Library of Experimental Phase Relations (LEPR): A database and Web portal for experimental magmatic phase equilibria data

Geochemistry Geophysics Geosystems

3

G

Technical Brief Volume 9, Number 3 14 March 2008 Q03011, doi:10.1029/2007GC001894

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

ISSN: 1525-2027

Click Here for

Full Article

Library of Experimental Phase Relations (LEPR): A database and Web portal for experimental magmatic phase equilibria data M. M. Hirschmann Department of Geology and Geophysics, University of Minnesota, 108 Pillsbury Hall, Minneapolis, Minnesota 55455, USA ([email protected])

M. S. Ghiorso OFM Research – West, 7336 24th Avenue NE, Seattle, Washington 98115, USA

F. A. Davis, S. M. Gordon, and S. Mukherjee Department of Geology and Geophysics, University of Minnesota, 108 Pillsbury Hall, Minneapolis, Minnesota 55455, USA

T. L. Grove, M. Krawczynski, E. Medard, and C. B. Till Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139-4307, USA

[1] We describe the Library of Experimental Phase Relations (LEPR), which is a new Web-based database of experiments pertaining to magmatic phase equilibria. The database includes information from experiments related to natural or near-natural magmatic silicate phase equilibria. At present, the database includes more than 6,600 experiments with liquids ranging in composition from komatiite to rhyolite performed at temperatures between 500 and 2500°C and at pressures up to 27.5 GPa. Here we describe the organization of data and metadata that are included in the database as well as the database structure and the implementation of its Web interface. We discuss a number of challenges and considerations related to construction of a database of experimental data that influence both the selection and presentation of the information incorporated in LEPR. Finally, we survey durations of experiments in LEPR and discuss the relationship between duration and the probability of crystal-liquid equilibrium in petrologic experiments. Components: 7687 words, 8 figures, 7 tables. Keywords: database; experimental petrology; igneous processes. Index Terms: 3630 Mineralogy and Petrology: Experimental mineralogy and petrology; 0525 Computational Geophysics: Data management; 3640 Mineralogy and Petrology: Igneous petrology. Received 12 November 2007; Revised 11 January 2008; Accepted 28 January 2008; Published 14 March 2008. Hirschmann, M. M., M. S. Ghiorso, F. A. Davis, S. M. Gordon, S. Mukherjee, T. L. Grove, M. Krawczynski, E. Medard, and C. B. Till (2008), Library of Experimental Phase Relations (LEPR): A database and Web portal for experimental magmatic phase equilibria data, Geochem. Geophys. Geosyst., 9, Q03011, doi:10.1029/2007GC001894.

Copyright 2008 by the American Geophysical Union

1 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

1. Introduction [2] In the past 40 years, experimental petrologists have collected large amounts of data pertaining to igneous phase equilibria. These data have been applied to a wide range of Earth and planetary science problems in the fields of petrology, geochemistry, geophysics, and tectonics. In many cases, these experimental data have found applications far beyond those that the original practitioners considered. For example, a study of partial melting of peridotite, applied originally to aid understanding of the origin of mid-ocean ridge basalts [e.g., Hirose and Kushiro, 1993], may later be applied to the petrology of an array of other lithologies [e.g., Francis et al., 1999; Mirnejad and Bell, 2006] and also may be of relevance to mantle trace element and isotopic evolution [e.g., Salters, 1996; Stracke et al., 2003; Huang et al., 2007] as well as to mantle flow, melt migration [e.g., Hart, 1993] and the seismological structure of the mantle and crust [e.g., Kelemen and Holbrook, 1995]. The same data may also be applied to calibration of empirical or theoretically based models of mantle melting, igneous phase equilibria, geothermometry, barometry and so forth or may be incorporated into parameterizations of petrologic phase relations [e.g., Hirschmann, 2000; Sugawara, 2000; Katz et al., 2003]. In many cases, the values of the data are increased when they can be compared to other experimental data or to compositions of natural rocks, glasses, or minerals. Therefore, experimental studies of mineral/melt phase equilibria have a collective impact that exceeds the specific problems addressed in the original studies, and hence they contribute to an overall understanding of the Earth and planetary sciences. As the quantity of such data increases and as the potential applications of the data proliferate, there is increasing need for practitioners to have ready access to large amounts of experimental data. [3] In the Earth sciences there is expanding appreciation of the value of readily accessible online databases of experimental and analytical data [Staudigel et al., 2003]. In recent years, databases have been constructed or are being constructed to compile chemical compositions of rocks and sediments, including PetDB [Lehnert et al., 2000], GEOROC [Schramm et al., 2006], SedDB [Lehnert et al., 2007], and NAVDAT [Carlson et al., 2001], as well as magnetic rock and mineral properties (MAGIC [Koppers et al., 2005]), geochronologic data (CHRONOS [Cox and Richard, 2005]), and others. In many respects, these database efforts are

10.1029/2007GC001894

modeled after the large databases of seismic data managed by IRIS [Ahern, 2003], which have proved so important to geophysical studies. Collectively, the databases are an essential aspect of the relatively new endeavor known as Geoinformatics. Here we present a description of LEPR (Library of Experimental Phase Relations), an online database of experimentally determined igneous phase relations. [4] The LEPR database is a descendant of earlier databases compiled for the purposes of calibration of thermodynamic models of magmatic silicate liquids [Ghiorso and Carmichael, 1980; Ghiorso et al., 1983, 2002; Ghiorso and Sack, 1995]. The original intended purpose for LEPR was to update the database for the latest version of MELTS, known as xMELTS [Ghiorso et al., 2007]. However, during the compiling of this database, we recognized the broader impact of a publicly available database for magmatic phase equilibria. Further, we recognized that database applications, such as calibration of thermodynamic or empirical models, would benefit from inclusion of analytical and experimental uncertainties and other metadata, including experimental duration or analytical technique, which allow users to evaluate data quality. Consequently, rather than simply add data to previously compiled compendia of phase relations used to calibrate past versions of MELTS, we returned to the original publications to compile LEPR. [5] As mentioned above, LEPR is being used to calibrate the xMELTS thermodynamic model for calculation of igneous phase equilibria. But importantly, it is our hope that there may be a wide range of additional applications for LEPR. Public availability of LEPR may facilitate calibration of competing phase equilibria models, based on alternative thermodynamic formulations or on other empirical relations. Additionally LEPR may be used to do the following: [6] 1. Calibrate thermometers and barometers applicable to igneous phase equilibria. [7] 2. Calibrate models for partial melting of key lithologies, ranging from peridotite to amphibolite. [8] 3. Calibrate models for igneous liquid lines of descent. [9] 4. Compare rocks sampled in the field with experiments performed on similar bulk composition or with experiments that produced similarcomposition product phases. 2 of 15

