of the most mobile elements in olivine and tends to be re-distributed by postcumulus to hydrothermal processes. This study reports coupled Li-P oscillatory and sector zoning in cumulus olivine (Fo &amp;gt; 80 mol%) from magmatic Ni-Cu deposits in northwest China. The close to 1:1 relationship between Li + Na cation and P cation in olivine suggests P substitutes for slowly diffusing Si at the tetrahedral site and Li (Na) substitutes for Mg at the octahedral metal site. The preservation of such coupled Li-P zoning suggests that the charge-balanced–controlled substitution of Li+P5+ for Mg2+Si4+, producing a member of the Li(Fe)PO4 structure, may cause sluggish Li diffusion in olivine. The Li-P zoning is uncorrelated to other elements (Fe, Mg, Cr, Al, Ca, Ti, V, Ni, Zn, Mn, Co, and Sc). The Ca abundances in cumulus olivine are significantly depleted relative to those in volcanic olivine and modeled for olivine using the rhyolite–MELTS program, whereas the Mg, Fe, Mn, Zn, Ni, and Co contents show no signs of depletion. The Cr and Al contents in high Fo olivine are lower than those in volcanic olivine with a similar Fo value. The depletions in Ca, Cr, and Al (both divalent and trivalent cations) have been attributed to post-crystallization re-equilibration processes. We suggest that the selective Ca-Cr depletion in olivine resulted from re-equilibration between olivine and clinopyroxene–orthopyroxene–spinel–melts at the postcumulus stage, whereas the Al content variation in olivine is likely controlled by olivine–spinel (pyroxene) re-equilibration. Olivine fractionation modeling results reveal that Co content in both olivine and sulfides changes slightly during evolution, whereas the olivine Ni/Co ratio decreases dramatically from 30 (at Fo90) to 4 (Fo80). These are consistent with the variation of Ni and Co contents in olivine from sulfide-barren rocks but cannot explain the strong positive Ni-Co correlation observed in olivine from the sulfide-bearing rocks. The considerable increase in the Ni/Co ratio in olivine coexisting with the high Ni tenor sulfide compared with the slight increase of the Ni/Co ratio in olivine coexisting with moderate-low Ni tenor sulfide strongly suggests that the Ni and Co contents and Ni/Co ratio in olivine from the mineralized rocks were the results of olivine–sulfide interaction. The olivine Fe/Zn and Mn/Zn ratios show little difference between sulfide-barren and sulfide-rich rocks, but these ratios decrease considerably with the decreasing Fo values. Overall, the study suggests that elements (Ca, Cr, Al, Mg, Fe, Ni, Co, Zn, etc.) diffuse faster than P have been re-distributed in cumulus olivine by re-equilibration processes, modifying the elemental content and inter-element ratios in olivine. The characteristics of the crystallization, particularly the source recorded in cumulus olivine, may be obscured and overprinted by postcumulus processes.

Timescales of crystal mush mobilization in the Bárðarbunga-Veiðivötn volcanic system based on olivine diffusion chronometry

