| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology Volume 42 Number 10 Pages 1789-1811 2001
© Oxford University Press 2001
Melting Systematics of Modally Variable, Compositionally Intermediate Peridotites and the Effects of Mineral Fertility
DEPARTMENT OF GEOLOGICAL SCIENCES, 1272 UNIVERSITY OF OREGON, EUGENE, OR 97403-1272, USA
Received February 11, 2000; Revised typescript accepted February 20, 2001
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Piston-cylinder experiments were performed at 10 kbar to investigate the near-solidus partial melting systematics of modally variable peridotites. Starting materials consisted of compositionally intermediate (i.e. containing moderate incompatible element abundances) minerals, separated from a spinel lherzolite xenolith from Mt. Noorat, SE Australia, and recombined to create five starting mixtures varying in their proportions of olivine, orthopyroxene, clinopyroxene (Cpx), and spinel. These modes match those of starting materials made with fertile (FER) minerals from a different xenolith, investigated in a companion study. A layer of vitreous carbon spheres provided a melt sink in the experiments. Solidus temperatures for the five peridotites are similar and estimated to be
1260 ± 10°C based on the zero-F (melt fraction) intercepts of F vs temperature curves. Melt productivity (
F/T) varies significantly between the starting materials, but unlike the FER runs, the overall melt productivity does not correlate with the initial Cpx abundance. This suggests that the compositional fertility of Cpx plays a greater role in determining F at a given temperature than simply its overall source abundance. Low-F glasses have elevated SiO2 and Na2O contents relative to average basalts, as high as 53·9 and 2·5 wt %, respectively, but are less enriched than in equivalent FER melts. KEY WORDS: experimental petrology; mantle melting; melt sink; partial melt; peridotite
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Mid-ocean ridge basalt (MORB) is the most voluminous rock type on the Earth’s surface. The well-accepted model for MORB generation involves low-pressure fractionation of magmas generated by polybaric melting of adiabatically upwelling mantle beneath mid-ocean spreading centers. Because the slope of the mantle adiabat is steeper than that of the mantle solidus (McKenzie, 1984
; McKenzie & Bickle, 1988), a rising parcel of hot peridotite will cross its solidus and begin to melt. The resulting melt is buoyant relative to residual solid phases and can rise faster than the upwelling solids once an interconnected pathway develops. Isotopic and trace element evidence in MORBs and abyssal peridotites suggests that this can occur at melt fractions (F) as low as 0·001 (i.e. 0·1%) (Salters & Hart, 1989; Johnson et al., 1990). These melts aggregate throughout the melting column and are eventually emplaced as oceanic crust, with the solid residuum forming abyssal peridotite. Although the processes occurring between the onset of melting and the ultimate eruption of MORBs are too complex to simulate experimentally, data generated by equilibrium batch partial melting experiments, like those presented here, are integral to any attempt to model the process mathematically (e.g. Klein & Langmuir, 1987; McKenzie & Bickle, 1988; McKenzie, 1989; Niu & Batiza, 1991; Kinzler & Grove, 1992a, 1992b; Langmuir et al., 1992; Ghiorso et al., 1994; Ghiorso & Sack, 1995).
Existing models of mantle melting (e.g. references above) predict significant variability in the compositions and amounts of melt generated from source regions with different bulk compositions. This reflects the modal heterogeneity of the peridotite sources and their variable concentrations of incompatible components that would partition preferentially into a coexisting basaltic liquid. We use the term fertility here in reference to the concentration of these components, with fertile (FER) peridotites containing greater abundances of them than intermediate (INT) or depleted peridotites. Together, fertility and mode affect melt productivity by controlling the depth at which melting begins and the rate of melt production with further ascent. Klein & Langmuir (1987) first noted trends in MORB compositions that correlate with axial depth and inferred extent of melting. They reasoned that fertile mantle, with its lower-temperature solidus, begins melting deeper and melts to a larger ultimate extent than depleted mantle with its higher-temperature solidus.
To investigate the effects of modal heterogeneity, we performed 10 kbar melting experiments using compositionally intermediate peridotitic assemblages, varying the relative abundances of orthopyroxene, clinopyroxene, and spinel, with a constant amount (50 wt %) of olivine. To address the effects of mineral fertility between peridotites of identical initial mode, we compare our data with those of the companion project of Pickering-Witter & Johnston (2000), along with the MM3 composition of Baker & Stolper (1994) and Hirschmann et al. (1998), which used identical modes, but compositionally fertile minerals. Thus, we can directly compare five different starting materials with identical initial mineral modes, but varying in fertility. We recognize that our starting mineral compositions, and those of Pickering-Witter & Johnston (2000) and Baker & Stolper (1994) probably differ, as a result of subsolidus re-equilibration, from those at the 10 kbar solidus of the studied compositions. Thus, we expect, and show, that the starting materials re-equilibrate as the solidus is approached with consequent small shifts in the modes relative to the initial mixtures. As a result, our comparison of findings for any particular initial mode in this study with the identical initial mode from the study of Baker & Stolper (1994) or Pickering-Witter & Johnston (2000) will not be perfect, as they probably re-equilibrate to differing degrees. Nonetheless, we believe that such comparisons can still yield valuable insight into the effect of mineral fertility on the melting behavior of peridotites with similar modes at near-solidus conditions.
Although peridotite partial melting experiments are conceptually simple, experimental difficulties, principally Fe loss to noble metal capsules and quench modification of melt compositions, have hampered the collection of high-quality chemical composition data, particularly for the melt phase, for decades. These problems are most severe for low-F melts, which is regrettable, as it is these melts that are thought to contribute to the aggregate melts that ultimately erupt as MORBs (e.g. Waff & Bulau, 1979; Cooper & Kohlstedt, 1986; von Bargen & Waff, 1986; Daines & Richter, 1988; Langmuir et al., 1992). Baker & Stolper (1994) and Pickering-Witter & Johnston (2000) devoted substantial discussion to these and other experimental problems and the historical development of techniques to circumvent them, and the reader is referred to these sources for further details.
In this study, we used a variation of the diamond aggregate technique developed by Baker & Stolper (1994) and Kushiro and co-workers (e.g. Hirose & Kushiro, 1993), utilizing vitreous carbon (vc) spheres in lieu of diamonds (Wasylenki et al., 1996). In short, a layer of vc spheres (80–100 µm diameter) was placed above the finely ground peridotite within a graphite-lined Pt capsule. The inner graphite capsule eliminates Fe loss to the Pt outer capsule, and the porosity within the layer of vc spheres provides a ‘sink’ for melt that physically separates it from the peridotite residuum, eliminating the development of quench overgrowths on mineral grains.
|
EXPERIMENTAL AND ANALYTICAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Starting materials and bulk compositions
Starting materials for the experiments consist of hand-picked mineral separates of olivine (Ol), orthopyroxene (Opx), clinopyroxene (Cpx), and spinel (Sp) from a compositionally intermediate [Cpx: Na2O 0·53 wt %, Al2O3 3·10 wt %, CaO 21·96 wt %; Ol: mg-number 90·9; i.e. intermediate to modeled ‘primitive’ upper mantle (e.g. Ringwood, 1979
; Hart & Zindler, 1986; McDonough & Frey, 1989) and ‘depleted’ mantle compositions (e.g. Sen, 1982; Wasylenki, 1999)], spinel lherzolite xenolith collected from Mt. Noorat, SE Australia. Electron microprobe analyses of the starting mineral compositions are given in Table 1 and were acquired using the same analytical procedures as described below for experimental run products. Xenolith mineral grains were picked under a binocular microscope and those with visible inclusions were rejected. The mineral separates were washed in warm, dilute HCl and rinsed with distilled water. The coarse grains were crushed and repicked to ensure selection of inclusion-free grains. The separates were then ground individually under ethanol in an agate mortar and pestle to fine powders (
40 µm) and stored in a drying oven at 110°C.
