Downloaded from geology.gsapubs.org on September 17, 2010
Reading the mineral record of fluid composition from element
partitioning
Vincent J. van Hinsberg*, Artasches A. Migdisov, and Anthony E. Williams-Jones
Hydrothermal Geochemistry Group, Department of Earth and Planetary Sciences, McGill University, Montreal, Quebec H3A 2A7, Canada
ABSTRACT
Earth is the “blue planet,” with more than 70% of its surface covered in water and the
equivalent of up to four oceans of water in its interior. This abundance of water has a profound
impact on the processes that shape our planet as well as the development of the organisms that
inhabit it. To understand this impact, it is necessary to know the properties and compositions
of this fluid. At present, this information is largely unavailable, because direct samples of fluid
are rare, especially for early Earth and Earth’s interior, and other estimators are semiquantitative, at best. Here we propose a different approach in which the composition of the fluid
is reconstructed from that of minerals, based on the characteristic trace element partitioning
between minerals and aqueous fluids. We show experimentally that this partitioning is systematic and obeys lattice-strain theory. It depends strongly on element complexation in the
fluid, but this dependence is predictable and can be accommodated. Unlike fluids, minerals
with preserved compositions are readily available in the geological record, and this approach
therefore provides a powerful and widely applicable tool to reconstruct a quantitative record
of fluid composition for the full range of Earth environments and for its earliest history.
INTRODUCTION
An understanding of the chemical behavior
of volatiles, and aqueous fluids in particular,
is key to interpreting the processes operating
in Earth’s interior and on its surface, because
of their profound impacts on the physical and
chemical properties of Earth materials. Volatiles
in the mantle allow plate tectonics to operate by
lowering the strength of the upper mantle and
facilitating melting (Mei and Kohlstedt, 2000;
Gerya et al., 2008; Médard and Grove, 2008).
In the crust, aqueous fluids are the main agents
of element transport. However, this transport is
highly selective, as the capacity to solvate individual elements is sensitive to the fluid properties, especially temperature and ligand activity
(Kesler, 2005). The most obvious manifestation
of this dependence is the selective enrichment
of elements to form ore deposits. Element redistribution during water-rock interaction at midocean ridges is also a key process determining
ocean water chemistry (Komiya et al., 2008).
It is likely that changes in ocean water chemistry over geological time have had an important impact on the evolution of life, because the
viability of organic synthesis pathways crucially
depends on the availability of specific trace elements for use as catalytic constituents in the
associated enzymes (Anbar and Knoll, 2002;
Russell et al., 2005; Komiya et al., 2008; Konhauser et al., 2009).
In all these environments, our understanding requires knowledge of the composition of
the associated fluids. Whereas we have direct
access to such fluids on Earth’s surface at the
present time, and can determine their composi*E-mail: [email protected].
tions directly, reliable fluid samples are exceedingly difficult to acquire for Earth’s interior and
earlier stages of its history. Fluid inclusions in
minerals are currently the main source of compositional information, but challenges in their
analysis and nontrivial interpretation (Heinrich
et al., 2003; Rickers et al., 2004) makes them
an unsatisfactory data source for all except the
major elements. Similarly, the metasomatic
imprint of fluids on their host rocks (e.g.,
Spandler et al., 2003) yields only semiquantitative information, and forward thermodynamic
modeling is severely limited by a lack of data.
These methods, moreover, are suited mainly to
determining the major element composition,
whereas it is the trace element signature of the
fluid that is the most sensitive indicator of processes and conditions.
Here we show that the composition of fluids,
and their trace element content in particular,
can be reconstructed from the mineral record.
At equilibrium, elements partition characteristically between minerals and fluid, and minerals thereby record a chemical signature of
the composition of the fluid (Fig. 1). When
partition coefficients are known, it is therefore
possible to reconstruct fluid composition from
mineral chemistry. The availability of minerals
with preserved compositions from deep within
Earth (Spandler et al., 2003) and ages up to
4 Ga (Compston and Pidgeon, 1986) allows
fluid compositions to be obtained from well into
Earth’s interior and for most of its history. Furthermore, because minerals are our main source
of information on pressure and temperature in
and on Earth, as well as our main chronometers,
their added capacity to provide information on
fluid composition would make them a univer-
sal tool to gain insights into the processes that
shape our planet and the changes therein over
geological time.
