-
Available online at www.sciencedirect.com
www.elsevier.com/locate/gca
Geochimica et Cosmochimica Acta 75 (2011) 1621–1641
Rhenium–osmium isotope and platinum-group elements in theXinjie
layered intrusion, SW China: Implications forsource mantle
composition, mantle evolution, PGE
fractionation and mineralization
Hong Zhong a,⇑, Liang Qi a, Rui-Zhong Hu a, Mei-Fu Zhou b,
Ti-Zhong Gou a,Wei-Guang Zhu a, Bing-Guang Liu c, Zhu-Yin Chu c
a State Key Laboratory of Ore Deposit Geochemistry, Institute of
Geochemistry, Chinese Academy of Sciences, 46 Guanshui Road,
Guiyang 550002, Chinab Department of Earth Sciences, University
of Hong Kong, Hong Kong, China
c Institute of Geology and Geophysics, Chinese Academy of
Sciences, Beijing 100029, China
Received 5 February 2010; accepted in revised form 4 January
2011; available online 13 January 2011
Abstract
The Xinjie mafic–ultramafic layered intrusion in the Emeishan
large igneous province (ELIP) hosts Cu–Ni–platinum groupelement
(PGE) sulfide ore layers within the lower part and Fe–Ti–V
oxide-bearing horizons within the middle part. The majormagmatic
Cu–Ni–PGE sulfide ores and spatially associated cumulate rocks are
examined for their PGE contents and Re–Osisotopic systematics. The
samples yielded a Re–Os isochron with an age of 262 ± 27 Ma and an
initial 187Os/188Os of0.12460 ± 0.00011 (cOs(t) = �0.5 ± 0.1). The
age is in good agreement with the previously reported U–Pb zircon
age, indicat-ing that the Re–Os system remained closed for most
samples since the intrusion emplacement. They have
near-chondriticcOs(t) values ranging from �0.7 to �0.2, similar to
those of the Lijiang picrites and Song Da komatiites.
Exceptionally,two samples from the roof zone and one from upper
sequence exhibit radiogenic cOs(t) values (+0.6 to +8.6), showing
minorcontamination by the overlying Emeishan basalts.
The PGE-rich ores contain relatively high PGE and small amounts
of sulfides (generally less than 2%) and the abundanceof Cu and PGE
correlate well with S, implying that the distribution of these
elements is controlled by the segregation andaccumulation of a
sulfide liquid. Some ore samples are poor in S (mostly
-
1622 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
magma in the ELIP were generated from a plume. This comprised
recycled Neoproterozic oceanic lithosphere, includingdepleted
peridotite mantle embedded with geochemically enriched domains. The
ascending magmas thereafter interacted withminor (possibly
-
Fig. 1. Simplified geological map of the Xinjie layered
intrusion (modified after Zhong et al., 2004). Inset a shows the
distribution of thelayered intrusions in the Pan-Xi area (modified
after Zhong et al., 2002). Insert b illustrates distribution of
major terranes in China and thePan-Xi area (modified after Chung
and Jahn, 1995). Abbreviations: NCB = North China block; YZB =
Yangtze block; SG = Songpan-Ganze accretionary complex; QT =
Qiangtang; LS = Lhasa; HI = Himalayan; TAR = Tarim; MON = Mongolia;
QD = Qaidam;WB = West Burma; STM = Shan-Thai-Malay; IC =
Indochina.
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1623
basaltic lava, sills, dikes and small intrusions, occur in
thewestern margin of the Yangtze block (Li et al., 2003,2006; Zhou
et al., 2006). The basement is overlain by athick sequence (>9
km) of Sinian (610–850 Ma) to Permianstrata composed of clastic,
carbonate, and meta-volcanicrocks (SBGMR, 1991).
The ELIP comprises the Emeishan continental floodbasalts and
spatially associated intrusions. The Emeishanbasalts are exposed
over a rhombic area of�2.5 � 105 km2, with the volcanic succession
ranging fromseveral hundred meters to 5 km in thickness. The
volcanicrocks consist predominantly of tholeiites and andesitic
bas-
-
1624 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
alts, with minor flows and tuffs of trachyte and rhyolite inthe
uppermost sequence (Chung and Jahn, 1995; Xuet al., 2001; Xiao et
al., 2004; Zhang et al., 2006). The bas-alts are divided into
high-Ti and low-Ti groups that areconsidered to have been derived
from different mantlesources (Xu et al., 2001; Xiao et al., 2004).
It is noteworthythat minor picrites associated with the high-Ti
basalts havebeen identified in the Pan-Xi and Lijiang areas (Chung
andJahn, 1995; Zhang et al., 2006), whereas some komatiiticrocks
interbedded with low-Ti olivine basalts were docu-mented in the
Song Da region of northern Vietnam (Hanskiet al., 2004).
Magnetostratigraphic data and field observa-tions suggest that the
bulk of the Emeishan volcanic se-quence formed within 1–2 million
years (Huang andOpdyke, 1998; Ali et al., 2002). Recent SHRIMP and
TIMSU–Pb dating of zircons from silicic ignimbrite,
mafic/ultra-mafic intrusions and diabasic dikes indicate that the
ELIPwas voluminously erupted at �260 Ma, consistent withthe
end-Guadalupian (end Middle Permian) stratigraphicage (Zhou et al.,
2002, 2005; Guo et al., 2004; Zhong andZhu, 2006; He et al.,
2007).
The Pan-Xi area is located in the inner zone of the ELIP(Xu et
al., 2004), which is considered the impact site of therising plume
head (He et al., 2003). The area comprises N–Strending,
fault-controlled, massive basalts, numerous spa-tially associated
mafic–ultramafic intrusions, granites, andsyenites. The ore-bearing
mafic–ultramafic intrusions de-scribed here are exposed along a 300
km-long and 10–30 km-wide belt, constituting the most important
metallo-genic district for Fe–Ti–V and Ni–Cu–(PGE) metals in
Chi-na. Giant Fe–Ti–V oxide deposits occur in several
relativelylarge layered intrusions (13–60 km2), including the
Panzhi-hua, Hongge, Baima and Taihe intrusions (Fig. 1; Yao andDu,
1993; Zhong et al., 2002, 2003, 2004, 2005; Zhou et al.,2005,
2008). Ni–Cu–(PGE) sulfide deposits are hosted inthe Limahe,
Jinbaoshan and Zhubu intrusions (Wanget al., 2005; Tao et al.,
2007, 2008; Zhou et al., 2008). Incontrast, both Fe–Ti–V oxide and
Ni–Cu–PGE mineraliza-tion were discovered in the Xinjie layered
intrusion (Zhonget al., 2004; Wang et al., 2008).
3. PETROGRAPHY OF THE XINJIE INTRUSION
The �260 Ma Xinjie intrusion (Zhou et al., 2002), is asill-like,
7.5 km long, 1–1.5 km wide, and 1200 m thickultramafic–mafic
layered body, which intruded the latePermian Emeishan flood basalts
(Fig. 1). Field investiga-tions reveal that the syenitic intrusions
always cut the Xinjieintrusion and adjacent Emeishan basalts. The
Xinjie ultra-mafic–mafic intrusion exhibits well-developed igneous
lay-ering and has been divided into three cycles containing
sixlithological zones (A–F, Fig. 2). Overall, the intrusion ismost
ultramafic in its lower part and each cycle includesnumerous layers
starting with the most ultramafic cyclicunits at the base followed
by progressively evolved cyclicunits. Cycle I is about 400 m thick
and comprises, fromthe bottom to top, peridotite, plagioclase
peridotite, olivineclinopyroxenite, plagioclase clinopyroxenite,
gabbro andquartz-bearing gabbro. Cycle II is 190 m thick and
mainlyconsists of plagioclase-bearing peridotite, olivine
clinopy-
roxenite, gabbro and quartz-bearing gabbro, while CycleIII is
greater than 600 m thick, and is dominated by plagio-clase
clinopyroxenite, gabbro and quartz diorite. A �20 m-thick
fine-grained gabbroic and olivine-gabbroic MarginalUnit is at the
base of the intrusion and in contact withthe country rocks (Mao and
Sun, 1981), containing up toseveral percent hornfelsed and
partially digested inclusionsof the underlying Emeishan basalts.