|
Starting materials are labeled ‘INT’, indicating the intermediate composition xenolith from which they are constructed, plus the letter A, B, C, D, or E, signifying which mineral mix they represent. The ratios of different minerals were not necessarily chosen to represent plausible mantle rocks, but were chosen to exaggerate the effects that an individual mineral has on the composition of partial melts derived from the rock and to match the initial modes studied by Pickering-Witter & Johnston (2000) using fertile minerals. The individual mineral powders were mixed in five different proportions to form the starting materials for the experiments. The modal mixes (Ol:Opx:Cpx:Sp) consisted of: (1) a ‘typical’ spinel lherzolite assemblage equivalent modally to MM3 of Baker & Stolper (1994) and FER-A of Pickering-Witter & Johnston (2000), INT-A (0·50:0·30:0·17:0·03); (2) a Cpx-enriched mode, INT-B (0·50:0·07:0·40:0·03); (3) a harzburgitic mode, INT-C (0·50:0·46:0·01:0·03); (4) a spinel-enriched mode, INT-D (0·50:0·30:0·10:0·10); (5) an equal-pyroxene mode, INT-E (0·50:0·235:0·235:0·03). The mineral mixes were thoroughly ground and mixed under ethanol in an agate mortar and pestle to fine powders (10 µm with rare grains up to 30 µm) and stored in an oven at 110°C.
The initial modes of the starting materials are given in Table 2, as are the bulk compositions calculated from the starting mineral compositions. Table 2 also contains the bulk compositions for MM3 (Baker & Stolper, 1994) and FER-B, -C, -D, and -E of Pickering-Witter & Johnston (2000) for reference. The MM3 and FER- starting materials were mixed from the same compositionally fertile xenolith from Kilbourne Hole, New Mexico. The fertile mixes of MM3, FER-B, -C, -D, and -E are modal equivalents to INT-A, -B, -C, -D, and -E, respectively, and thus provide a means of comparison of the effects of fertility on constant mode peridotites, but it should be recalled that some, and probably unequal, modal readjustments are expected as the solidus temperature is approached at the 10 kbar experimental pressure.
|
Experimental methods
Experiments were performed in 1/2 inch (1·27 cm) end-loaded solid-medium piston-cylinder apparatus with CaF2-based furnace assemblies as described by Pickering et al. (1998). Considerable care was exercised to ensure that the thermocouple and capsule were located within the furnace hot spot, defined as the region within which temperature does not vary by more than 10°C from the peak temperature. Sample position was carefully measured before and after each experiment (Pickering et al., 1998).
Sample capsules consisted of graphite-lined Pt ‘ash cans’ (Sneeringer & Watson, 1985) 3·5 mm in diameter and 5 mm tall. As noted earlier, a layer of vitreous carbon spheres (80–100 µm diameter) was used as a melt sink (e.g. Wasylenki et al., 1996), replacing the layer of diamonds used by Baker & Stolper (1994) and Hirose & Kushiro (1993). The strength of the spheres results in a porosity at the 10 kbar run pressure, and wetting properties enhance the ability of the melt to enter the space between spheres. The melt quenches to glass rinds on the spheres, which are several to tens of microns thick and easily analyzable by electron microprobe. The melt in the sphere layer is far enough away from any quench growth or residual crystals to be unmodified. Figure 1 is a backscattered electron (BSE) image of a dissected experimental run illustrating the glass rinds around the vc spheres. The glassy spheres are easily polished, which makes them a preferred alternative to diamonds. They are also reported (L. E. Wasylenki, unpublished data, 1993; Robinson et al., 1998) to sequester any moisture that might be present in the charge.
|
Starting materials were loaded on top of the sphere layer, covered with a small diameter graphite lid, placed inside the Pt capsule, and held in an oven at 110°C for several hours. The hot capsules were then covered with a graphite disk to ensure isolation from the Pt lid that was welded on top while the capsule was still hot. The sealed capsules were run with the vc sphere layer above the peridotite. Water contents of the starting minerals and run products were not measured and we can only assume starting materials were nominally anhydrous. Small quantities of H2O, P2O5, Cl, F, etc., if present, could lower the solidus temperature of the starting mixes slightly.
The sphere layer constitutes 10–30 wt % of the sample. It is difficult to estimate the initial porosity among the spheres for each run. Slight variations in the cavity volume of the graphite capsule, overall depth of the sphere layer, and initial packing geometry create variability in initial porosity from run to run. Additionally, intrusion of graphite and peridotite into the pore space, deformation of spheres, and slight variations in sphere size and distribution decrease the initial inter-sphere porosity. As a result, the porosity available for melt influx is considered to be less than the melt fraction for nearly all experiments, with possible exceptions being the lowest melt fraction runs.
All runs were performed at 10 kbar (1 GPa) and were maintained to within 100 bar using nominal Heise gauge values with no correction applied for possible friction effects. The furnace assembly was pressurized to 10 kbar at room temperature before power was applied and the assembly was heated to the target temperature. Thermal expansion resulted in the final pressure adjustment being a decrease in line pressure for all experiments (hot, piston-out). Temperature was maintained to within ±5°C of the set point using Eurotherm 808 temperature controllers and W–Re5/W–Re26 thermocouples relative to an Omega electronic ice point (0°C). No pressure correction was applied to thermocouple e.m.f. On the basis of previous work (Pickering et al., 1998), we estimate the thermal gradient to be <10°C over the entire sample capsule. Experimental temperatures with each starting material ranged from 1270°C to 1390°C, which includes near-solidus melting as well as melt fractions up to 28 wt %, depending on the starting mix, with supplementary runs increasing coverage. Run durations varied between 48 and 166 h and were chosen based on previous work in the literature as well as run times used in our laboratory (e.g. Pickering-Witter & Johnston, 2000). Run conditions are reported in Table 3.
|
Power consumption for the experiments was monitored and increased at most 1% during the first 18 h and then remained stable. Thermocouples were inspected after each experiment, and oxidation was not observed. Together these observations indicate that thermocouple drift was minimal.
Although oxygen fugacity was not measured for these experiments, the graphite capsule and furnace tend to buffer the oxygen fugacity below the graphite–C–O equilibrium boundary in the wüstite stability field for these experiments (Taylor & Green, 1989; Ulmer & Luth, 1991).
When power to the experiment is turned off, the rapid quench rates (200°C/s) of the piston-cylinder apparatus allow for the preservation of glass free of quench crystals around the vc spheres in the capsules. After quench, the furnace assemblies were dissected and the capsule position and assembly parts were measured. All experiments maintained the capsule position within the 10°C thermal hot spot of our furnace assembly (Pickering et al., 1998). Run products were then mounted in epoxy, sawn in half, and polished for analysis. Polished halves have the advantage of preserving the spatial distribution of spheres, melt quenched to glass, and residual crystalline phases. The polished samples were then carbon coated in preparation for microprobe analysis.
Analytical techniques
Chemical analysis and imaging of the experimental run products were performed at the electron microbeam facility at the University of Oregon. The phase compositions were obtained using a CAMECA SX-50 electron microprobe utilizing wavelength-dispersive X-ray detectors and PAP data reduction algorithms (e.g. Pouchou & Pichoir, 1991) in the xMAS software package. Natural and synthetic minerals and glasses were used as standards. The microprobe was operated in fixed mode with a 15 keV acceleration potential, a 20 nA beam current and 1 µm spot size for crystalline phases, and a 10 nA beam current for glass. Glass analyses were performed with as large a beam diameter as possible. Owing to the limited thickness of the glass rinds and pockets, beam diameters were typically 3–7 µm, occasionally 2 and up to 10 µm, with a 1 µm spot necessary on rare occasions.
