Available online at www.sciencedirect.com
Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
www.elsevier.com/locate/gca
Petrogenesis of olivine-phyric shergottite Yamato
980459, revisited
Tomohiro Usui a,*, Harry Y. McSween Jr. a, Christine Floss b
a
Planetary Geoscience Institute, Department of Earth and Planetary Sciences, University of Tennessee, Knoxville, TN 37996-1410, USA
b
Laboratory for Space Sciences and Physics Department, Washington University, St. Louis, MO 63130, USA
Received 13 June 2007; accepted in revised form 15 January 2008; available online 26 January 2008
Abstract
Primitive magmas provide critical information on mantle sources, but most Martian meteorites crystallized from fractionated melts. An olivine-phyric shergottite, Yamato 980459 (Y-980459), has been interpreted to represent a primary melt,
because its olivine megacrysts have magnesian cores (Fo84–86) that appear to be in equilibrium with the Y-980459 whole-rock
composition based on Fe–Mg partitioning. However, crystal size distribution (CSD) plots for Y-980459 olivines show a size
gap, suggesting a cumulus origin for some megacrysts. Because melting experiments using the Y-980459 whole-rock composition have been used to infer the thermal structure and volatile contents of the Martian mantle, the interpretation that this
rock is primitive should be scrutinized.
We report major, minor and trace element compositions of Y-980459 olivines and compare them with results from melting
experiments (both hydrous and anhydrous) and thermodynamic calculations. Cores of the olivine megacrysts have major and
minor element contents identical to those of the most magnesian olivines from the experiments, but they differ slightly from
those of thermodynamic calculations. This is probably because the Y-980459 whole-rock composition lies near the limit of the
range of liquids used to calibrate these models. The megacryst cores (Fo80–85) exhibit minor and trace element (Mn–Ni–Co–
Cr–V) characteristics distinct from other olivines (megacryst rims and groundmass olivines, Fo < 80), implying that the megacryst cores crystallized under more reduced conditions (IW + 1).
Y-980459 contains pyroxenes with orthopyroxene cores mantled by pigeonite and augite. We also found some reversely
zoned pyroxenes that have augite cores (low-Mg#) mantled by orthopyroxenes (high-Mg#), although they are uncommon.
These reversely zoned pyroxenes are interpreted to have grown initially as atoll-like crystals with later crystallization filling
in the hollow centers, implying disequilibrium crystallization at a moderate cooling rate (3–7 °C/h). The calculated REE pattern of a melt in equilibrium with normally zoned pyroxene is parallel to those of glass and the Y-980459 whole-rock as well as
other depleted olivine-phyric shergottites, suggesting that Y-980459 was derived from a depleted mantle reservoir.
Considering the CSD patterns of Y-980459 olivines, we propose that the olivine megacrysts are cumulus crystals which
probably formed in a feeder conduit by continuous melt replenishment, and the parent melt composition was indistinguishable from the Y-980459 whole-rock with 0–2 wt% of H2O and 0–5 wt% of CO2. The final magma pulse entrained these cumulus olivines and then crystallized groundmass olivines and pyroxenes. Although Y-980459 contains small amounts of cumulus
olivine (<6 vol%), we conclude that the Y-980459 whole-rock composition closely approximates a Martian primary melt
composition.
Ó 2008 Elsevier Ltd. All rights reserved.
1. INTRODUCTION
*
Corresponding author. Fax: +1 865 974 2368.
E-mail address: [email protected] (T. Usui).
0016-7037/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.gca.2008.01.011
Primitive magmas (unfractionated liquids formed by direct partial melting) provide critical constraints on their
mantle source regions, but terrestrial experience suggests
1712
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
that magmas are generally modified during ascent and
emplacement. Most Martian magmatic rocks, as represented by Martian meteorites, have also been fractionated.
The most promising primitive Martian meteorites are the
olivine-phyric shergottites (Goodrich, 2002). Their relatively high bulk-rock Mg#s (=molar ratio of Mg/[Mg+Fe])
support the inference that they may be primary. In contrast,
other shergottites are clearly fractionated; e.g. the basaltic
shergottites have lower Mg#s and contain cumulus pyroxenes, and the lherzolitic shergottites are cumulate gabbros
(McSween and Treiman, 1998).
As implied by their name, olivine-phyric shergottites are
characterized by megacrysts of olivine (and small amounts
of orthopyroxene) in a fine-grained groundmass of olivine,
pyroxene and plagioclase (now shock-metamorphosed to
maskelynite). There has been considerable disagreement
about the origin of the olivine megacrysts. In most olivine-phyric shergottites (e.g. NWA 1068, DaG 476, SaU
005, EETA79001 lithology-A, Dhofar 019), these grains
have Mg#s that are not in equilibrium with the host rock,
leading to their interpretation as xenocrysts or cumulate
crystals from similar parental magmas rather than as phenocrysts (McSween and Jarosewich, 1983; Wadhwa et al.,
2001; Taylor et al., 2002; Barrat et al., 2002b; Goodrich,
2003). However, one texturally distinctive olivine-phyric
shergottite, Yamato 980459 (Y-980459), contains olivine
megacrysts with magnesian olivine cores (Fo84–86) that appear to be in equilibrium with the Y-980459 whole-rock
composition based on Fe–Mg exchange partitioning. Consequently, a number of authors (e.g. Greshake et al.,
2004; Ikeda, 2004; McKay et al., 2004; Mikouchi et al.,
2004) have interpreted these megacrysts as phenocrysts
and hypothesized that the meteorite represents a primary
melt. However, this interpretation may be inconsistent with
the observation that the crystal size distribution (CSD) pattern of Y-980459 olivines shows a distinctive size gap between megacrysts and groundmass olivines and is similar
to those of other olivine-phyric shergottites that may contain xenocrystic or cumulus olivine megacrysts (Greshake
et al., 2004; Lentz and McSween, 2005). Since numerous
melting experiments were recently carried out using the
Y-980459 whole-rock composition as a primary melt to infer the thermal structure and volatile contents of the Martian mantle (Dalton et al., 2005, 2007; Musselwhite et al.,
2006; Norris and Herd, 2006; Draper, 2007), this interpretation should be thoroughly investigated.
The primary objective of this paper is to reassess
whether Y-980459 really represents a melt composition.
For this purpose, we carefully examine the origin of the
Y-980459 olivine megacrysts based on (1) Fe–Mg (Toplis,
2005) and Ca (Libourel, 1999) partitioning models, (2) results from anhydrous (Musselwhite et al., 2006) and hydrous (Draper, 2007) melting experiments on Y-980459
compositions and (3) trace element (Mn–Ni–Co–Cr–V) systematics for the Y-980459 olivines. Moreover, we report
two types of pyroxene grains with distinct zoning patterns
and rare earth element (REE) compositions. Although
our interpretation is that Y-980459 contains some cumulus
olivines, its whole-rock composition is a close approximation to a Martian primary melt composition so the infer-
ences drawn from it about the Martian mantle from
experiments are valid. Finally, we propose a petrogenetic
model for Y-980459.
2. ANALYTICAL METHODS
A polished thin section of Y-980459 (51-2) was examined to study its petrographic characteristics by optical
microscope and electron probe microanalyzer (EPMA) at
the University of Tennessee. Modal proportions of minerals
were measured by multipoint counting analysis using an
optical microscope (total points = 630), and the proportion
of olivine megacrysts was obtained by digital image analysis
(total points = 58,264). Mineral and glass compositions
were determined using wavelength dispersive spectrometry
(WDS), with an accelerating potential of 15 kV, a 20 nA
beam current (10 nA for glasses), 1 lm beam size (5 lm
for glasses), and standard ZAF (PAP) correction procedures for all silicate minerals. Peak and background counting times were 20–30 s, and data quality was maintained by
analyzing a combination of synthetic oxide and natural
mineral standards. Drift was within counting error
throughout every analytical session. Analytical errors (1r)
are due to counting statistics. Detection limits (3r background) are typically <0.03 for SiO2, TiO2, Al2O3, MgO,
CaO, Na2O, K2O and P2O5, and <0.05–0.1 for FeO,
MnO, Cr2O3, NiO and V2O5.
Concentrations of REE and other trace elements were
determined using the modified Cameca ims-3f ion microprobe at Washington University in St. Louis, following
the technique of Zinner and Crozaz (1986b). Analyses were
made using an O primary beam of 10–20 nA intensity (20–
40 lm beam diameter) and energy filtering at low mass resolution to remove complex molecular interferences. The
resulting mass spectrum was deconvolved in the mass
ranges of K–Ca–Sc–Ti, Rb–Sr–Y–Zr and Ba–REE to remove simple molecular interferences that are not eliminated
with energy filtering (Alexander, 1994; Hsu, 1995). Sensitivity factors for the REE are as given by Zinner and Crozaz
(1986a) and those for other elements are from Hsu (1995)
and Floss et al. (1998). Absolute concentrations were determined using sensitivity factors relative to Si, with SiO2 concentrations determined by EPMA for the specific spots
selected for ion probe analysis. Analytical errors (1r) are
due to counting statistics only. All analysis spots were observed by optical microscope and EPMA before and after
the analyses to examine whether contamination with other
phases accidentally occurred during the measurements. We
also monitored the secondary ion intensities of all cycles for
each individual analysis to check for contamination by
small inclusions within the grains.
3. RESULTS
3.1. Petrography and major element compositions
Y-980459 (51-2) shows dominantly a porphyric texture
with large olivine crystals set in a fine-grained groundmass of olivines, pyroxenes, chromite and dark glassy
materials (Figs. 1 and 2). The dark materials are glassy
Petrogenesis of olivine-phyric shergottite Y-980459
mesostasis filled with dendritic olivines and augites, as
well as sulfide droplets. No plagioclase/maskelynite and
phosphate are observed in the groundmass, unlike in
other olivine-phyric shergottites. The modal proportions
of minerals are 12 vol% olivine (6.2 vol% for megacrysts
that have high-Fo (>80) cores, described later), 58 vol%
1713
pyroxene, 23 vol% mesostasis and 1 vol% spinel. These
values are almost within the range of the published values from other Y-980459 thin sections: 8.7–26 vol% olivine, 48–52.7 vol% pyroxene and 25–37 vol% mesostasis
(Greshake et al., 2004; Ikeda, 2004; Mikouchi et al.,
2004).
Fig. 1. Photomicrograph (plane-polarized light) of Y-980459 showing four olivine megacrysts and pyroxenes that dominate the groundmass.
Dark materials are glassy mesostasis containing dendritic olivines and pyroxenes. Boxes identify olivine megacrysts pictured in Fig. 2.
Fig. 2. Back-scattered electron (BSE) images of the four olivine megacrysts in Fig. 1: (a) Ol-1; (b) Ol-2; (c) Ol-3; (d) Ol-4. The shading shows
that megacrysts are zoned from Mg-rich homogeneous cores (dark) to more Fe-rich mantles (brighter). Olivine megacrysts have narrow Ferich rims (white) in contact with mesostasis areas. Abbreviations: Ol, olivine; Px, pyroxene; Ch, chromite; Gl, glass; S, sulfide. X–Y in (d)
shows the microprobe traverse in Fig. 5.
