Roger L. Nielsen

COMAGMAT: a Fortran program to model MAGMA differentiation processes

Abstract The purpose of this paper is to present a general model (COMAGMAT) for the calculation of equilibrium temperatures and phase relations at a given extent of crystallization or melting in natural magmatic systems. This model is based on a set of empirical expressions that describe mineral-melt equilibria for major and trace elements in terms of pressure (up to 12 kbar), temperature, and liquid compositions for systems ranging from primitive basalts to dacites. These expressions are in the form of empirically calibrated mineral-melt geothermometers for Olivine, Augite, Pigeonite (Opx), Plagioclase, Ilmenite, and Magnetite used to develop the algorithm simulating multiply saturated magmatic melts. The results of the program are in the form of calculated liquid lines of descent, plus the equilibrium mineral proportions and compositions. The phase equilibria calculations form the core of a model that allows the user to simulate processes ranging from simple isobaric crystallization to in situ differentiation processes resulted from crystal settling, and polybaric fractionation.

Experimentally determined mineral-melt partition coefficients for Sc, Y and REE for olivine, orthopyroxene, pigeonite, magnetite and ilmenite

Our current lack of understanding of the partitioning behavior of Sc, Y and the REE (rare-earth elements) can be attributed directly to the lack of a sufficiently large or chemically diverse experimental data set. To address this problem, we conducted a series of experiments using several different natural composition lavas, doped with the elements of interest, as starting compositions. Microprobe analyses of orthopyroxene, pigeonite, olivine, magnetite, ilmenite and co-existing glasses in the experimental charges were used to calculate expressions that describe REE partitioning as a function of a variety of system parameters. Using expressions that represent mineral-melt reactions (versus element ratio distribution coefficients) it is possible to calculate terms that express low-Ca pyroxene-melt partitioning behavior and are independent of both pyroxene and melt composition. Compositional variations suggest that Sc substitution in olivine involves either a paired substitution with Al or, more commonly, with vacancies. The partitioning of Sc is dependent both on melt composition and temperature. Our experimentally determined olivine-melt REE Ds (partition coefficients) are similar to, but slightly higher than those reported by McKay (1986) and support their conclusions that olivines are strongly LREE depleted. Y and REE mineral/melt partition coefficients for magnetite range from 0.003 for La to 0.02 for Lu. Ilmenite partition coefficients range from 0.007 for La to 0.029 for Lu. These experimental values are two orders of magnitude lower than many of the published values determined by phenocryst/matrix separation techniques.

Major- and trace-element magnetite-melt equilibria

The effects of composition and temperature on the partitioning behavior of Sc, Ni, V, U and Th, and the high field strength elements (HFSE) Zr, Nb, Ta and Hf between magnetite and natural silicate melts were evaluated from doped experiments on natural mafic- to intermediate-composition lavas at 1-atm pressure. Composition was found to be the strongest controlling factor on partitioning behavior. The partition coefficients (D) for Zr, Nb, Hf and Ta correlate with DTi and are similar to one another for any given magnetite-melt pair, but vary from 2 in titanomagnetite. DSc is higher than DZr, DNb, DHf and DTa, and also correlated to DTi. In contrast, Ni is more compatible in Al-Cr-rich magnetites than in titanomagnetites, but is compatible in all magnetite-melt pairs in our experiments. V is generally more compatible than Zr, Nb, Ta and Hf, but its behavior is complicated by its multiple valence states. U and Th are incompatible (D < 0.035) in all magnetites. Expressions were derived to describe the relationship between D and the most strongly correlated parameters, oxide Fe/Mg ratio and Al, and DTi. These patterns of behavior are consistent with the observed miscibility gap between the spinel-group end-members. Equilibrium constants for spinel end-member-melt reactions were parameterized in terms of temperature, pressure and composition. These expressions can be used to predict the temperature and composition of equilibrium spinels. These major-element constraints can also be used to predict spinel-melt partition coefficients using the expressions describing DHFSE as a function of DTi and composition.

Terminology for trace-element partitioning

A self-consistent terminology for partitioning data is presented. Ratios of the concentration of a component in two phases are termed partition coefficients and given the symbol D. Ratios of partition coefficients are termed exchange coefficients and given the symbol KD. The prefix “bulk” implies that these coefficients are weighted according to the proportions of coexisting phases. Bulk partition and bulk exchange coefficients are denoted by D and KD, respectively.

