Donald G. Fraser

WATER CHEMISORPTION AND RECONSTRUCTION OF THE MGO SURFACE

Cavendish Laboratory, University of Cambridge, Madingley Road, Cambridge CB3 0HE(December 6, 1994)The observed reactivity of MgO with water is in apparent conflict with theoretical calculationswhich show that molecular dissociation does not occur on a perfect (001) surface. We have per-formed ab-initio total energy calculations which show that a chemisorption reaction involving areconstruction to form a (111) hydroxyl surface is strongly preferred with ∆E = −90.2 kJ mol

Pb, Sr and 10Be isotopic studies of volcanic rocks from the Northern Mariana Islands. Implications for magma genesis and crustal recycling in the Western Pacific

We report Sr and Pb isotope analyses for an extensive suite of volcanic rocks from the N. Mariana arc together with Sr and Pb isotope analyses of sediments from the nearby Mariana and Nauru basins. In addition ten of the most recent volcanic samples were analysed for 10Be. The Sr isotope compositions cluster tightly around 87Sr86Sr = 0.7035 being slightly but significantly higher than the Pacific ocean floor basalts on either side of the arc and agreeing well with previous data. In contrast, the large number of new Pb isotopic data presented significantly extends the observed range of Pb isotope compositions for volcanic rocks from the Mariana arc to more radiogenic compositions. The concentrations of 10Be were very low (< 0.5 × 106 atom g−1). These new data require either that the Pb and Sr isotopic compositions of the Mariana sub-arc mantle be substantially different from those of the mantle source of ocean floor basalts on either side of the arc, or that the enrichment in radiogenic Pb and Sr relative to the values observed in Pacific ocean floor basalts be related to the subduction process. We prefer the latter hypothesis in which radiogenic Sr and Pb in ocean floor sediments are added to M.O.R.B. type mantle either by direct assimilation of the sediments in partial melts or, more probably, by transfer in a fluid phase into the zone of magma production. The low 10Be concentrations observed suggest the removal of at least the top few metres of sediment during subduction.

Zonation patterns of skarn garnets: Records of hydrothermal system evolution

Chemically zoned skarn garnets provide a continuous record of hydrothermal processes in lower Paleozoic sedimentary rocks within the contact aureole around the Drammen granite in the Oslo rift, southern Norway. Major and trace element zonation profiles, the latter obtained using a scanning high-resolution proton microprobe, reveal early infiltration-controlled growth of relatively grossular rich garnets, the major and trace elements compositions being buffered by local mineral dissolution. Subsequent rapid, epitaxial growth of andradite-rich garnet on grossular-rich cores marks the onset of vigorous and focused fluid flow along high-permeability zones. During this later stage, the hydrothermal fluid composition was to a large extent externally controlled, and the andradite precipitating from these fluids was characterized by high As and W contents. The zonation patterns of the andradite-rich garnets record at least five intermittent growth periods, with rapid andradite precipitation from fluid batches with high f O 2 , and progressively decreasing As and W contents. Thin layers, poor in Fe, As, and W, but relatively high in Al and Mn, represent periods of slow growth rates between the major pulses of hydrothermal fluids. The marked rimward decrease in the As and W contents of the garnets may reflect influx of meteoric waters or exhaustion of these elements in the hydrothermal fluid reservoir caused by boiling-controlled distillation processes at depth.

O, S, Sr, and Pb isotope variations in volcanic rocks from the Northern Mariana Islands: implications for crustal recycling in intra-oceanic arcs

Quaternary lavas from the Northern Mariana Islands have respective O- and S-isotope ranges ofδ18O = +5.7 to +6.6 (‰ SMOW) andδ34S = +2.0to+20.7 (‰ CDT). Chemically evolved andesites and dacites with meanδ18O = +6.3 ± 0.2 are slightly enriched in18O with respect to unfractionated basalts of<53%SiO2 with meanδ18O = +6.0 ± 0.1. This18O enrichment can be explained in terms of differentiation of parental mafic magmas havingδ18O values between +5.7 to +6.2‰ through closed-system crystal fractionation because the lavas from all nine islands of the arc define a coherentδ18OSiO2 trend. The S-isotope composition of oxidized magmas is not modified extensively through the degassing of SO2; therefore, the meanδ34S value of ca. +11‰ for the Mariana lavas is considered to be representative of their source region. The enrichment of18O and34S in Mariana Arc parental magmas relative to ocean floor basalts withδ18O ca. + 5.7‰ andδ34S = ca.0.3‰ is attributed to the recycling of18O- and34S-rich crustal components (sediment withδ18O = ca. + 25‰ and seawater sulfate withδ34S = ca. +20‰ into the upper mantle source region for these arc magmas. This interpretation is consistent with enrichments of radiogenic Sr and Pb in the same lavas relative to ocean-floor basalts erupted either side of the arc, which are presumed to share a common upper mantle source. This enrichment is considered to reflect the mixing of two components, one having a typical upper mantle composition and the other having a more radiogenic character similar to that of western Pacific pelagic sediments.

