Available online at www.sciencedirect.

com

Chemical Geology 250 (2008) 16 – 28

www.elsevier.com/locate/chemgeo

PGE and Os-isotopic variations in lavas from Kohala Volcano, Hawaii:

Constraints on PGE behavior and melt/crust interaction
M. Jamais a,b,⁎, J.C. Lassiter c , G. Brügmann a,b
a
Max-Planck Institut für Chemie, Postfach 3060, 55020 Mainz, Germany
b
Institut für Geowissenschaften, Universität Mainz, Becherweg 21, 55099 Mainz, Germany
c
Department of Geological Sciences, Jackson School of Geosciences, University of Texas at Austin, Austin, TX 78751, USA
Received 26 May 2007; received in revised form 17 January 2008; accepted 20 January 2008

Editor: S.L. Goldstein

Abstract

We have examined Re, Platinum-Group Element (PGE) and Os-isotope variations in suites of variably fractionated lavas from Kohala Volcano,
Hawaii, in order to evaluate the effects of melt/crust interaction on the mantle isotopic signature of these lavas. This study reveals that the behavior
of Os and other PGEs changes during magma differentiation. The concentrations of all PGEs strongly decrease with increasing fractionation for
melts with MgO b 8 wt.%. Fractionation trends indicate significantly higher bulk partition coefficients for PGEs in lavas with less than 8 wt.%
MgO (DPGE = 35–60) when compared to values for more primitive lavas with MgO N 8 wt.% (DPGE ≤ 6). This sudden change in PGE behavior
most likely reflects the onset of sulfur saturation and sulfide fractionation in Hawaiian magmas at about 8 wt.% MgO.
The Os-rich primitive lavas (≥ 8 wt.% MgO, N 0.1 ppb Os) display a narrow range of 187Os/188Os values (0.130–0.133), which are similar to
values in high-MgO lavas from Mauna Kea and Haleakala Volcanoes and likely represent the mantle signature of Kohala lavas. However, Os-
isotopic ratios become more radiogenic with decreasing MgO and Os content in evolved lavas, ranging from 0.130 to 0.196 in the shield-stage
Pololu basalts and from 0.131 to 0.223 in the post-shield Hawi lavas. This reflects assimilation of local oceanic crust material during fractional
crystallization of the magma at shallow level (AFC processes). AFC modeling suggests that assimilation of up to 10% upper oceanic crust could
produce the most radiogenic Os-isotope ratios recorded in the Pololu lavas. This amount of upper crust assimilation has a negligible effect on the
Sr and Nd-isotopic compositions of Kohala lavas. Thus, these isotopic compositions likely represent the composition of the mantle source of
Kohala lavas.
© 2008 Elsevier B.V. All rights reserved.

Keywords: Osmium isotopes; PGE; Bulk partition coefficient; Crustal contamination; Sulfides; Hawaii

1. Introduction
Ocean Island Basalts (OIB) provide important information on
the chemical and physical characteristics of their mantle sources.
However, the composition of an ascending magma may be
modified during magma ascent and storage. Wallrock assimila-
tion may occur at various depths during magma ascent (e.g.,
lithospheric mantle, oceanic crust, volcanic edifice) (Reisberg et
al., 1993; Martin et al., 1994; Roy-Barman and Allegre, 1995;
⁎ Corresponding author. Max-Planck Institut für Chemie, Postfach 3060,
55020 Mainz, Germany. Tel.: +49 6131305394; fax: +49 6131371051.
E-mail addresses: jamais@mpch-mainz.mpg.de (M. Jamais),
lassiter1@mail.utexas.edu (J.C. Lassiter), bruegman@mpch-mainz.mpg.de
(G. Brügmann).
0009-2541/$ - see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.chemgeo.2008.01.028
Widom and Shirey, 1996; Caroff et al., 1999) and several studies
have proposed that much of the chemical and isotopic hetero-
geneity recorded within individual ocean islands reflects interac-
tion of melts with lithospheric material (Halliday et al., 1995; Eiler
et al., 1996; Class and Goldstein, 1997). In order to extract
accurate information regarding the mantle sources of ocean island
basalts, it is therefore necessary to constrain the extent to which
melt/crust interaction has affected isotope ratios in individual
basalt suites.
Osmium isotopes are very sensitive to melt/crust interaction
because of the extreme fractionation of the Re/Os ratio between
the earth's crust and mantle. Because oceanic crust is
characterized by much higher Re/Os ratios than mantle
peridotite, old oceanic crust such as that surrounding the
 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
Fig. 1. Map of Kohala Volcano showing sample locations. Figure modified from Feigenson et al. (1983) and Spengler and Garcia (1988). Pololu volcanics represent the
shield-building stage and Hawi volcanics represent the post-shield-building stage of activity (Stearn and Macdonald, 1946).
Hawaiian Islands possesses much more radiogenic Os-isotope
ratios than mantle-derived magmas. Osmium concentrations are
generally very low in crustal material (∼ 10–30 ppt) and the
sensitivity of Os-isotope ratios as a monitor of contamination
largely depends on the Os content in the magma. Previous
studies reported that Os content decreases during Hawaiian
magma differentiation (Martin et al., 1994), so that the
sensitivity of Os-isotope ratios to crustal assimilation increases
with magma fractionation. In contrast, the differences in Sr- and
Nd-isotope ratios between oceanic crust and mantle peridotite
are usually small, making these isotopic systems less sensitive
to crustal assimilation by ocean island basalts.
In this study, we report Os and other Platinum-Group Element
(PGE) concentrations as well as Os-isotope ratios for a suite of
variably fractionated lavas from Kohala Volcano, Hawaii.
Covariations between indices of fractional crystallization, PGE
concentrations, and Os-isotope ratios are utilized to constrain the
contrasting behavior of the PGEs during early and late stages of
fractional crystallization as well as quantify the amount of
crustal assimilation that accompanied fractional crystallization.
This permits a better constrain on the effects of crustal
assimilation on the mantle signature of OIB.
2. Magmatic evolution of Kohala Volcano
Kohala Volcano is the northernmost and the oldest of the five
major shield volcanoes of Big Island, Hawaii (Fig. 1). Kohala is
an extinct volcano, 35 km long and 24 km wide, and consists of an
eroded dome. The northern flank of the mountain is particularly
eroded, with numerous large river valleys, which provide access to
a relatively long stratigraphic sequence.
17
Lava flows are 455 to 62 ka old (McDougall and Swenson,
1972) and were erupted during two stages of volcanic activity
(Stearn and Macdonald, 1946): the shield-building stage named
“Pololu volcanics”, which forms roughly 99% of the subaerial
volume of the volcano (455 to 260 Ka), and a post-shield-building
stage named “Hawi volcanics” that overlaps the first event (232 to
62 ka) (Spengler and Garcia, 1988). These two units are separated
by an erosional unconformity on the wet northeastern slopes of
the mountain and by a red soil horizon on the dry western slopes.
Pololu volcanics are dominantly tholeiitic in composition, but
grade into alkalic basalts within the upper 100 m of the unit. Hawi
volcanics are highly-fractionated alkaline lavas ranging in
composition from hawaiites to trachytes (Stearn and Macdonald,
1946). Strontium and Nd isotopes in tholeiitic and alkalic lavas
from the Pololu volcanics are undistinguishable, and previous
studies have suggested that the transition from tholeiitic to alkalic
volcanism in the Pololu unit can be explained by a gradual
decrease in the degree of partial melting of a single source
(Hofmann and Feigenson, 1983; Hofmann et al., 1987). In
contrast, the alkaline Hawi lavas from the post-shield-building
stage of volcanism are isotopically and compositionally distinct
from the underlying Pololu lavas (Hofmann et al., 1987; Lanphere
and Frey, 1987; Spengler and Garcia, 1988).
Phenocryst assemblage, major element and transition-metal
variations suggest that olivine is the major crystallizing phase in
tholeiites and that aluminous clinopyroxene dominate over
olivine in the Pololu alkali basalts (Feigenson et al., 1983;
Hofmann et al., 1987). Major and trace element variations in the
Hawi lavas indicate that plagioclase, clinopyroxene, Ti-magne-
tite, olivine and apatite have crystallized in the post-shield-
building stage of Kohala volcanism (Spengler and Garcia, 1988).
18 M. Jamais et al. / Chemical Geology 250 (2008) 16–28