The petrographic fea-tures of the different rock types within each
individual cyclehave been described in detail by Zhou (1982) and
Zhonget al. (2004).
In the Xinjie intrusion, the Fe–Ti oxide ore layers occurmainly
at the top of Cycles I and II, composed of Ti-bear-ing chromite,
Ti-bearing chrome-magnetite, magnetite andilmenite (Fig. 2),
whereas the stratiform-type PGE mineral-ization is located in the
Marginal Unit and the lower part ofthe intrusion where it is
associated with disseminated cop-per and nickel sulfides with
interbedded thin Ti-bearingchrome-magnetite and Ti-bearing chromite
layers (Luo,1981; Zhu et al., 2010). In this study, our samples
comefrom borehole ZK411 that intersected Cycle I of the
Xinjieintrusion, which comprises a rock package about 380 mthick
and includes the main PGE mineralization occur-rences. The location
of borehole ZK411 is shown inFig. 1 and the positions of the
samples are given in Table1. Four major PGE-enriched sulfide ore
layers (PGE Layer1 to Layer 4; Fig. 2) were discovered in borehole
ZK411,although an additional layer of PGE mineralization alsooccurs
within the uppermost unit of this drill hole. Thedominant
PGE-bearing layered sequence has been dividedinto four units (Fig.
2). It should be pointed out that asthe samples were collected
primarily to study the PGE min-eralization occurrences, the
proportion of mineralized sam-ples is not representative of the
core as a whole, as moresamples were taken in mineralized sections.
In the followingdiscussion, we will focus on the main Cu–Ni–PGE
sulfidemineralization occurrences, which are hosted by
plagioclaseperidotite and plagioclase clinopyroxenite in the lower
partof Cycle I. These rocks consist of cumulus olivine in
modalamounts of up to 50%, titanaugite (15–60%), Ti-bearingchromite
and/or chrome-magnetite (5–15%), and intercu-mulus titanaugite
(5–40%) and plagioclase (10–30%).
The stratiform PGE mineralization in the Xinjie intru-sion
occurs in the form of disseminated PGE-rich sulfides.The sulfide
content within the PGE mineralization zoneranges from 0.1% to 1%,
and locally, up to 2%. The domi-nant base-metal sulfides (BMS)
comprise chalcopyrite (50–60%), pyrrhotite (20–25%), and
pentlandite (15–20%). Sper-rylite and Pd–Pt–Bi–Te minerals
(merenskyite, moncheite,and michenerite) are present in the
PGE-enriched layers.The contents of the Fe–Ti oxides correlated
with PGE min-eralization vary from 5% to 15%, and in places, up to
20%.These platinum-group minerals (PGMs) are commonlyassociated
with the BMS, or magnetite coexisting withBMS in the PGE
mineralization zone (Zhu et al., 2010).
4. ANALYTICAL METHODS
Platinum, Pd, Ir, and Ru were determined by isotopedilution
(ID)-ICP-MS using an improved Carius tube tech-
-
Fig. 2. Variations of Mg#, Cr/FeOT, Cr/TiO2, Cu/Zr, Cu/Pd, Pt,
Pd, Pt + Pd, Cu, Ni, S, and Pt/S with depth in the main PGE
mineralizationhorizon within the Xinjie intrusion. Stratigraphy of
the Xinjie intrusion is modified after Zhong et al. (2004).
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1625
nique (Qi et al., 2007). The mono-isotope element Rh wasmeasured
by external calibration using a 194Pt spike asthe internal standard
(Qi et al., 2004). Ten grams of rockpowder and appropriate amount
of enriched isotope spikesolution containing 101Ru, 105Pd, 193Ir,
194Pt were digestedwith �35 ml aqua regia in a 75 ml Carius tube,
which wasplaced in a sealed, custom-made, high pressure
autoclave
filled with water. The internal pressure of the Carius tubeis
balanced by the external pressure produced by the waterwhen heated.
Thus, this method not only avoids a possibleexplosion of the Carius
tube but also allows for relativelyhigh-temperature (300 �C)
digestion, a greater volume ofaqua regia (35 ml) and a larger
sample mass (10 g). Afterdigestion at 300 �C for 10 h, the solution
was transferred
-
Table 1Major, trace and highly siderophile element distribution
in the cumulates and sulfide ores from the Xinjie intrusion.
Sample Cyclicunit
Depth(m)
MgO(%)
FeOT(%)
TiO2(%)
Cr(ppm)
Zr(ppm)
Y(ppm)
Cu(ppm)
Ni(ppm)
S(ppm)
Ir(ppb)
Ru(ppb)
Rh(ppb)
Pt(ppb)
Pd(ppb)
HZK411-02 UpperUnit
77.0 24.71 20.42 3.67 2858 100 8.89 283 746 530 3.12 1.87 2.30
306 103
HZK411-04 78.0 26.92 17.60 3.13 2308 87.4 9.93 352 751 560 1.33
1.41 0.38 301 186HZK411-05 78.5 21.23 13.41 3.00 2045 97.9 15.5 243
521 280 0.90 0.92 0.45 263 183HZK411-06a 79.0 19.98 11.16 4.00 1360
98.6 15.9 75 421 100 7.59 6.71 9.60 445 507HZK411-42 129.73 22.23
14.09 5.17 1466 87.6 13.5 971 1192 1900 1.16 1.74 0.08 1.80
9.30HZK411-114 Unit 4 257.8 24.08 14.40 5.33 2272 109 11.7 379 1089
580 0.62 1.79 0.52 10.2 13.2HZK411-129 265.8 26.65 16.80 4.33 2842
92.6 10.7 754 1203 1600 2.64 4.24 1.28 32.8 36.0HZK411-156 278.3
19.09 15.30 5.00 1967 176 16.2 890 936 680 4.60 5.15 4.54 41.0
56.2HZK411-157 278.8 17.17 12.45 5.00 1505 153 18.0 666 753 2500
9.46 7.71 5.72 236 308HZK411-159a 279.5 16.67 10.40 3.27 1216 117
18.3 1428 769 2300 11.6 8.47 6.69 410 410HZK411-161 281.3 15.64
9.98 4.17 1627 125 19.5 1958 771 2300 12.6 10.1 7.32 185
225HZK411-162a 281.9 15.12 9.88 3.10 1345 139 17.5 2071 790 2000
32.6 22.2 57.6 470 944HZK411-163 282.4 14.31 9.59 4.07 1479 132
19.0 1574 653 3900 15.2 9.31 9.51 262 276HZK411-165 283.2 13.86
9.69 4.67 1178 138 19.3 1544 546 3800 9.00 6.04 5.50 347
311HZK411-166a 283.6 14.06 10.13 4.83 746 130 20.1 3415 777 3700
19.5 16.2 11.3 635 291HZK411-167 284.1 13.79 9.71 5.33 768 116 18.1
2937 749 4000 6.58 5.43 3.83 278 169HZK411-170 Unit 3 285.2 16.65
13.78 5.30 3069 129 15.8 921 710 760 2.89 3.70 2.04 141
132HZK411-171 285.8 19.25 13.00 4.83 2147 108 12.3 1157 846 130
3.76 4.33 2.26 123 101HZK411-178 289.7 18.15 11.60 5.33 1911 119
15.0 284 656 150 0.55 0.97 0.30 9.10 7.90HZK411-179 290.2 17.4
10.94 5.30 1589 108 13.7 383 660 280 1.20 1.08 0.98 53.2
53.5HZK411-180 290.8 15.17 8.38 4.17 1621 108 17.4 479 451 340 4.44
2.52 7.27 167 241HZK411-182 291.6 14.17 10.96 4.67 1649 219 16.0
1510 782 1900 1.83 2.89 1.21 45.6 90.0HZK411-183 292.2 14.52 8.61
3.83 1286 110 17.4 479 406 210 4.21 1.65 3.96 160 207HZK411-185
293.0 14.40 9.14 5.75 1168 125 19.6 316 355 160 3.61 1.53 3.08 142
160HZK411-188 294.8 13.41 9.09 5.00 929 115 18.7 278 347 100 5.24
1.87 4.53 108 91.1HZK411-191a 296.2 13.78 9.44 5.00 1132 118 19.0
1367 471 1300 17.0 8.63 12.9 416 525HZK411-193a 297.2 14.31 10.37
5.83 1054 121 16.9 2984 624 1300 30.7 15.8 22.9 707 1023HZK411-194a