To estimate the extent of Na loss during glass analyses, we analyzed INT-B6 glass for Na2O, SiO2, Al2O3, and K2O using a 2 nA beam current and a 2 µm spot multiple times for durations of 2, 5, 10, 20, 40, and 60 s at a new location for each analysis. A linear least-squares fit to the weight percent Na2O data plotted vs analysis duration yields a zero-time intercept of 1·692 wt % Na2O. The original glass analyses yield an average value of 1·597 wt % or a difference of 5·6% relative. This difference is our best estimate for the relative Na2O loss for our glass analyses. This +5·6% relative correction was applied to all glass analyses as a Na2O correction factor. There was no significant change in measured K2O content, nor was significant SiO2 or Al2O3 ‘grow-in’ observed (e.g. Morgan & London, 1996). Because large beam sizes (20 µm) were not feasible and Na loss increases with decreasing beam size (e.g. Morgan & London, 1996), our estimate represents the minimum loss associated with the analysis and therefore the reported Na2O values are minimum amounts.
Attainment of equilibrium
The use of the diamond aggregate technique and the variation of it used here has been a source of controversy. In particular, the high SiO2 and Na2O contents of low-F glasses have been debated in the literature (e.g. Baker & Stolper, 1994; Baker et al., 1996; Falloon et al., 1996, 1997, 1999; Hirschmann et al., 1998). Time series, reversal, and two-stage experiments by Baker & Stolper (1994), half-reversal and convergence experiments by Wasylenki (Wasylenki et al., 1996; Wasylenki, 1999), and two-stage experiments by Pickering-Witter & Johnston (2000) in our laboratory demonstrate that the experimental technique produces high-quality, low melt fraction, equilibrium melt compositions. Additionally, systematic variations in phase proportions and compositions with rising temperature in our experiments and identical results (within uncertainty) for repeated experiments (e.g. INT-E1 and -E2, both at 1300°C) support the conclusion that we can achieve near-equilibrium partial melt compositions in our runs.
Olivine grains in our study are unzoned and are compositionally homogeneous throughout the charge. Pyroxene grains, however, often display compositional differences between cores and rims, and unreacted cores of especially large (
20 µm) pyroxene grains are common even in the longest duration runs. Compositional gradients are not unique to natural starting materials and are observed in experiments utilizing oxide mixes for starting materials as well (e.g. Walter & Presnall, 1994; Robinson et al., 1998). Typical ranges in composition are exhibited in the INT-E runs. Cpx core compositions for INT-E1 (1300°C) were richer in CaO (20·07 wt %) compared with the average rim compositions reported in Table 4 (16·33 wt %). FeO and MgO concentrations were lower in the core, leading to a small decrease in mg-number [100Mg/(Mg + Fe), molar] from 92·1 to 91·3 from core to rim. Opx core CaO concentrations are lower than average rim concentrations (0·97 vs 2·50 wt %) and MgO and FeO decrease slightly from core to rim, but maintain constant mg-numbers (91·7). Estimates of core and rim volumes on the largest pyroxenes were made from BSE images and core volumes are estimated to be <10% even for the lowest temperature runs. Our goal was to analyze equilibrium rim compositions of the crystalline phases in our run products, but the compositional gradients between cores and rims of the pyroxenes in most charges necessitated the application of additional selection criteria.
|
Because the starting pyroxenes differed in composition from those on the solidus at 10 kbar we performed a flux-enhanced melting experiment to obtain a better estimate of the compositional starting point of our pyroxenes. This experiment was run at 10 kbar and 1235°C with a mixture of Opx and Cpx (49·0 and 50·0 wt %, respectively) doped with 1 wt % PbO (INT-PbO, Table 4). The PbO acts as a flux, lowering the solidus temperature of the pyroxene mixture. The resulting melt allows for relatively rapid equilibration of the pyroxenes by providing a pathway for diffusion. Although not truly subsolidus, the analyses of the pyroxene rims from the PbO-doped experiment are interpreted to represent the near-equilibrium values very near the 10 kbar solidus and, as expected, they differ somewhat from the composition of the pyroxene grains hand picked from the Mt. Noorat xenolith. Figure 2 is a plot of Al vs Ca cations (per six oxygens) showing the compositional shift of the starting pyroxenes as the solidus is approached. Starting Opx adjusts from higher Al and lower Ca, to lower Al and higher Ca content near the solidus, whereas Cpx compositions adjust in the opposite sense. Because no mechanism exists to buffer pyroxene Al content in the PbO-doped run, considerable scatter in Al content persists and Ca contents provided the primary selection criterion. Thus, near-equilibrium Cpx compositions were presumed to be those approaching the low-Ca end of the array, whereas Opx analyses from the high-Ca end of the array were selected to represent equilibrium compositions. Table 4 reports Cpx and Opx compositions averaged from the low- and high-Ca ends of the data array, respectively.
|
, multiple analyses of hand-picked pyroxenes. PbO–Px values are from a subsolidus run (10 kbar, 1235°C) of 49:50 wt % of Opx:Cpx mixed with 
, multiple core analyses of zoned PbO–pyroxenes. 
, multiple analyses of equilibrated pyroxene rims (
, average rim value), illustrating the change in Ca and Al values with equilibration. Equilibrium Opx are interpreted to approach higher Ca values whereas equilibrium Cpx approach lower Ca values. Significant spread in pyroxene Al content persists.
We used the Brey two-pyroxene thermometer (Brey & Köhler, 1990; Brey et al., 1990) for run products containing both pyroxenes as a secondary test of pyroxene composition selection. Pyroxene analyses were sorted in order of increasing Ca content. Cation values were then used to find the minimum temperature difference between the experimental and calculated temperatures to check the validity of our pyroxene selection. The (experimental – calculated) temperature difference was 15°C in all runs but INT-A3, -B2, -E7, and -E2, which were <50°C different (-41, -46, 38, and 27°C, respectively). We deem this agreement very good, as the thermometer uses only Na, Ca, Mg, and Fe cations without consideration of other elements in the pyroxenes.
Spinel grains are generally small and dispersed throughout the peridotite. Attempts were made to analyze rim compositions of the rare large (>10 µm) grains, but too few exist to fully assess the degree of homogeneity or inhomogeneity.
Crystalline phase distribution was generally homogeneous throughout the charges with occasional minor gradients in crystal distribution. The glass is concentrated toward the top of the capsule (in the sphere layer) but is also visible interstitially in the peridotite for higher melt fraction runs. For low melt fraction runs, melt in the peridotite was observable only in small pockets, which appear isolated from each other in 2D cross-sections at 1000x to 3000x magnification in BSE images. Only in some of the lowest-F runs is this considered potentially problematic. Generally, melt is connected throughout the charge allowing for ease of liquid–crystal equilibration (most experiments have olivine–melt Fe/Mg KD values of 0·34, after considering error; Table 3). In the lowest-F runs, where interconnected pathways may be few in number, the lack of complete equilibration is evident in high values of olivine–melt Fe/Mg KD values (e.g. 0·40 for INT-A1 and 0·42 for INT-B2, both at 1270°C, and 0·37 for INT-B7, at 1290°C). This lack of equilibration is especially apparent in the harzburgitic mode INT-C runs, as was also noted by Pickering-Witter & Johnston (2000) in their Opx-rich mix FER-C. Only the highest temperature run (INT-C3, 1390°C) with INT-C had an acceptable olivine–melt Fe/Mg KD (0·34; other INT-C runs range from 0·40 to 0·62). Only data from this single experiment are included in Tables 3 and 4. Data from all other runs with this extremely refractory starting material have been excluded from this paper and thus do not appear in Figs 3–9 or in the discussion.
|
|
|
|
standard deviations on average glass analyses. Error bars are smaller than symbol if not shown. Uncertainties in temperature are estimated to be ±10°C.
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Phase relations
Chemical compositions of the bulk starting materials reported in Table 2 and the experimental crystalline and glass phase compositions listed in Table 4 (wt % SiO2, Al2O3, FeO, MgO, CaO, Na2O, and Cr2O3) were used in least-squares mass balance calculations to determine phase proportions for each run product. Table 3 gives the results of these calculations using SIMPLEX3 (M. B. Baker, personal communiction, 1995), a program that utilizes the algorithm of Albarède & Provost (1977)
and propagates errors from analytical uncertainties in its calculations to provide error estimates for the phase abundances.