The key requirement for reading the mineral
record is precise partition coefficients for relevant pressure-temperature conditions and mineral and fluid chemistry. Notwithstanding experimental and analytical challenges, these data can
be obtained from experiments (e.g., Keppler,
1996; Brenan et al., 1995; Ayers, 1998; Stalder
et al., 1998). However, partitioning is highly
sensitive to variations in physicochemical conditions, with coefficients changing up to three
orders of magnitude and elements from compatible to incompatible (Stalder et al., 1998). This
precludes extrapolation of partition coefficients
beyond experimental conditions and, as obtaining coefficients for every plausible combination
of physicochemical conditions is a practical
impossibility, severely limits the applicability of
this approach to obtain information on aqueous
fluid composition at present.
In this paper, we present a new method for
interpolating and extrapolating mineral-fluid
partition coefficients among trace elements and
to conditions beyond those of available experimental data, based on lattice-strain theory, which
is used with great success to model partitioning
between minerals and silicate melts (Blundy and
Wood, 1994, 2003). Unlike mineral-melt partitioning, the speciation of elements in the aqueous fluid exerts a strong influence on their partitioning, and this has commonly been seen as
an insurmountable barrier to using lattice-strain
theory to evaluate fluid chemistry (e.g., Keppler,
1996; Brenan et al., 1995; Bau, 1996; Ayers,
Zn
Zn
Mn Mn
Pb
Zn
Mn
Mn
Pb
Zn
Mn
Zn
Pb
Mn
Mn
Zn Pb
Pb
Pb
Zn Pb
D(T, P, X) 5 Mn
5 Zn
1 Pb
Figure 1. Conceptual drawing of the mineralfluid element partitioning approach. As a
mineral grows in an aqueous environment,
its composition will reflect that of the surrounding fluid through characteristic partitioning of trace elements. After the mineral
is separated from the fluid, it can preserve
this information. Given partition coefficients
for appropriate physical and chemical conditions, D(T, P, X), the fluid composition can
be reconstructed.
© 2010 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected].
GEOLOGY,
September
2010
Geology,
September
2010;
v. 38; no. 9; p. 847–850; doi: 10.1130/G31112.1; 4 figures; Data Repository item 2010235.
847
Downloaded from geology.gsapubs.org on September 17, 2010
1998; Stalder et al., 1998). However, we show
here that element speciation effects are systematic and predictable, and can be accommodated.
PARTITIONING EXPERIMENTS
We have conducted partitioning experiments
involving fluorite and Cl-, NO3-, or SO4-bearing
aqueous solutions doped with trace concentrations of a wide range of elements (see the GSA
Data Repository1 for details). These experiments produced a well-developed, fine-grained
monomineralic crystalline precipitate of fluorite
(Fig. 2) with a trace element signature similar
to that of natural fluorite, including a strong
preference for the rare earth elements (REEs)
(Marshall et al., 1998; Hill et al., 2000). Good
reproducibility among duplicates and excellent
agreement between Ca concentrations in the
final solutions and those determined from fluorite solubility (Richardson and Holland, 1979)
show that fluorite composition reflects equilibrium (see the Data Repository for further details).
Digested fluorite and reacted solutions were analyzed to obtain partition coefficients. There is
good agreement between duplicate experiments
(Fig. 3) as well as between directly measured
XRD settings:
10o to 131o
step 0.018o
step time 277 s
8
fluorite
Cs-fluoride
REE-fluoride
Counts / 1000
4
2
Na-fluoride
REE-fluoride
fluorite
6
fluorite
fluorite fluorite
0
10
0.5 mm
30
50
2-Theta 90
110
130
Figure 2. The solid run product of the experiments consists of well-developed crystalline
monomineralic fluorite as is evident in this binocular microscope image and XRD spectrum
for a Cl-bearing experiment. Diffraction peaks characteristic of the highly refractory REEfluorides are absent, as are peaks for Cs-fluoride.
7
NO3 solution
Cl solution
6
partition coefficients and those obtained from
mass balance with the starting solution.
When plotted against ionic radius (for the fluorite cation coordination number of 8), the partition coefficients show a systematic decrease
away from the radius of Ca, the major cation
in fluorite (Fig. 3). This systematic behavior
follows lattice-strain theory as is evident from
the good lattice-strain fits to the data. Latticestrain theory quantifies the observation that elements that are similar in charge and radius will
exchange for one another, whereas elements
with large mismatch in these parameters will
not (Blundy and Wood, 1994, 2003). A mineral
will therefore selectively incorporate elements
according to the elastic and electric characteristics of its lattice sites. A stiff lattice site cannot
easily accommodate elements with deviating
radii, and partition coefficients will therefore
decrease sharply with increasing radius mismatch. Similarly, a high-capacitance site cannot
readily dissipate excess charge and will therefore shun elements with large charge mismatch.