Geochemistry Geophysics Geosystems

3

G

10.1029/2007GC001894

hirschmann et al.: library of experimental phase relations

Table 1. Primary Data in LEPR Description T P Phase Composition Comp-unc Fraction XH2O XCO2

Temperature, with uncertainty Pressure, with uncertainty Phase(s) produced during the experiment (for current list of entries, see legend of Figure 1) Composition of phases produced during an experiment, in wt%, with uncertainties. One or more of SiO2, TiO2, Al2O3, Cr2O3, FeO, Fe2O3, MnO, MgO, CaO, Na2O, K2O, NiO, CoO, P2O5, BaO, Li2O, SO3, H2O, SrO, REE oxides, V2O5, F, Cl, CO2 Uncertainties (as reported) on phase composition Proportional abundance of phase expressed as a mass fraction Mole fraction of H2O in the fluid phase Mole fraction of CO2 in the fluid phase

[10] 5. Evaluate phase equilibria properties, such as conditions of multiple saturation or solidus locations, of rocks or glasses sampled in the field. [11] 6. Plan new experimental studies and compare new experimental data to published experiments. [12] In this note, we document the schema behind LEPR and describe the scope of the database in its present form. We also discuss some of the considerations in evaluating the quality of the database and some of the obstacles associated with constructing a faithful representation of igneous phase equilibria based on published experimental data.

2. Scope and Database Schema [13] LEPR includes data from experiments documenting igneous phase equilibria of natural silicate rocks and of compositions with chemical complexity similar to that of natural silicate rocks. Experimental studies in simple systems, such as CaOMgO-Al2O3-SiO2 (CMAS) are not included at this time. Inclusion of such data is desirable, but incorporation of the extensive body of experiments on CMAS would require a considerable effort. Experimental studies conducted chiefly at subsolidus conditions are not included. [14] Entries of each experiment in the database include both experimental data and metadata, which is information about experimental methodologies. Data are entered as recorded in published papers and therefore not all data are available for all experiments. For example, some studies focused only on the identity of phases present as a function of temperature or pressure and report little or no data on the phase compositions. In some cases, the database includes some derivative information, which is information about experimental conditions or products that is inferred and not measured directly.

2.1. Experimental Data [15] Experimental data incorporated into LEPR include the composition of starting materials, the temperature and pressure of the experiments, the identity of the phases present at the end of the experiments and the compositions of those phases. In some cases, mass fractions of phases present in an experiment, measured directly by image analysis or point counting, are also reported. Additionally, the uncertainties in all of these quantities are included, as reported by the original authors. When experimental oxygen fugacities have been controlled by gas mixing at one atmosphere, they are also included. The primary data fields included in LEPR are summarized in Table 1.

2.2. Metadata [16] The metadata included in LEPR are summarized in Table 2. They include the laboratory where the experiments were performed and the names of the authors and the full citation of the paper from which the data were published. Also included are

Table 2. Metadata in LEPR Description Log10fO2 Author Lab Device Container Method ID Duration Phases Bulk Bulk-unc

Oxygen fugacity (fO2) or oxidation state of experiment; fO2 brackets; fO2 buffer Author, Title, Full journal citation, DOI Laboratory where experiment was done Experimental apparatus Sample container during experiment Analytical method(s) Sample name (author defined) Experiment duration Phase assemblage produced, even if analytical data are not reported Bulk composition of starting materials (major elements, in wt%; see list in Table 1.) Uncertainties (as reported) on bulk composition 3 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

Table 3. Controlled Vocabulary for Nonnumerical Log10fO2 Constraints

10.1029/2007GC001894

Table 5. Controlled Vocabulary for Experimental ‘‘Container’’ Metadata Entries

Entry

Explanation

Entry

Explanation

CCO Cu + Cu2O GCH HM IW MMO MW NNO QFM

Graphite-CO-CO2 buffer Copper-copper oxide buffer Graphite-methane buffer Hematite-magnetite buffer Iron-wu¨stite buffer MnO-Mn3O4 buffer Magnetite-wu¨stite buffer Nickel-nickel oxide buffer Quartz-fayalite-magnetite buffer

Ag AgPd Au AuPd AuPd/Pt

Silver capsule Silver-palladium alloy capsule Gold capsule Gold-palladium alloy capsule Double capsule of platinum surrounding one of gold-palladium Graphite capsule Iron capsule Platinum capsule ‘‘pretreated’’ with iron Molybdenum capsule Double capsule of platinum surrounding one of olivine Platinum capsule (or platinum wire at one atmosphere) Double capsule of platinum surrounding one of graphite Platinum-rhodium alloy capsule Rhenium capsule

the type of experimental apparatus used for the experiment, such as a piston cylinder apparatus or cold-seal pressure vessel, the experimental container, the duration of the experiment, and the analytical techniques used to determine phase compositions. A controlled vocabulary for description of LEPR metadata is provided in Tables 3, 4, 5, and 6.