The Sr- and Nd-isotope signatures of the shield-building lavas
of Kohala Volcano (Pololu series) are similar to those in lavas from
other “Kea-trend” volcanoes (e.g. Mauna Kea, Kilauea). Hawi
lavas from the post-shield-building stage extend to slightly more
depleted Sr and Nd-isotope signatures, i.e. higher 143Nd/144Nd and
lower 87Sr/86Sr values, although there is significant overlap
between the Pololu and Hawi ranges of isotopic values (Hofmann
et al., 1987; Lanphere and Frey, 1987). A similar shift to more
depleted isotopic compositions in post-shield-stage lavas has also
been observed at Haleakala, Mauna Kea and Kauai Volcanoes
(Chen and Frey, 1983, Feigenson, 1984; Chen and Frey, 1985;
Kennedy et al., 1991; Lassiter et al., 1996).
3. Samples and analytical method
The lava suite investigated here consists of a total of 15
samples from Kohala: 9 from the Pololu and 6 from the Hawi
lava flows. Petrographic descriptions and sample locations along
with major and trace element compositions are reported in
Feigenson et al. (1983). Strontium and Nd isotopic compositions
as well as transition-metal compositions on a subset of these
samples are presented in Feigenson et al. (1983) and Hofmann
et al. (1987).
The Kohala lavas, and the alkaline series in particular, have
variably fractionated compositions, with Mg numbers ranging
between 63 and 47 in the Pololu series and between 27 and 47 in
the alkaline Hawi series. The Pololu samples investigated here
range from relatively “primitive” tholeiites having MgO N 7 wt.%
(wt.%) to differentiated tholeiites and alkali basalts with MgO
between 7 and 5 wt.% (Table 1). All Hawi samples are highly
differentiated alkaline rocks with MgO content between 5 and
1 wt.%.
Osmium isotope compositions as well as Re, Os, Ir, Ru, and Pt
abundances were determined at the Max-Plank Institut (MPI) für
Chemie in Mainz. Two grams of whole rock powders was
dissolved in quartz vessels together with a mixed Re/PGE spike

Table 1
Re, PGE and isotopic compositions of Kohala lavas
Latitude a Longitude a MgO b 87
Sr/86Sr b 143
Nd/144Nd b 187
Os/188Os Os c Ir c Ru c Pt c Re c
(wt.%) (ppb) (ppb) (ppb) (ppb) (ppb)
Pololu lavas
Tholeiites
W4 20° 06′ 155° 35′ 8.75 0.703676 0.512956 0.1302 ± 4 0.117 0.079 (11) 0.035 (75) 2.63 (3) 0.315 (2)
48.97″ N 34.41″ W
Duplicate 0.1301 ± 2 0.238
W22 20° 06′ 155° 35′ 8.48 0.703601 0.512973 0.1305 ± 5 0.138 0.099 (9) 0.350 (29) 4.51 (2) 0.539 (1)
57.74″ N 16.63″ W
W11 20° 06′ 155° 35′ 7.57 0.703841 0.512944 0.1331 ± 3 0.219 0.213 (4) 0.320 (30) 3.22 (2) 0.613 (1)
49.60″ N 30.32″ W
Duplicate 0.1332 ± 2 0.204 0.217 (4) 0.305 (26) 3.51 (2) 0.599 (1)
W1 20° 07′ 155° 35′ 7.48 0.703672 0.512974 0.1316 ± 4 0.085 0.034 (23) 0.250 (36) 2.87 (3) 0.291 (3)
18.66″ N 54.91″ W
MG0 20° 02′ 155° 43′ 6.94 0.703689 0.512962 0.1545 ± 16 0.009 (2) 0.003 (77) – 1.38 (5) 0.212 (3)
16.71″ N 11.71″ W
KH16 20° 03′ 155° 44′ 6.43 0.703621 0.512969 0.1417 ± 4 0.021 (1)
16.47″ N 40.31″ W
MG3 20° 02′ 155° 43′ 5.94 0.703612 0.512966 0.1613 ± 5 0.007 (1) 0.052 (16) – 0.50 (13) 0.384 (2)
18.23″ N 11.13″ W
Duplicate 0.1616 ± 9 0.003 (13)
Alkali basalts
H1 20° 06′ 155° 34′ 5.29 0.1367 ± 11 0.009 (2)
57.25″ N 52.35″ W
MG1 20° 02′ 155° 43′ 5.25 0.703668 0.512972 0.1956 ± 10 0.008 (2) 0.011 (48) 0.020 (88) 0.10 (44) 0.611 (1)
17.21″ N 11.52″ W

Hawi lavas
KH5 20° 02′ 155° 43′ 4.93 0.1445 ± 4 0.011 (1) 0.003 (77) – 0.13 (37) 0.070 (10)
40.40″ N 03.70″ W
H3 20° 06′ 155° 34′ 4.29 0.703490 0.513012 0.1307 ± 4 0.036
57.25″ N 52.35″ W
KH13 20° 02′ 155° 43′ 3.25 0.1776 ± 9 0.008 (3)
23.93″ N 03.25″ W
KH1 20° 02′ 155° 41′ 1.73 0.703514 0.513000 0.2168 ± 10 0.003 (3)
17.52″ N 07.41″ W
KH19 20° 04′ 155° 45′ 1.16 0.703554 0.513018 0.2228 ± 7 0.005 (9)
20.64″ N 39.25″ W
a
Data from M.D. Feigenson (personal communication).
b
Data from Feigenson et al. (1983) and Hofmann et al. (1987).
c
Numbers into brackets are percentage of blank correction.
 M. Jamais et al. / Chemical Geology 250 (2008) 16–28 19

Fig. 2. Re, PGE–MgO variations in Hawaiian lavas. a)–d): Plots of Os, Ir, Ru and Pt as a function of MgO in Kohala lavas, together with previous data on Hawaiian
basalts, e) MgO–Pd variations in previous data on Hawaiian basalts. The new data emphasize a change in PGE behavior during differentiation of the Hawaiian magma
and a strong and simultaneous decline of all PGE concentrations in lavas with less than 8 wt.% MgO. Bulk DPGE values are graphically determined using FC trends
from a starting composition with 16 wt.% MgO in primitive compositions, and with 8 wt.% MgO in evolved compositions. In a), AFC trends are also calculated from
the same starting compositions (using R = 0.9) to see the impacts of crust assimilation processes on the inferred bulk DOs value. Parameter for oceanic crust is
Os = 21 ppt (see Fig. 5). f) MgO–Re variations indicating that Re concentrations do not clearly correlate with MgO. This is most likely due to the volatile nature of Re
and massive loss during subaerial erruption (see text). Additional data are from Bennett et al. (1996); Bennett et al. (2000); Brandon et al. (1999); Bryce et al. (2005);
Crocket (2000); Hauri et al. (1996); Hauri and Kurz (1997); Lassiter (2003); Lassiter and Hauri (1998); Martin et al. (1994); Mukhopadhyay et al. (2003); Tatsumi et al.
(1999).

(185Re, 190Os, 191Ir, 101Ru, 198Pt) and reverse aqua regia (4 ml
conc. HCl and 6 ml conc. HNO3) for 16 h at 300 °C and 100 bar
using a high-pressure asher (Meisel et al., 2001). Osmium was
extracted from the Re- and PGE-bearing solution as OsO4 into
liquid bromine and trapped in HBr (Brügmann et al., 1999). The
Os-fraction was further purified via micro-distillation (Birck et al.,
1997). Rhenium and the PGEs were then sequentially extracted
from the solution via anion exchange columns, using a technique
modified after Rehkämper and Halliday (1997) and described in
Büchl et al. (2004).
20 M. Jamais et al. / Chemical Geology 250 (2008) 16–28

Osmium isotopic compositions were measured by ion lavas and decreasing MgO, TiO2 and Sc content in Hawi lavas),
counting on a Finnigan MAT-262 thermal ionization mass consistent with the compatible behavior of Os during fractional
spectrometer in negative ion mode (NTIMS). Total Os analytical crystallization (Fig. 2a). The decrease of Os content is sharp in
blanks are less than 1 pg and represent generally less than 3% of Pololu basalts, with values decreasing from 250 ppt in lavas with
the total Os in the digested samples. However, some samples ∼8 wt.% MgO to 10 ppt in lavas with ∼5 wt.% MgO. This Os
analyzed in this study have extremely low-Os content (3–5 ppt) decrease is more gradual in the highly-fractionated Hawi lavas,
and have consequently higher uncertainties (with 3 to 13% blank with values ranging from 35 ppt in lavas with ∼5 wt.% MgO to
correction for Os and 2 to 5% blank corrections on Os-isotopic 3 ppt in lavas with ∼1 wt.% MgO (Fig. 2a).
ratios). In-run analytical uncertainties for Os-isotopic composi- The least fractionated tholeiites from Pololu volcanics (with
tion are reported for each sample in Table 1. Repeated analyses of MgO N 8 wt.%) have primitive mantle (PM)-normalized PGE
an in-house standard (made from ammonium chloroosmate patterns similar to those previously reported for relatively
bought from Johnson Matthey) containing 100 pg Os (n = 38) primitive tholeiites from Mauna Loa and Kilauea Volcanoes
yielded a 2σ external precision for the 187Os/188Os of better than (Tatsumi et al., 1999; Bennett et al., 2000; Crocket, 2000) (Fig.
0.3% (average 187Os/188Os = 0.10693). Duplicate analyses of 3a). These Hawaiian lavas are depleted in I-PGEs (Ir, Os and
several samples were performed, each starting with digestion of a Ru) relative to peridotites and komatiites but enriched in all
separate aliquot of sample powder. Duplicate analyses show good PGEs compared to MORB. In contrast, evolved Pololu and
agreement for 187Os/188Os (better than 0.6%), but significant Hawi lavas are depleted in all PGE compared to primitive
variation in Os concentrations (up to a factor of 2). Given to the Hawaiian tholeiites and the most differentiated compositions
good reproducibility of 187Os/188Os ratios, we suspect that the exhibit MORB-like abundances (Fig. 3b). Finally, rhenium
high variability of Os concentrations measured in our samples abundances in the Pololu lavas vary from 261 to 618 ppm, but
reflects a “nugget effect” resulting from inhomogeneous contrary to PGE abundances, are not clearly correlated with
distribution of Os-bearing phases, such as sulfides, in the lavas. MgO contents.
Although this within-sample heterogeneity is significant, it is
small relative to the total range in Os concentrations observed
within the sample suite. Therefore, this sample heterogeneity has
only a minor effect on the conclusions drawn in this study. The
results of a comparative analytical study of the serpentinite
reference material UB-N (Meisel et al., 2003) further constrain the
quality of Os data from the Mainz laboratory.
The Re, Ir, Ru and Pt abundances were measured by isotope
dilution on a Nu Plasma Multi-Collector ICPMS (MC-ICPMS) at
the MPI. External precision (2σ) of standard solution (10–25 ng/
g) measurements was 0.1–0.2% for 185Re/187Re, 198Pt/194Pt and
191 193
Ir/ Ir, and 0.6% for 101Ru/99Ru. The analyzed samples have
very low PGE concentrations compared to the standard and the
external precision (2σ) of sample measurements was 0.1–0.6%
for 185Re/187Re, 198Pt/194Pt and 191Ir/193Ir, and 0.3–2.2% for
101
Ru/99Ru. The total procedural blanks for Re, Ru, Ir, and Pt
were less than 0.015 ng, 0.30 ng, 0.02 ng, and 0.15 ng,
respectively. The percentages of blank correction are given in
Table 1. As for osmium, uncertainties are higher for samples with
low concentrations, particularly for Ir and Ru where blank
corrects are most significant. Platinum-group element concentra-
tions in two duplicate analyses agreed within 2% for Ir, 5% for Ru,
9% for Pt and 2% for Re.