Partition coefficients for elements not determined experimentally can be read from the
lattice-strain fit, because charges and radii of
the elements are largely known (Shannon,
1976). This fit is described by three parameters (Fig. 3): an ideal radius, r0, its associated
characteristic partition coefficient, D0, and a
measure of the elastic properties of the lattice
site, E, which is related to the Young’s modulus
1+
2+
3+
4+
SO4 solution
log D fluorite - fluid
5
4
3
Sb
2
Sb
1+
D0
Sb
1
0
-1
A
-2
0.5
0.7
0.9
1.1
1.3
Ionic radius (Å)
1.5
1.7
1.9 0.5
B
1+
rideal
0.7
Ca radius
Ionic radius (Å)
1.5
1.7
1.9 0.5
C
0.7
0.9
1.1
1.3
1.5
1.7
1.9
Ionic radius (Å)
Figure 3. Trace element partition coefficients for fluorite-fluid experiments conducted in Cl-, NO3-, and SO4-dominated aqueous solutions.
Symbols show experimental data (in duplicate for Cl experiments) with associated uncertainty, and curves show the lattice-strain fit through
these data. There is good agreement between the data and lattice-strain curves, indicating that the lattice-strain model provides an accurate
description of the observed partitioning behavior. Three characteristic values determine this fit: (1) the ideal radius of the mineral lattice site,
r0, which is close to that of Ca in our experiments as would be expected for fluorite, (2) the partition coefficient for this ideal radius, D0, and
(3) a measure of the elastic properties of the lattice site, E, which determines the width of the curve. Given these values, the partition coefficient for any element can be predicted from its charge and radius.
1
GSA Data Repository item 2010235, detailed methodology of the experiments, data handling and fitting, and thermodynamic modeling, as well as tables listing
the compositions of the experimental solutions, partition coefficients between fluorite and aqueous solution, analytical detection limits, lattice-strain fit parameters and
element radii and charges used in obtaining the lattice-strain fit, and calculated element speciation in the Cl-dominated experimental solution, is available online at
www.geosociety.org/pubs/ft2010.htm, or on request from [email protected] or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.
848
GEOLOGY, September 2010
Downloaded from geology.gsapubs.org on September 17, 2010
ated characteristic coefficient for each. With a
change in speciation, the contribution of each
species to bulk partitioning changes and hence
the bulk partition coefficient that is measured.
Consequently, differences in speciation among
elements adversely affect the ability of latticestrain theory to model partitioning. This effect
can be observed in our data for Sb, which occurs
dominantly as Sb(OH)30 whereas most other
3+ elements are F-speciated. Antimony consequently shows the strongest deviation from the
3+ lattice-strain fit (Fig. 3).
To incorporate speciation in a model of mineral-fluid partitioning therefore requires knowledge of the speciation of each element, which
we have determined from thermodynamic
modeling. This modeling shows that there is
no significant correlation between partitioning
and speciation for species involving the ligands
Cl, NO3, SO4, and O/OH, nor for the free ion.
Furthermore, the partitioning of elements in Cl,
NO3, and SO4 solutions is remarkably similar
(Fig. 4A). These observations lead us to conclude that partitioning of elements with fluorite
is controlled dominantly by F-species. Indeed,
the agreement between experimental data and
the lattice-strain fit improves significantly, in
particular for the 2+ elements, when only the
fraction of each element residing in F-species
is considered (Fig. 4B). Moreover, the recalculated partition coefficients for the 2+ elements
exceed those for the 3+ elements, as the latter
are dominated by F-species, whereas the former are not. This leads to an ideal charge for
the fluorite cation site of 1.9 when fitted, rather
than a value more than 3 for a fit of the original
data (Fig. 4C). A value close to 2 is expected
for this mineral and agrees with fluorite-melt
partitioning data (Marshall et al., 1998) where
(Blundy and Wood, 2003). To extrapolate to
conditions beyond the experimental data, only
the dependence of these three parameters on
the conditions needs to be known, rather than
the dependence for each individual trace element. Moreover, this dependence can, in part,
be independently constrained. For example, the
change in lattice site dimensions and hence r0
can be determined from structure analyses of
minerals at elevated temperatures and pressures
(e.g., Katsura et al., 2009), and changes in E can
be predicted from the change in elastic modulus
for the bulk mineral (Blundy and Wood, 2003)
as determined experimentally or in molecular
simulations (Li et al., 2006; Reichmann et al.,
2008). Finally, the changes in D0 with changing
conditions can be estimated from the changes in
the mineral’s solubility. However, a base set of
experimentally determined partition coefficients
for the mineral-fluid system of interest remains
necessary to establish a lattice-strain fit and act
as a starting point for extrapolation.
