共查询到20条相似文献,搜索用时 78 毫秒
1.
Olivine dissolution in basaltic melt 总被引:1,自引:0,他引:1
The main purpose of this work is to understand and quantify diffusive and convective olivine dissolution in basaltic melt. Crystal dissolution and growth in a magma chamber is often accompanied by the descent or ascent of the crystal in the chamber due to gravity. The motion induces convection that enhances mass transport. Such convective dissolution and growth rates have not been quantified before. MgO diffusivity in the melt (DMgO), MgO concentration of the interface melt (C0) and the effective thickness of the compositional boundary layer (δ) are necessary parameters to model the convective dissolution. Experiments of non-convective olivine dissolution in a basaltic melt were conducted at 1271-1480 °C and 0.47-1.42 GPa in a piston-cylinder apparatus. At specific temperature and pressure conditions, multiple experiments of different durations show that the interface melt reaches near-saturation within 2 min. Therefore, diffusion, not interface reaction, is the rate-controlling step for non-convective olivine dissolution in basaltic melt. The compositional profile length and olivine dissolution distance are proportional to the square root of experimental duration, consistent with diffusive dissolution. DMgO and C0 are obtained from the experimental results. DMgO displays Arrhenian dependence on temperature, but the pressure dependence is small and not resolved. C0 increases with increasing temperature and decreases with increasing pressure. Comparison with literature data shows that DMgO depends strongly on the initial melt composition, while C0 does not. δ is estimated from fluid dynamics. DMgO/δ, which characterizes the kinetic and dynamic aspects of convective crystal dissolution, is parameterized as a function of temperature, pressure, and olivine composition. Convective olivine dissolution rate in basaltic melt can be conveniently calculated from the model results. Application to convective crystal growth and xenolith digestion is discussed. 相似文献
2.
Veronika S. Heber Richard A. Brooker Simon P. Kelley 《Geochimica et cosmochimica acta》2007,71(4):1041-1061
Mineral-melt partition coefficients of all noble gases (min/meltDi) have been obtained for olivine (ol) and clinopyroxene (cpx) by UV laser ablation (213 nm) of individual crystals grown from melts at 0.1 GPa mixed noble gas pressure. Experimental techniques were developed to grow crystals virtually free of melt and fluid inclusions since both have been found to cause profound problems in previous work. This is a particularly important issue for the analysis of noble gases in crystals that have very low partition coefficients relative to coexisting melt and fluid phases. The preferred partitioning values obtained for the ol-melt system for He, Ne, Ar, Kr, and Xe are 0.00017(13), 0.00007(7), 0.0011(6), 0.00026(16), and , respectively. The respective cpx-melt partition coefficients are 0.0002(2), 0.00041(35), 0.0011(7), 0.0002(2), and . The data confirm the incompatible behaviour of noble gases for both olivine and clinopyroxene but unlike other trace elements these values show little variation for a wide range of atomic radius. The lack of dependence of partitioning on atomic radius is, however, consistent with the partitioning behaviour of other trace elements which have been found to exhibit progressively lower dependence of min/meltDi on radius as the charge decreases. As all noble gases appear to exhibit similar min/meltDi values we deduce that noble gases are not significantly fractionated from each other by olivine and clinopyroxene during melting and fractional crystallisation. Although incompatible, the partitioning values for noble gases also suggest that significant amounts of primordial noble gases may well have been retained in the mantle despite intensive melting processes. The implication of our data is that high primordial/radiogenic noble gas ratios (3He/4He, 22Ne/21Ne, and 36Ar/40Ar) characteristic of plume basalt sources can be achieved by recycling a previously melted (depleted) mantle source rather than reflecting an isolated, non-degassed primordial mantle region. 相似文献
3.