1714
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
3.1.1. Olivine
Olivine shows a bimodal grain size distribution with a
large gap. In a prior study of our thin section, most groundmass grains were measured to be <0.4 mm and four larger
grains (megacrysts) were >0.85 mm (Lentz and McSween,
2005). The olivine megacrysts occur as separate grains;
three of them (Ol-2, -3 and -4) have almost euhedral outlines (Fig. 2b–d) but the other one (Ol-1) is anhedral, and
is probably a composite grain of a few crystals (Fig. 2a).
Ol-1 contains pyroxene inclusions (<20 lm) near its prospective composite grain boundaries (Fig. 2a). An irregularly shaped glass inclusion occurs in Ol-2 (Fig. 2c).
Olivine in the groundmass generally has euhedral to subhedral outlines (Fig. 3a), and some grains have glass
inclusions.
Olivine megacrysts and groundmass olivines have different chemical characteristics. The megacrysts exhibit variable forsterite contents (Fo70–85) that partly overlap the
groundmass olivine compositions, whereas the groundmass
olivines show almost homogeneous Fo contents except for
the outermost rims (Table 1 and Fig. 4), although some
of them have irregular zoning patterns (Mikouchi et al.,
2004). Both have narrow (<10 lm) iron-rich rims (hereafter
referred to as outermost rim) in contact with mesostasis
areas, indicating sharp compositional boundaries; their
rim compositions are not distinguishable from each other
(Table 1). Minor element compositions have obvious relationships with Fo; CaO and MnO increase but Cr2O3 and
NiO decrease with decreasing Fo content (details for these
elements are described later). These features are consistent
with those previously reported (Ikeda, 2004; Mikouchi
et al., 2004).
Olivine megacrysts consist of relatively unzoned cores
and monotonically zoned rims; the cores always show higher Fo (85) and lower CaO (0.2 wt%) contents than the
rims. The area of an unzoned core is slightly broader for
CaO than for Fo (Fig. 5). This zoning feature is obviously
exhibited as a kink at around Fo80 in a CaO versus Fo dia-
Fig. 3. BSE images of pyroxenes and groundmass olivines. (a and b) Most pyroxene grains (Type-I) are zoned from orthopyroxene cores
(dark) to pigeonite mantles (less dark) and then narrow augite rims (white). Groundmass olivines are mostly homogeneous except for narrow
Fe-rich rims. Some pyroxenes are in contact with groundmass olivines and have not developed the augite rims. Dendritic crystals (Fe-rich
olivine and augite) occur in mesostasis areas. (c and d) Type-II pyroxenes have augite cores mantled by orthopyroxene/pigeonite, with rims of
augite. Type-II pyroxenes are generally smaller than Type-I pyroxenes. Abbreviations as in Fig. 2.
Table 1
Representative microprobe analyses of olivines in Y-980459
Olivine megacryst
Core
39.8
0.47
14.2
0.34
44.7
0.16
0.10
99.8
Atoms per 4 oxygens
Si
1.00
Cr
0.009
Fe2+
0.299
Mn
0.007
Mg
1.67
Ca
0.004
Ni
0.002
Total
3.00
Molar compositions
Fo
84.8
a
Rim
(1r)
0.1
0.03
0.1
0.02
0.1
0.01
0.01
38.8
0.24
21.0
0.46
39.6
0.24
0.10
100.4
Outermost rim
(1r)
0.1
0.20
0.2
0.02
0.1
0.01
0.01
1.00
0.005
0.453
0.010
1.52
0.006
0.002
3.00
77.1
36.6
0.13
32.6
0.65
29.0
0.34
0.08
99.4
1.01
0.003
0.753
0.015
1.19
0.010
0.002
2.99
61.3
(1r)
0.1
0.01
0.2
0.03
0.08
0.01
0.01
Ol-l core
39.4
0.28
18.7
0.42
40.9
0.27
0.03
99.9
1.01
0.006
0.399
0.009
1.56
0.007
0.001
2.99
79.6
(1r)
0.1
0.02
0.2
0.02
0.1
0.01
0.01
Ol-l rim
38.8
0.29
22.9
0.52
37.6
0.33
0.05
100.6
1.00
0.006
0.498
0.011
1.46
0.009
0.001
2.99
74.53
Core
(1r)
0.1
0.02
0.2
0.03
0.1
0.01
0.01
38.5
0.25
21.7
0.48
39.1
0.24
0.10
100.3
1.00
0.005
0.469
0.010
1.51
0.006
0.002
3.00
76.3
(1r)
0.1
0.02
0.2
0.03
0.10
0.01
0.01
Rim
Outermost rim
(1r)
38.6 0.1
0.20 0.02
23.6 0.2
0.51 0.03
37.5 0.10
0.31 0.01
n.d
100.8
35.9
0.12
36.6
0.67
26.7
0.36
n.d
100.3
1.00
0.004
0.514
0.011
1.45
0.009
n.d
2.99
1.00
0.003
0.854
0.016
1.11
0.011
n.d
3.00
73.9
56.5
(1r)
0.1
0.01
0.2
0.03
0.08
0.01
Petrogenesis of olivine-phyric shergottite Y-980459
Oxide wt%
SiO2
Cr2O3
FeOa
MnO
MgO
CaO
NiO
Total
Groundmass olivine
Total Fe as FeO. Fo, forsterite; n.d., not detected.
1715
1716
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
0.40
0.6
Megacrysts
Groundmass
Outermost rims
CaO (wt%)
CaO (wt%)
0.5
0.4
0.3
Mesostasis or outermost rims
0.2
(a) Ol-1
0.35
Groundmass
0.30
0.25
1σ
0.20
0.15
0.10
0.1
Megacrysts
0.40
0.0
(b) Ol-2
0.35
CaO (wt%)
0.6
Cr2 O3 (wt%)
0.5
Groundmass
0.4
Mesostasis or outermost rims
0.3
0.30
0.25
0.20
rim
1σ
0.15
0.10
core
0.2
0.40
Megacrysts
0.1
(c) Ol-3
40
50
60
70
80
90
100
Fo
Fig. 4. Summary of compositions of megacryst and groundmass
olivines and their outermost rims (see text for definition) in Y980459. (a) CaO and (b) Cr2O3 contents versus Fo content.
Compositional fields of olivines in another Y-980459 section
(Mikouchi et al., 2004) are shown as outlines.
CaO (wt%)
0.35
0.0
30
0.30
0.25
rim
0.20
1σ
0.15
core
0.10
0.40
(d) Ol-4
CaO (wt%)
0.35
86
84
rim
0.20
1σ
0.15
core
0.10
82
Fo
0.30
0.25
0.00
80
72
74
76
78
80
82
84
86
88
Fo
78
Fig. 6. CaO versus Fo contents of olivine megacrysts (a) Ol-1, (b)
Ol-2, (c) Ol-3 and (d) Ol-4. Compositions more magnesian than
Fo80 correspond to megacryst cores; Ol-1 contains no core
compositions. Vertical error bars (1r) show analytical reproducibility for CaO contents.
76
74
0.30
CaO (wt%)
1σ
0.25
gram (Fig. 6). The kink is readily observed in Ol-3 and -4, is
faint in Ol-2, but cannot be seen in the composite grain Ol-1
that has no high-Fo (>80) cores. Ol-1 has a major element
composition that is indistinguishable from those of the
other megacryst rims (Fo < 80) and the groundmass olivines (Table 1).
0.20
0.15
0
400
800
X
1200
1600
Y
Distance (μm)
Fig. 5. Zoning profile for Fo and CaO contents along the traverse
X–Y in Fig. 2d. Vertical error bar (1r) shows analytical reproducibility for CaO; analytical errors for Fo contents are approximately
the symbol size.
3.1.2. Pyroxene
Pyroxenes generally occur in the groundmass as prismatic euhedral to subhedral grains with sizes of <500 lm
in the major axes, but a few are included in olivine crystals
as small crystals (<50 lm) (Fig. 3a). Polysynthetic twinning
is commonly observed under cross-polarized light. Glass
inclusions appear to be absent from pyroxene grains.
Petrogenesis of olivine-phyric shergottite Y-980459
core
mantle
outermost rim
rim
50
50
Di
Hd
(a)
Mesostasis or outermost rim
1717
and are characterized by low ulvöspinel component. However, the Mg# is higher in the chromite inclusion in Ol-2
than in other chromites (Table 3).
Glass occurs as inclusions in olivine and as a matrix
phase. The compositions of glass inclusions are typically
higher in SiO2, MgO and CaO but lower in Al2O3 and
FeO relative to the matrix glasses (Table 4).
Groundmass
(b)&(c)
En
100
80
60
3.2. Trace element compositions
40
(b) Type-I
Fs
20
(c) Type-II
30
30
core
20
10
80
rim
10
mantle
core
90
20
rim
70
mantle
90
80
70
Fig. 7. Compositions of Y-980459 pyroxenes plotted in the system
En–Wo–Fs. (a) Type-I and Type-II pyroxenes, and their outermost
rims (see text for definition). Compositional fields of pyroxenes in
another section of Y-980459 (Mikouchi et al., 2004) are shown as
outlines. A quadrangle area defined by hatched lines is enlarged in
(b) Type-I and (c) Type-II pyroxenes. Arrows show compositional
zoning from core to rim.
Two types of chemical zoning are observed. Most
pyroxene grains (Type-I) are zoned from orthopyroxene
cores to pigeonite mantles and augite rims (Fig. 3a
and b). In contrast, some pyroxene grains (Type-II) have
augite cores mantled by orthopyroxene/pigeonite with
rims of augite (Figs. 3c, d and 7). The Type-II augite
cores (bright areas in Fig. 3c and d) show anhedral outlines. Greshake et al. (2004) reported some (rare) separate grains of augites and pigeonites; they may
correspond to the Type-II augite. Both type of pyroxenes exhibit trends that sharply bend at around En50Wo25Fs25 (Fig. 7). The Type-I trend has been observed
in Martian meteorites only in basaltic shergottites QUE
94201 and NWA 480, but it is common in lunar basalts
(McSween et al., 1996; Barrat et al., 2002a). Because
QUE 94201 and NWA 480 have abundant plagioclase,
the decrease of Wo components observed in these meteorites has been interpreted as a result of co-crystallization of plagioclase. However, since Y-980459 contains
no plagioclase, an explanation of the Type-I trend at
the rim cannot be that simple. Like the olivines, pyroxenes have narrow, iron-rich augite rims (as in olivine, we
will refer to these as the outermost rims) in contact with
mesostasis areas. The outermost rims are distinctly richer
in Wo and plot apart from the trends of major pyroxene
grains (Fig. 7a).