EFFECT OF SPECIATION ON
PARTITIONING
The distribution of elements between minerals and aqueous solutions is not controlled solely
by the mineral lattice as is the fundamental tenet
of lattice-strain theory, because the solution is
not a passive reservoir in which all elements are
equivalently available for incorporation in minerals. On the contrary, elements in aqueous solutions are distributed over a number of species,
consisting of metal-ligand complexes enclosed
in hydration shells, and this speciation varies
substantially among the elements. For a given
element, partitioning will take place between
the mineral lattice site and every aqueous species containing that element, with an associ-
7
3+
6
Cl
Cl
NO3
SO4
5
log D fluorite - fluid
9
Ligands:
4
Cl solution
speciation is much less of an issue. Given a
measure of element speciation, it is therefore
possible to account for its effects on partition
coefficients and regain, or use, the underlying
lattice-strain control. The application of this
method requires information on the major element composition of the fluid, especially its
ligand content, but this is available from other
sources, e.g., fluid inclusions.
CONCLUSIONS AND IMPLICATIONS
The results of this study demonstrate that mineral-fluid partition coefficients among elements
vary systematically in a manner that can be
predicted from lattice-strain theory. Adherence
to this theory allows partition coefficients to be
determined from interpolation and extrapolation
among elements of equal charge, and between
different charges. It also allows partitioning to
be described by a limited set of parameters, and
only the dependence of these parameters on
pressure, temperature, and mineral composition
needs to be determined for partition coefficients
to be extrapolated to conditions beyond those
for which they have been evaluated experimentally. The speciation of elements in the
fluid clearly impacts on partitioning, but this
effect can be accommodated when speciation is
known. However, even when speciation is not
known, interpolation among elements of equal
charge is still possible, because their speciation
is commonly similar (cf. the good lattice-strain
fit for noncorrected coefficients in Fig. 3). For
example, when partitioning for a few of the
REEs is known, coefficients for the other REEs
can be determined by interpolation, even when
there is no information on the aqueous fluid and
its ligands, because the speciation of the REEs
in aqueous fluids is similar.
1+
2+
3+
8
fluorite
D00
Speciation
corrected
7
Original fit
6
3
1+
2
5
1
4
0
3
-1
A
-2
0.5
0.7
0.9
1.1
1.3
Ionic radius (Å)
1.5
1.7
2
1.9 0.5
B
0.7
0.9
1.1
1.3
Ionic radius (Å)
1.5
1.7
1.9 0.5
C
Ideal charge
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
Cation charge
Figure 4. A comparison of the lattice-strain curves for Cl, NO3, and SO4 experiments shows that these are identical within uncertainty (A). This
independence of the partition coefficients from chloride, nitrate, and sulfate concentrations indicates that the species involved in element
partitioning with fluorite do not include any of these ligands. As discussed in the main text, F-species control the partitioning, and when
coefficients are corrected for the fraction of each element residing as F-species, the lattice-strain fit improves markedly (B). Furthermore,
the ideal charge shifts to 1.9 (C), which agrees with that expected for the fluorite structure.
GEOLOGY, September 2010
849
Downloaded from geology.gsapubs.org on September 17, 2010
The methodology presented in this paper
allows partition coefficients to be determined
for relevant physicochemical conditions and
therefore provides the means to read the record
of fluid chemistry stored in mineral compositions. Given the availability of marine mineral
precipitates for most of Earth’s history, our
results provide a method to reconstruct ocean
water chemistry back in time to that of early
Earth where life emerged, and explore the links
between changes in ocean water chemistry
and the evolution of marine organisms. Similarly, minerals from the subducting slab, transferred from deep within Earth by the actions
of plate tectonics, allow the compositions of
fluids released during subduction to be determined. This provides insights into the elements
released during slab dehydration and helps to
constrain the element fluxes to mantle and crust
associated with subduction. Although we have
focused predominantly on natural systems, the
methodology is equally applicable to man-made
environments. Mineral-fluid interaction is an
important process in industry from wastewater
treatment to preparation of drinking water and
ore beneficiation. In any process where minerals
form in the presence of a fluid, the composition
of this fluid is recorded by minerals, and their
analysis can unlock this record.