2.3. Derivative Information [17] Derivative information are quantitative inferences about the experiments that are not actually measured from the experiments, but which have been inferred indirectly. In some experiments, these may include the oxygen fugacity prevailing during the experiment, the proportions of phases present at the end of the experiment, and the volatile concentrations of liquids quenched from the experiments. [18] In some experiments, the oxygen fugacity was not strictly controlled with a gas mixture, but the experimental conditions place constraints on the prevailing fO2. For example, experiments conducted in double capsules with a buffering assemblage such as quartz, fayalite, and magnetite or Ni and NiO. Conversion of these buffers to oxygen fugacity at a particular temperature and pressure requires a calculation based on published thermo-

C Fe FePt Mo Oliv/Pt Pt PtC PtRh Re

dynamic data, as summarized in Table 7. Similarly, high-pressure experiments conducted in graphite capsules must reside at or below the oxygen fugacity defined by CCO, which is equilibrium between graphite and coexisting CO-CO2 vapor. In LEPR, such experiments are assumed to reside at CCO (Table 7), though in fact they are likely to be more reduced [Holloway et al., 1992]. [19] Experiments conducted in the presence of volatiles can cause special problems for recording

Table 6. Controlled Vocabulary for Analytical ‘‘Method’’ Metadata Entries Entry CO2 EDS EMP FTIR

Table 4. Controlled Vocabulary for Experimental ‘‘Device’’ Metadata Entries Entry

Explanation

1-atm Belt CSPV IHPV MA PC

One atmosphere furnace Belt apparatus Cold-seal pressure vessel Internally heated pressure vessel Multianvil apparatus Piston cylinder apparatus

ICP-MS IR KFT LECO SEM SIMS XRD

Explanation Carbon dioxide content determined by coulometric titration Energy dispersive X-ray analysis on a scanning electron microscope or electron microprobe Wavelength dispersive X-ray analysis on an electron microprobe Volatiles determined by Fourier-transform infrared spectroscopy Inductively coupled plasma mass spectroscopy Volatiles determined by infrared spectroscopy Karl Fischer titration method used for water analysis Total carbon determined using a LECO Carbon Analyser Scanning electron microscope Secondary ionization mass spectrometry (Ion microprobe) X-ray diffraction 4 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

Table 7. Oxygen Fugacity Buffers: log10 fO2 = A/T + B + C (P

10.1029/2007GC001894

1)/T + D ln (T)a

Algorithm Buffer CCO Cu + Cu2O HM IW MMO MW NNO QFM

A 21803 18162.2 25632 28776.8 29420.7 30396 25018.7 30686

B

C

4.325 12.855 14.620 14.057 92.025 3.427 12.981 82.760

0.171

D 0.6741

0.019 0.055 0.083 0.046 0.094

0.8853 11.517 2.0236 0.5117 10.620

Reference Jakobsson and Oskarsson [1994] O’Neill and Pownceby [1993a] Haas and Robie [1973] O’Neill and Pownceby [1993a] +0.00531848 T O’Neill and Pownceby [1993b] O’Neill [1988] O’Neill and Pownceby [1993a], Frost [1991] +0.004843 T O’Neill [1987]

a T in K, P in bars. Pressure terms (C = DV/(Rln(10))) are from Chou [1987] and are based on 0.1 MPa molar volume data of solids compiled by Robie et al. [1979].

and interpreting quenched liquid compositions. In some cases, volatile concentrations have been measured directly by FTIR, SIMS, or other methods, in which case, they are reported as data. In other cases, volatile concentrations have not been measured directly, but the experiments are known to be or assumed to be saturated with an H2O or CO2 vapor. In these cases, dissolved volatile concentrations can be calculated by assuming that the solubility is known at the temperature and pressure and melt composition, although this potentiality is not implemented in LEPR at this time. On the other hand, experiments conducted in the presence of mixed H2O-CO2 volatiles may be even more problematic, as the initial volatiles introduced during the experiment partition differentially between the vapor, the silicate melt, and/or other possible phases. Consequently, the composition of the equilibrium vapor may not be known and it is correspondingly more difficult to estimate volatile concentrations in the melt. Finally, in some volatile-bearing experiments that are not vapor-saturated, the only constraint on volatile concentrations in the liquid comes from microprobe totals. In these cases, we may use the deficits from 100% of those microprobe totals as an estimate of the concentration of dissolved volatiles, but the estimate is quite uncertain. [20] Many studies report mass fractions of phases present during experiments that have been calculated on the basis of mass balance between the bulk composition of the experiment and the compositions of the analyzed phases. When such proportions have been calculated by the authors, they are included in LEPR. On the other hand, when mass balances have not been reported, they are not recorded, though they potentially may be calculated by interested users from reported bulk proportions and phase compositions.

[21] Studies documenting the solubility of volatiles such as H2O or CO2 generally report the composition of the starting material, such as a basalt or rhyolitic glass, but apart from measured volatile concentrations, these studies commonly do not report the composition of the glass quenched from each experiment. In such cases, we have inferred that the glass composition is similar to that of the starting material, except that it also contains the volatile component reported from each experiment.

2.4. Comment on Equilibrium, Data Quality, and Database Accuracy [22] Importantly, not all experiments attempting to document igneous phase relations achieve equilibrium. Experimental durations of published experiments are not always sufficient to allow phases to come to equilibrium. Further, some experiments are subject to open-system processes such as reaction with containers or loss of volatiles from containers. Thus, experiments that are performed on a nominally volatile-free basis may nonetheless contain unquantified amounts of unexpected H2O, CO2 and other volatile species [Gaetani and Grove, 1998; Medard et al., 2008]. Although many studies present compelling evidence indicating that experiments have approached equilibrium, most experimental igneous phase relation studies are not reversed, meaning that evaluation of equilibrium criteria are commonly subjective. Rather than assuming whether particular experiments have achieved equilibrium, we have reserved judgment about equilibrium or data quality during compilation of LEPR. Rather, the metadata included in LEPR, such as experimental durations, container type and so forth, may allow users to make such judgments. Inclusion of data from a particular paper in LEPR does not indicate that the authors of this paper have made any judgment regarding 5 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