The influence of melt composition on the partitioning of REEs, Y, Sc, Zr and Al between forsterite and melt in the system CMAS 总被引:1,自引:0,他引:1
Partition coefficients for a range of Rare Earth Elements (REEs), Y, Sc, Al and Zr were determined between forsteritic olivine (nearly end-member Mg2SiO4) and ten melt compositions in the system CaO-MgO-Al2O3-SiO2 (CMAS) at 1 bar and 1400 °C, with concentrations of the trace elements in the olivine and the melt measured by laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The REEs and Sc were added at levels sufficient to ensure that concentrations in the olivine were well above the detection limits. The REE partition coefficients decrease with increasing silica in the melt, indicating strong bonding between REEO1.5 and SiO2 in the melt. The variation of as a function of ionic radius is well described by the Brice equation for each composition, although a small proportion of this variation is due to the increase in the strength of the REEO1.5-SiO2 interactions in the melt with ionic radius. Scandium behaves very similarly to the REEs, but a global fit of the data from all ten melt compositions suggests that deviates somewhat from the parabolas established by the REE and Y, implying that Sc may substitute into olivine differently to that of the REEs. In contrast to the behaviour of the large trivalent cations, the concentration of Al in olivine is proportional to the square root of its concentration in the melt, indicating a coupled substitution in olivine with a high degree of short-range order. The lack of any correlation of REE partition coefficients with Al in olivine or melt suggests that the REE substitution in olivine is charge-balanced by cation vacancies. The partition coefficient of the tetravalent trace element Zr, which is highly incompatible in olivine, depends on the CaO content of the melt. 相似文献
4.
V. Le Roux 《Geochimica et cosmochimica acta》2010,74(9):2779-17643
Oceanic basalts, such as mid-ocean ridge basalts (MORB) and ocean island basalts (OIB), are characterized by large isotopic and trace element variability that is hard to reconcile with partial melting of a peridotitic mantle alone. Their variability has been attributed to the presence of heterogeneities within the mantle, such as recycled crust, metasomatized material or outer core contribution. There have been few attempts to constrain the major element composition of those heterogeneities, most studies focusing on incompatible trace elements and radiogenic isotopes. Here, we report Zn, Mn and Fe systematics in mafic and ultramafic systems (whole-rocks and minerals) and we explore their use for detecting lithological heterogeneities that deviate from peridotitic mantle dominated by olivine and orthopyroxene. We suggest that Zn/Fe ratio is a particularly promising proxy. Zn/Fe fractionates equally between olivine, orthopyroxene and melt (e.g. the inter-mineral exchange coefficients ∼ is ∼0.9-1), and the distribution of Zn/Fe between minerals appears to be temperature-independent within error. In contrast, clinopyroxene and garnet are characterized by low Zn/Fe ratios compared to co-existing melt, olivine and orthopyroxene, that is, and are both <<1. These partitioning behaviors imply that Zn/Fe ratios are minimally fractionated during partial melting of peridotite and differentiation of primitive basalts, if differentiation is dominated by olivine control. Thus, the Zn/Fe ratios of primitive basalts preserve the Zn/Fe ratio of the primary parental magma, providing insight into the signature of the mantle source region. We also infer that Zn/Fe ratios in melts are unlikely to be fractionated by modal variations in peridotitic material but are highly fractionated if garnet and/or clinopyroxene are the main phases in the source during melting. Similar Zn/Fe ratios between MORB and average upper mantle confirm the lack of fractionation during peridotite melting. However, high Zn/Fe ratios of some OIB cannot be explained by peridotite melting alone, but instead require the presence of high Zn/Fe lithologies or lithologies that have bulk exchange coefficients < 1. All garnet-bearing or clinopyroxene-bearing lithologies, such as eclogites and garnet pyroxenites, fit the latter requirement. 相似文献
5.
Bjorn O. Mysen 《Geochimica et cosmochimica acta》2007,71(22):5500-5513
The influence on olivine/melt transition metal (Mn, Co, Ni) partitioning of substitution in the tetrahedral network of silicate melt structure has been examined at ambient pressure in the 1450-1550 °C temperature range. Experiments were conducted in the systems NaAlSiO4-Mg2SiO4- SiO2 and CaAl2Si2O8-Mg2SiO4-SiO2 with about 1 wt% each of MnO, CoO, and NiO added. These compositions were used to evaluate how, in silicate melts, substitution and ionization potential of charge-balancing cations affect activity-composition relations in silicate melts and mineral/melt partitioning.The exchange equilibrium coefficient, , is a positive and linear function of melt Al/(Al + Si) at constant degree of melt polymerization, NBO/T. The is negatively correlated with the ionic radius, r, of the M-cation and also with the ionization potential (Z/r2, Z = electrical charge) of the cation that serves to charge-balance Al3+ in tetrahedral coordination in the melts. The activity coefficient ratio, (γM/γMg)melt, is therefore similarly correlated.These melt composition relationships are governed by the distribution of Al3+ among coexisting Q-species in the peralkaline (depolymerized) melts coexisting with olivine. This distribution controls Q-speciation abundance, which, in turn, controls (γM/γMg)melt and . The relations between melt structure and olivine/melt partitioning behavior lead to the suggestion that in natural magmatic systems mineral/melt partition coefficients are more dependent on melt composition and, therefore, melt structure the more alkali-rich and the more felsic the melt. Moreover, mineral/melt partition coefficients are more sensitive to melt composition the more highly charged or the smaller the ionic radius of the cation of interest. 相似文献
6.