Pyroxene-melt equilibria

Abstract Using a literature survey of analyses of high-Ca pyroxene and co-existing silicate melt pairs and analyses of low-Ca pyroxene-silicate melt pairs, we have performed a thermodynamic analysis of pyroxene-melt equilibria. Three sets of mixing model pairs have been considered, based on two mixing models for liquid silicate solutions and two for pyroxene solid solutions. A modified version of a model developed by Bottinga and Weill (1972) for the mixing properties of silicate melts, in which the melt is considered to be composed of independent network-forming and network-modifying quasi-lattices, more successfully accounts for variations in melt composition than does a model which considers the melt to be composed of simple oxides which mix ideally. An empirical model for the mixing properties of pyroxenes, in which the M1 and M2 sites are considered to be equivalent and are combined as a hypothetical ‘M’ site, is as successful in accounting for variations in pyroxene composition at high temperatures as an ideal multisite mixing model. Using a variety of pyroxene-melt relations, and combinations of the mixing models outlined above, we have developed several pyroxene-melt and low-Ca pyroxene-high-Ca pyroxene geothermometers which have internally-consistent precisions of approximately ±20°C (1σ). One of the two-pyroxene geothermometers has been used to calculate ‘quench’ temperatures for a number of eucrites. Computed temperatures are subsolidus, and are consistent with independent geothermometers and with petrographic observations. The equations may also be used to calculate the composition of pyroxene crystallizing from a silicate melt of known composition, with or without independent knowledge of temperature. Internally consistent precisions vary, but are approximately ± 3 mol% Fs, ± 5 mol% En, and ±4 mol% Wo (all 1σ). These equations may have application in modeling the evolution of mineral compositions during differentiation of basaltic magmas, particularly terrestrial layered intrusions and the lunar magma ocean.

The partitioning of Sc, Y, and the rare earth elements between high-Ca pyroxene and natural mafic to intermediate lavas at 1 atmosphere

The effects of composition and temperature on the partitioning behavior of Sc, Y, and the rare earth elements (REEs) between high-Ca clinopyroxene and natural silicate melts were evaluated from doped experiments on natural mafic to intermediate composition lavas at 1 atmosphere pressure. Partition coefficients for these elements were found to be dependent on temperature and composition. The most important compositional parameters controlling clinopyroxene-melt partitioning for Y, Sc, and REEs are Al content of the liquid and pyroxene Ca content. Towards the goal of deriving expressions describing partitioning behavior, approximations were made of equilibrium constants for reactions involving a REE-Al component in the pyroxene. Regression of these equilibrium constants over the experimental temperature range (1180-1050°C) produced expressions which, when applied to the experimental glasses, reproduced the clinopyroxene trace element contents with precisions between 9 and 32% (1σ). The most important conclusion of this work is that pyroxene-melt partition coefficients for trivalent cations have different compositional dependencies than divalent cations because of the participation of Al in paired substitution. Values for high-Ca pyroxene Sc, Y, and REE partition coefficients have a range of over a factor of two between alkali basalts, andesites, and tholeiitic basalts (e.g., 0.2-0.7 for Sm at 1100°C). This represents a large proportion of the total range for D values from all mafic and intermediate magmas. Our contribution is to describe the parameters that control partitioning behavior. This will allow us to more accurately determine REE partitioning for specific systems.

High-field-strength element partitioning between pyroxene and basaltic to dacitic magmas

Abstract The effects of composition and temperature on pyroxene-melt partitioning of the high-field-strength elements (HFSE) — Ti, Zr, Nb and Ta — were evaluated from doped experiments on natural mafic to intermediate composition lavas at pressures from 0.1 MPa to 0.9 GPa (0.001 to 9 kbar). The HFSE partition coefficients (D) maintain similar relative relationships: DTi >DZr >DTa >DNb, but vary absolutely as a function of composition and temperature, often exhibiting a range of over a factor of 5 at a single temperature. For example, DZr ranges from 0.1 in a tholeiitic melt to 0.6 for a dacitic melt at 1100°C, 0.1 MPa. DZr, DTa and DNb for high- and low-Ca pyroxene can be described as linear functions of DTi. For high-Ca pyroxenes, the functions are DZr = 0.64DTi − 0.13, DTa = 0.14DTi − 0.02 and DNb = 0.04DTi − 0.01. The low-Ca pyroxene expressions are DZr = 0.60DTi − 0.06, DTa = 0.27DTi − 0.005 and DNb = 0.08DTi − 0.005. This linear relationship suggests similar substitution mechanisms for Ti and the other HFSE in both pyroxene and silicate melts. An expression was derived to calculate the Ti content of pyroxene based on the melt composition, Ca content of the pyroxene, temperature and pressure. This expression uses an approximation of the equilibrium constant for an exchange reaction of a Ti/Al-bearing component with a Ca-bearing component in the pyroxene. Over the experimental temperature range (1170-1070°C), the clinopyroxene Ti contents can be reproduced with a precision of ±20% (1σ).