4. Results

Rhenium and PGE abundances as well as Os-isotopic

compositions of Kohala lavas are reported in Table 1. Kohala
lavas have variable Os concentrations ranging from about 10 up
to 250 ppt in the Pololu basalts, and from 3 to 35 ppt in the Hawi
lavas. Concentrations in the Pololu basalts are within the range of
those previously obtained on variably fractionated basalts from
Haleakala Volcano (Martin et al., 1994). Osmium concentrations
of both Pololu and Hawi lavas decrease with indices of
fractionation (e.g., decreasing MgO, Ni and Cr content in Pololu
Fig. 3. Primitive-mantle normalized PGE patterns for Kohala lavas. Primitive-
mantle values are from McDonough and Sun (1995). MORB field and primitive
Hawaiian basalt field are also shown for comparison. The relatively primitive
tholeiites (N7–8% MgO) from Kohala Volcano display PGE patterns similar to
other primitive Hawaiian lavas. In contrast, evolved tholeiites, alkali basalt and
Hawi lavas have more depleted PGE compositions and their PGE patterns partly
overlap the MORB field. The primitive Hawaiian field is based on data reported
in Tatsumi et al. (1999) and Bennett et al. (2000) for high-MgO tholeiites from
Mauna Loa and Kilauea Volcano. The MORB field is based on data reported in
Tatsumi et al. (1999) for Pacific MORB.
 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
Fig. 4. 187Os/188Os ratios plotted as a function of MgO concentrations in Kohala
lavas. The field displayed by previous data obtained on lavas with both Loa- and
Kea-end-members are also reported for comparison. Osmium isotopic ratios are
very homogeneous in relatively primitive Kohala lavas and overlap the field of
other Kea-trend lavas. 187Os/188Os ratios are more radiogenic in evolved
compositions. Lavas from the post-shield stage of Kohala Volcano present a
very large range of 187Os/188Os, inversely correlated with MgO content.
Additional data are from Bennett et al. (1996); Bennett et al. (2000); Brandon et
al. (1999); Bryce et al. (2005); Hauri et al. (1996); Hauri and Kurz (1997);
Lassiter and Hauri (1998); Martin et al. (1994).
187
Os/188Os ratios range from 0.130 to 0.196 in the shield-
stage Pololu basalts and from 0.131 to 0.223 in the post-shield
Hawi lavas. In the Pololu Volcanics, tholeiites with more than
7 wt.% MgO span a narrow range of values (0.130–0.133) while
evolved tholeiites and alkalic basalts extend to more radiogenic
Os-isotopes ratios (0.137–0.196). In these evolved Pololu basalts
as well as in the Hawi lavas, Os-isotope ratios are inversely
correlated with MgO and Os concentrations (Figs. 4 and 5).
5. PGE behavior during magma evolution
5.1. PGE variation in Pololu basalts
Osmium concentrations decrease sharply during fractional
crystallization of the Pololu magmas. Together with published
data obtained on other Hawaiian basalts (Fig. 2a), these new
results emphasize a sudden change at about 8 wt.% MgO in the
slope of the MgO–Os correlation defined by the Hawaiian
trend. The Os content systematically decreases in the most
primitive compositions, but decreases much more rapidly in
lavas with less than 8 wt.% MgO. This break in slope indicates
a shift in the Os compatibility and suggests the appearance of a
new fractionating phase with a high affinity for Os.
Fig. 2b–d show bi-variate diagrams of MgO vs Ir, Ru and Pt
for Kohala tholeiites and previous literature data on Hawaiian
lavas. Palladium variations were previously studied on variably
fractionated lava series from Kilauea Volcano (Crocket, 2000)
and the obtained MgO–Pd variations are also reported in Fig.
2e. As with Os, results obtained on Kohala and Kilauea basalts
reveal a distinct change in the behavior of Ir, Ru, Pt, and Pd at
about 8 wt.% MgO. In more primitive Hawaiian lavas (N 8 wt.%
MgO), Ir, Ru, Pt, and Pd abundances are rather constant, though
we note a slight increase of Pt and Pd and a slight decrease of Ir
and Ru as the magma composition evolves from 16 to 8 wt.%
21
MgO. In contrast, all PGE concentrations decrease rapidly with
increasing fractional crystallization in lavas with b 8 wt.%
MgO. As will be discussed below (Section 5.4), this sudden
depletion of Ir, Os, Ru, Pt and Pd starting in lavas with b 8 wt.%
MgO most likely reflects the onset of sulfur saturation and
sulfide fractionation. Rhenium abundances in Kohala tholeiites
do not clearly correlate with either MgO or PGE concentrations
(Fig. 2f). However, much of the rhenium initially contained in
subaerial lavas is likely to escape during magma degassing prior
to eruption (Bennett et al., 2000; Lassiter, 2003; Norman et al.,
2004). Therefore, Re variations do not provide unique informa-
tion on the crystallization history of our lava set and will not be
discussed further.
Fig. 5. AFC modelling for the Kohala tholeiites. a) 1/Os–187Os/188Os variations in
Kohala lavas showing assimilation and fractional crystallization (AFC) modeling
for lavas with less than 8 wt.% MgO. AFC trends are calculated from a starting
composition with 8 wt.% MgO and 150 ppt Os, using a DOs = 50 and assuming
assimilation of either upper or lower crust material. Parameters for a 110 Ma-old
upper pacific crust are: 187Os/188Os= 0.821 and Os content= 21 ppt; parameters for a
110 Ma-old lower pacific crust are: 187Os/188Os = 0.371 and Os content=18 ppt (see
text Section 6.3); parameters for the plume-derived melt are 187Os/188Os= 0.133 and
Os content= 150 ppt. The 1/Os–187Os/188Os variations in Kohala tholeiites are best
reproduced assuming assimilation of upper crust material, for an assimilation/
fractionation ratio R =0.8. Small dots in the AFC trend indicate the percentage of
crystallisation from the starting composition with 8 wt.% MgO. Thus, this AFC
modeling suggests that up to 6% crystallization and up to 5% upper crust
assimilation could explain the 1/Os–187Os/188Os variations in Kohala tholeiites for a
range of composition between 8 and 5 wt.% MgO b) AFC modeling for lavas with
compositions between 16 and 8 wt.% MgO. AFC trends are calculated from a
primary melt composition with 600 ppt Os and 187Os/188Os =0.130, assuming upper
crust assimilation, and using a bulk DOs = 6 and different values of R. Osmium
compositions of relatively primitive Kohala tholeiites are best matched when
considering R = 0.2. The AFC modeling suggests that a magma with 8 wt.% MgO
and ∼150 ppt Os has experienced ∼24% olivine crystallization and consequently
∼5% of upper crust assimilation. The Os-isotopic ratios would also have slightly
increased from 0.130 to 0.133. Additional data are as in Fig. 4.
22 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
5.2. PGE behavior in the post-shield stage of Kohala Volcano
Osmium and other PGE concentrations are extremely low in
the highly differentiated Hawi Volcanics (Fig. 2). We note also
that the slope of the MgO–Os correlation defined by Hawi lavas
is less steep than that defined by the Pololu lavas. The overlap of
various Hawaiian lava series in the MgO–Os correlation (Fig.
2a) suggests a similar MgO–Os trend for most primitive
Hawaiian magmas. Based on this assumption, a more moderate
decline in Os content is suggested during the differentiation of
the evolved Hawi magma. Supporting the idea of sulfide
fractionation to explain the rapid decline of PGE in evolved
Pololu lavas (see Section 5.4), the more moderate Os decrease
in the Post-shield lavas could indicate that sulfide phases were
less abundant during differentiation of the Hawi magma than in
the Pololu lavas.
5.3. Bulk distribution coefficient calculation
The MgO–PGE correlations described by the whole set of
Hawaiian lavas (Fig. 2) allow us to calculate the bulk distribution
coefficients for PGE (DPGE) during both early and late stages of
magma differentiation. Clinopyroxene fractionation is important
in the most evolved Kohala lavas (Feigenson et al., 1983;
Hofmann et al., 1987). However, variations in Sc and CaO as a
function of MgO show that Clinopyroxene joins the liquidus
around 6.5% MgO, and then significantly below the abrupt
increase in DPGE values. Olivine is then the major silicate phase
crystallizing from Kohala tholeiites (see Section 2), and a
primary magma composition can be estimated using the model
of reverse olivine crystallization proposed by Danyushevsky et
al. (2000). For this, we considered that olivine in equilibrium
with the primary melt composition has forsterite content (Fo)
= 90, as proposed in Bennett et al. (2000). Starting from the
compositions of 22 Kohala tholeiites (reported in Feigenson et al.
(1983)), melt compositions were corrected for olivine fractiona-
tion by stepwise addition of olivine in 0.1% increments. After
each increment, compositions of the new melt and olivine in
equilibrium were re-evaluated (see Danyushevsky et al. (2000)).
Equilibrium olivine compositions at each step were calculated
assuming an olivine/liquid KD = 0.3 (Roeder and Emslie, 1970)
Ol L Ol L
where KD = (XFe /XFe) / (XMg /XMg) with X the mole fraction,
Ol = olivine and L = liquid. The average primary composition
obtained has ∼16 wt.% MgO, consistent with a primary Os
content of ∼600 ppt according to the Hawaiian MgO–Os trend
(Fig. 2a).
From this primary composition, fractional crystallization
trends have been calculated for different DOs values applying
the Rayleigh fractionation law:
Cm ¼ C0 Tð1  F ÞðD1Þ
where Cm is the Os concentration in the melt, Co is the
initial concentration of Os in the melt (600 ppt), F is the
degree of fractionation and D the bulk distribution coeffi-
cient for Os. The relationship between MgO content in the
melt and its degree of fractionation is given in the reverse
Table 2
Estimation of PGE partitioning during the Hawaiian magma differentiation
D for a magma Ir Os Ru Pt Pd
composition with
8 wt.% b MgO b 16 wt.% 2±1 6±3 4±2 b 0.5 b 0.5
MgO b 8 wt.% 60 ± 40 50 ± 20 40 ± 10 40 ± 10 35 ± 15
fractionation model. Fig. 2a shows that the MgO–Os
correlation defined by Hawaiian lavas with MgO N 8 wt.%
is best matched by fractional crystallization using a bulk
distribution coefficient for Os (DOs) of 6 (± 3). In the same
way, a DOs during the late stage of fractional crystallization
is estimated from the slope of the correlation defined by
Kohala tholeiites. Fractional crystallization trends have been
calculated from a starting composition with 8 wt.% MgO and
150 ppt Os. The MgO–Os correlation defined by evolved Kohala
tholeiites is best matched with a DOs = 50 (±20) (Fig. 2a). This
value is much higher than the value previously suggested by
Martin et al. (1994) for lavas from Haleakala Volcano (DOs = 10 ±
5). However, their calculations assumed a single value of DOs for
both primitive, high-MgO and evolved, low-MgO lavas. As
outlined before, Os data for Hawaiian lavas reveal a significant
increase in DOs during fractional crystallization of Hawaiian lavas
with b8 wt.% MgO. We note also that the DOs = 6 ± 3 calculated
for the most primitive Kohala tholeiites is in the range of DOs
values reported for komatiites (1–2.6, 1.7–2.8, 3.8–7.1, and 2–4,
respectively: Brügmann et al., 1987; Gangopadhyay and Walker,
2003; Puchtel et al., 2004; Gangopadhyay et al., 2005). These
komatiites present similar mineral assemblage (olivine± chromite)
and are consequently comparable in term of bulk distribution
coefficient. In summary, our calculations indicate that the DOs
during magma fractionation shifts from about 6 in primitive
magma compositions to about 50 in lavas with less than 8 wt.%
MgO (Fig. 2a).
The mineral assemblage crystallizing from the Hawi magma
is more complex than for the shield tholeiites and involves
clinopyroxene, olivine and plagioclase (Feigenson et al., 1983;
Spengler and Garcia, 1988). Because fractional crystallization
processes are more difficult to constrain in the limited number
of post-shield lavas examined in this study, a DOs value has not
been estimated for the Hawi lavas.
5.4. Evidence for sulfide fractionation in evolved
Hawaiian lavas
The partitioning of other PGEs in Hawaiian lavas has been
evaluated in the same way and the results are shown in Fig. 2b–
e and summarized in Table 2. Fractionation trends in evolved
compositions have been calculated from a starting melt
composition with 8 wt.% MgO. We note that the smaller
number of PGE data does not allow for as precise an estimate of
the starting composition as for Os. However, these calculations
also support important changes in the bulk partition coefficient
values for PGEs between primitive and evolved lava composi-
tions. As for Os, the estimated bulk partition coefficient values
for Ir, Ru, Pd and Pt are very high in evolved lavas (35–60)
 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
compared to values obtained on primitive lavas compositions
(b0.5–6).
The strong increase in PGE partitioning in lavas with less
than ∼ 8 wt.% MgO requires the arrival of a new phase in the
fractionating assemblage at this stage of magma evolution, and
suggests the onset of sulfur saturation and sulfide fractionation
in Hawaiian magma at around 8 wt.% MgO. Olivine
fractionation cannot account for this abrupt increase in DPGE
values, because olivine is present on the liquidus in both
primitive and evolved Hawaiian magmas. The positive correla-
tion between Cr and MgO observed along the entire suite of
Kohala lavas (see Section 2) further suggests that Cr-spinel
fractionation occurs in both primitive and evolved magmas and
can be ruled out for the same reason. Clinopyroxene
fractionation begins at around 6.5 wt.% MgO. However,
available data does not suggest a high affinity for PGEs in
clinopyroxene (c.f., Hill et al. (2000)). In contrast, Platinum-
group elements are extremely compatible in sulfide, with
sulfide/silicate melt partition coefficients in the range of 103–
105 (Keays and Campbell, 1981; Peach et al., 1990, Fleet and
Stone, 1991; Peach et al., 1994; Fleet et al., 1996; Roy-Barman
et al., 1998). Previous studies have also suggested that PGEs,
especially I-PGEs (e.g. Ir, Os, Ru), may partition into alloy
phases during magma fractionation (Ballhaus and Sylvester,
2000; Brenan and Andrews, 2001). However, there are many
different types of PGE alloys. Typically, these alloy phases
incorporate either I-PGEs or P-PGE, but not both simulta-
neously. Given the very different behavior of the I-PGEs and P-
PGEs in Hawaiian magmas with N 8 wt.% MgO (see Fig. 2),
simultaneous onset of I-PGE and P-PGE alloy fractionation at
the same stage of magma evolution (8 wt.% MgO) appears
unlikely. In contrast, sulfide liquid is the only phase known to
efficiently collect all PGEs simultaneously.
5.5. Sulfide proportion in evolved Hawaiian magma
After the magma has reached sulfur saturation, sulfide is
most likely to be the major host phase for all PGEs. It is
therefore possible to evaluate the proportion of sulfide present
in the Hawaiian magma using experimentally determined
sulfide/silicate melt KD values (Fleet et al., 1996; Roy-Barman
et al., 1998; Sattari et al., 2002) and the bulk distribution
coefficients for PGE calculated above. The bulk distribution
coefficient for any PGE is given by:
DPGE ¼ Xsulfide TKDPGE ðsulfide=meltÞ
þXðsilicateþoxideÞ TKDPGE ððsilicateþoxideÞ=meltÞ
where X represents the proportion of the different fractionating
phases in the mineral assemblage and KD PGE
represents the
partition coefficient between each mineral phase and silicate
melt for a given PGE. Osmium, Ir and Ru are moderately
compatible during the early stage of magma differentiation (Fig.
2, Table 2), whereas Pt and Pd are incompatible. Phases such as
olivine or Cr-spinel may control the I-PGE variations in the
primitive Hawaiian lavas (Barnes et al., 1985; Brügmann et al.,
23
1987; Sattari et al., 2002; Righter et al., 2004). However, precise
partitioning of noble metals in these phases remains a subject of
debate and the present data set does not allow for further constraint
on the possible I-PGE-bearing phase present during early stages of
magma differentiation. Assuming X (silicate + oxide) ≈ 1, and
assuming that the increase in bulk DPGE in lavas with b8 wt.%
MgO reflects the onset of sulfide segregation, the above equation
can be rewritten as follows:
DPGE ðMgO b8 wt:%Þ  DPGE ðMgO N 8 wt:%Þ
cXsulfide TKDPGE ðsulfide=meltÞ :
PGE 4 5
Given a range of KD (sulfide/melt) between 10 and 10 , we
estimate a sulfide proportion of 0.1–0.3% for a range of bulk
DPGE between 35 and 60 in lavas with b 8 wt.% MgO. This
estimation agrees well with what we should expect for a sulfur-
saturated and sulfide-fractionating magma. For example, if we
consider a typical sulfur content of ∼ 1000 ppm in Hawaiian
magma at sulfur saturation (see Mavrogenes and O'Neill
(1999)), 1% crystallization in the melt would produce
∼10 ppm over-saturation. If we consider a sulfur proportion
of ∼30 wt.% in magmatic sulfides, removal of this 10 ppm sulfur
from the melt would produce ∼ 30 ppm sulfide. Consequently,
1% fractionation of the magma at sulfur saturation would induce
∼30 ppm sulfide crystallization and a proportion of sulfide in
the crystallizing assemblage of 30 ⁎ 10− 6 / 1 ⁎ 10− 2 = 3 ⁎ 10− 3 or
0.3%.
We further note that confirmation of sulfide saturation in
Hawaiian lavas comes from the reported observance of included
sulfides in olivine phenocrysts (Dixon and Clague, 2001; Hauri,
2002; Davis et al., 2003). The evidence for sulfide fractionation
recorded in the subaerial Kohala lavas suggests that most
fractional crystallization occurred within the Kohala magma
chamber prior to magma degassing, and not during subaerial
eruption which would have resulted in the loss of sulfur through
degassing, thus preventing sulfide saturation from occurring.
6. Constraints on assimilation and fractional
crystallization (AFC) processes
6.1. Mantle signatures and contamination of the melt at
shallow level
Osmium isotopes in lavas from both the shield-building and
post-shield-building stages of Kohala Volcano are very
heterogeneous, ranging from values similar to those obtained
for other primitive Hawaiian basalts to highly radiogenic values
(Fig. 4).
In the Pololu volcanics, the less fractionated tholeiites have
homogeneous 187Os/188 Os ratios, with an average of ≈ 0.131.
This value is similar to values obtained on high-MgO lavas from
Haleakala, Kilauea and Mauna Kea Volcanoes (Martin, 1991;
Martin et al., 1994; Lassiter and Hauri, 1998; Brandon et al.,
1999), which are thought to represent the mantle source of these
Kea-trend volcanoes. In contrast, 187Os/188 Os ratios are scat-
tered and more radiogenic in evolved Pololu basalts having low-
24 M. Jamais et al. / Chemical Geology 250 (2008) 16–28