The effect of sulfur dissolved as sulfide (S2−) in silicate melts on the activity coefficients of NiO and some other oxides of divalent cations (Ca, Cr, Mn, Fe and Co) has been determined from olivine/melt partitioning experiments at 1400 °C in six melt compositions in the system CaO-MgO-Al2O3-SiO2 (CMAS), and in derivatives of these compositions at 1370 °C, obtained from the six CMAS compositions by substituting Fe for Mg (FeCMAS). Amounts of S2− were varied from zero to sulfide saturation, reaching 4100 μg g−1 S in the most sulfur-rich silicate melt. The sulfide solubilities compare reasonably well with those predicted from the parameterization of the sulfide capacity of silicate melts at 1400 °C of O’Neill and Mavrogenes (2002), although in detail systematic deviations indicate that a more sophisticated model may improve the prediction of sulfide capacities.The results show a barely discernible effect of S2− in the silicate melt on Fe, Co and Ni partition coefficients, and also surprisingly, a tiny but resolvable effect on Ca partitioning, but no detectable effect on Cr, Mn or some other lithophile incompatible elements (Sc, Ti, V, Y, Zr and Hf). Decreasing Mg# of olivine (reflecting increasing FeO in the system) has a significant influence on the partitioning of several of the divalent cations, particularly Ca and Ni. We find a remarkably systematic correlation between and the ionic radius of M2+, where M = Ca, Cr, Mn, Fe, Co or Ni, which is attributable to a simple relationship between size mismatch and excess free energies of mixing in Mg-rich olivine solid solutions.Neither the effect of S2− nor of Mg#ol is large enough by an order of magnitude to account for the reported variations of obtained from electron microprobe analyses of olivine/glass pairs from mid-ocean ridge basalts (MORBs). Comparing these MORB glass analyses with the Ni-MgO systematics of MORB from other studies in the literature, which were obtained using a variety of analytical techniques, shows that these electron microprobe analyses are anomalous. We suggest that the reported variation of with S content in MORB is an analytical artifact.Mass balance of melt and olivine compositions with the starting compositions shows that dissolved S2− depresses the olivine liquidus of haplobasaltic silicate melts by 5.8 × 10−3 (±1.3 × 10−3) K per μg g−1 of S2−, which is negligible in most contexts. We also present data for the partitioning of some incompatible trace elements (Sc, Ti, Y, Zr and Hf) between olivine and melt. The data for Sc and Y confirm previous results showing that and decrease with increasing SiO2 content of the melt. Values of average 0.01 with most falling in the range 0.005-0.015. Zr and Hf are considerably more incompatible than Ti in olivine, with and about 10−3. The ratio / is well constrained at 0.611 ± 0.016. 相似文献
7.
8.