297.7 15.29 10.78 4.83 906 95.9 14.6 2351 654 1200 36.6 18.1 36.9
680 1138HZK411-195a 298.2 12.94 9.70 4.25 822 100 15.5 3334 592
3600 19.9 9.47 23.0 547 852HZK411-196a 298.7 15.24 11.45 5.00 1039
88.8 14.3 5226 829 5800 35.9 19.1 46.1 875 1391HZK411-198 299.7
13.08 9.41 4.42 1212 128 19.1 355 357 220 5.47 2.01 5.49 169
192HZK411-199 Unit 2 300.1 15.08 12.48 5.83 1795 106 15.5 464 531
430 4.30 3.08 3.74 134 152HZK411-205 303.2 11.34 9.56 6.33 630 165
17.9 715 417 650 3.36 2.08 4.11 336 190HZK411-206 303.7 12.29 9.86
5.50 709 151 19.2 473 405 370 3.05 1.79 2.94 228 102HZK411-209
305.0 11.42 9.52 4.83 744 157 21.3 553 324 510 8.29 4.64 5.42 298
127HZK411-212a 306.5 12.60 10.20 6.50 1101 140 19.8 332 396 250
15.2 7.39 12.1 1236 566HZK411-215 308.1 12.95 9.79 7.67 1229 141
18.3 660 498 370 5.71 3.56 6.49 314 175HZK411-217a 309.0 12.52 8.48
4.17 1463 157 20.9 1494 579 1400 26.0 15.8 19.1 495 740HZK411-218
Unit 1 309.6 10.31 8.88 5.17 709 242 20.5 574 426 480 6.17 3.69
4.16 115 172HZK411-233 317.0 13.43 10.22 4.58 1119 105 16.9 6339
1395 3700 17.8 20.0 6.68 121 134HZK411-234 317.4 13.98 11.04 5.17
976 129 16.5 3774 872 4300 7.44 6.91 10.3 89.2 138
Line missing
1626H
.Z
ho
ng
etal./
Geo
chim
icaet
Co
smo
chim
icaA
cta75
(2011)1621–1641
-
HZ
K41
1-23
7a31
9.1
13.8
510
.35
5.17
1262
102
17.4
3114
730
3300
19.9
16.5
24.2
403
577
HZ
K41
1-23
9a32
0.0
14.4
812
.03
4.25
920
93.2
15.0
5476
1507
8700
72.1
63.9
55.8
1095
1758
HZ
K41
1-24
0a32
0.5
13.6
611
.57
4.83
551
113
14.8
1850
731
7200
16.2
11.1
14.3
356
591
HZ
K41
1-24
332
2.0
12.0
011
.08
4.92
967
118
17.2
615
401
6100
4.32
2.74
3.27
138
154
HZ
K41
1-24
632
3.5
12.8
611
.64
5.67
616
119
15.7
1099
571
5200
8.08
8.57
4.18
247
147
HZ
K41
1-24
732
4.1
11.1
711
.70
6.33
606
122
17.0
804
435
9700
8.79
7.76
3.73
271
146
aP
GE
-ric
hsu
lfid
eo
res.
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1627
to 50 ml centrifuge tube and then used for pre-concentrat-ing
PGE by Te-coprecipitation, as described in Qi et al.(2004). The
total procedural blanks were lower than0.003 ng/g for Ru, Rh and
Ir; 0.020 ng/g for Pd; and0.010 ng/g for Pt. The reference
standards, WPR-1,WGB-1 and TDB-1, were simultaneously used for
analyti-cal quality control. The results for WPR-1 are in
goodagreement with the certified values. The results of Ru,Rh, and
Ir for WGB-1 and TDB-1 are lower than the rec-ommended values, but
agree well with values reported byMeisel and Moser (2004). The
analytical results of the sam-ples are given in Table 1.
Re–Os isotopic compositions of the Xinjie intrusion (Ta-ble 2)
were determined at the State Key Laboratory ofLithospheric
Evolution, Institute of Geology and Geophys-ics, Chinese Academy of
Sciences (IGGCAS). The Cariustube digestion technique is similar to
those described byShirey and Walker (1995), which is reported in
detail byChu et al. (2009). Approximately 2 g of
homogenizedwhole-rock powders and appropriate amounts of
a187Re–190Os mixed spike were sealed in an externally cooled(�50
�C), single-use, Pyrex� borosilicate Carius tube, with3 ml of
purified concentrated HCl and 6 ml of purified con-centrated HNO3.
The Carius tubes were kept at �240 �C inan oven for 48–72 h. Osmium
was extracted from the aquaregia solution into CCl4 (Cohen and
Waters, 1996) and thenback-extracted into HBr, followed by
purification via mic-rodistillation (Birck et al., 1997). Re was
separated fromthe matrix and purified by anion exchange
chromatographywith about 0.6 ml resin (AG 1 � 8, 100–200 mesh).
Thesamples were loaded onto the columns in 0.8 mol/LHNO3, the
matrix elements were eluted with 0.8 mol/LHNO3 and 1 mol/L HCl, and
then the Re was collectedwith 8 mol/L HNO3. Os isotopic
compositions were mea-sured using a GV Isoprobe-T Mass Spectrometer
with neg-ative ion mode. Purified Os was loaded onto
platinumfilaments and Ba(OH)2 was used as an ion emitter. All
sam-ples were run with nine Faraday cups in static mode. TheOs
isotopic compositions and Os concentrations were ob-tained in one
mass spectrometric run. The measured Os iso-topic ratios were
corrected for mass fractionation using192Os/188Os = 3.08271 after
interference corrections, oxy-gen corrections and spike
subtractions. The isotope dilutionanalyses of Re were conducted on
a Neptune MC-ICP-MSusing a secondary electron multiplier in
peak-jumpingmode. Mass fractionations for Re were corrected using
aRe standard that was run alternately with the samples. To-tal
analytical blanks were 2 pg for Re and 3–5 pg for Oswith a
187Os/188Os ratio near 0.150. The reference valuesfor the standard
was 185Re/187Re = 0.5975. The in-run pre-cisions for Os isotopic
measurements were better than±0.2% (2rm) for all the samples.
During the period of mea-surements of our samples, the 187Os/188Os
ratio of John-son–Matthey standard of UMD was 0.11380 ± 4 (2r,n =
5). To calculate the age, Re–Os data were regressedusing the
ISOPLOT program (Ludwig, 2003) and assumingan error correlation
coefficient of 0.9. Error input wasdetermined by multiple analyses
of the in-house Os andRe standards to be 0.2% on Os isotopic
composition and1% on Re/Os ratio. The errors of two samples
(HZK411-
-
Table 2Re and Os isotopic data for the Xinjie intrusion.
Sample Re (ppb) Os (ppb) 187Re/188Os 187Os/188Os 2r
(187Os/188Os)i cOs(t)
HZK411-04 0.838 2.531 1.5983 0.14302 0.00008 0.13611
8.6HZK411-06 0.325 0.694 2.2600 0.14078 0.00031 0.13101
4.6HZK411-129 1.424 7.320 0.9376 0.13009 0.00003 0.12604
0.6HZK411-157 1.372 21.26 0.3110 0.12592 0.00003 0.12458
�0.6HZK411-162 0.689 46.94 0.0707 0.12503 0.00001 0.12472
�0.4HZK411-166 1.458 35.36 0.1986 0.12538 0.00001 0.12452
�0.6HZK411-167 1.123 13.55 0.3992 0.12622 0.00001 0.12449
�0.6HZK411-183 0.719 4.451 0.7780 0.12837 0.00010 0.12501
�0.2HZK411-191 0.892 24.04 0.1789 0.12547 0.00001 0.12470
�0.5HZK411-196 0.798 81.70 0.0471 0.12476 0.00001 0.12455
�0.6HZK411-205 0.274 3.504 0.3774 0.12629 0.00003 0.12465
�0.5HZK411-212 0.337 9.344 0.1737 0.12521 0.00001 0.12446
�0.7HZK411-217 0.658 49.93 0.0635 0.12497 0.00001 0.12470
�0.5HZK411-218 0.671 9.961 0.3246 0.12620 0.00001 0.12480
�0.4HZK411-234 1.302 41.88 0.1498 0.12534 0.00001 0.12469
�0.5HZK411-239 0.579 599.2 0.0047 0.12459 0.00001 0.12457
�0.6HZK411-243 1.304 12.08 0.5203 0.12689 0.00002 0.12464
�0.5HZK411-246 0.400 16.85 0.1142 0.12494 0.00007 0.12444 �0.7
1628 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
196 and 239) having quite high Os concentrations were as-sumed
to be 5% on Re/Os ratio.