10.1029/2007GC001894

Figure 1. Numbers of experimental phases tabulated in LEPR (4262 silicate liquids are not shown). Phase names are followed by a value in parentheses that indicates the number of database entries as of June 2007: amph, amphibole (269); ap, apatite (51); carb, carbonates (14); cpx, clinopyroxene (1577); cor, cordierite (6); fld, plagioclase (1270) and potassium feldspar (127), fluid (H2O, CO2, or mixed H2O-CO2 solution) (1545); grn, garnet (418); ilm, ilmenite (135); ky, kyanite (31); leu, leucite (19); mel, melilite (15); mica, biotite (172), muscovite (25), and phlogopite (2); nph, nepheline (14); olv, olivine (1585); opx, orthopyroxene (1080); per, (Fe2+, Mg)-perovskite (7); qtz, quartz (80); rut, rutile (25); sil, sillimanite (14); spn, spinel (739) and oxide (structural state not specified) (156); stis, stishovite (13); whit, whitlockite (62); zoi, zoisite (25). Other: armalcolite (2), unidentified Ca-Al-Si phase (5), Ca-perovskite (8), carbonate liquid (9), chevkinite (3), clay (1), coesite (1), corundum (4), cristobalite (1), epidote (4), ferrobustamite (1), ferropericlase (6), hollandite (4), montdorite (1), unidentified Na-K-Al-Si phase (3), sodalite (8), staurolite (5), sulfide (6), unidentified Ti-Si phase (1), titanite (7).

the quality of experiments in LEPR. In many cases, full evaluation of data quality requires careful reading of the original published papers, and we urge users to evaluate for themselves published studies within LEPR. [23] Finally, in all likelihood the LEPR database has errors. Some of these errors may derive from typographical or other types of errors in original publications. For example, we encountered literally hundreds of cases where the totals of microprobe analyses were markedly different (>0.3 wt.%) from tabulated totals. In such cases, we entered the original oxide values published in the tables unless clear errors in specific entries could be deduced. Other errors may have arisen owing to the myriad idiosyncratic formats in which experimental data are published, with key information in tables, table captions, online supplements, experimental methods and results text sections. Also, some data are occasionally corrected or revised in either published errata or in subsequent publications. This complexity naturally leads to some confusion dur-

ing data entry, and may have lead to errors. Finally, it is likely that we made some data entry errors during compilation of LEPR. We encourage users who encounter errors to notify us and there is a link on the LEPR home page to facilitate this.

3. Database Description [24] As of this writing (LEPR release 1.2.0, June 2007) LEPR contains experimental data extracted from 196 published studies, with work performed in 63 different laboratories. There are 6,663 distinct experiments tabulated, with major element compositional data on 4,262 liquids and a broad array of 9,612 additional phases as illustrated in Figure 1. The database will continue to grow as more experimental data become available, and it is possible for experimentalists to submit their published data for inclusion in the database through templates available on the LEPR Web portal. [25] The data are structured as a relational database with 105 tables. Figure 2 illustrates the mapping of 6 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

10.1029/2007GC001894

Figure 2. Table structure of the LEPR relational database. Table names are indicated in violet. Table entries in blue refer to database index variables. Metadata are indicated in red and are explained in Table 2. The controlled vocabularies associated with these metadata are provided in Tables 3, 4, 5, 6, and 7. Labels for data are shown in black, and entries are defined in Table 1. The ‘‘derived values’’ in the log10fO2 table are computed from the numerical value or buffer constraint (Table 7) on the oxygen fugacity entered during compilation. The derived values are (1) a numerical value of the log10fO2 constraint, (2) a minimum and (3) maximum log10fO2 value, and (4) – (11), log10fO2 buffer values for CCO, CuCuO (Cu-CuO), HM, IW, MMO (MnO-Mn3O4), MW, NNO, and QFM (see Table 7).

the database schema onto the table structure. Every experimental entry is uniquely identified in the ‘‘Experiment’’ table and cross-linked to an entry in the ‘‘Author’’ and ‘‘Laboratory’’ tables. More than one laboratory can be associated with a given experiment. Compositional data for phases associated with a given experiment are tabulated in a pair of tables, labeled generically in Figure 2 as ‘‘Phase’’ and ‘‘Phase Unc,’’ but in practice given names like ‘‘Olivine’’ and ‘‘Olivine_errors.’’ A complete list of phase table names is provided at the LEPR Web portal (see below); phase tables are added as necessary to accommodate database updates. The table labeled ‘‘log fO2’’ in Figure 2 contains primarily derived data (see above) related to oxygen fugacity constraints. A more detailed discussion of these derived quantities is provided in the figure legend. All of the entries labeled in blue in Figure 2 are arbitrary indices whose pur-

pose is to bind the relational database structure. They are not data or metadata and generally of no interest to database users. [26] There are three versions of the table containing compositions of experimental liquids in LEPR. The version that represents data as reported in the published literature is named ‘‘Liquid.’’ Two other versions, ‘‘LiquidAnhy’’ and ‘‘LiquidHydr’’ are derivative tables, in the sense discussed previously. Entries in all three tables are identical if the experiment associated with the liquid is volatile free. Because the conventions for reporting the compositions of volatile-bearing experimental liquids are not standardized (see above) the raw data (in the table Liquid) may be difficult to compare for volatile-bearing compositions. Consequently, we have prepared the table LiquidHydr to report the compositions of all volatile-bearing 7 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

10.1029/2007GC001894

Figure 3. Temperatures and pressures plotted for experimental data in the LEPR database. Uncertainties are not indicated.

liquids in such a way that the sum of all oxides, including H2O and CO2, sums to the reported or inferred analytical total. By contrast, the table LiquidAnhy reports liquid compositions for volatile-bearing liquids to sum to 100% on a volatilefree basis, with volatile concentrations provided (if reported) in addition to this sum. The data in LiquidAnhy allows users to compare liquid oxide concentrations on a volatile-free basis. Users of LEPR can choose which of these derived liquid data tables to use for their inquiries. The default access corresponds to the table LiquidHydr. [27] Experiments span a broad range of temperature-pressure space, as illustrated in Figure 3. Reported experimental liquid compositions reflect an equally broad spectrum of compositions, with an expected concentration of results in mafic systems, principally basalts, but with excellent representation in highly silicic systems and more alkali-rich magma types (Figure 4). In addition to experimental data on a wide range of liquids, LEPR documents experimental work on mineral solid solutions that exhibit extreme compositional variation. The data plotted from LEPR in Figure 5 is an illustration of this variability, where a simple MgO versus SiO2 diagram is used to demonstrate compositional variation in tabulated olivines, pyroxenes, garnets and amphiboles and liquids.