AbstractThe timescales of magmatic processes within a volcanic system may be variable over a volcano’s geological history. Crystals reflect environmental perturbations under which they grew, and compositional gradients quenched inside crystals on eruption can be exploited to extract timescales of magmatic processes. Here, we use multi-element diffusion chronometry in olivine macrocrysts to recover their residence time in a melt that ultimately erupted at the surface. The macrocrysts were mobilized by the carrier melt from mushy layers in the magma reservoir, and diffusion timescales likely reflect the time interval between mush disaggregation, ascent, and eruption. To unravel the evolution of mush disaggregation timescales with time, we target early-Holocene, middle-Holocene, and historical magmatic units erupted in the Bárðarbunga-Veiðivötn volcanic system in Iceland’s Eastern Volcanic Zone. Macrocryst contents vary between samples; early-Holocene samples are highly phyric (10–45 vol% macrocrysts) and contain gabbroic nodules, whereas middle-Holocene (5–15 vol%) and historical units (5–10 vol%) tend to be generally less phyric. Early-Holocene olivine macrocrysts have core compositions in the range Fo84–87, while middle-Holocene and historical samples record a wider range in core compositions from Fo80 to Fo86.5. Olivine rims are in chemical equilibrium with their carrier liquid and are slightly more evolved in early-Holocene units (Fo76–81) compared to middle-Holocene (Fo78–80) and historical (Fo81–83) units. Diffusion chronometry reveals that the timescale between mush dis-aggregation and eruption has changed over time, with timescales getting shorter approaching recent times. Early-Holocene olivine macrocrysts dominantly record Fe-Mg diffusion timescales between 200–400 days, while middle-Holocene and historical units typically record timescales of about 70 and 60 days, respectively. Barometric studies suggest that melts and crystals are likely stored and gradually transferred throughout an interconnected multi-tiered system that ultimately culminates in a mid-crustal reservoir(s) at about 6.8–7.5 ± 2.5 km depth, where final disaggregation by the carrier liquid took place. We argue that, as a result of extensional processes enhanced by rifting events, well-mixed melts got drawn into mid-crustal reservoir(s), causing crystal mush loosening and mobilization. In addition, we propose that more energy in the form of heat and/or melt supply was required in the early-Holocene to break up the dense mush fabric and convert it into an eruptible magma. Conversely, as evidenced by the diverse macrocryst content of the historical units and by the lack of gabbroic nodules, the system has become characterized by a less compact mush fabric since at least the middle-Holocene, such that fresh injection of melt would easily loosen and mobilize the mush, resulting in an eruption within a couple of months. This study provides evidence that along axial rift settings, rifting-related processes can help to “pull the mush apart” with no requirement for primitive magma injection as an eruption trigger. Furthermore, we provide evidence that in the Bárðarbunga-Veiðivötn volcanic system specifically, the time between mush disaggregation and eruption has decreased considerably with time, indicating shorter warning times before imminent eruptions.

From Explosive Vent Opening to Effusive Outpouring: Mineral Constraints on Magma Dynamics and Timescales at Paricutin Monogenetic Volcano

Abstract Paricutin volcano is the youngest monogenetic cone of the Michoacán–Guanajuato volcanic field (Mexico), with an excellent historical eruption record from February 1943 to March 1952. The magma emitted during the 9 years of activity was compositionally zoned from basaltic andesite to andesite. This eruption has been considered a classic example of magma differentiation controlled by crustal assimilation combined with fractional crystallization. However, a recent geochemical study of the eruption products points to mantle source heterogeneity and fractional crystallization as the key processes involved in the compositional variability of the magmas. Here we present a detailed petrological characterization of the minerals [olivine, plagioclase, pyroxene, Cr-spinel, and (Ti-)magnetite] to shed light on the processes that led to the chemical evolution of the eruptive products. Our sample set includes the early tephra from the first weeks or months of eruption and the whole sequence of lava flows that followed. The mineral assemblage, their texture, and chemical composition show a systematic evolution between the products from the opening of the explosive vent to the effusive stage. The early tephra are basaltic andesites with oscillatory-zoned olivine and plagioclase, zoned Cr-spinel, and rare pyroxene xenocrysts. In contrast, later erupted tephra and post-January 1944 lavas are basaltic andesites and andesites with normally zoned olivine, Cr-spinel inclusions in equilibrium with the host olivine, and frequent orthopyroxene (after December 1947) with minor chemical zoning. Mineral chemistry data and olivine diffusion timescales, together with whole-rock geochemistry, suggest a convective magma regime with large temperature and oxygen fugacity gradients characterized by short timescales (few days) during the opening stage, followed by a steadier magma regime with longer timescales (few months) and including periodic magma recharge, mixing, and fractional crystallization. In addition, the mineralogical evidence we have gathered does not support considerable crustal assimilation at Paricutin. This study shows that monogenetic eruptions are far from being geochemically simple, and instead involve multiple magma batches with complex storage and mixing stages before eruption.