In contrast to Baker & Stolper (1994), who calculated melt proportions utilizing both core and rim pyroxene analyses, we use only our best estimates of equilibrium rim compositions, determined as described above and reported in Table 4, in mass balance calculations. Calculations of phase proportions utilizing core pyroxene compositions result in a slight decrease in the calculated melt proportion. For example, our INT-E1 melt proportion decreases from 1·48 to 1·36 wt % using averaged core analyses. Therefore, utilizing a multiple fit of core and rim compositions, as used by Baker & Stolper (1994), would yield a slight decrease in average melt fraction.
As expected from the PbO runs, we observed small and consistent shifts in relative phase proportions and mineral compositions resulting from subsolidus re-equilibration in the two mixtures for which we have such runs. Starting materials INT-A and -E both have subsolidus experiments (1255 and 1245°C, respectively), which help to quantify the degree to which phase proportions change through subsolidus compositional shifts. Calculated phase abundances in these runs indicate that the abundance of Ol increased by 2·5 wt % (absolute), Cpx increased by 4 wt %, Opx decreased by 6·3 wt %, and Sp decreased by 0·5 wt %.
Figure 3 shows plots of percent melt vs temperature for starting modes A, B, D, and E for the INT (filled symbols) and MM3/FER experiments (open symbols). The error bars include the ±10°C temperature uncertainties and uncertainties in calculated glass abundances. Although all of the nominally anhydrous starting materials began with exactly the same mineral compositions, we expect that differing amounts of subsolidus re-equilibration with associated shifts in composition occurred in each mix. Thus, phase compositions at the solidus probably differ somewhat from one mix to another, suggesting that each might have a somewhat different solidus temperature. Polynomial curve fits to the percent melt vs temperature data do not yield meaningful zero melt intercepts and therefore make quantifying the solidus difficult. Hence, the melt productivity curves in Fig. 3 are hand-drawn estimates. Although single curves, lacking abrupt breaks in slope, can be fitted to the data if the uncertainties are considered, the curve for INT-A has a break in slope at 1345°C coincident with the consumption of Cpx to allow for a concave-up form at low temperature, similar to the curves for the other starting materials. The intercept of the estimated curves with the zero melt percent line suggests similar solidus temperatures of 1260 ± 10°C. This may seem surprising considering that the range in Na2O content in the bulk compositions would be expected to significantly alter the composition of low-F melts and thus result in different solidus temperatures for the different bulk compositions. However, because the variation in Na2O content of our starting materials is achieved by varying the amount of Cpx, we are also changing the compatibility of Na2O near the solidus. Hirschmann (2000) and Pickering-Witter & Johnston (2000) described the counterbalancing effect on solidus temperatures of increasing the concentration of an incompatible component, such as Na2O, while at the same time increasing the DNapx-liq by increasing the amount of Cpx in the starting material. These effects cancel to a large degree and minimize the ability of Na to alter the solidus temperature. Indeed, the alkali concentration of our near-solidus liquids correlates inversely with the bulk alkali concentrations of our starting materials (Hirschmann, 2000).
Above the solidus, the melt productivity curves are concave-up for each mix, indicating that melt productivity (F/T) increases with rising temperature and progressive melting. Lherzolitic mode INT-D produces the most melt with a maximum of 28 wt % at 1390°C, whereas harzburgitic mode INT-C experiences the least amount of melting with only 3 wt % melt present at 1390°C, the only acceptable run obtained with this starting material.
Figure 4 shows plots of phase proportions vs temperature for the different starting materials. As expected, olivine is present in all experiments. Pyroxene abundance is largely governed by its initial concentration in the starting materials. Orthopyroxene remains present throughout the melting intervals except for the low-Opx mode B where it is absent above 1270°C and in INT-E where Opx is absent in the 1390°C run. Clinopyroxene abundance decreases with rising temperature, except in mode E where its abundance remains essentially constant. Clinopyroxene is stable throughout the entire melting interval investigated for mode B, the Cpx-enriched starting mix, and mode E, the equal-pyroxene mode. Spinel concentrations decrease systematically with rising temperature. Spinel is consumed in mix B above 1360°C, but remains present to 1390°C in all other modes. The mode-specific behavior is detailed below.
Mode INT-A is equivalent to the fertile MM3 mode of Baker & Stolper (1994) and includes a 1255°C experiment in addition to the standard temperature interval 1270–1390°C. The variations in phase abundances with temperature are shown in Fig. 4a. The melt curve in Fig. 3 is slightly concave-up, reflecting a small increase in melt productivity with rising temperature. The 1255°C run appears to be melt-free. Melt plus a Sp-lherzolite assemblage persists up to 1330°C, above which melt coexists with Sp-harzburgite assemblage through the remaining studied temperature interval. Melt production increases with temperature from 0 to 12·5 wt % at 1330°C with Cpx present and to 17·4 and 21·3 wt % at 1360 and 1390°C, respectively, after Cpx-out. The Ol and Opx abundances remain essentially constant over this same temperature interval and spinel decreases in abundance but remains present.
The Cpx-enriched mode INT-B (Fig. 4b) produced a Sp-lherzolite assemblage at the initial temperature of 1270°C with <2 wt % melt. By 1290°C, Opx is exhausted, resulting in a Sp-wehrlite assemblage to 1345°C, above which Sp is exhausted. The percent melt vs temperature curve for this mode in Fig. 3 is also concave-up. With rising temperature, the abundance of olivine remains constant whereas that of Cpx decreases. After spinel is exhausted (above 1345°C and 6·3 wt % glass), the calculated Cpx abundance drops more quickly with rising temperature, with a corresponding increase in melt fraction. Opx is present at 1270°C but is absent in higher-temperature runs. Melt proportion reaches 20 wt % at 1390°C.
The spinel-enriched mode INT-D (Fig. 4c) has a concave-up trend in melt percent vs temperature and has the highest melt productivity of the investigated starting materials. The INT-D runs contain a Sp-lherzolite plus melt assemblage from 1270 to 1300°C, above which Cpx is exhausted, and the remaining temperature interval (1330–1390°C) is typified by a Sp-harzburgite plus melt assemblage. Olivine abundance varies little, and Opx content increases slightly after Cpx-out (>1300°C) and then decreases. The spinel abundance decreases steadily, but spinel remains present at 1390°C. Melt proportion increases to a maximum of 28 wt % at 1390°C, the highest F of the investigated starting materials.
The equal-pyroxene mode INT-E (Fig. 4d) also displays a concave-up melt percent vs temperature curve (Fig. 3) and includes a subsolidus experiment at 1245°C. Again, Ol varies little, and Cpx, which is present throughout the investigated temperature interval, remains nearly constant in abundance. Opx abundance decreases to 1360°C, where mass balance calculations yield a 0 wt % abundance although some grains were observed in the charge. This discrepancy is probably due to complicated pyroxene phase relations as the crest of the pyroxene miscibility gap is approached. Figure 5 shows a plot of temperature vs molar Ca/(Ca + Mg + Fe) for the residual pyroxenes from all runs, with the coexisting pyroxenes in INT-E connected with horizontal lines outlining the projected pyroxene miscibility gap defined by our experiments (e.g. Lindsley, 1980, 1983; Nickel & Brey, 1984; see also Pickering-Witter & Johnston, 2000). By 1390°C, the miscibility gap is exceeded and the residual assemblage consists of a single pyroxene with a molar Ca/(Ca + Mg + Fe) value intermediate to those of the two pyroxenes in the 1360°C run. The melt proportion reaches 20 wt % at 1390°C. Sp is present to at least 1390°C.
The near-constant abundance of Cpx and decreasing abundance of Opx with rising temperature in mode INT-E contrasts with the pyroxene phase relations observed in other modes, as well as with the well-accepted notion (e.g. Baker & Stolper, 1994) that Cpx is consumed with rising temperature by melting to a significantly greater degree than other peridotite phases. This difference in behavior is clear when the pyroxene abundance variations are compared for modes INT-A and INT-E, both of which contain both pyroxenes over relatively large temperature intervals. In particular, we see from Table 3 that in INT-A the abundance of Cpx decreases with rising temperature whereas that of Opx remains essentially constant, whereas just the opposite behavior is seen in INT-E. Controls on pyroxene abundances in experiments like ours are complex and include at least inter-pyroxene solubility relations (i.e. the miscibility gap) and complex melting equilibria that consume the different pyroxenes (and other solids) at differing rates with rising temperature.