Major elements were measured by wet chemical analysesat the
Center of Analysis and Test, Institute of Geochemis-try, Chinese
Academy of Sciences (CATIGCAS), with theanalytical precision better
than 5%. Sulfur was analyzedat the CATIGCAS by Leco induction
furnace-titration,with the accuracy better than 10%. Trace elements
weredetermined using a VG PQ Excell inductively coupled plas-ma
mass spectrometer (ICP-MS) at the University of HongKong. The
powdered samples (50 mg) were dissolved inhigh-pressure Teflon
bombs using HF + HNO3 mixturefor 48 h at �190 �C (Qi et al., 2000).
Rh was used as aninternal standard to monitor signal drift during
counting.The international standards AMH-1, GBPG-1 and OU-6were
used for analytical quality control. The analytical pre-cision is
generally better than 5% for trace elements. Thecontents of
selected major and trace elements are listed inTable 1. The
complete major and trace element dataset ispresented in Electronic
Annex.
5. RESULTS
5.1. Variations in major and trace elements
All the studied samples from the Xinjie intrusion
arecharacterized by high but variable MgO (10.3–26.9 wt.%),FeOT
(8.38–20.4 wt.%) and TiO2 (3.00–7.67 wt.%) contents(Table 1).
Variations in Mg-number (Mg#), Cr/FeOT andCr/TiO2 ratios with
stratigraphic height are shown for themain PGE mineralized horizon
of the Xinjie intrusion inFig. 2. The samples from cyclic Units 1,
3 and 4 displayconsistently increasing trends in Mg# followed by
steadilydecreasing trends upwards, whereas those from cyclic Unit2
exhibit a decreasing trend in Mg# upsection. It is interest-ing
that abrupt reversals of Mg# occur at the tops of Units1, 2 and 3
(Fig. 2a). The samples from Units 1–4 have var-iable Cr contents of
551–1262 ppm, 630–1795 ppm, 822–
3069 ppm, and 746–2842 ppm, respectively. In contrast,the
cumulate rocks and PGE ores from Upper Unit exhibithigh Cr contents
(1360–2858 ppm; Table 1). Notably, muchhigher Cr/FeOT and Cr/TiO2
ratios occur immediately be-low the PGE Layers 2, 3 and 4, while
those in cyclic Unit 1are relatively constant and low (Fig. 2b and
c).
5.2. Variations in chalcophile elements and PGE
As shown in Fig. 2, four major PGE-rich sulfide layersoccur near
or at the bottom of each cyclic unit. The Cu,Ni, PGE and sulfur
contents of the PGE-enriched layersand their host lithologies vary
by two to three orders ofmagnitude (Table 1). Sulfur concentrations
are highest inUnit 1 (3300–9700 ppm), with the exception of one
sample(480 ppm S) in the uppermost part of this unit. Apart fromone
ore sample (HZK411–217) with a S content of1400 ppm, most host
rocks and PGE-rich sulfide ores with-in Unit 2 contain
significantly less sulfur (250–650 ppm).The PGE ores in Unit 3 have
much higher sulfur concentra-tions (1200–5800 ppm) than the
cumulate rocks (100–760 ppm), with one exceptional rock sample
containing1900 ppm S. The cumulate rocks and PGE-rich sulfide
oreswithin Unit 4 are mostly rich in sulfur (1600–4000 ppm) ex-cept
two rock samples with 580–680 ppm S. In addition, thesamples from
Upper Unit contain significantly lower sulfurfrom 100 to 560 ppm,
with the exception of one sample(HZK411-42) having 1900 ppm S.
Copper concentrations exhibit a similar distribution pat-tern as
sulfur (Fig. 2), with elevated Cu concentrations inthe PGE-rich
sulfide ores and sulfide-enriched cumulaterocks. Apart from two ore
samples (HZK411-212 andHZK411-06) having significantly lower Cu
contents (332and 75 ppm), the ores from the PGE Layers 1, 3 and
4and one from the PGE Layer 2 (HZK411-217) are charac-terized by
high Cu contents between 1367 and 5476 ppm(Table 1; Fig. 2i). In
contrast, the S-poor (
-
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1629
mostly 100 ppb; Fig. 2h). Moreover, it is notable thatabundant
PGE-enriched rocks (several hundred ppbPt + Pd) and two ore samples
(HZK411-06 and HZK411-212) are relatively poor in S (100–760 ppm)
and Cu (243–1157 ppm, mostly
-
Fig. 4. Cr vs. Ir, Ru, Pt, and Pd for the Cr-rich,
sulfide-bearing and sulfide-poor samples from the Xinjie
intrusion.
1630 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
shaped PGE patterns, characterized by significant Ni deple-tion,
obvious Pd enrichment relative to Cu, and a markedfractionation
between Os, Ir and Ru on the one and Rh,Pt and Pd on the other.
These samples have Cu/Pd ratios(148–8710) similar to or lower than
that of the mantle(1000–10,000; Barnes et al., 1993). In contrast,
several sam-ples from Unit 1 (HZK411-233 and HZK411-234), Unit
3(HZK411-171, HZK411-178 and HZK411-182), Unit 4(HZK411-114,
HZK411-129, HZK411-156, HZK411-166and HZK411-167) and Upper Unit
(HZK411-42) exhibitobvious Pd depletion relative to Cu (Fig. 3),
which haveCu/Pd ratios (11,455–104,773) greater than that of
themantle. Interestingly, the sharp increases of Cu/Pd ratiosoccur
immediately above the PGE Layers 1, 3 and 4(Fig. 2e). It is also
noteworthy that most ores and cumulaterocks within Units 1–4 show
obviously negative Ru anom-alies (Fig. 3a–d), with the exception of
three samples(HZK411-114, HZK129 and HZK411-178).
Throughout the section of the intrusion examined, Ir,Ru, Pt and
Pd of the Cr-rich, sulfide-poor and sulfide-bear-ing samples
exhibit poor correlations with Cr (Fig. 4a–d).In contrast, the
variations in Cu, Pt, Pd, Ir and Ru concen-trations correlate well
with S, whereas Ni poorly correlateswith S (Table 1; Fig. 5).
5.3. Re–Os isotope
Re–Os isotope data for the Xinjie intrusion are reportedin Table
2 and plotted on the Re–Os isochron diagram inFig. 6. The Xinjie
cumulate rocks and PGE-rich sulfide oresexhibit high but variable
Os contents (0.694–81.7 ppb), withsample HZK411-239, at 599 ppb
being an exception. Incontrast, Re concentrations are low in the
range of 0.274–1.46 ppb. Most samples have near-chondritic initial
Os iso-topic compositions, with cOs(t) values (corrected to 259
Ma)ranging from �0.7 to �0.2 and 187Re/188Os ratios varying
from 0.005 to 0.778 (Table 2). Exceptions include the
twouppermost samples from the Upper Unit (HZK411-04and HZK411-06)
and HZK411-129 from Unit 4 (Table 2)which have slightly higher
initial 187Os/188Os (0.1260–0.1361), with radiogenic cOs(t) values
(+0.6 to +8.6) andhigher 187Re/188Os ratios (0.938–2.260). All
data, exceptfor the three uppermost samples and sample
HZK411-183which have slightly elevated 187Re/188Os (0.778–2.260),
de-fine an isochron age (MSWD = 0.8) of 262 ± 27 Ma(Fig. 6) that is
consistent, within the uncertainty, with theSHRIMP U–Pb zircon age
of the Xinjie intrusion(259 ± 3 Ma; Zhou et al., 2002). The
calculated initial187Os/188Os of 0.12460 ± 0.00011 (cOs(t) = �0.5 ±
0.1) isapproximately chondritic for Permian and attests to mini-mal
crustal contamination.