S. M. Eggins 《Contributions to Mineralogy and Petrology》1993,114(1):79-100
Key aspects of magma generation and magma evolution in subduction zones are addressed in a study of Ambae (Aoba) volcano, Vanuatu. Two major lava suites (a low-Ti suite and high-Ti suite) are recognised on the basis of phenocryst mineralogy, geochemistry, and stratigraphy. Phenocryst assemblages in the more primitive low-Ti suite are dominated by magnesian olivine (mg 80 to 93.4) and clinopyroxene (mg 80 to 92), and include accessory Cr-rich spinel (cr 50 to 84). Calcic plagioclase and titanomagnetite are important additional phenocryst phases in the high-Ti suite lavas and the most evolved low-Ti suite lavas. The low-Ti suite lavas span a continuous compositional range, from picritic (up to 20 wt% MgO) to high-alumina basalts (<5 wt% MgO), and are consistent with differentiation involving observed phenocrysts. Melt compositions (aphyric lavas and groundmasses) in the low-Ti suite form a liquid-line of descent which corresponds with the petrographically-determined order of crystallisation: olivine + Cr-spinel, followed by clinopyroxene + olivine + titanomagnetite, and then plagioclase + clinopyroxene + olivine + titanomagnetite. A primary melt for the low-Ti suite has been estimated by correcting the most magnesian melt composition (an aphyric lava with 10.5 wt% MgO) for crystal fractionation, at the oxidising conditions determined from olivine-spinel pairs (fo2 FMQ + 2.5 log units), until in equilibrium with the most magnesian olivine phenocrysts. The resultant composition has 15 wt% MgO and an mg
Fe2 value of 81. It requires deep (3 GPa) melting of the peridotitic mantle wedge at a potential temperature consistent with current estimates for the convecting upper mantle (T
p 1300°C). At least three geochemically-distinct source components are necessary to account for geochemical differences between, and geochemical heterogeneity within, the major lava suites. Two components, one LILE-rich and the other LILE- and LREE-rich, may both derive from the subducting ocean crust, possibly as an aqueous fluid and a silicate melt respeetively. A third component is attributed to either differnt degrees of melting, or extents of incompatible-element depletion, of the peridotitic mantle wedge. 相似文献
9.
We performed a series of experiments at high pressures and temperatures to determine the partitioning of a wide range of trace elements between ilmenite (Ilm), armalcolite (Arm) and anhydrous lunar silicate melt, to constrain geochemical models of the formation of titanium-rich melts in the Moon. Experiments were performed in graphite-lined platinum capsules at pressures and temperatures ranging from 1.1 to 2.3 GPa and 1300-1400 °C using a synthetic Ti-enriched Apollo ‘black glass’ composition in the CaO-FeO-MgO-Al2O3-TiO2-SiO2 system. Ilmenite-melt and armalcolite-melt partition coefficients (D) show highly incompatible values for the rare earth elements (REE) with the light REE more incompatible compared to the heavy REE ( 0.0020 ± 0.0010 to 0.069 ± 0.010 for ilmenite; 0.0048 ± 0.0023 to 0.041 ± 0.008 for armalcolite). D values for the high field strength elements vary from highly incompatible for Th, U and to a lesser extent W (for ilmenite: 0.0013 ± 0.0008, 0.0035 ± 0.0015 and 0.039 ± 0.005, and for armalcolite 0.008 ± 0.003, 0.0048 ± 0.0022 and 0.062 ± 0.03), to mildly incompatible for Nb, Ta, Zr, and Hf (e.g. 0.28 ± 0.05 and : 0.76 ± 0.07). Both minerals fractionate the high field strength elements with DTa/DNb and DHf/DZr between 1.3 and 1.6 for ilmenite and 1.3 and 1.4 for armalcolite. Armalcolite is slightly more efficient at fractionating Hf from W during lunar magma ocean crystallisation, with DHf/DW = 12-13 compared to 6.7-7.5 for ilmenite. The transition metals vary from mildly incompatible to compatible, with the highest compatibilities for Cr in ilmenite (D ∼ 7.5) and V in armalcolite (D ∼ 8.1). D values show no clear variation with pressure in the small range covered.Crystal lattice strain modelling of D values for di-, tri- and tetravalent trace elements shows that in ilmenite, divalent elements prefer to substitute for Fe while armalcolite data suggest REE replacing Mg. Tetravalent cations appear to preferentially substitute for Ti in both minerals, with the exception of Th and U that likely substitute for the larger Fe or Mg cations. Crystal lattice strain modelling is also used to identify and correct for very small (∼0.3 wt.%) melt contamination of trace element concentration determinations in crystals.Our results are used to model the Lu-Hf-Ti concentrations of lunar high-Ti mare basalts. The combination of their subchondritic Lu/Hf ratios and high TiO2 contents requires preferential dissolution of ilmenite or armalcolite from late-stage, lunar magma ocean cumulates into low-Ti partial melts of deeper pyroxene-rich cumulates. 相似文献
10.