Our point in presenting Figures 3–5 is only to demonstrate that LEPR is an excellent resource for investigating the systematic variation of the compositions of magmatic phases, and for evaluating hypotheses regarding compositional covariation in response to changes in intensive thermodynamic variables. Although the data in LEPR do not provide a comprehensive account of experimental relations in all igneous rocks, they comprise a thoroughly representative account of the experimental literature on igneous phase relations of natural systems. LEPR’s deficiencies in coverage reflect those of the literature, which in turn reflect the complexities of conducting experiments and the scientific interests of the experimental community.

4. User Interface to LEPR [28] LEPR is hosted as a mySQL relational database served at ofm-research.org. There are several methods of accessing and querying the LEPR database, including (1) a Web portal, (2) direct SQL queries to the open port on the database server, and (3) indirect SQL queries generated by third party software designed to provide a user interface to database inquiry (e.g., Microsoftk Office and CocoaMySQL). 8 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

10.1029/2007GC001894

Figure 4. Liquid compositions plotted for experimental data in the LEPR database. Uncertainties are not indicated. The space is partitioned according to the igneous rock classification diagram of Le Maitre et al. [2002]: 1, foidite; 2, picro-basalt; 3, basalt; 4, basaltic andesite; 5, andesite; 6, dacite; 7, rhyolite; 8, trachyte/trachydacite; 9, phonolite; 10, tephriphonolite; 11, phonotephrite; 12, tephrite basanite; 13, trachybasalt; 14, basaltic trachyandesite; 15, trachyandesite.

Figure 5. Concentrations of MgO and SiO2 of various phases in the LEPR database, illustrating the span of compositions included. Uncertainties are not indicated. 9 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

10.1029/2007GC001894

Figure 6. LEPR Web portal page.

[29] The Web portal to LEPR has the URL http:// lepr.ofm-research.org. The portal permits a user to generate an account to access the database and offers a number of options for constructing queries to search LEPR, to download data subsets as Excel workbooks or XML files, and to plot LEPR data in a variety of formats. The initial portal page is illustrated in Figure 6. At the most primitive level, the user can display all experiments performed at a particular laboratory, or all experiments from a particular literature reference. At a more advanced level, a user can restrict searches based upon criteria such as experimental technique and run duration, or can specify temperature and pressure

ranges or the presence of certain phases in run products (including the compositions of those phases). A portion of the advanced search Web page is illustrated in Figure 7. The advanced search options, which permit selection on the basis of both data and metadata, provide enormous flexibility in defining data subsets. Results of searches are displayed in either tabular form or graphically. In the tabular format results are organized in a summary table and links allow details of each experiment to be examined. Experiments can be selected from the summary table for download as Excel workbooks or XML text files. The Excel workbooks are organized with multiple worksheets mimicking

Figure 7. Web form for configuring an ‘‘Advanced Search’’ at the LEPR Web portal. 10 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

the internal structure of the relational database, as described above. The XML-based download files follow a standardized schema, which is available for retrieval at the Web portal. Advanced search results can be displayed on simple X-Y graphs where the abscissa and ordinate are user selectable. These graphs may be downloaded in a variety of formats, including two bitmap (PNG, JPG) and two resolution independent (SVG, and PDF) versions. [30] The Web portal allows users to access the most recent version of the database or previous releases; the former is the default. This feature permits users to lock the results of a LEPR query to a specific version (or time stamp) regardless of subsequent additions of new entries to LEPR. Consequently, publication of results derived from LEPR may be characterized by version number and query description with the knowledge that subsequent inquiries will generate identical data subsets. [31] Alternative to the Web portal, LEPR may be accessed directly at port 3306 at http://www.ofmresearch.org. The database can be interrogated on this open port using SQL queries produced from third party interface software, like CocoaMySQL or MySQL Query Browser or from user written software that rely on open source database drivers that are packed with the MySQL distribution. Commonly, software for this purpose is written in C, C++ or Java. Access is restricted to users with established LEPR user accounts, previously created at the Web portal. [32] Finally, LEPR may be queried from Microsoftk Office (usually from Excel) running on Windows XP and Vista Intel platforms. Results of queries to LEPR may be displayed directly and manipulated in Excel workbooks. In principle, a graphical user interface to LEPR could be developed in the context of an Excel portal. Further documentation on this mode of database access may be obtained at the Microsoftk Office Web site.

5. Discussion 5.1. Included and Excluded Metadata [33] Construction of LEPR required certain choices regarding the metadata from each experiment that should and should not be included. The guiding philosophy in designing LEPR was to include sufficient information to allow users to evaluate the quality and applicability of data without making collation intractable. Not all of these choices