The high Os and low Re contents and chondritic Oscomponent of
the Xinjie intrusion are quite similar to thoseof the Lijiang
picrites (Fig. 7;
-
Fig. 5. Covariations of Cu, Ni, and PGE with sulfur content in
the Xinjie intrusion. Cu and PGE exhibit positive correlations with
S, whereasNi shows poor correlation with S.
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1631
et al., 1999), Stillwater complex (cOs(t) = 2.0–16.4; Horanet
al., 2001) and Noril’sk-Talnakh intrusion (cOs(t) = 1.9–71; Walker
et al., 1994; Arndt et al., 2003) (Fig. 7), whichhave been proposed
to reflect outer core or recycled oceaniccrust contributions
(Walker et al., 1994), or crustal assimi-lation processes
(McCandless et al., 1999; Horan et al.,2001; Arndt et al., 2003;
Tao et al., 2007, 2010).
Fig. 6. 187Re/188Os vs. 187Os/188Os isochron figure for the
Xinjieintrusion. Insert includes all the analyzed samples but four
sampleshaving slightly radiogenic Os compositions are excluded from
ageplotting. Analytical uncertainties are the size of the symbols
orsmaller.
Fig. 7. cOs(t) vs. Os concentration for magmatic sulfides
showingdata for Cu–Ni–PGE mineralization from the Xinjie (this
study),Jinbaoshan (Tao et al., 2007), Limahe (Tao et al., 2010),
NT:Noril’sk-Talnakh (Walker et al., 1994; Arndt et al., 2003),
Bushveld(McCandless et al., 1999) and Stillwater (Horan et al.,
2001)intrusions. Data for the Lijiang picrites, Song Da komatiites
andEmeishan basalts (HTB: high-Ti basalt; LTB: low-Ti basalt)
arefrom Zhang et al. (2008), Hanski et al. (2004) and Xu et al.
(2007).Samples with cOs(t) > +120 are not plotted in this figure
for clarity.
-
1632 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
6. DISCUSSION
6.1. Estimation of a Xinjie parental melt composition
The parental magma compositions of the Xinjie intru-sion can be
obtained using the compositions of the chilledmargins. Previous
study has shown that one fine-grainedolivine-bearing gabbro sample
(CSXJ26) from the MarginalUnit within the Xinjie intrusion has high
MgO (14.6%),FeOT (15.5%), and TiO2 (3.6%) contents (Zhong et
al.,2004), This sample is also characterized by the highest
Yconcentration of 17 ppm (Fig. 8a), indicating that it
containdistinctly more trapped liquid than other samples in
theMarginal Unit. Sample CSXJ26 can be then taken as a mix-ture of
the primary magma and cumulus olivine. The mostprimitive olivine
observed in the Xinjie chilled margins con-tains 84 mol percent Fo
(Mao and Sun, 1981). Fig. 8b illus-trates the compositions of the
Xinjie parental meltestimated from the method of Chai and Naldrett
(1992).The coexisting liquid calculated using the ratio of
(FeO/MgO)olivine/(FeO/MgO)liquid = 0.3 (Roeder and Emslie,1970)
contains 13.9 wt.% MgO and 15.8 wt.% FeO. The cal-culated MgO
content and 17 ppm Y of the initial liquid areapplied to model the
fractionating liquid changes usingMELTS (Ghiorso and Sack, 1995).
The model line roughlyaccounts for the compositional variations in
the high-Ti
Fig. 8. (a) Plot of MgO vs. Zr in the Xinjie chilled margin
samplesand Emeishan basalt wall rocks (Zhong et al., 2004); (b)
modelingof primary olivine and coexisting liquid composition. Point
A is thecomposition of olivine Fo84 for the Xinjie chilled margins
(Maoand Sun, 1981). Point B is the compositions of a
fine-grainedolivine-bearing gabbro from the Xinjie chilled margins
(Zhonget al., 2004). Point C is the estimated compositions of
trappedliquid in equilibrium with olivine Fo 84 for the Xinjie
intrusion.
Emeishan basalt wall rocks of the Xinjie intrusion(Fig. 8a;
Zhong et al., 2004), suggesting that the Xinjieparental magma could
share a common mantle source withthe nearby high-Ti basalts.
The Xinjie cumulate rocks are also enriched in
highlyincompatible lithophile elements (Zhong et al.,
2004).Therefore, the mantle source of the Xinjie parental
magmashould have unusually high FeO and TiO2 contents andenrichment
of highly incompatible elements. The composi-tions are comparable
to those of the coeval Lijiang picritesin the ELIP, which are
enriched in MgO (12.3–27.0%),FeOT (11.6–17.6%) and TiO2
(1.14–2.36%; Zhang et al.,2006). The Xinjie intrusion also exhibits
similar primitivemantle-normalized PGE distribution patterns (Fig.
3) tothose of the Lijiang picrites (Zhang et al., 2005),
character-ized by enrichment of Pt and Pd relative to Os, Ir and
Ru.Moreover, the near-chondritic initial Os isotope values formost
Xinjie samples (Table 2) show no effects of crustalcontamination,
similar to those of the Lijiang picrites(Zhang et al., 2008; Fig.
7).
It has been suggested that the high temperature, highmagnesium
komatiitic and picritic magmas injected intothe upper crust are the
only types of magmas that can formthe major magmatic Ni–Cu–PGE
sulfide deposits in theworld, which are generally S-undersaturated
due to high de-grees of partial melting or derivation from a S-poor
plumesource (Keays, 1995; Arndt et al., 2005). The modeling
ofNaldrett (2010) has also shown that only a magma gener-ated by
high degree of melting (P15%) would be rich inPGE and Cu, implying
that the Xinjie parental magmawas derived through high-degree
partial melting of theELIP mantle source. As demonstrated above,
the occur-rence of Cu–Ni–PGE and Fe–Ti–V mineralization withinthe
Xinjie intrusion thus requires that the parental magmasare not only
enriched in Fe and Ti but also in magnesium,incompatible elements
and PGE, which have similarities tothe characteristics of the
ferropicritic magmas (e.g., Brüg-mann et al., 2000; Hanski et al.,
2001).
6.2. PGE behavior during the evolution of the Xinjie magma
The layered series of the Xinjie intrusion are cumulaterocks and
as such the minerals they accumulate dominantlycontrol their major
and trace element compositions. In theCr-rich, sulfide-poor and
sulfide-bearing samples, Ir, Ru, Ptand Pd are negatively correlated
with Cr (Fig. 4), showingthat these elements are unlikely to be
controlled by chro-mite during crystal fractionation. Instead, the
PGE concen-trations exhibit broadly positive correlations with
S(Fig. 5), suggesting that disseminated sulfide could be themain
collector phase. This is confirmed by the observationthat the
Xinjie platinum group minerals (PGMs) are mostlyhosted in the base
metal sulfides (Zhu et al., 2010). Notably,the occurrence of
significantly higher Cr/FeOT and Cr/TiO2ratios immediately below
the PGE Layers 2, 3 and 4(Fig. 2b and c) reflects chromite or
Cr-spinel accumulationbefore sulfide deposition.
As pointed out by Li and Naldrett (1999) for the Voi-sey’s Bay
intrusion, the Cu/Zr ratio is a good measure ofchalcophile
depletion. Most of the analyzed samples (except
-
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1633
HZK411-06) have Cu/Zr ratios higher than 2.0, suggestingthat
cumulus sulfides are present almost throughout theexamined package.
Meanwhile, the increase of Cu/Zr ratiowith height in Unit 1 reflect
increasing sulfide/trapped sili-cate liquid ratios, whereas the
Cu/Zr ratios decrease up-ward in Units 3 and 4 imply decreasing
sulfide/trappedsilicate liquid ratios. In Unit 2, the Cu/Zr ratios
slightly de-crease from the PGE Layer 2, implying insignificant
varia-tions in sulfide/trapped liquid ratio. The Cu/Pd ratio
isparticularly sensitive to S-saturation because Pd has a
muchlarger partition coefficient than Cu (Barnes et al., 1993).