We present major and trace element and Sr-Nd-Pb and U-Th-Pa-Ra isotope data for a small sample suite of primarily post-glacial, mildly alkalic volcanic rocks from the Snaefellsjökull central volcano situated off the main rift systems in western Iceland. The volcanic rocks are crystal-poor and range from olivine alkali basalt to trachyte and show tight correlations of major and trace elements that are explained by fractional crystallization involving removal of olivine, plagioclase, clinopyroxene, Fe-Ti oxide and apatite. Sr-Nd-Pb isotopes are practically invariant, consistent with derivation from the same source region. During fractionation from primitive basalt to evolved trachyte, (230Th/232Th), (230Th/238U) and (231Pa/235U) decrease progressively at broadly constant (238U/232Th). A continuous closed-system fractionation model that assumes constant initial (230Th/232Th) in the basaltic precursor melt indicates that hawaiite was derived from olivine basalt by ∼50% fractional crystallization within and trachyte by ∼80% fractionation within . An overrepresentation of evolved basalts and hawaiites with young inferred magma ages in the dataset is consistent with the parental precursor to these magmas intruded into the sub-volcanic magma plumbing system as a consequence of lithospheric rebound caused by deglaciation. Lavas affected by apatite removal have higher (231Pa/235U) than predicted for simple radioactive decay, suggesting apatite significantly fractionates U from Pa. The proposed fractionation model consistently explains our U-series data assuming and and . If applicable, these D values would indicate that the effect of apatite fractionation must be adequately considered when assessing differentiation time scales using (231Pa/235U) disequilibria data. 相似文献
11.
Kevin Grant Jannick Ingrin Jean Pierre Lorand Paul Dumas 《Contributions to Mineralogy and Petrology》2007,154(1):15-34
The speciation and amount of water dissolved in nominally anhydrous silicates comprising eight different mantle xenoliths
has been quantified using synchrotron micro-FTIR spectroscopy. Samples studied are from six geographic localities and represent
a cross-section of the major upper mantle lithologies from a variety of tectonic settings. Clinopyroxene contains between
342 and 413 ppm H2O. Orthopyroxene, olivine and garnet contain 169–201, 3–54 and 0 to <3 ppm H2O, respectively. Pyroxenes water contents and the distribution of water between ortho- and clinopyroxene is identical regardless
of sample mineralogy (D
watercpx/opx = 2.1 ± 0.1). The total water contents of each xenolith are remarkably similar (113 ± 14 ppm H2O). High-resolution spectroscopic traverses show that the concentration and speciation of hydrous defects dissolved in each
phase are spatially homogeneous within individual crystals and identical in different crystals interspersed throughout the
xenolith. These results suggest that the amount of water dissolved in the silicate phases is in partial equilibrium with the
transporting melt. Other features indicate that xenoliths have also preserved OH signatures of equilibrium with the mantle
source region: Hydroxyl stretching modes in clinopyroxene show that garnet lherzolites re-equilibrated under more reducing
conditions than spinel lherzolites. The distribution of water between pyroxenes and olivine differs according to xenolith
mineralogy. The distribution of water between clinopyroxene and olivine from garnet peridotites (D
watercpx/oliv(gnt) = 22.2 ± 24.1) is a factor of four greater than mineral pairs from spinel-bearing xenoliths (D
watercpx/oliv(sp) = 88.1 ± 47.8). Such an increase in olivine water contents at the spinel to garnet transition is likely a global phenomenon
and this discontinuity could lead to a reduction of the upper mantle viscosity by 0.2–0.7 log units and a reduction of its
electrical resistivity by a factor of 0.5–0.8 log units. 相似文献
12.
Patches of glass with a second generation of small crystals of olivine, clinopyroxene, and spinel are abundant in hydrous peridotite mantle xenoliths with tabular equigranular textures from two maar-type volcanoes, Meerfelder Maar and Dreiser Weiher (West Eifel, Germany). The patches are similar in size to the main phases of the hosting peridotite. Their central part is often occupied by relics of pargasitic amphibole. Mass-balance calculations show that the patches were formed by surface controlled incongruent thermal breakdown of amphibole according to the reaction: amphibole olivine + clinopyroxene + spinel + melt. Simultaneously with the decomposition of amphibole, small crystals of olivine, clinopyroxene, and spinel grew radially from the patch/peridotite interface toward the centre of the patch. Apart from sector zoning of clinopyroxene, the crystals are virtually homogeneous and are separated from the amphibole by a seam of melt (glass). Secondary olivines reveal higher Mg-numbers, secondary clinopyroxenes higher Cr2O3 concentrations than olivines and clinopyroxenes, respectively, of the host peridotite. The silica contents of melts produced by the above breakdown reaction range from 48 to 52% SiO2 as a function of the composition of the parent amphiboles. Patches surrounded by primary olivines only reveal no reaction with the host peridotite. The variation of SiO2, MgO and CaO in melts from these patches is the result of minor precipitation of olivine and clinopyroxene during fast cooling. If patches are in contact with primary olivine and orthopyroxene, melts are additionally modified by the reaction liquid 1 + orthopyroxene liquid 2 + olivine + clinopyroxene resulting in more silica-rich compositions between 54 and 58%. For the rare glasses richer in silica, a more complex formation is required. Veinlets along grain boundaries are filled with glasses which are chemically identical to those from nearby patches. This suggests that the veinlets were filled by melts formed by amphibole breakdown during entrainment of the xenoliths to the host magmas. 相似文献
13.