The kinetics and mechanism of MgO dissolution

Abstract Reactions and atomic rearrangements at fluid–crystal interfaces play an important role in catalysis and in controlling the kinetics and mechanisms of dissolution. We have studied the attachment and reactions of water molecules at the MgO–water interface by combining measurements of 1 H and 2 D surface penetration and etch pit morphology with ab initio calculations. These studies show that the most common MgO cleavage surface, (001), is thermodynamically unstable when hydrated. Proton rearrangement on such surfaces precedes proton–cation exchange and provides a general mechanism for the detachment of ions during dissolution. The kinetics of dissolution are strongly influenced by the concentration of surface defects and a simple model based on the ab initio results predicts a dissolution rate of 10−10 mol cm−2 s−1 for a typical surface defect concentration of 0.1.

Thermodynamic Properties of Silicate Melts

One of the major current problems in experimental igneous petrology concerns the influence of additional components on liquidus boundaries observed in simple synthetic systems or in given natural rock compositions. The petrological significance of such effects may be substantial. For example, the addition of H20 to the anhydrous system diopside-forsterite-silica causes both the forsterite and enstatite primary phase volumes to move to more silica-rich compositions (Kushiro, 1972). Adding C02 has the opposite effect (Eggler, 1974) as has increasing total pressure (Kushiro, 1969). These results are summarized in Fig. 1.

Periclase surface hydroxylation during dissolution

Abstract Periclase (001) surfaces were etched in dilute acid at pH 2 and 4. X-ray reflectivity measurements on a reference crystal constrained the initial roughness of these surfaces to be approximately 30 A. The reference crystal and the crystal reacted at pH 2 were analyzed by Elastic Recoil Detection Analysis (ERDA) for proton penetration. After reaction the etched sample showed proton penetration to a depth of at least 5000 A while the reference crystal showed no significant proton inventory. Down to 900 A, the H Mg ratio in the etched sample was approximately 2, consistent with near-surface protonation of the MgO to form hydroxylated brucite-like layers. Protonation is a far more likely mechanism to explain the proton profile than precipitation because this reaction was completed over 16 orders of magnitude below saturation with brucite. Formation of a hydroxylated near-surface layer on periclase during dissolution explains why the dissolution rates of periclase and brucite are identical in the pH range 2–5; the detachment rates are the same because the surface structures are the same. This suggests that even for this ionic solid in acid, the dissolution reaction involves a two-step mechanism with a rapid single protonation step of near-surface oxygen atoms and a slower, rate determining second protonation step. In general, product phases such as brucite are likely to be better developed under natural weathering conditions of near-neutral pH because the second step of protonation (and thus full hydration of the detaching cation, e.g., Mg+2) is much slower than in acid. Our proposed protonation mechanism relates field observations of the periclase weathering reaction to laboratory dissolution, hydration, and dehydration experiments.

Activities of trace elements in silicate melts

Abstract A theoretical model has been developed which describes the amphoteric character of oxides in silicate melts. This has been used to account for the increased stability of the higher oxidation states of altervalent trace elements in silicate melts with increasing basicity and to derive a general expression for the estimation of trace element activities in silicate melts.

Determination of activities in silicate melts by Knudsen cell mass spectrometry—I. The system NaAlSi3O8-KAlSi3O8

The mixing properties of aluminosilicate melts in the pseudobinary system NaAlSi3O8-KAlSi 3O8 have been determined by measuring the compositions of their saturated vapours by hightemperature Knudsen cell mass spectrometry. The melts mix close to ideally over most of the composition range with small positive deviations from ideality for K-rich compositions. These may be related to incipient partial ordering of melt constituents into leucite-like and SiO2-like structures above the feldspar liquidus.

Activity measurements by Knudsen cell mass spectrometry—the system Fe-Co-Ni and implications for condensation processes in the solar nebula

Abstract We report the use of new apparatus for the direct determination of the activities of components in minerals and melts by the measurement of their intrinsic vapour pressures at high temperatures using a Knudsen-cell mass spectrometer combination. The activity coefficients of Fe, Ni and Co have been determined using this technique in binary and ternary Fe-Co-Ni alloys over the temperature range 1200–1650°C. The results show negative deviations from ideality in the Fe-Ni system and both positive and negative deviations in Fe-Co. Activity coefficients determined in the ternary system were used to calculate the compositions of Fe-Ni-Co alloys in equilibrium with a gas of chondritic composition and the results are consistent with the compositions of metallic particles in Ca, Al rich inclusions in chondritic meteorites.