Inthe present study, most samples have Cu/Pd ratios close toor
lower than that in primitive mantle (Fig. 2e), furtherindicating
that they contain varying amounts of cumulussulfides. In contrast,
several samples above the PGE Layers1, 3 and 4 having Cu/Pd ratios
significantly higher than themantle value, suggesting their
crystallization from differen-tiated magma that had experienced
prior sulfide liquid seg-regation. The sharp increases in Cu/Pd
ratio above thesePGE Layers (Fig. 2e) support the derivation of at
leastsome of the PGE within the sulfide-enriched layers fromthe
overlying magma.
As shown above, the marked PGE depletion in thecumulate rocks
upsection and downsection of the mainPGE-rich layers is lacking in
the Xinjie intrusion (Pt + Pdmostly >100 ppb; Fig. 2h). The
rocks overlying the Meren-sky Reef within the Bushveld Complex are
depleted in PGErelative to Ni and Cu, which is interpreted as a
result ofPGE extraction from the overlying magma (Maier andBarnes,
1999; Barnes and Maier, 2002). The absence of con-sistent PGE
depletion in most of the Xinjie host rocks indi-cates that they
formed in an open magmatic system. It mayoccur in a dynamic conduit
system (Li et al., 2000; Evans-Lamswood et al., 2000) due to metal
upgrading of earlyformed sulfide melt by continued influx of the
later, fresh,sulfide-unsaturated and PGE-undepleted magma (Kerrand
Leitch, 2005). Another important observation thatsome PGE-rich ores
and PGE-enriched cumulate rocksfrom various units within the Xinjie
intrusion are poor inS (
-
Table 3Modeling of the Ni, Cu and PGE compositions of the
sulfides in the Xinjie intrusion.
Ni (%) Os (ppb) Ir (ppb) Ru (ppb) Rh (ppb) Pt (ppb) Pd (ppb) Cu
(%)
Parental magmaa 0.025 0.70 0.42 0.30 0.43 10.3 8.2 0.008D
sulfide liquid/silicate liquidb 300 35,000 35,000 35,000 35,000
35,000 35,000 1000R = 8000 7.23 4559 2735 1954 2800 67,078 53,402
7.112% sulfide 0.145 91 55 39 56 1342 1068 0.142R = 5000 7.08 3063
1838 1313 1882 45,072 35,882 6.672% sulfide 0.142 61 37 26 38 901
718 0.1330.5% sulfide 0.035 15.3 9.2 6.6 9.4 225 179 0.0330.2%
sulfide 0.014 6.1 3.7 2.6 3.8 90 72 0.013R = 1000 5.78 681 409 292
418 10,024 7980 4.000.5% sulfide 0.029 3.4 2.0 1.5 2.1 50 40
0.020
a Ni, Cu and PGE compositions of the parental magma are similar
to those of the pyroxene-phyric basalts in the ELIP (Zhang et al.,
2005;Zhong et al., 2006).
b Assuming Dsulfide liquid/silicate liquid of 35,000 for the PGE
(Peach et al., 1994; Fleet et al., 1999), 1000 for Cu, and 300 for
Ni (Francis, 1990).
Fig. 9. Re/Os ratio vs. Os concentration for magmatic
sulfides(recalculated 100% sulfide according to Barnes and
Lightfoot,2005) in the Xinjie intrusion. Data for the Lijiang
picrites and SongDa komatiites are from Zhang et al. (2008) and
Hanski et al.(2004). Modeling of the sulfide liquid R factor
enrichment processis conducted using Re concentration of 0.15 ppb
and Os concen-tration of 0.70 ppb, similar to those of the
pyroxene-phyric basaltsin the ELIP (Zhang et al., 2005, 2008) and
assuming that DRe is1000 and DOs is 35,000 (similar to Lambert et
al., 2000).
1634 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
fractionation dominated by olivine with minor clinopyrox-ene and
Cr-spinel. We therefore take the Ni, Cu and PGEcompositions of the
least-contaminated pyroxene-phyricbasalts in the ELIP (Zhang et
al., 2005; Zhong et al.,2006) to represent the original composition
of the Xinjie sil-icate liquid (Table 3). The Dsulfide
liquid/silicate liquid of 35,000for the PGE (Peach et al., 1994;
Fleet et al., 1999), 1000 forCu, and 300 for Ni (Francis, 1990) is
used for modeling.
The compositions of the whole-rocks in Unit 1 and Unit3 may be
modeled as assuming that they contain 0.5% sul-fides at R = 1000
and 2% sulfides at R = 8000, respectively(Fig. 3a and c). Using a
sulfide liquid formed in equilibriumwith the melt at R = 5000 and
allowing 0.2–2% cumulussulfides (Fig. 3b) it is possible to model
the whole-rockcompositions in Unit 2. The compositions of the
samplesin Unit 4 are similar to the model compositions
containing0.5–2% sulfides at R = 5000. The modeled sulfide
contentspresented here are consistent with the petrographic
obser-vation for the analyzed samples. Modeling of the sulfide
li-quid R factor enrichment process is also conducted usingRe
concentration of 0.15 ppb and Os concentration of0.70 ppb and
assuming that DRe is 1000 and DOs is 35,000(similar to Lambert et
al., 2000). The calculated results(Fig. 9) show that most Cu–Ni
sulfides (100% sulfide) inthe Xinjie intrusion correspond to R
factors between 1000and 10,000, comparable to those shown in Fig.
3. The mod-erate to high R factors (1000–8000) suggested for the
dis-seminated sulfides in the Xinjie PGE ores and silicaterocks
indicate that the sulfide droplets interacted with alarge volume of
magma in an open-system magmachamber.
As mentioned above, some samples overlying the PGELayers 1, 3
and 4 possess Cu/Pd significantly above mantlelevels and thus are
depleted in PGE relative to Cu (Fig. 3).Their compositions can be
explained by small amounts ofsulfide extraction, which are
estimated from the followingmass balance equation (Barnes et al.,
1993; Thériaultet al., 2000): S = 100(Cs/CL � 1)/(D � 1), where
Cs, CLand D are the same as the above expression and S repre-sents
the amounts of segregated sulfides in weight percent.Assuming that
the initial magma had a Cu/Pd value of9800 and Dsulfide
liquid/silicate liquid for Cu and Pd are 1000and 35,000,
respectively (Table 3), calculations show that
0.0004–0.028% sulfides may have been removed from themagma prior
to the formation of these samples.
6.4. Petrogenesis of the PGE-rich sulfides
The much higher Mg#, Cr/FeOT and Cr/TiO2 ratiosimmediately below
the PGE Layers 2, 3 and 4 define thesharp geochemical reversals
(Fig. 2a–c), providing the clearevidence for an open-system
behavior characterized bymagma replenishment episodes in the Xinjie
intrusion.Moreover, the marked Cu/Zr ratio decrease from eachPGE
ore layer within Units 2, 3 and 4 (Fig. 2d), indicatesthat sulfide
accumulated from each new magma pulse. Fur-thermore, the lack of
consistent PGE depletion with heightpresented here suggests that
the sulfides have segregatedfrom successive surges of fertile
magma. Collection of met-als by sulfides formed at moderate to high
R factors (1000–8000) occurred widely within these cyclic units,
also imply-
-
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1635
ing that the Xinjie PGE mineralization formed in an extre-mely
dynamic system. New pulses of magma apparently flo-wed through the
Xinjie magma chamber, which wouldinteract with pools of sulfide
that had accumulated earlierand transfer chalcophile elements from
the magma to thissulfide. The fact that the basalts immediately
overlyingthe Xinjie intrusion display significant PGE
depletion(Zhong et al., 2006) strengthens the hypothesis, that
thesehave been derived from the parental magma that was
inequilibrium with the Xinjie PGE-rich sulfides. The
Xinjieintrusion shows a close relationship to the
N–S-trendingfaults and intruded the Emeishan basalts (Fig. 1),
indicatingthat the locus for the sill-like intrusion may have acted
as adynamic conduit system for the magmas.