ACKNOWLEDGMENTS
This study was funded by a Rubicon Postdoctoral
Fellowship from the Netherlands Organisation for Scientific Research to van Hinsberg, and a Natural Sciences and Engineering Research Council of Canada
Discovery research grant and Le Fonds québécois de
la recherche sur la nature et les technologies (FQRNT)
team research grant to Williams-Jones. We thank the
Ceramics Research Centre of Corus RD&T for access
to X-ray diffraction equipment. This study has benefited from discussions with Bob Linnen and Bernie
Wood, and we thank Jon Blundy and two anonymous
reviewers for their comments and suggestions.
REFERENCES CITED
Anbar, A.D., and Knoll, A.H., 2002, Proterozoic
ocean chemistry and evolution: A bioinorganic
bridge?: Science, v. 297, p. 1137–1142, doi:
10.1126/science.1069651.
Ayers, J., 1998, Trace element modeling of aqueous fluid–peridotite interaction in the mantle
wedge of subduction zones: Contributions to
Mineralogy and Petrology, v. 132, p. 390–404,
doi: 10.1007/s004100050431.
Bau, M., 1996, Controls on the fractionation of isovalent trace elements in magmatic and aqueous
systems: Evidence from Y/Ho, Zr/Hf, and lanthanide tetrad effect: Contributions to Mineralogy and Petrology, v. 123, p. 323–333, doi:
10.1007/s004100050159.
850
Blundy, J., and Wood, B., 1994, Prediction of
crystal-melt partition coefficients from elastic moduli: Nature, v. 372, p. 452–454, doi:
10.1038/372452a0.
Blundy, J., and Wood, B., 2003, Partitioning of trace
elements between crystals and melts: Earth and
Planetary Science Letters, v. 210, p. 383–397,
doi: 10.1016/S0012-821X(03)00129-8.
Brenan, J.M., Shaw, H.F., Ryerson, F.J., and Phinney,
D.L., 1995, Mineral-aqueous fluid partitioning of trace elements at 900°C and 2.0 GPa:
Constraints on the trace element chemistry of
mantle and deep crustal fluids: Geochimica et
Cosmochimica Acta, v. 59, p. 3331–3350, doi:
10.1016/0016-7037(95)00215-L.
Compston, W., and Pidgeon, R.T., 1986, Jack Hills,
evidence of more very old detrital zircons in
Western Australia: Nature, v. 321, p. 766–769,
doi: 10.1038/321766a0.
Gerya, T.V., Connolly, J.A.D., and Yuen, D.A., 2008,
Why is terrestrial subduction one-sided?: Geology, v. 36, p. 43–46, doi: 10.1130/G24060A.1.
Heinrich, C.A., Pettke, T., Halter, W.E., Aigner-Torres, M., Audétat, A., Günther, D., Hattendorf,
B., Bleiner, D., Guillong, M., and Horn, I.,
2003, Quantitative multi-element analysis of
minerals, fluid and melt inclusions by laserablation inductively-coupled-plasma massspectrometry: Geochimica et Cosmochimica
Acta, v. 67, p. 3473–3497, doi: 10.1016/S00167037(03)00084-X.
Hill, G.T., Campbell, A.R., and Kyle, P.R., 2000, Geochemistry of southwestern New Mexico fluorite
occurrences: Implications for precious metals
exploration in fluorite-bearing systems: Journal
of Geochemical Exploration, v. 68, p. 1–20, doi:
10.1016/S0375-6742(99)00047-3.
Katsura, T., Shatskiy, A., Manthilake, M.A.G.M.,
Zhai, S., Fukui, H., Yamazaki, D., Matsuzaki,
T., Yoneda, A., Ito, E., Kuwata, A., Ueda, A.,
Nozawa, A., and Funakoshi, K.-I., 2009, Thermal expansion of forsterite at high pressures
determined by in situ X-ray diffraction: The
adiabatic geotherm in the upper mantle: Physics of the Earth and Planetary Interiors, v. 174,
p. 86–92, doi: 10.1016/j.pepi.2008.08.002.
Keppler, H., 1996, Constraints from partitioning experiments on the composition of subductionzone fluids: Nature, v. 380, p. 237–240, doi:
10.1038/380237a0.