3.1.3. Other phases
Chromite is the only spinel group mineral present in Y980459. It occurs both as inclusions in olivine and in the
groundmass. No chromite inclusions occur in cores of the
olivine megacrysts but one (10 lm) was found in Ol-2
(Fig. 2b). The chromite grains are generally homogeneous
3.2.1. Olivine
Fig. 8 illustrates variations of Ni, Co, Cr and V versus
Mn concentrations in all 4 olivine megacrysts (Ol-1, -2, -3
and -4) and in 4 groundmass olivine grains. As Mn behaves
incompatibly in olivine during igneous processes, it can
serve as a fractionation index (Karner et al., 2003). With
increasing Mn, Co increases monotonically, whereas Ni increases in high-Fo olivine megacrysts (cores) and decreases
in low-Fo megacrysts (rims) and in groundmass olivines. Cr
decreases monotonically with increasing Mn, while V is
roughly constant in the high-Fo megacrysts (cores) and
increases in other olivines. These trace element characteristics are strikingly displayed as kinks in Ni–Co and V–Cr
diagrams (Fig. 9). No distinction between the low-Fo rims
of olivine megacrysts and the groundmass olivines is observed. Abundances of these transition elements in olivines
(in particular, Cr) may be disturbed by minute spinel inclusions if they accidentally occur within the analyzed points.
However, examination of the individual analyses shows
that such inclusions did not affect the data presented here.
Other trace element data are listed in Table 5.
3.2.2. Pyroxenes
REE patterns of pyroxenes (Type-I and -II) are shown
in Fig. 10. Type-I pyroxenes have mostly LREE-depleted
patterns in which normalized abundances progressively increase from Nd to Lu, although enrichments of La and/or
Ce (and Pr for orthopyroxene) are observed (Fig. 10a).
REE abundances for Type-I pyroxenes increase from the
orthopyroxene core to the pigeonite mantle and then to
the augite rim. The La, Ce and Pr enrichments are most
pronounced in the orthopyroxene core (lowest REE abundance). These enrichments are probably due to redistribution during terrestrial alteration, which is known to
increase LREE contents heterogeneously in pyroxene
grains and produce V-shaped REE patterns as observed
in Type-I pyroxenes (Wadhwa et al., 2001; Crozaz et al.,
2003). Most augite outermost rims cannot be analyzed by
the ion microprobe because of their narrow widths. TypeII pyroxenes also have LREE-depleted patterns
(Fig. 10b). REE abundances for Type-II pyroxene decrease
from the augite core to the pigeonite mantle. Type-II augite
cores are distinctly enriched in REE abundances relative to
those of Type-I augite rims, even though their major element compositions are almost identical (Fig. 7). Moreover,
Type-II augite has a REE pattern that is less depleted in
LREE than that of Type-I augite; the normalized [Ce/Yb]
ratios are 0.13 and 0.07 for the Type-II augite core and
the Type-I augite rim, respectively. Other trace element
data for pyroxene are listed in Table 6.
1718
Table 2
Representative microprobe analyses of pyroxenes in Y-980459
Type-I
Type-II
Core
orthopyroxene
56.7
0.05
0.36
0.56
11.8
0.42
29.3
1.27
n.d
100.4
Atoms per 6 oxygens
Si
2.00
Ti
0.001
Al(IV)b
0.000
Al(VI)
0.015
Cr
0.015
Fe2+
0.348
Mn
0.013
Mg
1.54
Ca
0.048
Na
n.d
Total
3.98
Molar compositions
Mg#
81.6
En
79.6
Wo
2.5
Fs
18.0
a
b
(1r)
0.1
<0.01
0.01
0.03
0.1
0.02
0.08
0.02
51.7
0.29
2.31
0.70
17.9
0.64
17.4
8.11
0.09
99.2
1.95
0.008
0.049
0.053
0.021
0.565
0.020
0.981
0.328
0.006
3.98
63.5
52.4
17.5
30.1
(1r)
0.1
0.01
0.02
0.03
0.1
0.03
0.06
0.05
0.01
Rim
augite
52.3
0.25
1.77
0.94
13.8
0.56
16.7
12.9
0.12
99.2
1.96
0.007
0.039
0.039
0.028
0.431
0.018
0.935
0.518
0.009
3.98
68.4
49.6
27.5
22.9
Outermost rim
augite
(1r)
0.1
0.01
0.02
0.04
0.1
0.03
0.06
0.07
0.01
48.8
0.84
5.18
0.18
15.8
0.53
11.4
16.7
0.19
99.6
1.87
0.024
0.134
0.100
0.005
0.505
0.017
0.647
0.685
0.014
4.00
56.2
35.2
37.3
27.5
(1r)
0.1
0.02
0.53
0.02
0.1
0.03
0.05
0.08
0.01
Core
augite
52.7
0.30
1.96
0.76
14.7
0.57
16.2
12.9
0.13
100.2
1.96
0.008
0.036
0.050
0.022
0.458
0.018
0.899
0.513
0.009
3.98
66.3
48.1
27.5
24.5
Mantle
pigeonite
(1r)
0.1
0.01
0.02
0.04
0.1
0.03
0.06
0.07
0.01
53.6
0.14
0.71
0.57
16.2
0.61
21.2
6.88
0.06
100.0
1.98
0.004
0.020
0.011
0.017
0.501
0.019
1.17
0.272
0.004
3.99
69.9
60.1
14.0
25.8
(1r)
0.1
0.01
0.01
0.03
0.1
0.03
0.07
0.05
<0.01
Rim
pigeonite
51.3
0.28
2.08
0.43
19.4
0.69
17.5
7.71
0.07
99.5
1.94
0.008
0.057
0.036
0.013
0.615
0.022
0.987
0.313
0.005
4.00
61.6
51.5
16.3
32.1
(1r)
0.1
0.01
0.02
0.03
0.2
0.03
0.06
0.05
<0.01
Outermost rim
augite
47.3
0.98
6.69
0.13
19.2
0.61
10.4
13.9
0.14
99.2
(1r)
0.1
0.02
0.04
0.01
0.2
0.03
0.05
0.07
0.01
1.83
0.029
0.169
0.136
0.004
0.620
0.020
0.597
0.575
0.010
3.99
49.1
33.3
32.1
34.6
Total Fe as FeO.
Tetrahedrally coordinated aluminum, Al (IV), is obtained by assuming that Si and Al (IV) occupy the tetrahedral site. En, enstatite; Wo, wollastonite; Fs, ferrosilite; n.d., not detected.
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
Oxide wt%
SiO2
TiO2
Al2O3
Cr2O3
FeOa
MnO
MgO
CaO
Na2O
Total
Rim
pigeonite
Petrogenesis of olivine-phyric shergottite Y-980459
1719
Table 3
Representative microprobe analyses of chromites in Y-980459
Inclusion in olivine megacryst (Ol-2)
Oxide wt%
SiO2
TiO2
Al2O3
Cr2O3
Fe2O3a
FeO
MnO
MgO
CaO
V2O3
Total
0.18
0.54
5.86
60.8
1.46
18.0
0.34
9.35
0.07
0.41
96.9
Atoms: per 4 oxygens
Si
Ti
Al
Cr
Fe3+
Fe2+
Mn
Mg
Ca
V
Total cation
Molar compositions
Mg#
Cr#
Ulv
Chr
Sp
Mag
(1r)
0.01
0.02
0.04
0.3
0.04
0.1
0.02
0.05
<0.01
0.03
Inclusion in groundmass olivine
0.10
0.79
8.69
56.6
1.93
21.6
0.36
7.75
n.d.
0.58
98.4
0.006
0.014
0.241
1.68
0.038
0.526
0.010
0.487
0.003
0.011
3.01
48.1
87.4
1.4
84.5
12.1
1.9
(1r)
<0.01
0.02
0.05
0.3
0.05
0.2
0.02
0.05
Groundmass
0.13
0.45
6.04
60.7
1.44
21.9
0.45
7.13
n.d.
0.41
98.6
0.03
0.003
0.020
0.353
1.54
0.050
0.624
0.011
0.399
n.d
0.016
3.02
0.004
0.012
0.249
1.68
0.038
0.639
0.013
0.372
n.d
0.011
3.02
39.0
81.4
2.1
77.7
17.8
2.5
36.8
87.1
1.2
84.4
12.5
1.9
(1r)
0.01
0.02
0.04
0.3
0.04
0.2
0.02
0.05
0.03
a
Calculated from stoichiometry (Carmichael, 1967). Ulv, ulvöspinel; Chr, chromite; Sp, spinel; Mag, magnetite; Cr#, molar Cr/
[Cr + Al] 100; n.d., not detected.
and the Y-980459 whole-rock (discussed later in Fig. 13).
Other trace element data for glass are listed in Table 6.
Table 4
Representative microprobe analyses of glasses in Y-980459
Inclusion in groundmass olivine
Oxide wt%
SiO2
TiO2
Al2O3
Cr2O3
FeOa
MnO
MgO
CaO
Na2O
K2O
P2O5
56.6
1.29
14.3
0.09
8.0
0.27
2.15
13.7
1.70
0.06
1.04
Total
99.2
a
(1r)
0.2
0.04
0.1
0.02
0.2
0.03
0.04
0.1
0.05
0.01
0.03
Matrix
51.3
1.03
17.3
0.22
15.2
0.33
0.65
8.76
2.19
0.06
1.23
(1r)
0.2
0.03
0.1
0.03
0.2
0.04
0.02
0.1
0.05
0.01
0.03
98.2
Total Fe as FeO.
3.2.3. Glass
Glass is enriched in REE and other incompatible trace
elements. The glass has a LREE-depleted pattern, which
is almost parallel to those of the Type-II core (Fig. 10b)
4. DISCUSSION
4.1. Origin of olivine megacrysts
The origin of olivine megacrysts has been investigated
by assessing chemical equilibrium between megacryst cores
and the whole-rock composition (presumed to represent a
liquid composition in previous studies). The test of equilibrium has previously been examined based on the ratio
of the partition coefficients of iron and magnesium, called
the exchange coefficient ðKOlMelt
FeMg Þ. However, this approach has two potential problems. One is that employing
the compositions of the olivine core and the whole-rock to
test for chemical equilibrium ideally requires that the core
should have been chemically isolated from the system so
that it experienced no elemental redistribution after crystallization. Since the olivine megacrysts have homogeneous cores in terms of their Fe/Mg ratio (Fig. 5), the
effect of elemental redistribution should be considered.
Secondly, if an olivine is equilibrated with a melt the
Fe–Mg distribution yields a KOlMelt
FeMg value identical to that
1720
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
300
1600
250
Co (ppm)
Ni (ppm)
1400
1200
1000
200
150
Megacrysts (>80 Fo)
Megacrysts (<80 Fo)
Groundmass
800
600
100
4000
55
3500
50
V (ppm)
Cr (ppm)
3000
2500
2000
1σ
40
35
1500
1000
2500
45
3000
3500
4000
4500
5000
5500
Mn (ppm)
30
2500
3000
3500
4000
4500
5000
5500
Mn (ppm)
Fig. 8. Trace element contents of Y-980459 olivines for (a) Ni, (b) Co, (c) Cr and (d) V, versus Mn contents. High-Fo olivines occur only as
the cores of olivine megacrysts. An error bar (1r) in (c) shows analytical reproducibility for Cr, and those for Mn, Ni, Co and V are
comparable to or smaller than the symbols (see Table 5). Cr was measured by electron microprobe, whereas other analyses were made by ion
microprobe.
of the equilibrium condition, but this cannot be reversed.