In the sulfide collection model, the mixing of the residentmagma
and a new injection of magma has been invoked toexplain
sulfide-enriched horizons in layered intrusions(Campbell et al.,
1983; Naldrett et al., 1986; Barnes andMaier, 2002). In the case of
the Xinjie samples, the abovediscussion indicates that the PGE
Layer 1 was generatedby the normal accumulation of PGE-rich
sulfides, whereasthe PGE Layers 2, 3 and 4 were formed by multiple
re-charges of magma and resultant sulfide segregation. Irvine(1977)
demonstrated that the mixing of a new liquid withthe resident magma
can initiate the precipitation of exten-sive stratiform chromite
and/or sulfide enrichments in largelayered intrusions. In the
Xinjie intrusion, several thin Ti-bearing chromite and Cr-bearing
magnetite layers occurimmediately below the PGE Layers 2, 3 and 4,
consistentwith their much higher Cr/FeOT and Cr/TiO2. We
thereforesuggest that a fresh injection of magma mixed
vigorouslywith the resident magma in the Xinjie chamber, leadingto
the crystallization of Cr-spinel and slightly decreasingFeO in the
melt. Subsequently, relatively abundant magne-tite and ilmenite
grains formed by early crystallization fromthe magma (Wang et al.,
2008), which significantly removedFeO from the melt. This process
resulted in a significant de-crease in the solubility of sulfide in
the ferropicritic melts(Haughton et al., 1974) and thus caused
separation of sul-fide which thereafter scavenged Cu, Ni and PGE
from thesilicate magma. The sulfide droplets settled and
accumu-lated at specific stratigraphic horizons to form the PGEore
layers in the Xinjie intrusion. Subsequent reaction ofthe sulfides
with new magma gave rise to the high R-factorsmentioned above. As a
result, multiple layers of the Xinjiemineralized rocks can be
explained by periodic magma re-charge and mixing. Every
replenishment event fed fromthe bottom stirred up the previously
accumulated sulfideliquids, and the entrapped sulfides equilibrated
with thenew hybrid composition of the silicate magma (Li et
al.,2009). Some of the Xinjie samples have lost S, which couldhave
happened when the cumulus pile was reheated by afresh injection of
magma. Consequently, successive batchesof magma passing through the
Xinjie chamber dissolvedFeS, leaving the residual sulfide
progressively more en-riched in Ni, Cu and PGE.
In summary, PGE Layer 1 originated from the accumu-lation of
PGE-rich sulfides, whereas PGE Layers 2–4formed from mixing between
evolved and primitive ferropi-critic magmas, all of which had
experienced multiple mag-
ma replenishment events. The locus for the Xinjieintrusion acted
as a magma conduit that processed suffi-cient amounts of silicate
liquid, thus providing a potentialmechanism for concentrating PGE
in the disseminatedsulfides.
6.5. The Xinjie and ELIP mantle source in a global context
The above demonstration links the generation of thePGE-rich
disseminated sulfides to the evolution of theparental ferropicritic
magmas, suggesting that the plume-derived ferropicritic magmas
produced not only the Cu–Ni–PGE mineralization but also the Fe–Ti–V
oxide ore-bearing layers in the Xinjie intrusion. The Re–Os age
ofthe Xinjie intrusion is in good agreement with the data re-ported
from the zircon U–Pb isotope system, implying thatthe Re–Os system
remained closed for most samples afterthe intrusion emplacement.
Three samples from UpperUnit have slightly radiogenic cOs(t) values
(+0.6 to +8.6),suggesting that they were subjected to minor
contaminationwith the overlying basalts. Previous studies have also
indi-cated that the floor and roof rocks in the Xinjie
intrusionwere slightly contaminated by the Emeishan basalts of
thecontact zone (Zhong et al., 2004; Zhang et al., 2009), as
evi-denced by the Sr–Nd–O isotopic compositions as well assome
trace elemental ratios (e.g., Rb/La and Ba/Th). TheXinjie
mafic–ultramafic rocks have initial Os isotope ratiosthat are
within error of the initial ratios for the Lijiang pi-crites (Zhang
et al., 2008) and Song Da komatiites (Hanskiet al., 2004) in the
ELIP, suggesting that the intrusion pre-serves the primary Os
isotope characteristics of plume-de-rived magmas. A rather constant
initial 187Os/188Os in theXinjie samples throughout the dominant
sequence and aslightly subchondritic cOs(t) of �0.5 ± 0.1 imply
that theEmeishan plume source evolved with a long-term
nearlychondritic Re/Os ratio.
In order to trace the 187Os/188Os isotopic evolution ofthe
mantle over geological time, a compilation of initialOs isotopic
compositions for ancient plume-derived komat-iites and picrites,
along with the data of this study, is pre-sented in Fig. 10 to
reflect those of contemporary mantlereservoirs. Data for most
Archean and Paleoproterozoickomatiites including 3.46-Ga Pilbara
komatiites (Bennettet al., 2002), 2.88-Ga Ruth Well komatiites
(Meisel et al.,2001), 2.72-Ga Alexo and Pyke Hill komatiites
(Gangopad-hyay and Walker, 2003; Puchtel et al., 2004), 2.70-Ga
Kam-balda komatiites (Foster et al., 1996), and 2.43-Ga
Vetrenykomatiites (Puchtel et al., 2001a) indicate roughly
chon-dritic Os isotopic compositions and their derivation
frommantle sources that evolved with time-integrated
near-chondritic Re/Os ratios. However, the 2.8-Ga Kostomuk-sha
komatiites (Puchtel et al., 2001b), 2.7-Ga Belingwekomatiites
(Walker and Nisbet, 2002), and 1.98-Ga Pech-enga ferropicrites
(Walker et al., 1997) exhibit high positivecOs(t), implying that
they were derived from a mantle sourcewith a long-term
suprachondritic Re/Os ratio and requiredthe ancient incorporation
of the outer core material (Walk-er et al., 1995). Additionally,
the 2.7-Ga Boston Creekkomatiites have subchondritic Os isotopic
compositionsthat were suggested to be derived from
subcontinental
-
Fig. 10. Initial cOs vs. age (in Ga) for presumed
plume-derivedkomatiites, picrites and Xinjie intrusion. The
horizontal dashed linerepresents the evolution of the chondritic
average. Abbreviations:RW: Ruth Well; Bos: Boston Creek; Kam:
Kambalda; Ale: Alexo;PH: Pyke Hill; Vet: Vetreny; Gor: Gorgona;
Cur: Curacao. Shownfor comparison is the range of compositions that
have beenreported for modern ocean island basalts (OIBs; compiled
byWalker et al., 1997). See text for references.
1636 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
lithospheric mantle (Walker and Stone, 2001). It is notablethat
radiogenic Os signatures of the 1.1-Ga Keweenawanpicrites were
inferred to originate from mantle containingrecycled oceanic crust
from late Archean subduction (Shi-rey, 1997).
In contrast, the database for Phanerozoic picrites andkomatiites
yielded an increasing complex picture of mantleevolution. For
example, the 260-Ma picrites (Zhang et al.,2008), komatiites
(Hanski et al., 2004) and Xinjie mafic–ultramafic intrusion (this
study) in the ELIP indicate aslightly subchondritic Os isotopic
composition, of whichrespective source will be discussed later. The
250-Ma Sibe-rian picrites however have suprachondritic cOs(t),
consistentwith their derivation from a mantle plume that
experiencedcore–mantle interaction (Horan et al., 1995; Walker et
al.,1995). Moreover, the 89-Ma Gorgona komatiites show alarge range
in cOs(t) (Walker et al., 1999; Brandon et al.,2003), whereas the
contemporaneous Curacao picrites havevery uniform initial cOs
values (Walker et al., 1999). The Osisotopic results indicate a
heterogeneous plume, comprisingone Os reservoir with a composition
very similar to thechondritic average and one with long-term
enriched Re/Os. The 187Os and 186Os-enriched component for the
Gorg-ona and Curacao komatiites and picrites would require
amechanism that could transfer Os from the outer core tothe lower
mantle (Walker et al., 1999; Brandon et al.,2003). In comparison,
it was proposed that the 60-Ma WestGreenland and Baffin Island
picrites, that were derivedfrom the onset of the proto-Iceland
plume with chondriticto slightly suprachondritic Os isotopic
compositions(Schaefer et al., 2000; Dale et al., 2009), reflect the
mixingof depleted MORB mantle, recycled oceanic crust and
high3He/4He primitive mantle. The large variations in187Os/188Os
ratios for the picrites from the Hawaiian volca-nic centers require
long-term differences in the Re/Os ratiosof the source regions.