Os content (less than 30 ppt) (Fig. 4). The variation is also
correlated with indices of fractionation. Figs. 4 and 5a clearly
show that 187Os/188Os in the Pololu basalts increase with
decreasing MgO content and Os concentrations, reflecting a
progressive contamination of these lavas during fractional
crystallization. In the previous section, we discussed the rapid
decline in Os and PGE concentrations with decreasing MgO for
lavas below ∼8 wt.% MgO, most likely due to sulfide seg-
regation. This drop in Os content during fractional crystallization
should have enhanced the sensitivity of Os-isotopes in evolved
lavas to any contamination with radiogenic materials. This does
not imply that the primitive lavas have not experienced this
process. However, assimilation of even low-Os components is
more easily recognized in the evolved lavas because of their
strong Os depletion. We note also that Os concentrations in
evolved lavas are comparable with the low-Os concentrations
reported for oceanic crust materials (1–35 ppt) (Lassiter and
Hauri, 1998; Blusztajn et al., 2000; Peucker-Ehrenbrink et al.,
2003). The radiogenic 187Os/188Os ratios recorded in evolved
Pololu basalts are most likely due to coupled processes of oceanic
crust assimilation and fractional crystallization (AFC processes).
6.3. Contaminant parameters
According to studies of abyssal peridotites, lithospheric mantle
material possesses very low 187Os/188Os ratios (∼0.125) compared
to Hawaiian basalts (N 0.129) (Martin, 1991; Roy-Barman and
Allegre, 1994; Snow and Reisberg, 1995). Consequently, interac-
tions between the Hawaiian plume and upper mantle material
cannot be responsible for the radiogenic Os-isotopic ratios recorded
in the evolved Kohala lavas. In contrast, the old oceanic crust
surrounding the Hawaiian Islands (110 ± 2 Ma, Waggoner, (1993))
should possess highly radiogenic Os-isotopes and low-Os
concentrations and appears to be the most suitable candidate for
contaminating Kohala magma during its evolution. Rhenium is
moderately incompatible but Os is compatible during partial
melting processes, so that crustal materials possess low-Os
concentrations and high Re/Os ratios compared to the Earth's