Near-primary melt inclusions in anorthite phenocrysts from the Galapagos Platfrom

Abstract Partially crystalline melt inclusions in anorthite phenocrysts from a Galapagos seamount have been rehomogenized in a series of heating experiments in order to accurately determine their initial compositions. They are primitive tholeiites, high in CaO and Al 2 O 3 , and low in TiO 2 and alkalies. We infer that they were entrapped at a temperature close to 1270°C. At this temperature, the trapped melt composition is multiply saturated with olivine, plagioclase and spinel at 1 atmosphere pressure, but it is also near-primary requiring addition of less than 5% olivine plus or minus plagioclase and spinel in order to be in equilibrium with mantle olivine. The inclusions occur in association with, and are potentially parental to, a suite of MORB-like pillow lavas which range in MgO content from 10 to 8 wt% and which could have been derived by up to 40% fractionation of olivine plus plagioclase from the inclusion composition. Although the host anorthites are remarkably homogeneous and unzoned and the inclusions themselves are uniform in major element composition, minor element contents vary by a factor of two or more. Unlike many other plagioclase megacrysts, these anorthites must have crystallized under conditions in which the major element compositions were buffered but minor (and presumably trace) elements were free to vary. Such conditions could be achieved by liquid-solid interactions either in the melting regime or, perhaps more likely, in a liquid-crystal mush within the oceanic crust. These melt inclusions are not unique. They appear to belong to a class of primitive, high CaO and Al 2 O 3 MORB that occur in association with high-An plagioclase in a variety of oceanic settings, usually where magma supply is low. We propose that such magmas are derived from the shallowest, most depleted part of a mantle melt column, and that they constitute one end member of a spectrum of primary magmas which gives rise to the global array of MORB compositions.

Magnetite–melt HFSE partitioning

Abstract Results from doped, hydrous experiments on natural mafic-to intermediate-composition lavas at 2–5 kbar pressure were combined with existing 1 atm data to evaluate the effects of composition and temperature on the partitioning behavior of the high field strength elements (HFSE), Zr, Nb, Ta and Hf between magnetite and natural silicate melts. Magnetite composition was found to be the strongest controlling factor on partitioning behavior. The partition coefficients ( D ) for Zr, Nb, Hf, and Ta correlate with D Ti , Ti and Al content of the magnetite, temperature and pressure. The partition coefficients for the HFSE are similar to one another for any given magnetite–melt pair, but range from 2 in titanomagnetite. In addition, the relationship between Ti and the HFSE changes as a function of pressure and temperature, with the HFSE becoming more incompatible relative to Ti at lower temperatures and/or higher pressures. This change in the relationship between D Ti and D HFSE with temperature and pressure means that the expressions presented in Nielsen et al. (1994) [Nielsen, R.L., Forsythe, L.M., Gallaghan, W.E., Fisk, M.R., 1994. Major and trace element magnetite–melt partitioning. Chem. Geol. 117, 167–191.] are not valid for hydrous, aluminous systems. Expressions were derived to describe the relationship between D HFSE and temperature, pressure, Fe 2+ /Mg exchange, Ti/Al ratio of the magnetite, and D Ti . These expressions reproduce the input data within 35–50% (1 σ ) over a range extending from highly incompatible to compatible (

Geochemical characteristics of basaltic glasses from theamar andfamous axial valleys, Mid-Atlantic Ridge (36°–37°N): Petrogenetic implications

The axial valley of the Mid-Atlantic Ridge from 36° to 37°N was intensively sampled by submersible during thefamous andamar projects. Our research focussed on the compositional and isotopic characteristics of basaltic glasses from theamar valley and thenarrowgate region of thefamous valley. These basaltic glasses are characterized by: (1) major element abundance trends that are consistent with control by multiphase fractionation (olivine, plagioclase and clinopyroxene) and magma mixing, (2) near isotopic homogeneityδ18O= 5.2to6.4,87Sr/86Sr= 0.70288to0.70299 and206Pb/204Pb= 18.57to18.84, and (3) a wide range of incompatible element abundance ratios; e.g., within theamar valley chondrite-normalizedLa/Sm ranges from 0.7 to 1.5 andLa/Yb from 0.6 to 1.6. These ratios increase with decreasing MgO content. Because of the limited variations in isotopic ratios of Sr, Nd and Pb, it is plausible that these compositional variations reflect post-melting magmatic processes. However, it is not possible to explain the correlated variation in MgO content and incompatible element abundance ratios, such asLa/Sm andZr/Nb, by fractional crystallization or more complex processes such as boundary layer fractionation. These geochemical trends can be explained by mixing of parental magmas that formed by very different extents of melting. In particular, the factor of three variation in Ce content in samples with ∼ 2.1% Na2O and 8% MgO requires a component derived by < 1% melting. If the large variations in abundance ratios of incompatible elements reflect the melting process, a large, long-lived magma chamber was not present during eruption of theseamar lavas. The geological characteristics of theamar valley and the compositions ofamar lavas are consistent with episodic volcanism; i.e., periods of magma eruption were followed by extensive periods of tectonism with little or no magmatism.