We refer to Fig. 5 in an attempt to understand the differing behavior of the pyroxenes in INT-A and INT-E. We show with the vertical lines our calculated ‘bulk pyroxene’ compositions, which represent the weighted averages of the abundances of the two pyroxenes in subsolidus runs INT-A6 and INT-E4 in conjunction with the compositions of the near-solidus pyroxenes as determined in the PbO-bearing, two-pyroxene experiment described earlier. Considering only inter-pyroxene solubility relations for the moment, we see by application of the lever rule in Fig. 5 that the Cpx/Opx ratio remains relatively constant with rising temperature for INT-A but rises dramatically for INT-E. This, of course, is simply a consequence of the fact that the lever for INT-E plots toward the ‘diopside’ leg of the miscibility gap relative to the INT-A lever. Thus, the line length representing the proportion of the total pyroxenes accounted for by Opx decreases much more rapidly with rising temperature in INT-E than INT-A. This would suggest that inter-pyroxene solubility relations, considered alone, will result in a rapid increase in the Cpx/Opx ratio of the total pyroxene in INT-E with rising temperature, with relatively little change in this ratio over the same temperature interval in INT-A. Walter (1999) described a similar situation for fertile lherzolites near the solidus whereby with rising temperature, Cpx becomes more enstatitic, and the overall abundance of Cpx increases at the expense of Opx. Similarly, by changing the composition of Cpx alone, a correspondingly significant change in the Cpx/Opx ratio occurs (Walter, 1999).
This analysis represents a gross simplification of what actually occurs with rising temperature because it does not consider the effects of melting equilibria, which simultaneously exert other controls on phase abundances and compositions. Indeed, the fact that both the Cpx and Opx compositions in run INT-E7 (1360°C) plot to the left of the ‘bulk pyroxene’ lever for INT-E clearly demonstrates that more than just inter-pyroxene solubility relations control pyroxene compositions and relative abundances. On the basis of previous work on peridotite melting (e.g. Kinzler & Grove, 1992a, 1992b; Baker & Stolper, 1994; Pickering-Witter & Johnston, 2000) we can confidently consider the prime effect of progressive melting on the pyroxene abundances to be a decrease in the abundance of Cpx with relatively little change in the abundance of Opx. If we now superimpose this effect on the expected variation in the Cpx/Opx ratio for INT-A and INT-E deduced above, we arrive at the following qualitative expectations. Our initial observation with INT-A is that the relative proportions of the two pyroxenes, as a result of inter-pyroxene solubility relations, should remain relatively constant. As Cpx is simultaneously and preferentially consumed by progressive melting, however, the expectation from the combined processes would be a decrease in the abundance of Cpx, with the abundance of Opx remaining relatively constant, as observed in the data (see Table 3). With INT-E, however, we see a rapid increase in the abundance of Cpx with a concomitant decrease in the abundance of Opx with rising temperature, owing to inter-pyroxene solubility relations. Superimposing the anticipated effect of melting leads to the expectation that the combined processes will result in a decrease in the Opx abundance with rising temperature, and a relatively constant abundance of Cpx, as the consumption of Cpx by melting processes is balanced to a large degree by the production of Cpx from Opx as temperature rises through the pyroxene miscibility gap. Again, this is what the data show (Table 3). Although it is gratifying that our analysis appears to be consistent with our data, we believe it is fortuitous to some degree because the final result represents the balance between two processes that produce or consume pyroxenes at rates that are difficult to quantify and will vary from one peridotite to another (e.g. melt productivity and rate of Cpx consumption).
The overall melt productivity of these starting materials does not always correlate with the initial Cpx abundance. Langmuir et al. (1992) suggested that veins of Cpx equal in composition to that in the host peridotite would increase the extent of melting. In a polybaric melting scenario, Gaetani & DeLong (1995) modeled Cpx veins by increasing the modal abundance of Cpx of a fertile source mantle and indeed found that more melt was produced for an equal decompression step. The resulting melt was compositionally identical to the original, following Walter & Presnall (1994), who showed that varying the relative proportions of equilibrium phases has no effect on the composition of those phases and simply drives the melting reaction toward more melt. In contrast, simply increasing the relative Cpx abundance in the intermediate composition peridotites studied here does not necessarily result in greater melt production. Part of the discrepancy is probably due to the compositional difference between the starting Cpx and the ‘true’ Cpx equilibrated at the pressure and temperature of the experiment. Nevertheless, Mode INT-C is exceptionally refractory as a result of the lack of Cpx and it produces the least melt. Mode INT-D is the highest melt producer and has the second lowest Cpx content, but has the highest Sp content. If, as suggested by the curve in Fig. 3c, INT-D has the lowest solidus temperature of the investigated modes, the higher melt fraction of INT-D would be consistent with the greater overstepping of the solidus.
Phase compositions
Glass
Figure 6 shows plots of the weight percent of selected oxides in the glass versus percent melt. MgO, FeO, and Cr2O3 concentrations generally increase with rising temperature and percent melt, whereas Al2O3, Na2O, and K2O contents decrease reflecting their relative incompatibility; they are enriched at low melt fraction and decrease in concentration through dilution with progressive melting. Melt K2O contents of as high as 1·8 and 1·6 wt % are observed in the lowest-F runs in INT-E and INT-D, despite the fact that no K2O was detected in the starting materials. However, if we consider K2O to be perfectly incompatible (D = 0) and calculate the abundance in the starting material (C0) that would be necessary to generate the measured melt concentration (CL, Table 4) given the calculated F values (Table 3) using the batch melting equation (C0 = CLF, assuming D = 0), we obtain values of 0·02 wt % with no systematic difference from one mode to another. Such low amounts could easily go undetected without making special efforts. Although not shown, TiO2 also behaves as an incompatible component, ranging from 0·4 wt % at low melt fraction to 0·1 wt % at higher melt fractions. CaO abundances increase continuously with percent melt for modes B and E for which Cpx persists at least to 1390°C, but decrease in modes A and D, after Cpx is exhausted, as a result of dilution. The very low CaO content in the single plotted INT-C run simply reflects the very low initial Cpx abundance (1%) in this starting material. Melt SiO2 contents are highest at low melt fraction and decrease with progressive melting, as seen previously in fertile composition peridotites (e.g. Baker & Stolper, 1994; Baker et al., 1995; Hirschmann et al., 1998; Falloon et al., 1999; Pickering-Witter & Johnston, 2000).
At F >0·05 (i.e. 5%), the Al2O3 content is highest in the INT-D glasses, reflecting the higher Al2O3 availability from the elevated spinel content of the mix. Figure 7 shows that melt mg-numbers and cr-numbers generally increase with melt fraction (and temperature).
Olivine
The mg-numbers of olivine range from 91 to 93 with rising temperature and CaO contents vary little (0·1–0·5 wt % depending on bulk CaO content). Olivine Cr2O3 contents generally range from 0·1 wt % to as high as 0·5 wt %, but reach 1·0 wt % in the highest temperature run with spinel-rich INT-D. These high Cr contents probably reflect the presence of substantial Cr2+ as a result of the reducing environment created by the graphite inner capsules.
Clinopyroxene
Figure 8 shows a plot of CaO contents vs temperature for residual Cpx. CaO contents remain nearly constant for INT-B and -D, in contrast with modes A and E, in which CaO contents decrease with increased melting. This is another expression of the fact that with rising temperature, the Cpx in modes A and E, coexisting with Opx, tracks the ‘diopside’ leg of the pyroxene miscibility gap (e.g. Lindsley, 1980, 1983; Nickel & Brey, 1984; Fig. 5). Clinopyroxene is present in both INT-B and INT-E throughout the investigated temperature interval, but the extended coexistence of Opx in INT-E leads to a decrease in the ‘wollastonite component’ of the Cpx with rising temperature, with the highest-temperature Cpx being pigeonite (see Bertka & Holloway, 1993; Longhi & Bertka, 1996). In contrast, INT-B loses Opx at relatively low temperature (above 1270°C), and the ‘wollastonite component’ in the Cpx remains high.