10.1029/2007GC001894

were straightforward and reasonable practitioners could have adopted different strategies. Experiments pertaining to magmatic phase equilibria are conducted for a wide range of motivations, methodologies, and with a diversity of experimental philosophies. They are also published in different journals that provide different opportunities to include methodological detail. The result is a diverse literature that is rife with idiosyncrasies and no database can hope to capture all of the important facts required to evaluate the applicability of all experiments. Such evaluation will always require reading the original published literature. [34] There are a number of metadata not incorporated in LEPR that merit some additional discussion. For example, there are no pressure calibration data and we have not compiled the type of thermocouple employed for the experiment. Such information might better allow users to account for experimental differences resulting from interlaboratory inconsistencies in the methods of temperature and pressure estimates. On the other hand, the laboratory location for each experiment is recorded, which may facilitate consideration of any such inconsistencies. [35] Additionally, LEPR does not include metadata on the type of starting material employed. Some experiments are conducted with powders of natural rock materials, whereas others are conducted with glasses, gels, mixtures of oxides, mixtures of natural minerals, or some combination of these. Experimental results and/or times required to achieve equilibrium may vary accordingly. Also, some experiments may be conceived to investigate reaction between minerals and melts or between rocks and melts, and these experiments result in partial or local equilibration being achieved only in certain regions of the charge. In general, it has been difficult to incorporate such experiments in the LEPR schema; in some cases they are not included, in others they are entered in such a fashion that the data may potentially reflect equilibrium assemblages. Further, LEPR does not specify when experiments begin with compositionally heterogeneous sample geometries, including the so-called sandwich configuration [Stolper, 1980]. Neither can LEPR provide spatial information about phase occurrences in capsules, as is especially important for experiments conducted in strong temperature gradients [Zhang and Herzberg, 1994]. Finally, LEPR does not incorporate information when experiments were conducted under variable temperature conditions; i.e., in some cases experiments 11 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

10.1029/2007GC001894

Figure 8. Experimental duration (in hours) plotted against inverse temperature for experimental data compiled in LEPR. All experiments are plotted as black circles. Colored circles discriminate phases. The lines on the ‘‘olivine’’ panel refer to characteristic times for Fe2+/Mg diffusion over 10 microns for wet (blue) [Hier-Majumder et al., 2005] and dry (red) [Chakraborty, 1997] conditions. See text for discussion.

are ‘‘reversed,’’ with temperature held above or below the temperature of interest for a period of time, and then a cooling or heating step leads to a dwell at the target temperature [Baker and Stolper, 1994]. In LEPR, only the temperature and duration of the final equilibration step are recorded.

5.2. Applications of Metadata: Attainment of Equilibrium [36] One of the features of the LEPR database is that it allows exploration of the variety of experimental procedures prevailing in the experimental petrology community. An important aspect of an experimental protocol is the duration of heating. For those experiments seeking to document magmatic phase equilibria, the experimental duration should be sufficient to achieve equilibrium between the minerals and the melt. [37] In many individual studies, experimental time series are conducted to establish the time required for experiments to achieve steady state, which may be an indication of approach to equilibrium. More broadly, experimental durations may be selected for a host of other reasons. For example, steady temperature and pressure conditions cannot be maintained indefinitely in many experimental devi-

ces. Also, thermocouples may oxidize or otherwise change their characteristics during long duration experiments, leading to temperature drift in solidmedia apparatuses. Additionally, open-system effects may arise during long experiments owing to reaction with containers, volatile loss through containers or exhaustion of volatile or oxygen fugacity buffers. In some cases, experimental durations may be too short to permit mineral/melt equilibration. [38] As illustrated in Figure 8, durations of mineral/silicate liquid experiments vary by more than 2 orders of magnitude at a given temperature. Depending on the object of the experiment and on the dimensions of the crystals present, as well as on the details of the chemical composition of the experimental charge, including the composition of the melt, and the presence or absence of volatiles, significantly different experimental run durations may be justifiable. However, the careful consumer of data in LEPR may wish to compare the durations of experiments of interest to those summarized in Figure 8. When durations are on the shorter end of the span of times in Figure 8, there is greater incentive to investigate whether the experiments achieved equilibrium. 12 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

[39] An interesting aspect of the body of experimental data is that virtually all experiments are short relative to characteristic equilibration times of the crystals typically present. In Figure 8, experimental durations for experiments containing olivine and silicate liquid are compared to the characteristic diffusion time for Fe-Mg equilibration of 10 micron crystals (defined as x2/2D, where x is the characteristic crystal width and D is the diffusion coefficient for Fe-Mg interdiffusion) under dry conditions [Chakraborty, 1997] and at conditions of 300 MPa H2O pressure [Hier-Majumder et al., 2005]. Calculated equilibration times would be orders of magnitude longer for virtually any other major element component of common igneous minerals, such as Ca-Mg diffusion or Al diffusion in pyroxene or CaAl-NaSi diffusion in feldspar [Brady and Yund, 1983; Brady and McCallister, 1983; Grove et al., 1984; Sautter et al., 1988]. [40] The short duration of many experiments compared to crystal equilibration times could suggest that a large number of experiments have been too brief to allow Fe-Mg equilibration of olivine crystals with melts and are far too brief to allow equilibration with the interiors of pyroxenes, feldspars, garnets, and other common igneous minerals. Fortunately, in many cases, the rate controlling process for establishing mineral/melt equilibrium is not lattice diffusion but the much more rapid diffusion in the liquid from which the crystal is growing [Kirkpatrick, 1981]. Also, though it is often not stated, many experimentalists analyze the edges of mineral grains that are directly in contact with the silicate melt. So, the entire grain may not be in equilibrium with the melt, but it is assumed that the outer few microns approach an equilibrium value. For partial melting studies employing natural rocks or minerals as starting materials, crystal-melt equilibration does occur in part by lattice diffusion, but is also promoted by more rapid dissolution-reprecipitation processes [Lo Cascio et al., 2004]. Thus, mineral-melt equilibrium may have been achieved in many magmatic phase equilibria experiments, but the short duration of experiments compared to some transport processes underlines the need for caution when planning and interpreting these studies.

6. Recommendations on Citing LEPR Query Results [41] When citing data compilations generated from querying the LEPR database, it is our recommendation that the original experimental papers be

10.1029/2007GC001894

referenced. Occasionally space considerations or editorial policy may limit this practice, but every effort should be made to credit primary sources. And, it should be borne in mind that references cited in electronic supplements are not properly credited citations to an author’s work. [42] We anticipate that the majority of LEPR users will utilize the UI portal and download an Excel workbook of data subsets. In addition to documenting the exact query specification (as well as LEPR version number), we recommend that these workbooks should be designated as electronic supplements to any publication resulting from analysis of these data.