Linking fluid dynamics and olivine crystal scale zoning during simulated magma intrusion

The compositional zoning styles of natural crystals produced during magma intrusion can be used to investigate the structure of magmatic plumbing systems and its relation to expressions of volcanic unrest (seismic, deformation, volatiles). However, magma intrusion is a progressive, dynamic process and yields non-monotonic heterogeneities in physio-chemical variables such as complex spatial variations in temperature and liquid composition with time. Such changes in variables are difficult to incorporate in models of crystal zoning in natural systems. Here we take another approach by integrating the results of a numerical multiphase simulation of melt arrival in an olivine-rich reservoir with models of chemical re-equilibration of olivine. We evaluate the diversity of chemical zoning styles and the inferred time scales using Fe–Mg diffusion in olivine for a limited range of system geometries and time-composition-temperature values. Although our models are still a large simplification of the processes that may occur in natural systems we find several time-dependent and systematic relations between variables that can be used to better interpret natural data. The proportions of zoned and unzoned crystals, the zoning length scales, and the calculated diffusion times from the crystals, vary with time and the initial position of the crystal in the reservoir. These relationships can be used, for example, to better constrain the plumbing structure and dynamics of mafic eruptions from monogenetic volcanoes by detailed studies of changes in the zoning of the crystal cargo with eruptive sequence. Moreover, we also find that the time scales obtained from modeling of crystals at a single temperature and boundary condition tend to be shorter (> about 25%) than the residence time, which could also be tested in natural studies by combining crystal time scale records with monitoring datasets.

Reconciling the discrepancy between the dehydration rates in mantle olivine and pyroxene during xenolith emplacement

Hydrogen concentration profiles through olivine and pyroxene in peridotite xenoliths carried in rift basalts from northern Tanzania (Lashaine, Eledoi, and Kisite localities) show bell-shaped distributions, indicating that diffusive hydrogen loss has occurred in all minerals. Homogeneous major element concentrations and equilibration of hydrogen between the cores of olivine and coexisting pyroxene suggest that hydrogen loss resulted from diffusive degassing during host magma emplacement. For these samples, hydrogen diffusivities in olivine and coexisting pyroxene must be within the same order of magnitude, similar to experimentally determined diffusivities, but in contrast to previous observations made on xenolithic peridotites. We demonstrate here, for the first time using natural samples, that significant differences in activation energy is likely the primary parameter that causes the discrepancy between hydrogen diffusion in olivine and pyroxene observed in different suites of mantle xenoliths. Because hydrogen diffuses faster in olivine than in pyroxene as temperature increases, hydrogen loss in the Tanzanian mantle xenoliths must have occurred at relatively low temperatures (∼750 to ∼900 °C), whereas hydrogen loss observed in previous xenolith studies likely occurred at higher temperatures (∼950 to >1200 °C). Thus, the diffusive loss of hydrogen in the Tanzanian mantle xenoliths may have occurred at shallow depths or at the Earth’s surface.

Magma decompression rate correlates with explosivity at basaltic volcanoes — Constraints from water diffusion in olivine

Despite the strong theoretical basis for the importance of magma decompression rate in modulating the dynamics of explosive basaltic eruptions, few measurements exist. Here, we quantify magma decompression rate using the incomplete diffusive re-equilibration of water through olivine for two contrasting eruptions at Cerro Negro volcano in Nicaragua: the 1992 VEI 3 and 1995 VEI 2 eruptions. Both eruptions have the same basaltic composition (SiO2 ∼ 50 wt%) and maximum volatile concentrations (H2O ∼ 4.7 wt%). SIMS and NanoSIMS measurements of water in olivine-hosted melt inclusions (MIs) from both eruptions show a negative correlation between MI size and water loss, as predicted for diffusive re-equilibration. MIs from the 1995 eruption appear to have experienced more water loss compared to those from the 1992 eruption (up to 80–90% water loss in MIs < 50 μm), which is consistent with slower magma ascent.Quantifying the timescale for water diffusion during magma decompression requires an accurate constraint on H+ diffusivity through olivine. We perform dehydration experiments on oriented cuboids of olivine from Cerro Negro in a 1 atm. furnace at 800 and 1000 °C and at an oxygen fugacity 0.25 log units above the nickel-nickel oxide buffer (NNO + 0.25) in a sequence of heating steps. Between each heating step we measure the H+ concentration along each crystallographic direction by FTIR and model these profiles to determine the diffusivity. We find the fastest diffusion direction along a̱ and develop an Arrhenius relationship for the diffusivity along this direction: Da(m2/s)=9.6×10−6exp−125RT, where R is the gas constant 8.314 J/mol K; T is the temperature in K and the activation energy is 125 kJ/mol. These are the fastest rates ever measured for olivine dehydration, which may relate to the low forsterite content of the olivines (Fo ∼ 79).We have developed a 1D model for H+ diffusion along the fast direction since this dominates water loss (Da > 10 × Db or Dc). Using our new H+ diffusivity, we model water diffusion from MIs during magma decompression to determine ascent timescales. We find that magma from the VEI 3 eruption of 1992 ascended at a median rate of 0.007 MPa/s, which is ∼5 × faster than the median for the 1995 VEI 2 eruption, although both eruptions show a ∼ two orders of magnitude spread in decompression rate. Assuming linear decompression, this places a minimum bound of ∼6.5 h on magma ascent during the 1992 eruption and ∼35 h for the 1995 eruption. Monte Carlo simulations give a 2σ uncertainty on log 10(dP/dt) of ∼ 10% — much smaller than the order of magnitude spread in decompression rate within each eruption, which suggests that there is real variation in decompression rate for a given deposit. Our results support the hypothesis that higher decompression rates are correlated with more explosive eruptions and show that diffusion modelling of MI water loss is a powerful tool that is sensitive to changes in decompression between VEI 2 and 3 eruption intensities.