6.2. Assimilation in the post-shield lavas from Kohala Volcano

The highly-fractionated lavas from Hawi volcanics also

exhibit a very large range of Os-isotopic ratios, with extremely
radiogenic values in the most fractionated compositions
(187Os/188Os up to 0.223, see Fig. 4). The low-Os concentrations
(3–35 ppt) and the very good correlation displayed between
MgO contents and Os-isotopic ratios in these lavas suggest that
AFC processes are also responsible for their elevated
187
Os/188 Os ratios.
However, we observe a clear distinction in the MgO–187-
Os/188Os correlation between the shield and the post-shield
lavas from Kohala Volcano (Fig. 4). Although Hawi lavas
samples have more evolved compositions than all Pololu basalts
samples, the least fractionated samples have isotopic ratios
similar to those of the more primitive Pololu tholeiites. We saw
above that the Os decrease was more moderate during dif-
ferentiation of the evolved Hawi magma than it was during
differentiation of the evolved Pololu magma (with b8 wt.%
MgO). The sensitivity of Os-isotopes to contamination strongly
depends on the Os content in the lavas. Consequently, a more
moderate decrease in Os during fractional crystallization of the
Fig. 6. Effects of AFC processes on the Sr- and Nd-isotopic systems. a) 187Os/188Os
Hawi magma may have reduced the effect of AFC processes on versus 87Sr/86Sr and b) 187Os/188Os versus 143Nd/144Nd diagrams showing the
the Os-isotopes. The overlap of the Hawi lava trend and the effect of upper crust assimilation on the isotopic compositions of the Kohala
Pololu lava trend in the 1/Os–187Os/188Os variations diagram magma. Parameters for a 110 Ma-old upper pacific crust are 187Os/188Os = 0.821,
(Fig. 5a) suggests that primitive Hawi and Pololu melts are Os content = 21 ppt, 87Sr/86Sr= 0.703, Sr= 130 ppm, 143Nd/144Nd = 0.51313 and
derived from mantle sources with similar Os-isotope character- Nd = 7 ppm (see text Sections 6.3 and 6.5); parameters for the plume-derived melt
are 187Os/188Os = 0.130, Os content = 600 ppt, 87Sr/86Sr = 0.70365, Sr = 285 ppm,
istics and that the type and amount of assimilated material in 143
Nd/144Nd = 0.51297 and Nd = 20 ppm (see text Section 6.5). Previous data
both suites are broadly similar. obtained on Hawaiian lavas (with Os N 50 ppt) are also reported to emphasize the
187
In the following section, we examine the nature and quantity Os/188Os–87Sr/86 Sr trend and the 187Os/188Os–143Nd/144Nd trend displayed by
of material assimilated during evolution of the Pololu magmas. primitive Hawaiian lavas. The 187Os/188Os ratios in evolved Kohala lavas have
After reviewing the potential types of assimilated materials and been contaminated by local crust assimilation and are outliers from the Hawaiian
trend. Contrary to Os-isotopes, small amount of upper crust assimilation does not
their geochemical characteristics, the 1/Os–187 Os/188Os corre- significantly change the Sr- and Nd-isotope compositions of the melt. Additional
lation is used to constrain the extent of coupled assimilation/ data are as in Fig. 4 and Chen and Frey (1991); Feigenson et al. (1983); Hofmann et
fractional crystallization. al. (1987).
 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
mantle. With time, 187Re decay produces an increase in
187
Os/188Os ratios in the crust, resulting in highly radiogenic Os-
isotopic ratios compared to mantle-derived magmas.
In order to evaluate the effects of crustal assimilation on the Os-
isotope systematics of Kohala lavas, we first estimated the average
Os-isotope composition of typical upper and lower oceanic crust.
Peucker-Ehrenbrink et al. (2003) reported Re and Os concentra-
tions in a suite of oceanic crust samples from DSDP/ODP sites 504
and 417/418. For the purposes of AFC modeling, we calculated the
average [Os] and 187Os/188Os of 110 Ma oceanic crust by taking
the weighted mean [Re], [Os], and 187Os/188Os estimated by
Peucker-Ehrenbrink et al. (2003) (187Os/188Os = 0.213; Re/
Os = 353 and Os =21 ppt) and allowing for 110 Ma radiogenic
ingrowth. The obtained parameters for a 110 Ma upper crust are
187
Os/188Os=0.821 and Os content =21 ppt. An average lower
oceanic crust composition was further calculated using the Re, Os
and 187Os/188Os data reported by Blusztajn et al. (2000) for a suite
of lower oceanic crust samples from ODP site 735B. The average
187
Os/188Os composition was obtained here by taking the Os
content associated with each time-corrected Os-isotopic ratio into
account such as average 187 Os/ 188 Os = ∑ [Os ⁎ ( 187 Os/
188
Os)110 Ma-crust] / ∑ Os. The average Os concentration and
187
Os/188Os calculated in this manner for a 110 Ma lower oceanic
crust (Os content = 18 ppt and 187Os/188Os = 0.371) are similar to
the average composition of four gabbroic xenoliths from Hualalai
volcano reported by Lassiter and Hauri (1998) (Os content =
18 ppt and 187Os/188Os = 0.448). These xenoliths have depleted
Sr- and Nd-isotopic compositions indicative of a Pacific crust
origin, and thus provide an independent constraint on the
composition of the local Pacific crust beneath the Hawaiian
Islands. The following section evaluates the amount of assimila-
tion necessary to reproduce the Os-isotope trends observed in the
Kohala lavas, whether upper or lower crustal material is the likely
assimilant, and the effects of this assimilation on incompatible
element isotopic tracers (e.g., Sr- and Nd-isotopes). However, it
should be noted that the data reported by Peucker-Ehrenbrink et
al. (2003) and Blusztajn et al. (2000) for upper and lower crustal
sections reveal very large variations in both the Os concentrations
and isotopic compositions in different portions of upper and lower
oceanic crust.
6.4. AFC modeling: nature and proportion of the
assimilated material
The nature and proportion of assimilated material respon-
sible for the most radiogenic 187Os/188 Os ratios measured in
Table 3
AFC modeling suggestion for the evolution of the Kohala tholeiite composition within the magmatic chamber
MgO (wt.%) Os (ppt)
16 600
8 150
6 7
Assimilant 21
(upper oceanic crust)
Assimilant parameters used for this modeling are also shown.
25
Kohala lavas are investigated here using the AFC equations of
DePaolo (1981) with the assimilant compositions and DOs
values calculated above. Due to the limited constraints on Os
behavior during the post-shield stage of volcanism, the effects
of assimilation/fractional crystallization have only been calcu-
lated for the shield Kohala lavas.
The DOs values calculated previously did not consider the
possible effects of crustal assimilation on the observed MgO–
Os trends. Therefore, it is necessary to evaluate the effect of
AFC processes on the Os content of the melt and on the inferred
DOs value (see equation 6a of DePaolo (1981)). Fig. 2a shows
AFC trends (dashed lines) calculated for the same DOs values
previously inferred (plain lines), but incorporating assimilation
of oceanic crust material with 21 ppt Os (see Section 6.3) and a
maximum assimilation/fractionation mass ratio R = 0.9 (see
below). These calculations indicate that assimilation processes
have only negligible effect on MgO–Os trends and confirm our
previous estimates of DOs in the Hawaiian magma.
The 1/Os–187Os/188Os correlation obtained for evolved
Pololu lavas (MgO b 8 wt.%) have been modeled using the
method of DePaolo (1981). Fig. 5a shows AFC trends calculated
for a starting melt composition with 150 ppt of Os and
187
Os/ 188 Os = 0.133 (see below and Table 4) and using
DOs = 50. Assimilation of both upper or lower oceanic crust
materials is considered using different assimilation/fractionation
mass ratios R. As discussed in DePaolo (1981) and Lassiter and
Luhr (2001), thermal constraints require that this ratio R is b 1
under most conditions. Fig. 5a indicates that assimilation of
upper oceanic crust material can reproduce the radiogenic
187
Os/188Os ratios observed in most of Pololu basalts using a
ratio R ∼ 0.8. In contrast, assimilation of lower oceanic crust
material cannot reproduce the slope of the observed correlation
between 187Os/188Os ratios and Os content in Kohala basalts,
except for unreasonable values of R N 1. This may indicate that
the Kohala Volcano magma chamber was situated at shallow
levels, in the upper portions of the oceanic crust, rather than in the
lower gabbroic levels. This inference is in agreement with the
shallow depth of magma chambers observed for Mauna Loa and
Kilauea volcanoes (Decker et al., 1983; Rowan and Clayton,
1993). However, as previously noted the average composition of
both upper and lower oceanic crust is still fairly uncertain, and
the assimilation of lower crust material cannot be excluded. The
AFC modeling suggests that up to 6% crystallization is necessary
to produce the most radiogenic Kohala tholeiites starting with an
initial melt composition with 8 wt.% MgO (Fig. 6a). For an
assimilation/fractionation ratio R ∼ 0.8, this suggests the Pololu
187
Os/188Os % olivine fractionation % upper crust assimilation
0.130 0 0
0.133 24 5
0.162 24 + 6 5+5
0.821
26 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
magma assimilated up to 5% upper crust material as its
composition evolved from 8 to 5 wt.% MgO.
In order to estimate the total amount of material assimilated
during evolution of Kohala lavas, AFC processes are modeled
for more primitive lavas, ranging in composition from 16 to
8 wt.% MgO. Fig. 5b shows AFC trends calculated for a starting
melt composition with 600 ppt Os and 187Os/188Os = 0.130,
using DOs = 6 and considering assimilation of upper crust