Acknowledgments [43] We are grateful for the helpful comments of the referees, Keith Putrika, John Longhi, and Othmar Mu¨ntener, as well as those of G-Cubed editors Vincent Salters and Glenn Gaetani. Material support for this investigation was provided by the National Science Foundation through grants EAR-0608532 (Ghiorso), EAR-0439800 (Hirschmann), and EAR-0440045 (Grove). We apologize to those reviewers, readers, and database users who do not appreciate the name ‘‘LEPR,’’ but the movement toward geochemical databases was catalyzed in part by a GERM, and so it is only natural that one of the results should be a disease.

References Ahern, T. K. (2003), The FDSN and IRIS Data Management System: Providing easy access to terabytes of information, in The International Handbook of Earthquake and Engineering Seismology, Part B, pp. 1645 – 1657, Academic, New York. Baker, M. B., and E. M. Stolper (1994), Determining the composition of high-pressure mantle melts using diamond aggregates, Geochim. Cosmochim. Acta, 58, 2811 – 2827. Brady, J. B., and R. J. McCallister (1983), Diffusion data for clinopyroxenes from homogenization and self-diffusion experiments, Am. Mineral., 68, 95 – 105. Brady, J. B., and R. A. Yund (1983), Interdiffusion of K and Na in alkali feldspars—Homogenization experiments, Am. Mineral., 68, 106 – 111. Carlson, R. W., D. Walker, R. Black, A. Glazner, L. Farmer, and J. Grossman (2001), NAVDAT: A western North American volcanic and intrusive rock geochemical database, Geol. Soc. Am. Abstr. Programs, 33(6), 175. Chakraborty, S. (1997), Rates and mechanisms of Fe-Mg interdiffusion in olivine at 980 – 1300°C, J. Geophys. Res., 102, 12,317 – 12,331. Chou, I.-M. (1987), Oxygen buffer and hydrogen sensor techniques at elevated pressures and temperatures, in Hydrothermal Experimental Techniques, edited by G. C. Ulmer and H. C. Barnes, pp. 60 – 99, John Wiley, New York. Cox, S. J. D., and S. M. Richard (2005), A formal model for the geologic time scale and global stratotype section and point, compatible with geospatial information transfer standards, Geosphere, 1, 119 – 137. 13 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

Francis, D., J. Ludden, R. Johnstone, and W. Davis (1999), Picrite evidence for more Fe in Archean mantle reservoirs, Earth Planet. Sci. Lett., 167, 197 – 213. Frost, R. B. (1991), Introduction to oxygen fugacity and its petrologic importance, in Oxide Minerals: Petrologic and Magnetic Significance, Rev. Mineral., vol. 25, edited by D. Lindsley, pp. 1 – 9, Mineral. Soc. of Am., Washington, D. C. Gaetani, G. A., and T. L. Grove (1998), The influence of water on melting, Contrib. Mineral. Petrol., 131, 323 – 346. Ghiorso, M. S., and I. S. E. Carmichael (1980), A regular solution model for met-aluminous silicate liquids: Applications to geothermometry, immiscibility, and the source regions of basic magmas, Contrib. Mineral. Petrol., 71, 323 – 342. Ghiorso, M. S., and R. O. Sack (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, Contrib. Mineral. Petrol., 119, 197 – 212. Ghiorso, M. S., I. S. E. Carmichael, M. L. Rivers, and R. O. Sack (1983), The Gibbs free energy of mixing of natural silicate liquids: An expanded regular solution approximation for the calculation of magmatic intensive variables, Contrib. Mineral. Petrol., 84, 107 – 145. Ghiorso, M. S., M. M. Hirschmann, P. W. Reiners, and V. C. Kress III (2002), The pMELTS: A revision of MELTS for improved calculation of phase relations and major element partitioning related to partial melting of the mantle to 3 GPa, Geochem. Geophys. Geosyst., 3(5), 1030, doi:10.1029/ 2001GC000217. Ghiorso, M. S., M. M. Hirschmann, and T. L. Grove (2007), xMELTS: A thermodynamic model for the estimation of magmatic phase relations over the pressure range 0 – 30 GPa and at temperatures up to 2500°C, Eos Trans. AGU, 88(52), Fall Meet. Suppl., Abstract V31C-0608. Grove, T. L., M. B. Baker, and R. J. Kinzler (1984), Coupled CaAl-NaSi diffusion in plagioclase feldspar—Experiments and applications to cooling rate speedometry, Geochim. Cosmochim. Acta, 48, 2113 – 2121. Haas, J. L., Jr., and R. A. Robie (1973), Thermodynamic data for wu¨stite, Fe0.947O, magnetite, Fe3O4, and hematite, Fe2O3, Eos Trans. AGU, 54, 483. Hart, S. R. (1993), Equilibration during mantle melting—A fractal tree model, Proc. Natl. Acad. Sci. U. S. A., 90, 11,914 – 11,918. Hier-Majumder, S., I. M. Anderson, and D. L. Kohlstedt (2005), Influence of protons on Fe-Mg interdiffusion in olivine, J. Geophys. Res., 110, B02202, doi:10.1029/2004JB003292. Hirose, K., and I. Kushiro (1993), Partial melting of dry peridotites at high pressures—Determination of compositions of melts segregated from peridotite using aggregates of diamond, Earth Planet. Sci. Lett., 114, 477 – 489. Hirschmann, M. M. (2000), Mantle solidus: Experimental constraints and the effects of peridotite composition, Geochem. Geophys. Geosyst., 1(10), doi:10.1029/2000GC000070. Holloway, J. R., V. Pan, and G. Gudmundsson (1992), Highpressure fluid-absent melting experiments in the presence of graphite: Oxygen fugacity, ferric/ferrous ratio and dissolved CO2, Eur. J. Mineral., 4, 105 – 114. Huang, S. C., M. Humayun, and F. A. Frey (2007), Iron/manganese ratio and manganese content in shield lavas from Ko’olau Volcano, Hawai’i, Geochim. Cosmochim. Acta, 71, 4557 – 4569.