Argon storage and diffusion in Earth’s upper mantle

In this study, fine-grained polycrystalline olivine was doped with argon at static high pressure (0.30 ± 0.01 GPa) and high temperature (1050 ± 25 °C) conditions during 24 h in a Paterson press, and analysed using a step-heating extraction protocol coupled with noble gas mass spectrometry to investigate argon storage and diffusivity in Earth’s upper mantle. Our results show that a single diffusion mechanism controlled argon diffusion in our samples during the step heating experiments. Effective Ar diffusion in olivine has a low activation energy, implying that argon diffusivity is governed by both grain boundary and lattice diffusion. Mean values of lattice diffusion parameters obtained from our results and by reprocessing literature data are Ea = 166 ± 44 kJ mol−1 and D0 = 10−7.04 ± 1.13 m2·s−1, and grain boundary diffusion parameters determined from our data are Ea = 22 ± 5 kJ·mol−1 and D0 = 10−12.87 ± 0.3 m2·s−1. Isotopic diffusivity ratios were constant and close to the values determined by Graham’s law in the C-regime (i.e., bulk diffusion dominated by grain boundary diffusion) and A-regime (i.e., bulk diffusion controlled by grain boundary and lattice diffusion in proportion to the segregation of Ar between those sites), but varied in the B-regime (i.e., bulk diffusion controlled by both grain boundary and lattice diffusion in a complex manner), implying a higher isotopic fractionation in the kinetic B-regime. Extrapolation to typical mantle grain sizes implies that around 22% of the argon in the upper mantle can be stored at grain boundaries and that effective diffusion is mostly in the A-regime, suggesting a low isotopic fractionation and diffusivities faster than lattice diffusivities alone. The consideration of grain boundaries as a potential Ar storage site can modify equilibrium during partial melting and significantly enrich a liquid in Ar during fluid percolation. The grain size dependence of Ar storage and diffusivity highlights the underestimated role of grain boundaries in the upper mantle, especially in zones of reduced grain size (via dynamic recrystallization) possibly followed by fluid percolation and/or partial melting, such as in subduction zones or below oceanic ridges

Crystal and melt inclusion timescales reveal the evolution of magma migration before eruption

Volatile element concentrations measured in melt inclusions are a key tool used to understand magma migration and degassing, although their original values may be affected by different re-equilibration processes. Additionally, the inclusion-bearing crystals can have a wide range of origins and ages, further complicating the interpretation of magmatic processes. To clarify some of these issues, here we combined olivine diffusion chronometry and melt inclusion data from the 2008 eruption of Llaima volcano (Chile). We found that magma intrusion occurred about 4 years before the eruption at a minimum depth of approximately 8 km. Magma migration and reaction became shallower with time, and about 6 months before the eruption magma reached 3–4 km depth. This can be linked to reported seismicity and ash emissions. Although some ambiguities of interpretation still remain, crystal zoning and melt inclusion studies allow a more complete understanding of magma ascent, degassing, and volcano monitoring data.