Therefore, constraints from other elemental distributions
should be helpful. In this study, we focus on the calcium
partitioning between olivine and melt ðDOlMelt
Þ because
CaO
it has been intensively examined experimentally and thermodynamically (e.g. Roeder, 1974; Longhi et al., 1978;
Jurewicz and Watson, 1988; Libourel, 1999). Trace element abundances (Mn, Ni, Co, V and Cr) in olivines
are also used to further constrain the origin of the olivine
megacrysts.
We employed empirically parameterized thermodynamic models to compute an olivine composition that is
in equilibrium with the whole-rock composition (KOlMelt
FeMg
from Toplis, 2005; DOlMelt
from Libourel, 1999). ExperiCaO
mental studies (e.g. Kushiro and Walter, 1998; Toplis,
2005) have suggested that KOlMelt
FeMg is a function of temperature, pressure and the compositions of olivine and melt.
Melting experiments on the Y-980459 whole-rock composition (Dalton et al., 2005; Musselwhite et al., 2006) reported an olivine liquidus field of P = 1 bar to 12 kbar
and T = 1300 °C to 1500 °C. Under these P–T conditions,
the Toplis (2005) model yields KOlMelt
values ranging
FeMg
from 0.31 to 0.34. DOlMelt
is
very
sensitive
to
melt compoCaO
sition and Fo content of the olivine but not to oxygen
fugacity, temperature, or pressure (below 20 kbar) (Jurewicz and Watson, 1988). By introducing an empirically
parameterized Ca partitioning coefficient (DOlMelt
),
CaO
Libourel (1999) successfully reproduced the CaO contents
of olivines within ±15 relative % over a wide range of
temperature and melt compositions. By employing these
OlMelt
KOlMelt
values, CaO and Fo contents of
FeMg and DCaO
olivine in equilibrium with the whole-rock composition
are calculated to be 0.12 ± 0.02 (wt%) and 85–86, respectively (Fig. 11). Moreover, to evaluate the effect of elemental redistribution within the olivine megacrysts, we assume
that homogeneous parts of the olivine megacrysts
(Fo 85) are in equilibrium with a melt having a composition obtained by subtracting the homogeneous olivine
cores from the whole-rock composition. Although the
amount of homogeneous olivine cores cannot be determined, it must be less than the modal proportion of the
olivine megacrysts (6.2 vol%). The olivine compositions
that are in equilibrium with the subtracted whole-rock
compositions are shown as a dashed trajectory in Fig. 11.
The olivine compositions calculated to be in equilibrium
with the Y-980459 whole-rock composition are plotted in a
CaO versus Fo diagram along with the Y-980459 olivine
megacrysts and liquidus olivines from hydrous (2 wt%
of H2O and 5 wt% of CO2, Draper, 2007 ) and anhydrous
(Musselwhite et al., 2006) melting experiments (Fig. 11).
The liquidus olivines from both the hydrous and anhydrous
experiments plot within analytical error of the composi-
Petrogenesis of olivine-phyric shergottite Y-980459
1721
280
Abundance/CI-Chondrite
Type-I opx (core)
Co (ppm)
240
rim
200
160
Megacrysts (>80 Fo)
Megacrysts (<80 Fo)
Groundmass
120
600
800
1000
core
1200
1400
10
Type-I pig (mantle)
Type-I aug (rim)
1
0.1
1600
Abundance/CI-Chondrite
Ni (ppm)
3500
core
Cr (ppm)
3000
2500
2000
rim
1σ
1500
10
1
Type-II aug (core)
0.1
Type-II pig (mantle)
Glass (matrix)
La Ce Pr Nd
1000
Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Fig. 9. (a) Co versus Ni, and (b) Cr versus V, for Y-980459
olivines. Error bar (1r) shows analytical reproducibility for Cr.
Fig. 10. REE patterns of pyroxenes and a glass sample. (a) Type-I
pyroxenes and (b) Type-II pyroxenes and glass. REE abundances
are normalized to C1 chondrite (Anders and Grevesse, 1989). Error
bars (1r) show analytical reproducibility. Abbreviations: opx,
orthopyroxene; pig, pigeonite; aug, augite.
tional range of the olivine megacrysts, though one anhydrous datum shows higher CaO content than the megacrysts. The hydrous experiment yields relatively lower
CaO olivines than the anhydrous one, consistent with the
fact that DOlMelt
decreases with increasing water contents
CaO
(Feig et al., 2006). On the other hand, the calculated olivine
has lower CaO contents than the megacryst cores, but they
distantly overlap each other if the analytical error is considered. The slight difference between the calculated and measured CaO contents probably arises because the Y-980459
whole-rock composition lies near the limit of the range of
liquids used to calibrate the model (Libourel, 1999), where
the computed CaO contents are sometimes beyond the reported error of ±15 relative % (Fig. 10c in Libourel, 1999).
The chemical equilibrium for other minor elements,
Cr2O3 and MnO, is examined by comparing their abundances in olivine megacrysts with those in the experimentally synthesized olivines (Fig. 12). There are no
appropriate thermodynamic models to calculate Cr2O3
and MnO abundances. The olivines from the anhydrous
experiment plot within ranges of the olivine megacrysts
for both Cr2O3 and MnO. On the other hand, olivines from
30
35
40
45
V (ppm)
Table 5
Representative trace element abundances in olivines of Y-980459
Olivine megacrysts
Core
Sc
V
Mn
Co
Ni
Sr
Y
Zr
Groundmass olivines
Rim
Ol-1 core
Ol-1 rim
Core
Rim
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
4.0
33.6
2765
141
1313
0.61
0.10
0.26
0.3
0.7
9
3
21
0.06
0.01
0.04
6.2
37.4
4135
199
1313
0.69
0.15
0.34
0.4
0.8
12
4
21
0.07
0.02
0.03
7.4
39.1
4020
203
1384
0.32
0.19
0.21
0.3
0.7
11
3
19
0.04
0.02
0.02
9.4
39.5
4871
219
1076
0.48
0.20
0.28
0.3
0.7
13
4
18
0.06
0.02
0.03
6.8
35.1
4509
209
1281
0.33
0.12
0.17
0.3
0.7
12
4
20
0.05
0.01
0.02
9.4
43.5
5081
231
843
2.8
0.22
0.68
0.4
0.9
14
4
18
0.2
0.02
0.06
1722
Table 6
Representative trace element and REE abundances in pyroxenes and glasses of Y-980459
Type-II pyroxenes
Core
orthopyroxene
Sc
V
Co
Ni
Sr
Y
Zr
Ba
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
a
Rim
pigeonite
Rim
augite
Core
augite
Glass
Mantle
pigeonite
Groundmass
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
(ppm)
(1r)
9.9
90
90
370
0.34
0.31
0.77
0.13
0.02
0.06
0.008
0.012
a
—
—
—
0.005
0.041
0.012
0.059
0.006
0.062
0.011
0.4
1
2
9
0.04
0.02
0.05
0.02
0.01
0.01
0.003
0.006
—
—
—
0.003
0.009
0.004
0.010
0.003
0.016
0.005
36
234
89
272
0.51
1.7
0.52
0.061
0.034
0.027
0.003
0.11
0.029
0.011
0.10
0.021
0.34
0.053
0.23
0.047
0.17
0.033
0.7
2
2
9
0.06
0.1
0.05
0.018
0.007
0.008
0.003
0.02
0.017
0.004
0.03
0.008
0.03
0.011
0.02
0.009
0.03
0.011
41
225
93
318
3.2
4.5
3.6
0.48
0.052
0.14
0.015
0.20
0.13
0.047
0.28
0.077
0.71
0.16
0.38
0.053
0.53
0.074
1
2
3
11
0.2
0.3
0.2
0.06
0.010
0.02
0.006
0.03
0.03
0.009
0.05
0.015
0.06
0.02
0.04
0.011
0.06
0.025
34.4
180
117
496
16
9.1
14
1.6
0.18
0.47
0.087
0.59
0.41
0.20
0.76
0.23
1.7
0.41
1.1
0.14
1.0
0.12
0.8
1.3
2
10
0.4
0.3
0.5
0.1
0.02
0.05
0.009
0.04
0.04
0.01
0.09
0.02
0.1
0.04
0.1
0.02
0.1
0.03
57
312
93
234
2.1
3.9
2.3
0.26
0.068
0.077
0.026
0.16
0.16
0.051
0.30
0.094
0.60
0.12
0.42
0.079
0.47
—
1
3
3
21
0.2
0.3
0.3
0.06
0.014
0.023
0.008
0.03
0.04
0.012
0.10
0.025
0.06
0.03
0.06
0.016
0.09
—
3.3
6.4
54
156
65
26
48
4.2
0.40
1.4
0.26
1.6
1.1
0.67
2.5
0.46
4.0
0.83
2.5
0.35
2.2
0.55
1.1
0.3
2
8
1
0.6
1
0.2
0.04
0.1
0.03
0.1
0.1
0.04
0.2
0.05
0.2
0.06
0.1
0.03
0.2
0.06
Not reported.
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
Type-I pyroxenes
0.35
0.6
0.30
0.5
0.25
0.4
Cr2O3 (wt%)
CaO (wt%)
Petrogenesis of olivine-phyric shergottite Y-980459
0.20
1σ
0.15
Y98 ol (This study)
Exp. Draper (2007)
Exp. Musselwhite (2006)
Calc. T & L
0.10
Ol fractionation
(up to 6.2 vol%)
75
80
85
1σ
0.3
0.2
0.1
0.05
70
1723
0.0
70
90
75
Fo
the hydrous experiment show distinctly lower Cr2O3 contents
than those of the olivine megacrysts. The low Cr2O3 abundances are due to the presence of chromite that buffers chromium contents in the liquids (Hanson and Jones, 1998). The
hydrous run products contain abundant chromite inclusions
in olivine grains (Draper, 2007), whereas no chromite inclusions occur in the cores of the olivine megacrysts (except
for one chromite in Ol-2, Fig. 2b) or in the anhydrous run
products (Musselwhite et al., 2006). The difference in the
chromite crystallization sequence among experimental run
products and Y-980459 can be explained by the difference
in their redox states because chromite solubility increases
rapidly with decreasing the oxygen fugacity (fO2) (Barnes,
1998); the hydrous experiment was performed under more
oxidized condition (FMQ) than those for the anhydrous
experiment (IW + 0.7, Musselwhite et al., 2006) and the crystallization of Y-980459 (IW + 0.9, Shearer et al., 2006).
Therefore, hydrous experiments under reduced conditions
so that chromite does not co-crystallize with olivine would
likely produce olivines with Cr2O3 contents similar to the
megacryst cores. These observations suggest that the cores
of the olivine megacrysts (Fo 85) are in chemical equilibrium with a melt that has a composition like that of the
Y-980459 whole-rock.