Kesler, S.E., 2005, Ore-forming fluids: Elements,
v. 1, p. 13–18, doi: 10.2113/gselements.1.1.13.
Komiya, T., Hirata, T., Kitajima, K., Yamamoto, S.,
Shibuya, T., Sawaki, Y., Ishikawa, T., Shu, D.,
Li, Y., and Han, J., 2008, Evolution of the composition of seawater through geologic time, and
its influence on the evolution of life: Gondwana
Research, v. 14, p. 159–174, doi: 10.1016/j
.gr.2007.10.006.
Konhauser, K.O., Pecoits, E., Lalonde, S.V., Papineau, D., Nisbet, E.G., Barley, M.E., Arndt,
N.T., Zahnle, K., and Kamber, B.S., 2009, Oceanic nickel depletion and a methanogen famine before the Great Oxidation Event: Nature,
v. 458, p. 750–754, doi: 10.1038/nature07858.
Li, L., Weidner, D.J., Brodholt, J., Alfè, D., Price,
G.D., Caracas, R., and Wentzcovitch, R., 2006,
Elasticity of CaSiO3 perovskite at high pressure and high temperature: Physics of the Earth
and Planetary Interiors, v. 155, p. 249–259, doi:
10.1016/j.pepi.2005.12.006.
Marshall, A.S., Hinton, R.W., and Macdonald, R.,
1998, Phenocrystic fluorite in peralkaline
rhyolites, Olkaria, Kenya Rift Valley: Mineralogical Magazine, v. 62, p. 477–486, doi:
10.1180/002646198547855.
Médard, E., and Grove, T.L., 2008, The effect of
H2O on the olivine liquidus of basaltic melts:
Experiments and thermodynamic models: Contributions to Mineralogy and Petrology, v. 155,
p. 417–432, doi: 10.1007/s00410-007-0250-4.
Mei, S., and Kohlstedt, D.L., 2000, Influence of water on plastic deformation of olivine aggregates
1. Diffusion creep regime: Journal of Geophysical Research, v. 105, p. 21,457–21,469, doi:
10.1029/2000JB900179.
Reichmann, H.J., Sinogeikin, S.V., and Bass, J.D.,
2008, Single-crystal elastic properties of
(Mg0.987, Fe0.013)O to 9 GPa: American Mineralogist, v. 93, p. 1306–1311, doi: 10.2138/am
.2008.2717.
Richardson, C.K., and Holland, H.D., 1979, The
solubility of fluorite in hydrothermal solutions, an experimental study: Geochimica et
Cosmochimica Acta, v. 43, p. 1313–1325, doi:
10.1016/0016-7037(79)90121-2.
Rickers, K., Thomas, R., and Heinrich, W., 2004,
Trace-element analysis of individual synthetic
and natural fluid inclusions with synchrotron
radiation XRF using Monte Carlo simulations
for quantification: European Journal of Mineralogy, v. 16, p. 23–35, doi: 10.1127/0935-1221/
2004/0016-0023.
Russell, M.J., Hall, A.J., Boyce, A.J., and Fallick,
A.E., 2005, On hydrothermal convection systems and the emergence of life: Economic Geology and the Bulletin of the Society of Economic Geologists, v. 100, p. 419–438.
Shannon, R.D., 1976, Revised effective ionic radii
and systematic studies of interatomic distances
in halides and chalcogenides: Acta Crystallographica, v. A32, p. 751–767.
Spandler, C., Hermann, J., Arculus, R., and Mavrogenes, J., 2003, Redistribution of trace elements during prograde metamorphism from
lawsonite blueschist to eclogite facies; implications for deep subduction-zone processes: Contributions to Mineralogy and Petrology, v. 146,
p. 205–222, doi: 10.1007/s00410-003-0495-5.
Stalder, R., Foley, S.F., Brey, G.P., and Horn, I., 1998,
Mineral-aqueous fluid partitioning of trace elements at 900–1200°C and 3.0–5.7 GPa: New
experimental data for garnet, clinopyroxene,
and rutile, and implications for mantle metasomatism: Geochimica et Cosmochimica Acta,
v. 62, p. 1781–1801, doi: 10.1016/S0016-7037
(98)00101-X.
Manuscript received 10 February 2010
Revised manuscript received 16 April 2010
Manuscript accepted 29 April 2010
Printed in USA
GEOLOGY, September 2010