Spinel
With rising temperature, spinel mg-number decreases and cr-number increases. The INT-A and -E spinels display the widest range in both mg- and cr-numbers. Spinel persists in INT-B to only 1360°C, thus limiting its compositional range, and the relatively large spinel abundance in INT-D (10%) decreased its compositional sensitivity to changes in temperature and other phase compositions, accounting for its more limited range in cr-number and mg-number. The low mg-number of the starting spinel is undoubtedly another indication of re-equilibration of our starting xenolith to temperatures lower than the solidus at 10 kbar.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Comparison with results from fertile peridotites with similar modes
Solidus temperature
The estimated nominally anhydrous solidus temperatures for the INT modes (
1260 ± 10°C) are higher than those of the fertile equivalents (1240 ± 10°C for MM3 (Baker & Stolper, 1994) and 1250 ± 10°C for FER (Pickering-Witter & Johnston, 2000)). This, of course, reflects lower concentrations of incompatible components in the INT mixes. Walter & Presnall (1994) demonstrated a rise in solidus temperature with depletion of incompatible components in model systems and showed that the amount of melt produced before exhaustion of a phase decreased with depletion.
Melt productivity
Figure 3 shows the weight percent glass plotted vs temperature for the INT (filled symbols) and MM3 (Baker & Stolper, 1994) and FER (Pickering-Witter & Johnston, 2000) starting materials (open symbols). In each mode, the FER samples produced more liquid and plot above the INT sample at a given temperature. In all cases, the fertile runs have higher melt fractions than the intermediate equivalents at a given temperature, but the run with the highest percent melting at a given temperature is not always of the same starting mode. For example, the highest percent melting for the fertile runs is achieved by the Cpx-enriched, FER-B (Fig. 3b), whereas the highest percent melting for the intermediate compositions is the Sp-enriched mode, INT-D (Fig. 3c).
Figure 3 also illustrates that the melt fraction difference between the FER and INT curves correlates with the amount of initial Cpx. Mode D (Fig. 3c) has the lowest initial Cpx abundance (10%) and the melting curves are nearly coincident. Mode A (Fig. 3a) has 17% Cpx initially, and the melting curves are separate and parallel. The equal-pyroxene Mode E has 23·5% Cpx, and the curves diverge, and the curves for Cpx-enriched Mode B (Fig. 3b), with 40% initial Cpx, diverge to an even greater extent. Modes A and D also lose the same phases at comparable temperatures, whereas the FER and INT modes B and E lose phases at different temperatures. This increased disparity between the melting curves with increasing initial Cpx content makes clear that it is the compositional fertility of Cpx, rather than just its simple abundance, that is critical in determining a rock’s melt productivity.
Glass composition
To facilitate comparison of glass compositions between the INT and MM3/FER experiments, the concentrations of selected glass oxides are plotted versus percent melt in Fig. 9. In general, the trends are similar for each modally equivalent starting material, although they are offset from each other by amounts controlled by the crystalline phase assemblage stable at a given melt fraction, as well as the concentration of an oxide in the bulk composition of the starting material. The lower melt productivity of the INT samples tends to group the data together to the left in contrast to the wider spread in the FER results.
Glass SiO2 and Na2O contents are elevated at low melt fractions in all starting materials (Fig. 9). The INT glass Na2O concentrations are lower than the FER glass values, reflecting the more depleted nature of the INT starting materials. Much of the elevated SiO2 concentration at low melt fraction is attributed to the network-modifying capabilities of alkalis concentrated in low-F melts (e.g. Mysen & Kushiro, 1977; Hirschmann et al., 1988; Walter & Presnall, 1994; Baker et al., 1995; Walter et al., 1995). The lower bulk Na2O contents of the INT starting materials result in a smaller increase in Na2O and SiO2 at low melt fraction relative to fertile runs, consistent with the suggestion of Hirschmann et al. (1998) that depletion results in a decrease in low-F enrichment.
The melt FeO trends (shown in Fig. 9) of the modally equivalent FER/MM3 and INT samples follow similar trajectories. The FER samples reach higher melt fractions, but otherwise the trends overlap. The Cpx-poor mode C melts have the highest FeO contents with the single INT-C run plotting congruently on the FER-C trend. These elevated FeO contents probably reflect the fact that the fusible components in these nearly Cpx-free mixes are the fayalite and ferrosilite components of the Ol and Opx, respectively. MgO trends display more spread between modes with INT glass MgO values higher than FER glasses, consistent with the higher MgO contents and mg-numbers of the INT samples. The INT glasses are generally richer in CaO, but display similar concave-down trends, increasing with Cpx present and decreasing after Cpx-out. Melt Al2O3, Cr2O3, and TiO2 concentrations for FER/MM3 and INT glasses tend to emanate from similar initial concentrations. Cr2O3 contents increase approximately linearly with increased melting and Al2O3 decreases approximately linearly from initial near-solidus values of 18–20 wt %, but along trends with different slopes for the different starting material types (i.e. INT vs FER). TiO2 is low in the INT modes and decreases with progressive melting. In contrast, FER/MM3 glass TiO2 contents initially increase and then decrease beyond 10 wt % melt. Baker et al. (1995) attributed the decreased melt TiO2 contents they observed at low F in MM3 primarily to the change in Cpx composition (and thus, DTi) during early melting. TiO2 substitutes into Cpx via charge-coupled exchange and is dependent on the availability of charge-coupling cations (Baker et al., 1995). These cations (e.g. Na+, Fe3+, Al3+) behave incompatibly and are most abundant in Cpx at low F (Baker et al., 1995). The greater depletion of the INT samples results in a paucity of the charge-coupling cations and consequently, a similar decrease in melt TiO2 (or increase in DTi) at low F is not observed.
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Table 5 summarizes comparisons made between the modally equivalent FER/MM3 and INT experiments and highlights some of the consequences of changing the fertility of modally equivalent source rocks. The peridotite starting materials become more refractory with increased depletion, as seen in the
1250 ± 10°C solidus for the FER samples (Pickering-Witter & Johnston, 2000) compared with the 1260 ± 10°C solidus for the INT samples. The maximum melt fraction (at 1390°C) for the FER samples is 0·48 for Cpx-rich mode B, whereas the Sp-rich mode D is the most productive (F = 0·28) of the INT peridotite compositions. Harzburgitic mode C is least productive for both the FER and INT samples, and this Opx-enriched starting mix proved extremely refractory for the INT runs. One key observation is that the extent of melting for the FER samples generally correlates with the initial abundance of Cpx in the starting material. This is not the case for the INT runs in which the maximum F was achieved with INT-D, the mode with the lowest Cpx content. This shows that the compositional fertility of the Cpx apparently plays a greater role in controlling the extent of melting than simply its overall abundance in the source rock. For example, an intermediate composition peridotite with 40 wt % Cpx would still be less melt productive than a fertile peridotite with only 10 wt % Cpx, but three times higher bulk Na2O. However, the ultimate difference between melt curves of fertile and intermediate composition peridotite with equal, but low (i.e. 10–17 wt %), initial Cpx contents is small.
|
Another observation made in the INT experiments is that pyroxene miscibility relations dominate the stoichiometry of the melting reaction in mode E (see also Walter, 1999). Plotting the sum of the two pyroxenes vs temperature for modes A and E reveals the difference in pyroxene relations in FER and INT. The slopes of linear least-squares regressions of the data yield the rate at which pyroxene abundance changes with temperature. Whereas the mode A pyroxenes are consumed at similar rates (-0·18 and -0·17 wt %/°C for MM3 and INT-A, respectively), the FER-E pyroxenes are consumed much more rapidly (-0·26 wt %/°C) than the INT-E pyroxenes (-0·12 wt %/°C). This relatively slow loss of pyroxenes in INT-E allows us to readily observe the pyroxene solubility relations described earlier.