Trace element systematics (Mn–Ni–Co–Cr–V) also constrain the origin of the olivine megacrysts because these elements have significantly different partitioning behaviors
during olivine crystallization. Ni is a highly compatible element in olivine whereas Mn is incompatible, suggesting that
Ni contents in olivine should generally decrease with
increasing Mn contents. However, this pattern is not the
observed for the high-Fo cores of the olivine megacrysts,
85
90
0.6
0.5
MnO (wt%)
Fig. 11. CaO versus Fo contents of Y-980459 olivine megacrysts
(this study) and liquidus olivines from melting experiments on the
Y-980459 whole-rock composition (Musselwhite et al., 2006;
Draper, 2007). A vertical error bar (1r) on the right shows
analytical reproducibility for CaO; that for Fo is approximately the
size of the symbols. Half-filled square with error bars is the
calculated olivine composition in equilibrium with the Y-980459
whole-rock composition based on thermodynamic models (Calc. T
& L, Libourel, 1999; Toplis, 2005). A hatched line shows a
trajectory of equilibrium olivine compositions corrected for
elemental redistribution within the olivine grain (see text for
details).
80
Fo
0.4
0.3
0.2
70
1σ
Y98 ol (This study)
Exp. Draper (2007)
Exp. Musselwhite (2006)
75
80
85
90
Fo
Fig. 12. Minor element compositions of Y-980459 olivine megacrysts (this study) and liquidus olivines from melting experiments
on the Y-980459 whole-rock composition (Musselwhite et al., 2006;
Draper, 2007). (a) Cr2O3 and (b) MnO contents, versus Fo content.
Vertical error bars (1r) show reproducibility of electron microprobe analysis.
which show positive correlations in the Ni–Mn diagram
(Fig. 8a). The positive correlation requires that Ni behaves
OlMelt
as an incompatible element ðDNiO
< 1Þ. However, based
on a trace element partitioning model (Jones, 1995), Ni bevalue of <1.4; this value is
comes incompatible at a DOlMelt
MgO
distinctly lower than the calculated DOlMelt
value of 2.4 for
MgO
Y-980459 ([molar MgO in Fo85 olivine]/[molar MgO in the
whole-rock]). Another possible explanation for the Ni–Mn
positive correlation is that DOlMelt
increases during crystalNiO
lization of the megacryst cores, because olivine crystallization radically increases the DOlMelt
value (Jones, 1995).
NiO
OlMelt
At DOlMelt
¼
2:4,
D
is
calculated
to be 4.3 ± 2.0
MgO
NiO
(1r) based on the Jones model. Using Ni abundances in
the highest-Fo olivine core (1300 ppm, Fig. 8a) and in the
Y-980459 whole-rock (240 ppm, Dreibus et al., 2003) yields
DOlMelt
¼ 5:4, which is consistent with the Jones model
NiO
within error ðDOlMelt
¼ 4:3 2:0Þ. This suggests that the
NiO
olivine megacryst cores (Fo 85) are in equilibrium with
the Y-980459 whole-rock composition for Ni as well as
the other elements mentioned above. However, an increase
of DOlMelt
estimated by the Jones model is not enough to
NiO
1724
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
outpace Ni depletion of the melt. For example, equilibrium
crystallization of 5 vol% of Fo85 olivines from a melt having the Y-980459 whole-rock composition raises DOlMelt
NiO
from 4.3 ± 2.0 to 5.0 ± 2.0 (1r), whereas the increase of
the observed Ni concentrations in the high-Fo olivines
(1300–1500 ppm, Fig. 8a) should require an increase of
DOlMelt
from 5.4 to 8.1.
NiO
Y-980459 olivines show a positive correlation between
Co and Mn (Fig. 8b). This is inconsistent with the fact that
Co is moderately compatible ðDOlMelt
> 1Þ; Co should deCoO
crease with increasing Mn. Using Co abundances in the Y980459 whole-rock (56.2 ppm, Dreibus et al., 2003) and in
the highest-Fo olivine core (140 ppm, Fig. 8b) yields
DOlMelt
¼ 2:5. The Jones model predicts DOlMelt
¼
CoO
CoO
1:5 0:2ð1rÞ at the DOlMelt
¼
2:4,
which
is
also
greater
MgO
than 1. Positive correlations between Co and Mn are commonly observed in picritic basalts from Earth and Mars
(Karner et al., 2003), despite the compatible behavior of
Co. This correlation has been explained by an increase of
DOlMelt
that outpaces the depletion of Co in the melt (Herd
CoO
et al., 2001). However, this inference does not quantitatively
explain the increase of Co in Y-980459 olivines based on the
Jones model. For example, equilibrium crystallization of
5 vol% of Fo85 olivine from a melt with the Y-980459
whole-rock composition (as done for Ni above) slightly
raises DOlMelt
from 1.5 ± 0.2 to 1.7 ± 0.2 (1r), suggesting
CoO
that the DOlMelt
is almost constant within error. On the
CoO
other hand, the increase of the observed Co concentrations
in the high-Fo olivines (140–180 ppm, Fig. 8b) requires an
increase of DOlMelt
from 2.5 to 3.5. The discrepancy beCoO
tween the observed and modeled partitioning coefficient
(DOlMelt
and DOlMelt
) probably occurs because the liquids
NiO
CoO
used to calibrate the model (mostly terrestrial and lunar
rock compositions, data sources are in Beattie et al., 1991;
Jones, 1995) do not entirely cover the Y-980459 whole-rock
composition. Such a compositional effect on the partitioning
coefficient between olivine and melt was reported for lunar
high-Ti basalts (Jones, 1988). Moreover, Herd et al. (2001)
proposed that fO2 may play an important role in modifying
the melt structure and result in changing DOlMelt
.
CoO
Fig. 9a shows distinct Ni–Co systematics between highand low-Fo olivines. This distinct signature should be
examined by further experimental studies under appropriate conditions for the crystallization of Y-980459 (P, T,
fO2 and composition). Ni and Co concentrations in the
Y-980459 olivines may be affected by co-crystallization of
pyroxene, which would also be constrained by such experimental studies. Most olivine-phyric shergottites contain
olivines with Fo < 80 and show the negative correlation
for the Ni–Co systematic similar to low-Fo olivines in Y980459 (Shearer et al., 2007). Recently, a new olivine-phyric
shergottite (NWA 2046) that contains olivine megacrysts
with Fo > 80 was discovered. Further analyses of Co and
Ni in the high-Fo (>80) olivines will also aid in understanding Ni and Co partitioning during crystallization of Martian basalts (preliminary data have already been reported
by Shearer et al., 2007).
Partitioning coefficients of Cr and V between olivine and
melt are influenced by fO2, and the presence of spinel
phases that generally coexist with olivine dominates the dis-
tributions of these elements (Hanson and Jones, 1998; Canil
and Fedortchouk, 2001; Shearer et al., 2006). The observed
change in slope in the Cr–V diagram (positive for Fo > 80
and negative for Fo < 80 olivines, Fig. 9b) may be explained by co-crystallization of chromite with high-Fo olivine megacrysts (cores) but not with low-Fo megacrysts
(rims) and groundmass olivines. However, this conclusion
is inconsistent with the fact that chromite mostly occurs
in the groundmass or olivine megacryst rims and not as
inclusions in megacryst cores. This is also inconsistent with
the anhydrous experimental result that liquidus olivines do
not coexist with chromites and that these olivines have
Cr2O3 contents indistinguishable from those of the olivine
megacrysts (Fig. 12a). Changing the partitioning coefficient
of V ðDOlMelt
Þ for the high-Fo and low-Fo olivines may ofV
fer another explanation, because DOlMelt
increases with
V
decreasing fO2 (Canil and Fedortchouk, 2001). If that is
the case, the high-Fo olivine megacryst cores crystallized
at a lower fO2 than did the low-Fo olivines. By employing
a relationship between fO2 and DVOlMelt for the Y-980459
whole-rock composition (Shearer et al., 2005), the highFo olivine core (33.6 ± 0.7 ppm, Table 5) crystallized under
an fO2 of IW + 1.1. This is consistent with the previous result fO2 = IW + 0.9 for olivines containing 34 ppm V
(Shearer et al., 2005).
As noted earlier, CSD analysis provides physical constraints for the origin of the olivine megacrysts. CSD patterns of Y-980459 olivines show a distinctive size gap
between megacrysts (0.8–1.2 mm diameter) and groundmass olivines (<0.4 mm) (Greshake et al., 2004; Lentz and
McSween, 2005). Moreover, each group has a distinct linear
slope in the conventional CSD plot (ln [number of grains]
versus width), reflecting a slower cooling rate for the crystallization of olivine megacrysts than for the groundmass
olivines. A theoretical approach for interpretation of the
CSD pattern, based on the Johnson–Mehl–Avrami (JMA)
equation, yields information about crystal sizes (r: radius
in cm) in magma (Marsh, 1996, 1998). Assuming constant
nucleation (J0) and growth rates (G0) for spherical crystals,
the JMA-equation is rewritten as follows:
14
3
G0
:
r ¼ lnð1 f Þ p
J0
Although the assumption above is almost certainly not the
case for magma, this equation provides a rough estimate for
crystal size in magma at a given f (fraction of crystals in
magma). Assuming G0 = 1010 cm/s and J0 = 103 cm3/s
(a reasonable result for basaltic volcanism, e.g. Resmini
and Marsh, 1995) regarding a single cycle of emplacement
of a magma carrying zero crystals, r reaches 0.012 cm
(0.24 mm in diameter) when f reaches 0.20 (approximately
equal to the modal proportion of normative olivine in Y980459, 19 vol%, Greshake et al., 2004). Even considering
one magnitude faster growth rate G0 = 109 cm/s, the crystal size of olivine only reaches up to 0.43 mm in diameter.
These grain sizes are comparable to those of groundmass
olivines but distinctly smaller than the olivine megacrysts.
Thus, multiple melt replenishments would be required to
form olivine megacrysts with sizes greater than 1 mm. The
olivine megacrysts may have formed in a feeder conduit
Petrogenesis of olivine-phyric shergottite Y-980459
4.2. Origin of two types of pyroxenes: implication for Y980459 geochemical reservoir(s)
The parental magmas for various basaltic and olivinephyric shergottites sampled two distinct geochemical
reservoirs which produced correlations in their magmatic
oxidation states, oxygen and radiogenic isotope compositions, and trace element abundances. This correlation has
been interpreted as indicating the presence of a reduced,
low-d18O, less radiogenic, incompatible element-depleted
reservoir and an oxidized, high-d18O, more radiogenic,
incompatible element-rich reservoir (e.g. Borg et al., 1997;
Wadhwa, 2001; Herd et al., 2002). However, these reservoirs do not correlate with mineralogical and major element
compositions such as SiO2 contents or Mg# (Borg and Draper, 2003). For example, olivine-phyric shergottite NWA
1068 has unfractionated major element compositions but
is strongly enriched in incompatible elements (Barrat
et al., 2001), whereas basaltic shergottite QUE 94201, which
represents a fractionated melt, is depleted in incompatible
elements (Fig. 13a) (Kring et al., 2003). Three contrasting
models have been proposed regarding the location and mixing processes of the two geochemical reservoirs: assimilation of oxidized crust by mantle-derived, reduced magmas
(e.g. Wadhwa, 2001; Herd et al., 2002), mixing of two distinct mantle reservoirs during melting (Borg and Draper,
2003), or fluid-induced mantle metasomatism (Treiman,
2003).