Vladica Cvetkovi? Suzana Eri? Ma?a Radivojevi? Kristina ?ari? 《Mineralogy and Petrology》2012,106(3-4):131-150
The study focuses on clinopyroxene from mantle xenolith-bearing East Serbian basanites and suggests that dissolution of mantle orthopyroxene played an important role in at least some stages of the crystallization of these alkaline magmas. Five compositional types of clinopyroxene are distinguished, some of them having different textural forms: megacrysts (Type-A), green/colourless-cored phenocrysts (Type-B), overgrowths and sieve-textured cores (Type-C), rims and matrix clinopyroxene (Type-D), and clinopyroxene from the reaction rims around orthopyroxene xenocrysts (Type-E). Type-A is high-Al diopside that probably crystallized at near-liquidus conditions either directly from the host basanite or from compositionally similar magmas in previous magmatic episodes. Type-B cores show high VIAl/IVAl≥1 and low Mg# of mostly <75 and are interpreted as typical xenocrysts. Type-C, D and E are interpreted as typical cognate clinopyroxene. Type-D has Mg#<78, Al2O3?=?6–13?wt.%, TiO2?=?1.5–4.5?wt.%, and Na2O?=?0.4–0.8?wt.% and compositionally similar clinopyroxene is calculated by MELTS as a phase in equilibrium with the last 30?% of melt starting from the average host lava composition. Type-C has Mg#?=?72–89, Al2O3?=?4.5–9.5?wt.%, TiO2?=?1–2.5?wt.%, Na2O?=?0.35–1?wt.% and Cr2O3?=?0.1–1.5?wt.%. This clinopyroxene has some compositional similarities to Type-E occurring exclusively around mantle orthopyroxene. Cr/Al vs Al/Ti and Cr/Al vs Na/Ti plots revealed that Type-C clinopyroxene can crystallize from a mixture of the host basanite magma and 2–20?wt.% mantle orthopyroxene. Sieve-textured Type-C crystals show characteristics of experimentally produced skeletal clinopyroxene formed by orthopyroxene dissolution suggesting that crystallization of Type-C was both texturally and compositionally controlled by orthopyroxene breakdown. According to FeO/MgOcpx/melt modelling the first clinopyroxene precipitating from the host basanite was Type-A (T?~?1250?°C, p?~?1.5?GPa). Dissolution of orthopyroxene produced decreasing FeO/MgOmelt and crystallization of Type-E and sieve-textured Type-C clinopyroxene (0.3–0.8?GPa and 1200–1050?°C). The melt composition gradually shifted towards higher FeO/MgOmelt ratios precipitating more evolved Type-C and Type-D approaching near-solidus conditions (<0.3?GPa; ~950?°C). 相似文献
14.
Water diffusion in silicate melts is important for understanding bubble growth in magma, magma degassing and eruption dynamics of volcanos. Previous studies have made significant progress on water diffusion in silicate melts, especially rhyolitic melt. However, the pressure dependence of H2O diffusion is not constrained satisfactorily. We investigated H2O diffusion in rhyolitic melt at 0.95–1.9 GPa and 407–1629 °C, and 0.2–5.2 wt.% total water (H2Ot) content with the diffusion-couple method in a piston-cylinder apparatus. Compared to previous data at 0.1–500 MPa, H2O diffusivity is smaller at higher pressures, indicating a negative pressure effect. This pressure effect is more pronounced at low temperatures. Assuming H2O diffusion in rhyolitic melt is controlled by the mobility of molecular H2O (H2Om), the diffusivity of H2Om (DH2Om) at H2Ot ≤ 7.7 wt.%, 403–1629 °C, and ≤ 1.9 GPa is given by
DH2Om=D0exp(aX),