Combined Fe-Mg chemical and isotopic zoning in olivine constraining magma mixing-to-eruption timescales for the continental arc volcano Irazú (Costa Rica) and Cr diffusion in olivine

Combined Fe-Mg chemical and isotopic zoning in olivine constraining magma mixing-to-eruption timescales for the continental arc volcano Irazú (Costa Rica) and Cr diffusion in olivine

Constraints on oxygen fugacity within metal capsules

Experiments were conducted with olivine encapsulated or wrapped in five different metals (Pt, Ni, $$\hbox {Ni}_{70}\hbox {Fe}_{30}$$ , Fe, and Re) to determine the oxygen fugacity in the interior of large capsules used for deformation and seismic property experiments. Temperature ( $$1200\,^{\circ }\hbox {C}$$ ), pressure (300 MPa), and duration (24 h) were chosen to represent the most common conditions in these experiments. The oxygen fugacity was determined by analysing the Fe content of initially pure Pt particles that were mixed with the olivine powder prior to the experiments. Oxygen fugacities in the more oxidizing metal containers are substantially below their respective metal-oxide buffers, with the $$\hbox {fO}_2$$ of sol–gel olivine in Ni about 2.5 orders of magnitude below Ni–NiO. Analysis of olivine and metal blebs reveals three different length-, and hence diffusive time scales: (1) Fe loss to the capsule over $$\sim$$ 100 $$\upmu \hbox {m}$$ , (2) $$\hbox {fO}_2$$ gradients at the sample–capsule interface up to 2 mm into the sample, and (3) constant interior $$\hbox {fO}_2$$ values with an ordering corresponding to the capsule material. The inferred diffusive processes are: Fe diffusion in olivine with a diffusivity $$\sim \,10^{-14}\,\hbox {m}^2\hbox {/s}$$ , diffusion possibly of oxygen along grain boundaries with a diffusivity $$\sim \,10^{-12}\,\hbox {m}^2\hbox {/s}$$ , and diffusion possibly involving pre-existing defects with a diffusivity $$\sim \,10^{-10}\,\hbox {m}^2\hbox {/s}$$ . The latter, fast adjustment to changing $$\hbox {fO}_2$$ may consist of a rearrangement of pre-existing defects, representing a metastable equilibrium, analogous to decoration of pre-existing defects by hydrogen. Full adjustment to the external $$\hbox {fO}_2$$ requires atomic diffusion.

Volatile loss during homogenization of lunar melt inclusions

Volatile abundances in lunar mantle are critical factors to consider for constraining the model of Moon formation. Recently, the earlier understanding of a “dry” Moon has shifted to a fairly “wet” Moon due to the detection of measurable amount of H2O in lunar volcanic glass beads, mineral grains, and olivine-hosted melt inclusions. The ongoing debate on a “dry” or “wet” Moon requires further studies on lunar melt inclusions to obtain a broader understanding of volatile abundances in the lunar mantle. One important uncertainty for lunar melt inclusion studies, however, is whether the homogenization of melt inclusions would cause volatile loss. In this study, a series of homogenization experiments were conducted on olivine-hosted melt inclusions from the sample 74220 to evaluate the possible loss of volatiles during homogenization of lunar melt inclusions. Our results suggest that significant loss of H2O could occur even during minutes of homogenization, while F, Cl and S in the inclusions remain unaffected.We model the trend of H2O loss in homogenized melt inclusions by a diffusive hydrogen loss model. The model can reconcile the observed experimental data well, with a best-fit H diffusivity in accordance with diffusion data explained by the “slow” mechanism for hydrogen diffusion in olivine. Surprisingly, no significant effect for the low oxygen fugacity on the Moon is observed on the diffusive loss of hydrogen during homogenization of lunar melt inclusions under reducing conditions. Our experimental and modeling results show that diffusive H loss is negligible for melt inclusions of >25 μm radius. As our results mitigate the concern of H2O loss during homogenization for crystalline lunar melt inclusions, we found that H2O/Ce ratios in melt inclusions from different lunar samples vary with degree of crystallization. Such a variation is more likely due to H2O loss on the lunar surface, while heterogeneity in their lunar mantle source is also a possibility. A similar size-dependence trend of H2O concentrations was also observed in natural unheated melt inclusions in 74220. By comparing the trend of diffusive H loss in the natural MIs and in our homogenized MIs, the cooling rate for 74220 was estimated to be ∼1 °C/s or slower.