material. The 1/Os– 187 Os/ 188 Os correlation described by
primitive tholeiites from Kea-trend Volcanoes is reproduced
for R = 0.2. This value is much lower than the previous ratio
required for evolved lavas with b 8 wt.% MgO. The assimila-
tion/fractionation ratio R is expected to correlate with the
temperature of the surrounding wallrocks. The hotter the
wallrocks, the less heat is consumed in melting wallrock
material, resulting in higher R values. The inferred increase in R
in the more evolved magmas may reflect an elevation of the
wallrock temperature as the magma evolved, consistent with
gradual heating of the walls of the magma chamber over time.
The AFC trend in Fig. 5b also suggests that ∼ 24% olivine
crystallization is necessary for the magma to evolve from 16 to
8 wt.% MgO, corresponding to an additional ∼ 5% assimilation
of upper crust material. This assimilation would increase the
187
Os/188 Os ratio from 0.130 to 0.133 (Table 3).
In summary, AFC modeling applied to primitive and evolved
melt compositions suggests that up to 10% assimilation of local
upper crust material is responsible for the most radiogenic
187
Os/188Os values observed in Kohala tholeiites. In the
following section, we evaluate the effects of this assimilation
on Sr- and Nd-isotopic systems.
6.5. Impacts of oceanic crust assimilation on Sr- and Nd-
isotopic compositions
In order to examine the effects of assimilation on the Sr- and
Nd-isotopic compositions of Kohala lavas, we have performed
the same AFC modeling using Sr- and Nd-isotopic data.
Parameters for the local upper Pacific crust were estimated
from the Sr- and Nd-isotopic data reported by King et al.
(1993) and Waggoner (1993) for Pacific mid-oceanic ridge
basalts (MORB) from ODP site 843, located approximately
400 km west of Hawaii (87Sr/86Sr = 0.703, Sr = 130 ppm,
143
Nd/144 Nd = 0.51313 and Nd = 7 ppm, see calculation of Os
parameters for the lower oceanic crust Section 6.3). Bulk
distribution coefficients for Sr and Nd of 0.1 were assumed and
parameters for the plume-derived melt composition were
calculated from data obtained for Kohala tholeiites with
MgO N 8 wt.% (Hofmann et al., 1987). Average values obtained
are (87Sr/86Sr)melt = 0.70365, Sr = 300 ppm, (143 Nd/144Nd)-
melt = 0.512966 and Nd = 20 ppm. As for the osmium models,
we used an assimilation/fractionation ratio R = 0.2 until the
magma reaches 8 wt.% MgO and R = 0.8 afterwards. The
resulting AFC trends are plotted in Fig. 6 and indicate that, in
contrast to Os-isotopic ratios, 10% assimilation in the Pololu
magma would not significantly change the 87Sr/86Sr and
143
Nd/144 Nd composition of the primitive Hawaiian melt. The
average Sr isotopic value would decrease 0.0014% from
0.70365 to 0.70364, whereas the average Nd isotopic value
would increase 0.00078% from 0.512966 to 0.512970. These
minor changes are well within the measurement error typically
obtained for these isotopic ratios and are consequently negligible
compared to isotopic heterogeneities measured in primitive
Kohala tholeiites (87Sr/86Sr = 0.70360–0.70374 and 143Nd/
144
Nd = 0.51296–0.51300, see Fig. 6). Therefore, interactions
between the Kohala magma and local oceanic crust material are
not responsible for the Sr- and Nd-isotopic heterogeneities
observed within the Kohala lavas. The Sr- and Nd-isotopic
compositions of even the most evolved Hawaiian compositions
are likely a signature of their mantle source. These results further
indicate that the isotopic differences between Kea- and Loa-
trend volcanoes are unlikely to reflect large extents of crustal
assimilation in Kea-trend magmas, as previously suggested
(Eiler et al., 1996).
7. Conclusion
Correlations between MgO and PGE concentrations in
variably fractionated Kohala and other Hawaiian lavas indicate
a pronounced increase in bulk distribution coefficients for both
P-PGEs and I-PGEs in lavas with b8 wt.% MgO. Bulk
distribution coefficients range from b 1 (for P-PGEs) up to ∼ 6
(for I-PGEs) in lavas with N8 wt.% MgO, but increase to ∼35–
60 for both I-PGEs and P-PGEs in lavas with b 8 wt.% MgO.
This increase in bulk D values most likely reflects the onset of
sulfur saturation and sulfide fractionation in more evolved lavas
as a result of fractional crystallization, and indicates that primary
Hawaiian magmas are sulfur undersaturated.
Osmium isotopes (187Os/188 Os) increase with deõcreasing
MgO and Os concentrations. These correlations are consistent
with contamination of Kohala magmas with Pacific oceanic
crust via coupled assimilation/fractional crystallization. AFC
modelling suggests that assimilation of up to 10% upper oceanic
crust is required to produce the most radiogenic Os-isotope
values observed in the most evolved Kohala tholeiites. Although
Os-isotopes in evolved Kohala lavas are significantly impacted
by assimilation of oceanic crust, this degree of assimilation has a
negligible effect on Sr- and Nd-isotopes. Strontium and Nd
isotopic heterogeneity observed in suites of Kohala lavas reflects
isotopic heterogeneity inherent to the Hawaiian plume rather
than shallow-level contamination of plume-derived lavas during
magma ascent and storage.
Acknowledgements
We wish to thank A.W. Hofmann and M.D. Feigenson for
kindly providing the samples in this study. We are grateful to K.
Zentel for her assistance in the chemistry room and H. Philipp and
M. Griselin for their help during acquisition of these data. We also
thank S.L. Goldstein, R.J. Walker and an anonymous reviewer for
their constructive criticisms of the initial version of the manuscript.
This work was further supported by the Graduierten Kolleg
“Stoffbestand und Entwicklung von Mantel und Kruste” at the
Johannes Gutenberg University of Mainz and by the Max-Planck
Institut.
 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
References
Ballhaus, C., Sylvester, P., 2000. Noble metal enrichment processes in the
Merensky Reef, Bushveld Complex. J. Petrol. 41 (4), 545–561.
Barnes, S.J., Naldrett, A.J., Gorton, A.P., 1985. The origin of the fractionation of
Platinum-Group elements in terrestrial magmas. Chem. Geol. 53, 303–323.
Bennett, V.C., Esat, T.M., Norman, M.D., 1996. Two mantle-plume components
in Hawaiian picrites inferred from correlated Os–Pb isotopes. Nature 381,
221–224.
Bennett, V.C., Norman, M.D., Garcia, M.O., 2000. Rhenium and platinum
group element abundances correlated with mantle source components in
Hawaiian picrites: sulphides in the plume. Earth Planet. Sci. Lett. 183,
513–521.
Birck, J.L., Roy-Barman, M., Capmas, M., 1997. Re–Os isotopic measurements
at the femtomole level in natural samples. Geostand. Newsl. 21, 19–27.
Blusztajn, J., Hart, S.R., Ravizza, G., Dick, H.J.B., 2000. Platinum-group
elements and Os isotopic characteristics of the lower oceanic crust. Chem.
Geol. 168, 113–122.
Brandon, A.D., Norman, M.D., Walker, R.J., Morgan, J.W., 1999. 186Os–
187Os systematics of Hawaiian picrites. Earth Planet. Sci. Lett. 174, 25–42.
Brenan, J.M., Andrews, D., 2001. High-temperature stability of Laurite and Ru–
Os–Ir alloy and their role in PGE fractionation in mafic magmas. Can.
Mineral. 39, 341–360.
Brügmann, G.E., Arndt, N.T., Hofmann, A.W., Tobschall, A.J., 1987. Noble metal
abundances in komatiite from Alexo, Ontario, and Gorgona Island, Columbia.
Geochim. Cosmochim. Acta 51, 2159–2169.
Brügmann, G.E., Yang, J.H., Snow, J.E., Rehkämper, M., Mezger, K., 1999.
Gemeinsame Bestimmung der Os-Isotopie und der Edelmetall-konzentra-
tionen in Peridotiten, Basalten, Sulfiden und Eisenmeteoriten. Kolloquium
des DFG-Schwerpunktprogramms ‘Ocean Drilling Programm and Deep Sea
Drilling Project' Tagungsprogramm und Abstracts, pp. 21–22.
Bryce, J.G., DePaolo, D.J., Lassiter, J.C., 2005. Geochemical structure of the
Hawaiian plume: Sr, Nd, and Os isotopes in the 2.8 km HSDP-2 section of
Mauna Kea volcano. Geochem. Geophys. Geosystem 6, (9).
Büchl, A., Brügmann, G.E., Batanova, V.G., Munker, C., Hofmann, A.W., 2004.
Melt percolation monitored by Os isotopes and HSE abundances: a case study
from the mantle section of the Troodos Ophiolite. Earth Planet. Sci. Lett. 204
(385–402).
Caroff, M., Guillou, H., Lamieux, M., Maury, R.C., Guille, G., Cotten, J., 1999.
Assimilation of ocean crust by hawaiitic and mugearitic magmas: an
example from Eiao (Marquesas). Lithos 46, 235–258.
Chen, C.Y., Frey, F.A., 1983. Origin of Hawaiian tholeiites and alkalic basalt.
Nature 302, 785–789.
Chen, C.Y., Frey, F.A., 1985. Trace element and isotopic geochemistry of lavas
from Haleakala Volcano, east Maui, Hawaii: implications for the origin of
Hawaiian basalts. J. Geophys. Res. 90, 8743–8768.
Chen, C.Y., Frey, F.A., 1991. The tholeiite to alkalic basalt transition at
Haleakala Volcano, Mauai, Hawaii. Contrib. Mineral. Petrol. 106, 183–200.
Class, C., Goldstein, S.L., 1997. Plume–lithosphere interactions in the ocean
basins: constraints from the source mineralogy. Earth Planet. Sci. Lett. 150,
245–260.
Crocket, J.L., 2000. PGE fresh basalt, hydrothermal alteration products, and
volcanic incrustations of Kilauea Volcano. Geochim. Cosmochim. Acta 64
(10), 1791–1807.
Danyushevsky, L.V., Della-Pasqua, F.N., Sokolov, S., 2000. Re-equilibration of
melt inclusions trapped by magnesian olivine phenocrysts from subduction-
related magmas: petrological implications. Contrib. Mineral. Petrol. 138,
68–83.
Davis, M.G., Garcia, M.O., Wallace, P., 2003. Volatiles in glasses from Mauna
Loa Volcano, Hawai'i: implications for magma degassing and contamina-
tion, and growth of Hawaiian volcanoes. Contrib. Mineral. Petrol. 144,
570–591.
Decker, R.W., Koyanagi, R.Y., Dvorak, J.J., Lockwood, J.P., Okamura, A.T.,