10.1029/2007GC001894

Jakobsson, S., and N. Oskarsson (1994), The system C-O in equilibrium with graphite at high pressure and temperature: An experimental study, Geochim. Cosmochim. Acta, 58, 9 – 17. Katz, R. F., M. Spiegelman, and C. H. Langmuir (2003), A new parameterization of hydrous mantle melting, Geochem. Geophys. Geosyst., 4(9), 1073, doi:10.1029/2002GC000433. Kirkpatrick, R. J. (1981), Kinetics of crystallization of igneous rocks, Rev. Mineral., 8, 321 – 395. Kelemen, P. B., and W. S. Holbrook (1995), Origin of thick, high-velocity igneous crust along the U.S. East Coast Margin, J. Geophys. Res., 100, 10,077 – 10,094. Koppers, A., et al. (2005), The Magnetics Information Consortium (MagIC) online database: Uploading, searching and visualizing paleomagnetic and rock magnetic data, Eos Trans. AGU, 86(52), Fall Meet. Suppl., Abstract GP33A0088. Le Maitre, R. W., et al. (2002), Igneous Rocks: A Classification and Glossary of Terms, Recommendations of the International Union of Geological Sciences, Subcommission of the Systematics of Igneous Rocks, Cambridge Univ. Press, New York. Lehnert, K., Y. Su, C. H. Langmuir, B. Sarbas, and U. Nohl (2000), A global geochemical database structure for rocks, Geochem. Geophys. Geosyst., 1(5), doi:10.1029/ 1999GC000026. Lehnert, K., S. Goldstein, A. Johansson, R. W. Murray, N. Pisais, S. Vinayagamoorthy, and B. Djapic (2007), SedDB—A new information system to facilitate use of marine sediment geochemistry in science and education, MARGINS Newsl., 18, 9 – 11. Lo Cascio, M., Y. Liang, and P. C. Hess (2004), Grain-scale processes during isothermal-isobaric melting of lherzolite, Geophys. Res. Lett., 31, L16605, doi:10.1029/2004GL020602. Medard, E., C. A. McCammon, and T. L. Grove (2008), Oxygen fugacity, temperature reproducibility, and H2O content for nominally dry piston-cylinder experiments using graphite capsules, Am. Mineral, in press. Mirnejad, H., and K. Bell (2006), Origin and source evolution of the Leucite Hills lamproites: Evidence from Sr-Nd-Pb-O isotopic compositions, J. Petrol., 47, 2463 – 2489. O’Neill, H. S. C. (1987), Quartz-fayalite-iron and quartzfayalite-magnetite equilibria and the free energy of formation of fayalite (Fe2SiO4) and magnetite (Fe3O4), Am. Mineral., 72, 67 – 75. O’Neill, H. S. C. (1988), Systems Fe-O and Cu-O: Thermodynamic data for the equilibria Fe-‘‘FeO’’, Fe-Fe 3 O 4 , ‘‘FeO’’-Fe3O4, Fe3O4-Fe2O3, Cu-Cu2O and Cu2O-CuO from EMF measurements, Am. Mineral., 73, 470 – 486. O’Neill, H. St. C., and M. L. Pownceby (1993a), Thermodynamic data from redox reactions at high temperatures. I. An experimental and theoretical assessment of the electrochemical method using stabilized zirconia electrolytes, with revised values for the Fe-‘‘FeO’’, Co-CoO, Ni-NiO, and Cu-Cu2O oxygen buffers, and new data for the W-WO2 buffer, Contrib. Mineral. Petrol., 114, 296 – 314. O’Neill, H. St. C., and M. L. Pownceby (1993b), Thermodynamic data from redox reactions at high temperatures. II. The MnO-Mn3O4 oxygen buffer, and implications for the thermodynamic properties of MnO and Mn3O4, Contrib. Mineral. Petrol, 114, 315 – 320. Robie, R. A., B. Hemingway, and J. R. Fisher (1979), Thermodynamic properties of minerals and related substances at 298.15 K and 1 bar (105 Pa) pressure and at higher temperatures, U.S. Geol. Soc. Bull., 1452.

14 of 15

Geochemistry Geophysics Geosystems

3

G

hirschmann et al.: library of experimental phase relations

Salters, V. J. M. (1996), The generation of mid-ocean ridge basalts from the Hf and Nd isotope perspective, Earth Planet. Sci. Lett., 141, 109 – 123. Sautter, V., O. Jaoul, and F. Abel (1988), Aluminum diffusion in diopside using the 27Al (P, g) 28Si nuclear reaction—Preliminary results, Earth Planet. Sci. Lett., 89, 109 – 114. Schramm, B., K. P. Jochum, B. Sarbas, and U. Nohl (2006), GEOROC and GeoReM—Linking the information of two geological databases, Geochim. Cosmochim. Acta, suppl., 70(18), A565. Staudigel, H., et al. (2003), Electronic data publication in geochemistry, Geochem. Geophys. Geosyst., 4(3), 8004, doi:10.1029/2002GC000314.

10.1029/2007GC001894

Stolper, E. (1980), A phase diagram for mid-ocean ridge basalts: Preliminary results and implications for petrogenesis, Contrib. Mineral. Petrol., 74, 13 – 27. Stracke, A., A. Zindler, V. J. M. Salters, D. McKenzie, J. Blichert-Toft, F. Albare`de, and K. Gro¨nvold (2003), Theistareykir revisited, Geochem. Geophys. Geosyst., 4(2), 8507, doi:10.1029/2001GC000201. Sugawara, T. (2000), Empirical relationships between temperature, pressure, and MgO content in olivine and pyroxene saturated liquid, J. Geophys. Res., 105, 8457 – 8472. Zhang, J., and C. Herzberg (1994), Melting experiments on anhydrous peridotite KLB-1 from 5.0 to 22.5 GPa, J. Geophys. Res., 99, 17,729 – 17,742.

15 of 15