Two types of pyroxene zoning are observed in Y-980459,
and they have different REE patterns (Figs. 7 and 10). Mineral REE patterns have been used to identify mineral origins (phenocryst, xenocryst or others) by comparing them
with whole-rock REE patterns, because the mineral REE
patterns reflect those of the melts they equilibrated with
(e.g. Wadhwa et al., 2001). In this section, we examine
the origin of the two types of pyroxenes and then discuss
the geochemical reservoir(s) for producing Y-980459.
Fig. 13b compares measured REE compositions of Y980459 whole-rock and glass with the calculated REE patterns for melts in equilibrium with Type-I and -II augites.
The REE pattern of glass is almost identical to that of
Abundance/CI-chondrite
100
Basaltic shergottite
"enriched"
10
Basaltic shergottite
"depleted"
QUE94201
EETA79001-B
1
Olivine-phyric shergottite
"depleted"
Olivine-phyric shergottite
"enriched"
NWA1068
0.1
100
Abundance/CI-chondrite
where fresh primary melts were continuously supplied, promoting growth of the megacrysts as cumulus crystals.
Summarizing the discussion above, we conclude that the
olivine megacrysts represent cumulus grains which probably formed in a feeder conduit by continuous melt replenishment under reduced conditions (IW + 1). The parent
melt compositions were indistinguishable from the Y980459 whole-rock. Even if the parental melt contains
H2O (0–2 wt%) and CO2 (0–5 wt%), such a hydrous melt
could be also parental to the olivine megacrysts, although
its liquidus temperature (1250 °C at 10 kbar, Draper,
2007) would be significantly lower than that of an anhydrous magma (1550 °C) at the same pressure (Musselwhite
et al., 2006). Greshake et al. (2004) previously suggested
that the olivine megacrysts represent cumulate crystals from
the same magma as the groundmass olivines and pyroxenes.
This study corroborates their suggestion based on more
comprehensive petrologic and geochemical considerations.
1725
10
Melts in equilibrium with
Type-II augite
Type-I augite
1
Y-980459
Glass
Whole-rock
0.1
La Ce Pr Nd
Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Fig. 13. Chondrite-normalized REE compositions for shergottites.
(a) Whole-rock compositions of olivine-phyric shergottite NWA
1068 and basaltic shergottites QUE 94201 and EETA79001
lithology-B. (b) REE compositions of Y-980459 glass and wholerock and calculated melts in equilibrium with Type-I and -II
augites. Shaded and non-patterned areas show REE fields of
geochemically enriched basaltic shergottites (Los Angeles, Shergotty, Zagami and NWA 480) and depleted olivine-phyric shergottites (SaU 005, Dhofar 019, EETA79001 lithology-A, Dar al
Gani 476), respectively. Data are from Lodders (1998), Dreibus
et al. (2000), Rubin et al. (2000), Barrat et al. (2001, 2002a), Neal
et al. (2001) and Kring et al. (2003). REE abundances are
normalized to C1 chondrite (Anders and Grevesse, 1989).
the whole-rock but has somewhat higher absolute concentrations. This is consistent with the observation that Y980459 contains no phosphate minerals that dominate the
whole-rock REE pattern, as is the case for other shergottites (Lundberg et al., 1990). The melt compositions are calculated by using augite/melt partitioning coefficients for
REE ðDAugMelt
Þ appropriate for shergottite compositions
REE
(McKay et al., 1986). These authors found that DAugMelt
REE
exhibits a strong positive correlation with pyroxene wollastonite content. The DAugMelt
values for Type-I and -II augREE
ites used are those for a wollastonite content of Wo28
(Table 2). La and Ce abundances in the Type-I augite were
not used in this calculation because they are not magmatic
but have been affected by terrestrial alteration. The calculated melt for Type-I augite shows a REE pattern similar
to those for glass and the whole-rock along with other depleted olivine-phyric shergottites, although it lacks La and
1726
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
served in Y-980459 in either this or other studies (Greshake
et al., 2004; Ikeda, 2004; Mikouchi et al., 2004; Lentz and
McSween, 2005). Furthermore, major element compositions of the Type-II augite cores (En48Wo28Fs25, Table 2)
fall inside the pyroxene solvus for T > 1200 °C at
P = 5 kbar (Lindsley, 1983), which are unlikely crustal
conditions.
The REE-enrichment of the Type-II core can be explained by kinetic effects that increase apparent DAugMelt
,
REE
resulting in erroneous estimations for the REE compositions of the parental melts. The Type-II pyroxenes are reversely zoned; they have low-Mg# augite cores mantled
by high-Mg# orthopyroxenes or pigeonites (Table 2 and
Fig. 7). Such reverse zoning has been found in terrestrial
(e.g. Ohnenstetter and Brown, 1992) and Martian (Zagami,
Treiman and Sutton, 1992) pyroxenes that grew under disequilibrium conditions. These pyroxenes first grew as partially hollow crystals; later crystallization then filled in
these hollows as cores. Y-980459 pyroxenes crystallized at
a moderate cooling rate (3–7 °C/h, Lentz and McSween,
2005). Dynamic crystallization experiments (Lofgren,
1980) suggested that this cooling rate is high enough to produce pyroxenes with complex chemical zoning and morphology, although a pyroxene of a given shape is not
always characteristic of a single cooling rate. Kinetic effects
on trace element distributions during crystal growth have
Ce. On the other hand, the calculated melt for Type-II augite has a REE pattern similar to those of LREE-enriched
shergottites.
These observations suggest that the normally zoned
Type-I pyroxenes (common in Y-980459) represent phenocrysts; they crystallized from an evolved melt that experienced crystal fractionation from the Y-980459 parental
melt. The Type-II augite cores (uncommon in Y-980459)
might be interpreted to be xenocrysts. However, the xenocrystic origin of the Type-II augite cores is inconsistent with
petrographic and geochemical observations. If the Type-II
augite cores are xenocrysts, they could originate from (1)
an enriched magma mixed into the main Y-980459 magma
(magma mixing), or from (2) crustal wallrock of the magma
chamber or conduit which were incorporated into the main
Y-980459 magma (crustal assimilation). For the former
case (1), the enriched melt (parental to the Type-II augites)
should have affected the REE patterns of the whole-rock
and the glass, because it has LREE abundances more than
ten times greater than those of the whole-rock and glass
(Fig. 13b). However, such REE signatures are not seen in
either the whole-rock or the glass REE pattern; they exhibit
depleted REE patterns similar to those of the other depleted shergottites. For the latter case (2), Y-980459 should
contain crustal materials as xenocrysts (e.g. Fe-oxides,
phosphate, or maskelynite), but they have not been ob-
Depth
+10 km
Stage-3
Stage-3
0 km
Stage-2
Stage-1
-50 km
Stage-2
melt migration
Stage-1
Reduced, depleted
mantle reservoir
olivine
Type-I pyroxene
-100 km
Type-II pyroxene
Fig. 14. Schematic illustration of a multi-stage model for the petrogenesis of Y-980459. Stage-1: Primary melts with compositions
indistinguishable from the Y-980459 whole-rock were extracted from the reduced, depleted mantle reservoir and continuously supplied
through a feeder conduit. Olivine megacrysts crystallized from these melts. Stage-2: A magma pulse entrained cumulus olivine megacrysts and
then crystallized groundmass olivines and pyroxenes (Type-I and -II) during the magma ascent. Stage-3: Magma erupted as a thin lava flow to
produce dendritic olivine and pyroxene crystals. Thin ferroan rims on olivines and pyroxenes formed at this stage.
Petrogenesis of olivine-phyric shergottite Y-980459
been studied experimentally and theoretically (Albarède,
1972; Kennedy et al., 1993; Watson, 1996). Albarède
(1972) suggested that, when diffusion in the liquid is too
slow or crystal growth is too fast to maintain a homogenous liquid, the apparent partitioning coefficient approaches 1. For a highly incompatible element such as
REE (D 1), the apparent partitioning coefficient is several orders of magnitude greater than the equilibrium value
(Kennedy et al., 1993). In addition to the Wo component,
DAugMelt
also depends on Al contents in the liquid and
REE
the amount of tetrahedrally coordinated Al3+ in augite
(Gallahan and Nielsen, 1992; Schosnig and Hoffer, 1998).
However, because both Type-I and -II augites have similar
tetrahedral Al contents (Table 2), the Al effect on DAugMelt
REE
is expected to be much smaller than the kinetic effect.
Therefore, we interpret the Type-II pyroxenes as phenocrysts, and their cores represent later crystallization that
filled in the early crystallized hollow grains. The Type-II augite cores appear to have crystallized under a disequilibrium
condition in which the kinetic effects played an important
role. Thus, the calculated melt for the Type-II augite core
(Fig. 13b) does not represent the real melts from which
the Type-II cores crystallized.
Summarizing the discussion above, we conclude that
both Type-I and -II pyroxenes represent phenocrysts and
they crystallized from melts with the LREE-depleted patterns. Due to the moderate cooling rate (3–7 °C/h), most
pyroxenes (Type-I) have prismatic euhedral to subhedral
outlines, whereas other pyroxenes (Type-II) exhibit complex chemical zonings (Fig. 3). Considering the lack of tangible evidence for xenocrystic materials and the depleted
REE patterns of the glass and the whole-rock, it is likely
that Y-980459 was derived from the depleted mantle reservoir and was not affected by assimilation from an incompatible element-enriched reservoir.
4.3. Petrogenesis of Y-980459
Combining the petrographic and geochemical observations, we propose a three-stage scenario to account for olivine-phyric shergottite Y-980459 (Fig. 14). First, high-Mg#
(65) parental melts from the reduced, depleted mantle reservoir were continuously supplied to a magma chamber
through a feeder conduit where the olivine megacrysts
formed under reduced conditions (IW + 1), which is within the proposed range of Martian mantle conditions (IW to
IW + 1, Herd et al., 2002; Shearer et al., 2006) (Stage-1).
These primary melts were chemically indistinguishable from
the Y-980459 whole-rock composition. The melts could be
hydrous or anhydrous; if the latter, the primary melts must
have separated from their mantle source at a depth of
100 km (Musselwhite et al., 2006). Second, the final magma pulse entrained the cumulus olivine megacrysts and
crystallized the groundmass olivines and pyroxenes (both
Type-I and -II) in a discharging conduit during magma ascent (Stage-2). This is consistent with the quasi-linear CSD
patterns of groundmass minerals (Greshake et al., 2004;
Lentz and McSween, 2005), although Greshake et al.