Yamashita, K.M., Tanigawa, W.R., 1983. Seismicity and surface deformation of
Mauna Loa Volcano, Hawaii. EOS 64, 545–547.
DePaolo, D.J., 1981. Trace element isotopic effects of combined wallrock
assimilation and fractional crystallisation. Earth Planet. Sci. Lett. 53,
189–202.
27
Dixon, J.E., Clague, D.A., 2001. Volatiles in basaltic glasses from Loihi
Seamount, Hawaii: evidence for a relatively dry plume component. J. Petrol.
42 (3), 627–654.
Eiler, J.M., Farley, K.A., Valley, J.W., Hofmann, A.W., Stolper, E.M., 1996. Oxygen
isotope constraints on the source of Hawaiian volcanism. Earth Planet. Sci. Lett.
144, 453–468.
Feigenson, M.D., 1984. Geochemistry of Kauai volcanics and a mixing model
for the origin of Hawaiian alkali basalts. Contrib. Mineral. Petrol. 87,
109–119.
Feigenson, M.D., Hofmann, A.W., Spera, F.J., 1983. Case studies on the origin of
basalt II. The transition from tholeiitic to alkalic volcanism on Kohala Volcano,
Hawaii. Contrib. Mineral. Petrol. 84, 390–405.
Fleet, M.E., Stone, W.E., 1991. Partitioning of platinum-group elements in the
Fe–Ni–S system and their fractionation in nature. Geochim. Cosmochim.
Acta 55, 245–253.
Fleet, M.E., Crocket, J.L., Stone, W.E., 1996. Partitioning of platinum-group
elements (Ir, Os, Ru, Pt, Pd) and gold between sulfide liquid and basalt melt.
Geochim. Cosmochim. Acta 60, 2397–2412.
Gangopadhyay, A., Walker, R.J., 2003. Re–Os systematics of the ca. 2.7-Ga
komatiites from Alexo, Ontario, Canada. Chem. Geol. 196, 147–162.
Gangopadhyay, A., Sproule, R.A., Walker, R.J., Lesher, C.M., 2005. Re–Os
systematics of komatiites and komatiitic basalts at Dundonald Beach,
Ontario, Canada: evidence for a complex alteration history and implications
of a late-Archean chondritic mantle source. Geochim. Cosmochim. Acta 69
(21), 5087–5098.
Halliday, A., Lee, D.C., Tommasini, S., Davies, G.R., Paslick, C.R., Fitton, J.G.,
James, D.E., 1995. Incompatible trace elements in OIB and MORB and
source enrichment in the sub-oceanic mantle. Earth Planet. Sci. Lett. 133,
379–395.
Hauri, E., 2002. SIMS analysis of volatiles in silicate glasses, 2: isotopes and
abundances in Hawaiian melt inclusions. Chem. Geol. 183, 115–141.
Hauri, E.H., Kurz, M.D., 1997. Melt migration and mantle chromatography, 2: a
time-series Os isotope study of Mauna Loa Volcano, Hawaii. Earth Planet.
Sci. Lett. 153, 21–36.
Hauri, E., Lassiter, J.C., DePaolo, D.J., 1996. Osmium isotope systematics of
drilled lavas from Mauna Loa, Hawaii. J. Geophys. Res. 101, 11793–11806.
Hill, E., Wood, B.J., Blundy, J.D., 2000. The effect of Ca-Tschermaks
component on trace element partitioning between clinopyroxene and silicate
melt. Lithos 53, 203–215.
Hofmann, A.W., Feigenson, M.D., 1983. Case studies on the origin of basalt. I.
Theory and reassessment of Grenada basalts. Contrib. Mineral. Petrol. 84,
382–389.
Hofmann, A.W., Feigenson, M.D., Raczek, I., 1987. Kohala revisited. Contrib.
Mineral. Petrol. 95, 114–122.
Keays, R.R., Campbell, I.E., 1981. Precious metals in the Jimberlana Intrusion,
Western Australia: implications for the genesis of platiniferous ores in
layered intrusions. Econ. Geol. 76, 1118–1141.
Kennedy, A., Kwon, S.T., Frey, F.A., West, H.B., 1991. The isotopic
composition of postshield lavas from Mauna-Kea Volcano, Hawaii. Earth
Planet. Sci. Lett. 103, 339–353.
King, A.J., Waggoner, D.G., Garcia, M.O., 1993. Geochemistry and petrology
of basalts from leg 136, central pacific ocean. Proc. ODP, Sci. Results 136,
107–118.
Lanphere, M.A., Frey, F.A., 1987. Geochemical evolution of Kohala Volcano.
Hawaii. Contrib. Mineral. Petrol. 95, 100–113.
Lassiter, J.C., 2003. Rhenium volatility in subaerial lavas: constraints from
subaerial and submarine portions of the HSDP-2 Mauna Loa Drillcore. Earth
Planet. Sci. Lett. 214, 311–325.
Lassiter, J.C., Hauri, E.H., 1998. Osmium-isotope variations in Hawaiian lavas:
evidence for recycled oceanic lithosphere in the Hawaiian plume. Earth
Planet. Sci. Lett. 164, 483–496.
Lassiter, J.C., Luhr, J.F., 2001. Osmium abundance and isotope variations in mafic
Mexican volcanic rocks: evidence for crustal contamination and constraints on
the geochemical behavior of osmium during partial melting and fractional
crystallization. Geochem. Geophys. Geosyst. 2 2000GC000116.
Lassiter, J.C., DePaolo, D.J., Tatsumoto, M., 1996. Isotopic evolution of Mauna
Kea Volcano: results from the initial phase of the Hawaii Scientific Drilling
Project. J. Geophys. Res. 101, 11769–11780.
28 M. Jamais et al. / Chemical Geology 250 (2008) 16–28
Martin, C.E., 1991. Osmium isotopic characteristics of mantle-derived rocks.
Geochim. Cosmochim. Acta 55, 1421–1434.
Martin, C.E., Carlson, R.W., Shirey, S.B., Frey, F.A., Chen, C.Y., 1994. Os isotopic
variations in basalts from Haleakala Volcano, Maui, Hawaii: a record of magmatic
processes in oceanic mantle and crust. Earth Planet. Sci. Lett. 128, 287–301.
Mavrogenes, J.A., O'Neill, H.S.C., 1999. The relative effects of pressure,
temperature and oxygen fugacity on the solubility of sulfide in mafic
magmas. Geochim. Cosmochim. Acta 63 (7/8), 1173–1180.
McDonough, W.F., Sun, S.S., 1995. The composition of the Earth. Chem. Geol.
120, 223–253.
McDougall, I., Swenson, D.A., 1972. Potassium–Argon ages of lavas from
Hawi and Pololu volcanics Series, Kohala Volcano, Hawaii. Bull. Geol. Soc.
Am. 83, 3731–3738.
Meisel, T., Walker, R.J., Irving, A.J., Lorand, J.P., 2001. Osmium isotopic
compositions of mantle xenoliths: a global perspective. Geochim.
Cosmochim. Acta 65, 1311–1323.
Meisel, T., Fellner, N., Moser, J., 2003. A simple procedure for the determination of
platinum group elements and rhenium (Ru, Rh, Pd, Re, Os, Ir, Pt) using ID-
ICP-MS with an inexpensive on-line matrix separation in geological and
environmental materials. J. Anal. At. Spectrom. 18, 720–726.
Mukhopadhyay, S., Lassiter, J.C., Farley, K.A., Bogue, S.W., 2003. Geochem-
istry of Kauai shield-stage lavas: implications for the chemical evolution of
the Hawaiian plume. Geochem. Geophys. Geosyst. 4 (1).
Norman, M.D., Garcia, M.O., Bennett, S.L., 2004. Rhenium and chalcophile
elements in basaltic glasses from Ko'olau and Moloka'i volcanoes: magmatic
outgassing and composition of the Hawaiian plume. Geochim. Cosmochim.
Acta 68 (18), 3761–3777.
Peach, C.L., Mathez, E.A., Keays, R.R., 1990. Sulphide melt–silicate
distribution coefficient for noble metals and other chalchophile elements
as deduced from MORB: implication for partial melting. Geochim.
Cosmochim. Acta 54, 3379–3389.
Peach, C.L., Mathez, E.A., Keays, R.R., Reeves, S.J., 1994. Experimentally
determined sulfide melt–silicate melt partition coefficients for iridium and
palladium. Chem. Geol. 117, 361–377.
Peucker-Ehrenbrink, B., Bach, W., Hart, S.R., Blusztajn, J., Abbruzzese, T., 2003.
Rhenium–osmium isotope systematics and platinum group element concentra-
tions in oceanic crust from DSDP/ODP Sites 504 and 417/418. Geochem.
Geophys. Geosyst. 4.
Puchtel, I.S., Humayun, M., Campbell, A.J., Sproule, R.A., Lesher, C.M., 2004.
Platinum group element geochemistry of komatiites from the Alexo and Pyke
Hill areas, Ontario, Canada. Geochim. Cosmochim. Acta 68 (6), 1361–1383.
Rehkämper, M., Halliday, A.N., 1997. Development and application of new ion-
exchange techniques for the separation of the platinum group and other
siderophile elements from geological samples. Talanta 44, 663–672.
Reisberg, L.C., Zindler, A., Marcantonio, F., White, W.M., Wyman, D., Weaver,
B., 1993. Os isotope systematics in ocean island basalts. Earth Planet. Sci.
Lett. 120, 149–167.
Righter, K., Campbell, A.J., Humayun, M., Hervig, R.L., 2004. Partitioning of
Ru, Rh, Pd, Re, Ir, and Au between Cr-bearing spinel, olivine, pyroxene and
silicate melts. Geochem. Geophys. Geosyst. 68 (4), 867–880.
Roeder, P.L., Emslie, R.F., 1970. Olivine-liquid equilibrium. Contrib. Mineral.
Petrol. 29, 275–289.
Rowan, L.R., Clayton, R.W., 1993. The three-dimensional structure of Kilauea
Volcano, Hawaii, From travel time tomography. J. Geophys. Res. 98,
4355–4375.
Roy-Barman, M., Allegre, C.J., 1994. 187Os/188Os ratios of mid-ocean ridge
basalts and abyssal peridotites. Geochim. Cosmochim. Acta 58, 5043–5054.
Roy-Barman, M., Allegre, C.J., 1995. 187Os/186Os in oceanic island basalts:
tracing oceanic crust recycling in the mantle. Earth Planet. Sci. Lett. 129,
145–161.
Roy-Barman, M., Wasserburg, G.J., Papanastessiou, D.A., Chaussidon, M.,
1998. Osmium isotopic compositions and Os concentrations in sulfide
globules from basaltic glasses. Earth Planet. Sci. Lett. 154, 331–347.
Sattari, P., Brenan, J.M., Horn, I., McDonough, W.F., 2002. Experimental
constraints on the sulfide- and shromite-silicate melt: partitioning behavior of
rhenium and platinum-group elements. Econ. Geol. 97, 385–398.
Snow, J.E., Reisberg, L.C., 1995. Os isotopic systematics of the MORB mantle:
results from altered abyssal peridotites. Earth Planet. Sci. Lett. 133,
411–421.
Spengler, R.S., Garcia, M.O., 1988. Geochemistry of the Hawi lavas, Kohala
Volcano, Hawaii. Contrib. Mineral. Petrol. 99, 90–104.
Stearn, H.T., Macdonald, G.A., 1946. Geology and ground-water resources of the
island of Hawaii. Hawaii Div. Hydrogr. Bull. 9, 362.
Tatsumi, Y., Oguri, K., Shimoda, G., 1999. The behaviour of platinum-group
elements during magmatic differentiation in Hawaiian tholeiites. Geochem
Journal 33, 237–247.
Waggoner, D.G., 1993. The age and alteration of central Pacific Oceanic Crust near
Hawaii, site 843. Proc. ODP, Sci. Results 136, 119–132.
Widom, E., Shirey, S.B., 1996. Os isotope systematics in the Azores:
implications for mantle plume sources. Earth Planet. Sci. Lett. 142,
451–465.