The compositions of lowest-F melts also vary between modes. The near-solidus melts of the FER samples have a maximum Na2O content of 7·4 wt % (MM3, Hirschmann et al., 1998) whereas the INT-E glass has a maximum of 2·5 wt %. Similarly, FER-A glasses display a maximum SiO2 content of 55·5 wt % whereas INT-A has a maximum SiO2 of 53·9 wt %. In contrast, there are striking similarities between near-solidus Cpx compositions for the FER and INT samples, which have 18 wt % CaO, but which vary in Al2O3 contents from 6·5 to 6·9 wt % for FER Cpx and from 4·5 to 5·5 wt % for INT Cpx. It is interesting to note that near-solidus Cpx of depleted DMM1 (Wasylenki, 1999) contains 18·01 wt % CaO.
|
ACKNOWLEDGEMENTS |
|---|
The authors thank G. Gaetani, M. Walter and L. Wasylenki, whose thorough and constructive reviews considerably improved the manuscript. We also thank M. Hirschmann for helpful discussion of the results. Dave ‘honorary geologist’ Senkovich’s skills and ideas kept the laboratory operating more times than we can count, and we are grateful for M. Shaffer’s knowledge and patience in the microprobe laboratory. A very special thank you goes to J. Pickering-Witter for innumerable things. M. Baker generously provided time, training, and advice on the diamond aggregate technique. J. Blundy also provided helpful discussion and suggestions while on sabbatical leave at the University of Oregon. We also thank George Bergantz for timely and thorough editorial handling. This research was supported by National Science Foundation grants EAR-9506045 and EAR-9804913 to A.D.J. The electron microbeam facility at the University of Oregon was acquired with NSF grant EAR-8803960 and a matching grant from the Keck Foundation, and technician support was provided by NSF-EAR-9712115.
|
FOOTNOTES |
|---|
*Corresponding author. Present address: Department of Geology, Humboldt State University, Arcata, CA 95521, USA. Telephone: 707-826-3950. Fax: 707-826-5241. E-mail: bes21{at}axe.humboldt.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL AND ANALYTICAL...RESULTSDISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Albarède, F. & Provost, A. (1977). Petrological and geochemical mass-balance equations: an algorithm for least-square fitting and general error analysis. Computers and Geosciences 3, 309–326.
Baker, M. B. & Stolper, E. M. (1994). Determining the composition of high-pressure mantle melts using diamond aggregates. Geochimica et Cosmochimica Acta 58, 2811–2827.[Web of Science]
Baker, M. B., Hirschmann, M. M., Ghiorso, M. S. & Stolper, E. M. (1995). Compositions of near-solidus peridotite melts from experiments and thermodynamic calculations. Nature 375, 308–311.
Baker, M. B., Hirschmann, M. M., Wasylenki, L. E., Stolper, E. M. & Ghiorso, M. S. (1996). Quest for low-degree mantle melts—reply. Nature 381, 286.
Bertka, C. M. & Holloway, J. R. (1993). Pigeonite at solidus temperatures: implications for partial melting. Journal of Geophysical Research 98, 19755–19766.
Brey, G. P. & Köhler, T. (1990). Geothermobarometry in four-phase lherzolite II. New thermobarometers, and practical assessment of existing thermobarometers. Journal of Petrology 31, 1353–1378.
Brey, G. P., Köhler, T. & Nickel, K. G. (1990). Geothermobarometry in four-phase lherzolites I. Experimental results from 10 to 60 kb. Journal of Petrology 31, 1313–1352.
Cooper, R. F. & Kohlstedt, D. L. (1986). Rheology and structure of olivine-basalt partial melts. Journal of Geophysical Research 91, 9315–9323.
Daines, M. J. & Richter, F. M. (1988). An experimental method for directly determining the interconnectivity of melt in a partially molten system. Geophysical Research Letters 15, 1459–1462.
Falloon, T. J., Green, D. H., O’Neill, H. S. & Ballhaus, C. G. (1996). Quest for low-degree mantle melts. Nature 381, 285.
Falloon, T. J., Green, D. H., O’Neill, H. S. C. & Hibberson, W. O. (1997). Experimental tests of low degree peridotite partial melt compositions: implications for the nature of anhydrous near-solidus peridotite melts at 1 GPa. Earth and Planetary Science Letters 152, 149–162.
Falloon, T. J., Green, D. H., Danyushevsky, L. V. & Faul, U. H. (1999). Peridotite melting at 1·0 and 1·5 GPa: an experimental evaluation of techniques using diamond aggregates and mineral mixes for determination of near-solidus melts. Journal of Petrology 40, 1343–1375.
Gaetani, G. A. & DeLong, S. E. (1995). Petrogenesis of basalts from the Blanco Trough, northeast Pacific: inferences for off-axis melt generation. Journal of Geophysical Research 100, 4197–4214.
Ghiorso, M. S. & Sack, R. O. (1995). Chemical mass transfer in magmatic processes IV. A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid–solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to Mineralogy and Petrology 119, 197–212.[Web of Science]
Ghiorso, M. S., Hirschmann, M. M. & Sack, R. O. (1994). MELTS: software for thermodynamic modeling of magmatic systems. EOS Transactions, American Geophysical Union 75, 571–576.
Hart, S. R. & Zindler, A. (1986). In search of bulk Earth composition. Chemical Geology 57, 247–267.
Hirose, K. & Kushiro, I. (1993). Partial melting of dry peridotites at high pressures: determination of compositions of melts segregated from peridotite using aggregates of diamond. Earth and Planetary Science Letters 114, 477–489.[Web of Science]
Hirschmann, M. M. (2000). The mantle solidus: experimental constraints and the effects of peridotite composition. Geochemistry, Geophysics, and Geosystems 1, paper number 70.
Hirschmann, M. M., Baker, M. B. & Stolper, E. M. (1998). The effect of alkalis on the silica content of mantle-derived melts. Geochimica et Cosmochimica Acta 62, 883–902.
Johnson, K. T. M., Dick, H. J. B. & Shimizu, N. (1990). Melting in the oceanic upper mantle: an ion microprobe study of diopside in abyssal peridotites. Journal of Geophysical Research 95, 2661–2678.
Kinzler, R. J. & Grove, T. L. (1992a). Primary magmas of mid-ocean ridge basalts 1. Experiments and methods. Journal of Geophysical Research 97, 6885–6906.
Kinzler, R. J. & Grove, T. L. (1992b). Primary magmas of mid-ocean ridge basalts 2. Applications. Journal of Geophysical Research 97, 6907–6926.
Klein, E. M. & Langmuir, C. H. (1987). Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92, 8089–8115.
Langmuir, C. H., Klein, E. L. & Plank, T. (1992). Petrological systematics of mid-ocean ridge basalts: constraints on melt generation beneath ocean ridges. In: Morgan, J. P., Blackman, D. K. & Sinton, J. M. (eds) Mantle Flow and Melt Generation at Mid-Ocean Ridges. Geophysical Monograph, American Geophysical Union, 71, 183–280.
Lindsley, D. H. (1980). Phase equilibria of pyroxenes at pressures >1 atmosphere. In: Prewitt, C. T. (ed.) Mineralogical Society of America, Reviews in Mineralogy 7, 289–307.[Abstract]
Lindsley, D. H. (1983). Pyroxene thermometry. American Mineralogist 68, 477–493.[Abstract]
Longhi, J. & Bertka, C. M. (1996). Graphical analysis of pigeonite–augite liquidus equilibria. American Mineralogist 81, 685–695.[Abstract]
McDonough, W. F. & Frey, F. A. (1989). Rare earth elements in upper mantle rocks. In: Lipin, B. R. & McKay, G. A. (eds) Geochemistry and Mineralogy of Rare Earth Elements. Mineralogical Society of America, Reviews in Mineralogy 21, 223–253.