(2004) proposed that orthopyroxene cores crystallized at
depth, possibly in a magma chamber. Third, this magma
1727
erupted as a thin lava flow, where rapid cooling formed
quench crystals of dendritic olivines and pyroxenes and thin
Fe-rich rims around olivines and pyroxenes, and prevented
crystallization of plagioclase and Ca-phosphate (Stage-3).
4.4. Can Y-980459 be used to represent a primary melt?
One of the most important issues for researchers studying Martian meteorites is whether Y-980459 represents a
primary melt. We have reported that Y-980459 contains
small amounts of cumulus olivines but the cores of the
cumulus olivine megacrysts are chemically in equilibrium
with a melt having a composition indistinguishable from
the Y-980456 whole-rock. This might be explained by a
quasi-reversible process in which cumulus olivines were entrained by a magma pulse into the fractionated magma that
lost the cumulus olivines (Marsh, 1996). Marsh (1996)
intensively examined this process for basaltic magmas in
vigorous and large volume systems like Hawaii. He suggested that this process is generally not entirely reversible
because cumulus grains represent a wide range of conditions of cooling and fractionation and they continuously
strive to reach equilibrium (e.g. Fe/Mg) with the local prevailing magma composition. This can be seen in other olivine-phyric shergottites containing cumulus olivine
megacrysts that are too iron-rich to be in equilibrium with
their whole-rock compositions. Since the Y-980459 wholerock composition has high-Mg# (65) and is chemically
in equilibrium with the cores of the olivine megacrysts, we
conclude that the Y-980459 whole-rock composition can
be used as a Martian primary melt in experimental studies.
5. CONCLUSIONS
Olivine-phyric shergottite Y-980459 contains magnesian
olivine megacrysts set in a groundmass of fine-grained, less
magnesian olivine, pyroxenes, chromite and mesostasis.
The olivine megacrysts have relatively homogeneous cores
with compositions of Fo (80–85) and CaO (0.2 wt%),
which are the most magnesian among the Martian meteorites. Pyroxenes are classified into two types: Type-I pyroxene has an orthopyroxene core mantled by pigeonite with
a rim of augite, and Type-II pyroxene has a subcalcic augite
core mantled by orthopyroxene and pigeonite.
Based on thermodynamic models for Ca–Fe–Mg partitioning between olivine and melt (Libourel, 1999; Toplis,
2005) and the results from melting experiments on anhydrous and hydrous Y-980459 whole-rock compositions
(Musselwhite et al., 2006; Draper, 2007), chemical equilibrium between cores of the olivine megacrysts and the Y980459 whole-rock composition is rigorously examined.
Accordingly, the cores of the olivine megacrysts are in
chemical equilibrium with both anhydrous and hydrous
Y-980459 whole-rock compositions, under reduced conditions so that chromite does not coexist with the olivine
megacrysts. The magnesian cores (Fo80–85) of the olivine
megacrysts and the groundmass olivines (Fo < 80) have distinct trace element systematics, as revealed in Ni–Co and
Cr–V diagrams. These observations suggest that the cores
of the olivine megacrysts crystallized at a lower-fO2
1728
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
(IW + 1) than did the groundmass olivines. Examining the
olivine CSD patterns based on the JMA-equation suggests
that the olivine megacrysts are cumulus grains that probably formed in a feeder conduit by continuous melt replenishment. The parent melt compositions for megacrysts are
indistinguishable from the anhydrous and hydrous Y980459 whole-rock compositions.
Although Type-II pyroxene cores and Type-I pyroxene
rims have almost the same major element compositions
and plot in the field of augite, their trace element compositions are distinct; Type-II augite cores are enriched in REE
abundances relative to Type-I augite rims. The calculated
composition of a melt in equilibrium with the Type-I augite
rim shows a LREE-depleted pattern, which is similar to
those for Y-980459 glass and whole-rock, as well as other
reduced, depleted reservoir shergottites. The calculated melt
composition for the Type-II augite core has a relatively flat
REE pattern like those of the oxidized, enriched shergottites. However, the calculated melt composition does not
represent compositions of melts from which the Type-II
cores crystallized, because the Type-II cores appear to have
crystallized under a disequilibrium condition in which kinetic effects played an important role. Considering the depleted glass and whole-rock REE patterns and absence of
tangible evidence for xenocrystic materials, Y-980459 was
derived from the depleted mantle reservoir and was not affected by assimilation from an incompatible element-enriched reservoir.
We propose a petrogenetic model for Y-980459, which
consists of 3 stages: (1) continuous replenishment of primary melts produced cumulus olivine megacrysts in a feeder conduit, (2) final magma pulse entrained the cumulus
megacrysts and then crystallized groundmass olivines and
pyroxenes during magma ascent in a discharging conduit
and (3) the stage-2 magma erupted as a thin lava flow so
as to form quench crystals and ferroan rims of olivine
and pyroxenes and prevent crystallization of plagioclase
and Ca-phosphate. Although Y-980459 contains minor
cumulus grains, they have hardly changed the bulk-rock
so that the Y-980459 whole-rock composition can be used
as a Martian primary melt for experimental purposes.
ACKNOWLEDGMENTS
We thank A. Patchen for assistance with electron microprobe
analyses. J. Day, L.A. Taylor and T. Kuritani provided insightful
discussions, and D.S. Draper loaned experimental charges and provided mineral compositions in them. We are grateful to A. Treiman, C.K. Shearer and J.A. Barrat for constructive reviews, and
D.W. Mittlefehldt for editorial handling. This research was supported in part by NASA Cosmochemistry Grants NNG06GG36G
to H.Y.M. and NNX07AI45G to C.F.
REFERENCES
Albarède F. (1972) Kinetic disequilibrium in trace-element partitioning between phenocrysts and host lava. Geochim. Cosmochim. Acta 36, 141.
Alexander C. M. O. D. (1994) Trace element distributions within
ordinary chondrite chondrules: implications for chondrule
formation conditions and precursors. Geochim. Cosmochim.
Acta 58, 3451–3467.
Anders E. and Grevesse N. (1989) Abundances of the elements:
meteoritic and solar. Geochim. Cosmochim. Acta 53, 197–214.
Barnes S. J. (1998) Chromite in komatiites, 1. Magmatic controls
on crystallization and composition. J. Petrol. 39, 1689–1720.
Barrat J. A., Blichert-Toft J., Nesbitt R. W. and Keller F. (2001)
Bulk chemistry of Saharan shergottite Dar al Gani 476.
Meteorit. Planet. Sci. 36, 23–29.
Barrat J. A., Gillet P., Sautter V., Jambon A., Javoy M., Gopel C.,
Lesourd M., Keller F. and Petit E. (2002a) Petrology and
chemistry of the basaltic shergottite North West Africa 480.
Meteorit. Planet. Sci. 37, 487–499.
Barrat J. A., Jambon A., Bohn M., Gillet P., Sautter V., Gopel C.,
Lesourd M. and Keller F. (2002b) Petrology and chemistry of
the picritic shergottite North West Africa 1068 (NWA 1068).
Geochim. Cosmochim. Acta 66, 3505–3518.
Beattie P., Ford C. and Russell D. (1991) Partition-coefficients for
olivine-melt and ortho-pyroxene-melt systems. Contrib. Mineral. Petrol. 109, 212–224.
Borg L. E. and Draper D. S. (2003) A petrogenetic model for the
origin and compositional variation of the Martian basaltic
meteorites. Meteorit. Planet. Sci. 38, 1713–1731.
Borg L. E., Nyquist L. E., Taylor L. A., Wiesmann H. and Shih C.Y. (1997) Constraints on Martian differentiation processes from
Rb–Sr and Sm–Nd isotopic analyses of the basaltic shergottite
QUE 94201. Geochim. Cosmochim. Acta 61, 4915–4931.
Canil D. and Fedortchouk Y. (2001) Olivine-liquid partitioning of
vanadium and other trace elements, with applications to
modern and ancient picrites. Can. Mineral. 39, 319–330.
Carmichael I. S. E. (1967) The iron-titanium oxides of salic
volcanic rocks and their associated ferromagnesian silicates.
Contrib. Mineral. Petrol. 14, 36–64.
Crozaz G., Floss C. and Wadhwa M. (2003) Chemical alteration
and REE mobilization in meteorites from hot and cold desert.
Geochim. Cosmochim. Acta 67, 4727–4741.
Dalton, H. A., Musselwhite, D. S., Kiefer, W. K. and Treiman, A.
H. (2005) Experimental petrology of the basaltic shergottite
Yamato 980459: implications for the thermal structure of the
Martian mantle? Lunar Planet. Sci. XXVI. Lunar Planet. Inst.,
Houston #2142 (abstr.).
Dalton, T. G., Sharp, T. G. and Holloway, J. R. (2007)
Investigation of the effects of water on a Martian mantle
composition. Lunar Planet. Sci. XXVIII. Lunar Planet. Inst.,
Houston #2102 (abstr.).
Draper, D. S. (2007) Water-undersaturated near-liquidus phase
relations of Yamato 980459: preliminary results. Lunar Planet.
Sci. XXVIII. Lunar Planet. Inst., Houston #1447 (abstr.).
Dreibus, G., Haubold, R., Huisl. W. and Spettel, B. (2003)
Comparison of the chemistry of Yamato 980459 with DaG 476
and SaU 005. In International Symposium-Evolution of Solar
System Materials: A New Perspective from Antarctic Meteorites, pp. 19–20.
Dreibus G., Spettel B., Haubold R., Jochum K. P., Palme H., Wolf
D. and Zipfel J. (2000) Chemistry of a new shergottite: Sayh al
Uhaymir 005. Meteorit. Planet. Sci. 35 (Suppl.), A49.
Feig S. T., Koepke J. and Snow J. (2006) Effect of water on
tholeiitic basalt phase equilibria: an experimental study
under oxidizing conditions. Contrib. Mineral. Petrol. 152,
611–638.
Floss C., James O. B., McGee J. J. and Crozaz G. (1998) Lunar
ferroan anorthosite petrogenesis: clues from trace element
distributions in FAN subgroups. Geochim. Cosmochim. Acta
62, 1255–1283.
Gallahan W. E. and Nielsen R. L. (1992) The partitioning of Sc, Y,
and the rare-earth elements between high-Ca pyroxene and
Petrogenesis of olivine-phyric shergottite Y-980459
natural mafic to intermediate lavas at 1-atmosphere. Geochim.
Cosmochim. Acta 56, 2387–2404.
Goodrich C. A. (2002) Olivine-phyric Martian basalts: a new type
of shergottite. Meteorit. Planet. Sci. 2002, B31–B34.
Goodrich C. A. (2003) Petrogenesis of olivine-phyric shergottites
Sayh al Uhaymir 005 and Elephant Moraine A79001 lithology
A. Geochim. Cosmochim. Acta 67, 3735–3771.
Greshake A., Fritz J. and Stoffler D. (2004) Petrology and shock
metamorphism of the olivine-phyric shergottite Yamato
980459: evidence for a two-stage cooling and a single-stage
ejection history. Geochim. Cosmochim. Acta 68, 2359–2377.