McKenzie, D. (1984). The generation and compaction of partially molten rock. Journal of Petrology 25, 713–765.
McKenzie, D. (1989). Some remarks on the movement of small melt fractions in the mantle. Earth and Planetary Science Letters 95, 53–72.[Web of Science]
McKenzie, D. & Bickle, M. J. (1988). The volume and composition of melt generated by extension of the lithosphere. Journal of Petrology 29, 625–679.
Morgan, G. B. & London, D. (1996). Optimizing the electron microprobe analysis of hydrous alkali aluminosilicate glasses. American Mineralogist 81, 1176–1185.[Abstract]
Mysen, B. O. & Kushiro, I. (1977). Compositional variations of coexisting phases with degree of melting of peridotite in the upper mantle. American Mineralogist 62, 843–865.[Abstract]
Nickel, K. G. & Brey, G. (1984). Subsolidus orthopyroxene–clinopyroxene systematics in the system CaO–MgO–SiO2 to 60 kb: a re-evaluation of the regular solution model. Contributions to Mineralogy and Petrology 87, 35–42.
Niu, Y. & Batiza, R. (1991). An empirical method for calculating melt compositions produced beneath mid-ocean ridges: application for axis and off-axis (seamounts) melting. Journal of Geophysical Research 96, 21753–21777.
Pickering-Witter, J. M. & Johnston, A. D. (2000). The effects of variable mineral proportions on the melting systematics of fertile peridotite assemblages. Contributions to Mineralogy and Petrology 140, 190–211.
Pickering, J. M., Schwab, B. E. & Johnston, A. D. (1998). Off-center hot spots: double thermocouple determination of the thermal gradient in a 1·27 cm (1/2 in.) CaF2 piston-cylinder furnace assembly. American Mineralogist 83, 228–235.[Abstract]
Pouchou, J.-L. & Pichoir, F. (1991). Quantitative analysis of homogeneous and stratified microvolumes applying the model ‘PAP’. In: Heinrich, K. F. & Newberry, D. E. (eds) Electron Probe Quantitation. New York: Plenum.
Ringwood, A. E. (1979). Origin of the Earth and Moon. Berlin: Springer, 295 pp.
Robinson, J. A. C., Wood, B. J. & Blundy, J. D. (1998). The beginning of melting of fertile and depleted peridotite at 1·5 GPa. Earth and Planetary Science Letters 155, 97–111.
Salters, V. J. M. & Hart, S. R. (1989). The Hf paradox and the role of garnet in the source of mid-ocean ridge basalts. Nature 342, 420–422.
Sen, G. (1982). Composition of basaltic liquids generated from a partially depleted lherzolite at 9 kbar pressure. Nature 299, 336–338.
Sneeringer, M. A. & Watson, E. B. (1985). Milk cartons and ash cans: two unconventional welding techniques. American Mineralogist 70, 200–201.[Abstract]
Taylor, W. R. & Green, D. H. (1989). The role of reduced C–O–H fluids in mantle partial melting. In: Kimberlites and Related Rocks: Proceedings of the 4th International Kimberlite Conference, Vol. 1. Oxford: Blackwell Scientific, pp. 592–602.
Ulmer, P. & Luth, R. W. (1991). The graphite–COH fluid equilibrium in P, T, fO2 space: an experimental determination to 30 kbar and 1600°C. Contributions to Mineralogy and Petrology 106, 265–272.[Web of Science]
von Bargen, N. & Waff, H. S. (1986). Permeabilities, interfacial areas and curvatures of partially molten systems: results of numerical computations of equilibrium microstructures. Journal of Geophysical Research 91, 9261–9276.
Waff, H. S. & Bulau, J. R. (1979). Equilibrium fluid distribution in an ultramafic partial melt under hydrostatic stress conditions. Journal of Geophysical Research 84, 6109–6114.
Walter, M. J. (1999). Comments on ‘Mantle melting and melt extraction processes beneath ocean ridges; Evidence from abyssal peridotites’ by Yaoling Niu. Journal of Petrology 40, 1187–1193.
Walter, M. J. & Presnall, D. C. (1994). Melting behavior of simplified lherzolite in the system CaO–MgO–Al2O3–SiO2–Na2O from 7 to 35 kbar. Journal of Petrology 35, 329–359.
Walter, M. J., Sisson, T. W. & Presnall, D. C. (1995). A mass proportion method for calculating melting reactions and application to melting of model upper mantle lherzolite. Earth and Planetary Science Letters 135, 77–90.
Wasylenki, L. E. (1999). Partial melting of depleted peridotite in the Earth’s upper mantle and implications for generation of mid-ocean ridge basalts. D.Phil. Dissertation, California Institute of Technology, 158 pp.
Wasylenki, L. E., Baker, M. B., Hirschmann, M. M. & Stolper, E. M. (1996). The effect of source depletion on equilibrium mantle melting. EOS Transactions, American Geophysical Union 77, F847.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Sakuyama, K. Ozawa, H. Sumino, and K. Nagao Progressive Melt Extraction from Upwelling Mantle Constrained by the Kita-Matsuura Basalts in NW Kyushu, SW Japan J. Petrology, April 1, 2009; 50(4): 725 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Spandler, G. Yaxley, D. H. Green, and A. Rosenthal Phase Relations and Melting of Anhydrous K-bearing Eclogite from 1200 to 1600{degrees}C and 3 to 5 GPa J. Petrology, April 1, 2008; 49(4): 771 - 795. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Falloon, D. H. Green, L. V. Danyushevsky, and A. W. McNeill The Composition of Near-solidus Partial Melts of Fertile Peridotite at 1 and 1{middle dot}5 GPa: Implications for the Petrogenesis of MORB J. Petrology, April 1, 2008; 49(4): 591 - 613. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. D. Putirka Thermometers and Barometers for Volcanic Systems Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 61 - 120. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Lustrino, L. Melluso, and V. Morra The geochemical peculiarity of "Plio-Quaternary" volcanic rocks of Sardinia in the circum-Mediterranean area Geological Society of America Special Papers, January 1, 2007; 418(0): 277 - 301. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Lustrino and E. Carminati Phantom plumes in Europe and the circum-Mediterranean region Geological Society of America Special Papers, January 1, 2007; 430(0): 723 - 745. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. MEDARD, M. W. SCHMIDT, P. SCHIANO, and L. OTTOLINI Melting of Amphibole-bearing Wehrlites: an Experimental Study on the Origin of Ultra-calcic Nepheline-normative Melts J. Petrology, March 1, 2006; 47(3): 481 - 504. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. PEARCE Mantle Preconditioning by Melt Extraction during Flow: Theory and Petrogenetic Implications J. Petrology, May 1, 2005; 46(5): 973 - 997. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. D. ASIMOW and J. LONGHI The Significance of Multiple Saturation Points in the Context of Polybaric Near-fractional Melting J. Petrology, December 1, 2004; 45(12): 2349 - 2367. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. LIU and H. ST. C. O'NEILL The Effect of Cr2O3 on the Partial Melting of Spinel Lherzolite in the System CaO-MgO-Al2O3-SiO2-Cr2O3 at 1{middle dot}1 GPa J. Petrology, November 1, 2004; 45(11): 2261 - 2286. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. W. SCHMIDT, D. H. GREEN, and W. O. HIBBERSON Ultra-calcic Magmas Generated from Ca-depleted Mantle: an Experimental Study on the Origin of Ankaramites J. Petrology, March 1, 2004; 45(3): 531 - 554. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. PERTERMANN and M. M. HIRSCHMANN Anhydrous Partial Melting Experiments on MORB-like Eclogite: Phase Relations, Phase Compositions and Mineral-Melt Partitioning of Major Elements at 2-3 GPa J. Petrology, December 1, 2003; 44(12): 2173 - 2201. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. E. WASYLENKI, M. B. BAKER, A. J. R. KENT, and E. M. STOLPER Near-solidus Melting of the Shallow Upper Mantle: Partial Melting Experiments on Depleted Peridotite J. Petrology, July 1, 2003; 44(7): 1163 - 1191. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||