Hanson B. and Jones J. H. (1998) The systematics of Cr3+ and Cr2+
partitioning between olivine and liquid in the presence of spinel.
Am. Mineral. 83, 669–684.
Herd C. D. K., Borg L. E., Jones J. and Papike J. J. (2002) Oxygen
fugacity and geochemical variations in the Martian basalts:
implications for Martian basalt petrogenesis and the oxidation
state of the upper mantle of Mars. Geochim. Cosmochim. Acta
66, 2025–2036.
Herd C. D. K., Karner J. M., Shearer C. K. and Papike J. J. (2001)
The effect of oxygen fugacity on Co and Ni partitioning in
olivine: insights into martian magmas. Meteorit. Planet. Sci. 36
(Suppl.), A37.
Hsu, W. (1995) Ion microprobe studies of the petrogenesis of
enstatite chondrites and eucrites. Ph.D. thesis, Washington
University.
Ikeda Y. (2004) Petrology of the Yamato 980459 shergottite.
Antarct. Meteorite Res. 17, 35–54.
Jones J. (1988) Partitioning of Mg and Fe between olivine and
liquids of lunar compositions: the roles of composition,
pressure and Ti speciation. Proc. Lunar. Sci. Conf. 14th, 561–
562.
Jones J. H. (1995) Experimental trace element partitioning. In Rock
Physics & Phase Relations: A Hand Book of Physical Constants,
vol. 3 (ed. T. J. Ahrens). American Geophysical Union,
Washington, DC, pp. 73–104.
Jurewicz A. J. G. and Watson E. B. (1988) Cations in olivine, Part
1: calcium partitioning and calcium–magnesium distribution
between olivines and coexisting melts, with petrologic applications. Contrib. Mineral. Petrol. 99, 176–185.
Karner J., Papike J. J. and Shearer C. K. (2003) Olivine from
planetary basalts: chemical signatures that indicate planetary
parentage and those that record igneous setting and process.
Am. Mineral. 88, 806–816.
Kennedy A. K., Lofgren G. E. and Wasserburg G. J. (1993) An
experimental study of trace element partitioning between
olivine, orthopyroxene and melt in chondrules: equilibrium
values and kinetic effect. Earth Planet. Sci. Lett. 115, 177–195.
Kring D. A., Gleason J. D., Swindle T. D., Nishiizumi K., Caffee
M. W., Hill D. H., Jull A. J. T. and Boynton W. V. (2003)
Composition of the first bulk melt sample from a volcanic
region of Mars: Queen Alexandra Range 94201. Meteorit.
Planet. Sci. 38, 1833–1848.
Kushiro I. and Walter M. (1998) Mg–Fe partitioning between
olivine and mafic–ultramafic melts. Geophys. Res. Lett. 25,
2337–2340.
Lentz R. C. F. and McSween H. Y. (2005) A textural examination
of the Yamato 9804569 and Los Angels shergottites using
crystal size distribution analysis. Antarct. Meteorite Res. 18, 66–
82.
Libourel G. (1999) Systematics of calcium partitioning between
olivine and silicate melt: implications for melt structure and
calcium content of magmatic olivines. Contrib. Mineral. Petrol.
136, 63–80.
Lindsley D. H. (1983) Pyroxene thermometry. Am. Mineral. 68,
477–493.
1729
Lodders K. (1998) A survey of shergottite, nakhlite and chassigny
meteorites whole-rock compositions. Meteorit. Planet. Sci. 33,
A183–A190.
Lofgren G. E. (1980) Experimental studies on the dynamic
crystallization of silicate melts. In Physics of Magmatic
Processes (ed. R. B. Hargraves). Princeton University Press,
Princeton, NJ, pp. 487–551.
Longhi J., Walker D. and Hays J. F. (1978) Distribution of Fe and
Mg between olivine and lunar basaltic liquids. Geochim.
Cosmochim. Acta 42, 1545–1558.
Lundberg L. L., Crozaz G. and McSween H. Y. (1990) Rare earth
elements in minerals of the ALHA77005 shergottite and
implications for its parent magma and crystallization history.
Geochim. Cosmochim. Acta 54, 2535–2547.
Marsh B. D. (1996) Solidification fronts and magmatic evolution.
Mineral. Mag. 60, 5–40.
Marsh B. D. (1998) On the interpretation of crystal size distributions in magmatic systems. J. Petrol. 39, 553–599.
McKay, G., Le, L., Schwandt, C., Mikouchi, T., Koizumi, E. and
Jones, J. (2004) Yamato 980459: the most primitive shergottite.
Lunar Planet. Sci. XXV. Lunar Planet. Inst., Houston #2145
(abstr.).
McKay G., Wagstaff J. and Yang S.-R. (1986) Clinopyroxene REE
distribution coefficients for shergottites: the REE content of the
Shergotty melt. Geochim. Cosmochim. Acta 50, 927–937.
McSween H. Y., Eisenhour D. D., Taylor L. A., Wadhwa M. and
Crozaz G. (1996) QUE94201 shergottite: crystallization of a
Martian basaltic magma. Geochim. Cosmochim. Acta 60, 4563–
4569.
McSween H. Y. and Jarosewich E. (1983) Petrogenesis of the
Elephant Moraine A79001 meteorite: multiple magma pulse on
the shergottite parent body. Geochim. Cosmochim. Acta 47,
1501–1513.
McSween H. Y. and Treiman A. (1998) Martian meteorite. In
Planetary Materials, vol. 36 (ed. J. J. Papike), p. 53. Reviews in
Mineralogy. Mineralogical Society of America, Washington,
DC.
Mikouchi T., Koizumi E., McKay G., Monkawa A., Ueda Y.,
Chokai J. and Miyamoto M. (2004) Yamato 980459: mineralogy and petrology of a new shergottite-related rock from
Antarctica. Antarct. Meteorite Res. 17, 13–34.
Musselwhite D. S., Dalton H. A., Kiefer W. S. and Treiman A. H.
(2006) Experimental petrology of the basaltic shergottite
Yamato-980456: implications for the thermal structure of the
Martian mantle. Meteorit. Planet. Sci. 41, 1271–1290.
Neal, C. R., Taylor, L. A., Ely, J. C., Jain, J. C., and Nazarov, M.
A. (2001) Detailed geochemistry of new shergottite, Dhofar
019. Lunar Planet. Sci. XXII. Lunar Planet. Inst., Houston
#1671 (abstr.).
Norris, J. R. and Herd, C.D.K. (2006) The Yamato 980459
liquidus at 10 to 20 kilobars. Lunar Planet. Sci. XXVII. Lunar
Planet. Inst., Houston #1787 (abstr.).
Ohnenstetter D. and Brown W. L. (1992) Overgrowth textures,
disequilibrium zoning, and cooling history of a glassy 4pyroxene boninite dyke from New-Caledonia. J. Petrol. 33,
231–271.
Resmini R. G. and Marsh B. D. (1995) Steady-state volcanism,
paleoeffusion rates, and magma system volume inferred from
plagioclase crystal size distributions in mafic lavas: Dome
Mountain, Nevada. J. Volcanol. Geotherm. Res. 68, 273–296.
Roeder P. L. (1974) Activity of iron and olivine solubility in
basaltic liquids. Earth Planet. Sci. Lett. 23, 397–410.
Rubin A. E., Warren P. H., Greenwood J. P., Verish R. S., Leshin
L. A., Hervig R. L., Clayton R. N. and Mayeda T. K. (2000)
Los Angeles: the most differentiated basaltic Martian meteorite.
Geology 28, 1011–1014.
1730
T. Usui et al. / Geochimica et Cosmochimica Acta 72 (2008) 1711–1730
Schosnig M. and Hoffer E. (1998) Compositional dependence of
REE partitioning between diopside and melt at 1 atmosphere.
Contrib. Mineral. Petrol. 133, 205–216.
Shearer, C. K., Borg, L. E., Papike, J. J., Chaklader, J., Symes, S.
J., Irving, A. J. and Herd, C. D. K. (2005) Do early liquidus
phases in olivine-phyric Martian basalts reflect the characteristics of their mantle sources? Insights from NWA 1110, NWA
1195, and NWA 2049. Lunar Planet. Sci. XXVI. Lunar Planet.
Inst., Houston #1193 (abstr.).
Shearer, C. K., Burger, P. V., Papike, J. J., Borg, L. E., Irving, A. J.
and Herd, C. D. K. (2007) Petrogenetic linkages among
Martian basalts. Implications based on trace element chemistry
of olivine. Lunar Planet. Sci. XXVIII. Lunar Planet. Inst.,
Houston #1140 (abstr.).
Shearer C. K., McKay G., Papike J. J. and Karner J. M. (2006)
Valence state partitioning between olivine-liquid: estimates of
the oxygen fugacity of Y980459 and application to other
olivine-phyric Martian basalts. Am. Mineral. 91, 1657–1663.
Taylor L. A., Nazarov M. A., Shearer C. K., McSween H. Y.,
Cahill J., Neal C. R., Ivanova M. A., Barsukova L. D., Lentz
R. C., Clayton R. N. and Mayeda T. K. (2002) Martian
meteorite Dhofar 019: a new shergottite. Meteorit. Planet. Sci.
37, 1107–1128.
Toplis M. J. (2005) The thermodynamics of iron and magnesium
partitioning between olivine and liquid: criteria for assessing
and predicting equilibrium in natural and experimental systems.
Contrib. Mineral. Petrol. 149, 22–39.
Treiman A. H. (2003) Chemical compositions of Martian basalts
(shergottites): some inferences on basalt formation, mantle
metasomatism, and differentiation in Mars. Meteorit. Planet.
Sci. 38, 1849–1864.
Treiman A. H. and Sutton S. R. (1992) Petrogenesis of the Zagami
meteorite—inferences from synchrotron X-ray (SXRF) microprobe and electron-microprobe analyses of pyroxenes. Geochim.
Cosmochim. Acta 56, 4059–4074.
Wadhwa M. (2001) Redox state of Mars’ upper mantle and crust
from Eu anomalies in shergottite. Science 291, 1527–1530.
Wadhwa M., Lentz R. C. F., McSween H. Y. J. and Crozaz G.
(2001) A petrologic and trace element study of Dar al Gani 476
and Dar al Gani 489: twin meteorites with affinities to basaltic
and lherzolitic shergottite. Meteorit. Planet. Sci. 36, 195–208.
Watson E. B. (1996) Surface enrichment and trace-element uptake
during crystal growth. Geochim. Cosmochim. Acta 60, 5013–
5020.
Zinner E. and Crozaz G. (1986a) Ion probe determination of the
abundances of all the rare earth elements in single mineral
grains. In Secondary Ion Mass Spectrometry, SIMS (eds. A.
Benninghoven, R. J. Colton, D. S. Simons and H. W. Werner).
Springer-Verlag, Berlin, pp. 444–446.
Zinner E. and Crozaz G. (1986b) A method for the quantitative
measurement of are earth elements by ion microprobe. Int. J.
Mass Spectrom. Ion Process. 69, 17–38.
Associate editor: David W. Mittlefehldt