Lassiter and Hauri (1998) attributedthe 187Os/188Os variations to
the addition of recycled oce-
anic mafic crust and/or sediments in the Hawaiian plumesource,
whereas Brandon et al. (1999) proposed thatcore–mantle interaction
could account for some of the187Os enrichment observed in some
volcanic centers. The30-Ma Ethiopian picrites have unradiogenic but
broadlychondritic Os isotopic compositions, implying that theywere
generated during the initial turbulent ascent of theAfar plume head
form pyroxene rich veins in a peridotitematrix (Rogers et al.,
2010). As shown above, the featureof the dataset indicates that Os
isotopic heterogeneity inplume-related materials is present as
early as 2.7 Ga. Themechanism for long-term Os isotopic
heterogeneity of var-ious plume sources will not be elucidated
until better under-standings of the Earth differentiation, mantle
dynamics, Re(Pt) and Os behavior during melting are achieved and
thereis a better knowledge of high-pressure,
high-temperaturepartitioning between mantle silicates and
metal.
Previous studies have shown that the Lijiang picrites,Song Da
komatiites and uncontaminated Xinjie rocks hadpositive initial eNd
values (+1.0 to +7.5; Hanski et al.,2004; Zhang et al., 2008,
2009), implying that their mantlesources recorded a long-term
depletion in LREE after previ-ous episodes of melt extraction. This
is true for the LREE-depleted Song Da komatiites as well as the
LREE-enrichedLijiang picrites and Xinjie samples, although the
lattershould also have experienced a subsequent LREE enrich-ment
event. The suggestion reconciles with Re depletion in-ferred from
the slightly subchondritic Os isotopiccompositions of the Lijiang
picrites, Xinjie rocks and someSong Da komatiites. Hanski et al.
(2004) proposed thatthe Song Da komatiitic rocks had been subjected
to 1–2%contamination with Proterozoic crust, which would have
re-duced the extent of the Re depletion. The mantle Re–Os
sys-tematics are much less sensitive to extraction of
low-degreemelts than the Sm–Nd systematics due to the less
incompat-ible behavior of Re compared to Nd (Puchtel et al.,
2004),consistent with the effect of prior melting events on
theSm/Nd and Re/Os ratios in the ELIP source. The dominant,slightly
subchondritic Os isotopic component for the ELIPis similar to that
of the depleted mantle source of midoceanridge basalts (DMM; Walker
et al., 2002). However, Herz-berg and O’Hara (1998) considered that
it was difficult toenvision a mechanism for heating the DMM to
sufficienttemperatures at the 3–4 GPa pressure necessary to
generatethe komatiites and picrites, particularly on the scale of
thevolumes estimated for the LIPs. The Os isotopic composi-tions of
the Emeishan plume source could also be recycledoceanic lithosphere
that experienced a similar depletion his-tory as the DMM. It has
been speculated that the depletedoceanic lithosphere that descended
to the core–mantle inter-face could then subsequently have become
entrained in anascending plume (Kerr et al., 1995).
Combining the above arguments, a simple model to con-sider for
generating the ELIP is the derivation of a meltfrom a plume that
consisted of recycled deep portion ofoceanic peridotitic
lithosphere. Previous studies have sug-gested subduction of oceanic
lithosphere underneath theYangtze Block during Neoproterzoic time
(e.g., Zhouet al., 2006). It might be expected that the komatiites
andpicrites with near-chondritic Os isotopic compositions were
-
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1637
derived from the recycled deep portion of oceanic litho-sphere.
Moreover, the melts that produced the Lijiang pi-crites and Xinjie
rocks have been termed ferropicritic,which are also enriched in
LREEs, TiO2, Zr and manyother incompatible trace elements (Zhong et
al., 2004;Zhang et al., 2006). These characteristics require the
exis-tence of geochemically enriched domains as dikes or veinsin
the depleted peridotite mantle (e.g., Niu and O’Hara,2003; Regelous
et al., 2003). Thus, relatively small degreemelts from the
heterogeneous mantle may contain a largerproportion of the more
readily fusible component, whichcarries the enriched chemical and
isotopic signatures forthe ferropicritic magmas. In contrast, the
komatiites inthe ELIP could have been generated from a greater
degreeof melting of incompatible trace element depleted,
refrac-tory mantle components. However, even samples with
con-siderable degree of ‘re-fertilization’ by metasomatic meltsdo
not show correspondingly enriched Ti concentrations(Hellebrand et
al., 2002). The TiO2 contents of most OIBare too high to be
generated from any plausible peridotiticsources, which can be
explained by the addition of smallamounts (generally less than 10%)
of recycled mafic crust(Prytulak and Elliott, 2007). We therefore
propose thatthe ferropicritic magmas in the ELIP were generated
froma plume, possibly containing a substantial portion of recy-cled
lithosphere. The plume-derived magmas thereafterinteracted with a
relatively small amount of subducted/al-tered oceanic crust
(probably
-
1638 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
APPENDIX A. SUPPLEMENTARY DATA
Supplementary data associated with this article can befound, in
the online version, at doi:10.1016/j.gca.2011.01.009.
REFERENCES
Ali J. R., Thompson G. M., Song X. Y. and Wang Y. (2002)Emeishan
basalts (SW China) and the ‘end-Guadalupian’
crisis:magnetobiostratigraphic constraints. J. Geol. Soc. 159,
21–29.
Arndt N. T., Czamanske G., Walker R. J., Chauvel C. andFedorenko
V. (2003) Geochemistry and origin of the intrusivehosts of the
Noril’sk-Talnakh Cu–Ni–PGE sulfide deposits.Econ. Geol. 98,
495–515.
Arndt N. T., Lesher C. M. and Czamanske G. K. (2005)
Mantle-derived magmas and magmatic Ni–Cu–(PGE) deposits. InEconomic
Geology 100th Anniversary Volume (eds. J. W.Hedenquist, J. F. H.
Thompson, R. J. Goldfarb and J. P.Richards). pp. 5–23.
Barnes S.-J. and Lightfoot P. C. (2005) Formation of
magmaticnickel-sulfide ore deposits and processes affecting their
copperand platinum-group element contents. In Economic Geology100th
Anniversary Volume (eds. J. W. Hedenquist, J. F. H.Thompson, R. J.
Goldfarb and J. P. Richards). pp. 179–213.7
Barnes S.-J. and Maier W. D. (1999) The fractionation of Ni,
Cuand the noble metals in silicate and sulphide liquids.
Geol.Assoc. Can. Short Course Notes 13, 69–106.
Barnes S.-J. and Maier W. D. (2002) Platinum-group elements
andmicrostructures of normal Merensky Reef from Impala plati-num
mines, Bushveld Complex. J. Petrol. 43, 103–128.
Barnes S.-J., Couture J.-F., Sawyer E. W. and Bouchaib C.
(1993)Nickel–copper occurrences in the Belleterre–Angliers Belt of
thePontiac Subprovince and the use of Cu–Pd ratios in
interpretingplatinum-group element distributions. Econ. Geol. 88,
1402–1418.
Barnes S.-J., Zientek M. L. and Severson M. J. (1997) Ni, Cu,
Auand platinum-group element contents of sulphides associatedwith
intraplate magmatism: a synthesis. Can. J. Earth Sci.
34,337–351.
Bennett V. C., Nutman A. P. and Esat T. M. (2002) Constraints
onmantle evolution from 187Os/188Os isotopic compositions ofArchean
ultramfic rocks from southern West Greenland(3.8 Ga) and Western
Australia (3.46 Ga). Geochim. Cosmo-chim. Acta 66, 2615–2630.
Birck J. L., Roy-Barman M. and Capmas F. (1997) Re–Os
isotopicmeasurements at the femtomole level in natural
samples.Geostand. Newslett. 20, 19–27.
Brandon A. D., Norman M. D., Walker R. J. and Morgan J. W.(1999)
186Os–187Os systematics of Hawaiian picrites. EarthPlanet. Sci.
Lett. 174, 25–42.
Brandon A. D., Walker R. J., Puchtel I. S., Becker H., HumayunM.
and Revillon S. (2003) 186Os–187Os systematics of GorgonaIsland
komatiites: implications for early growth of the innercore. Earth
Planet. Sci. Lett. 206, 411–426.
Brügmann G. E., Hanski E. J., Naldrett A. J. and Smolkin V.
F.(2000) Sulphide segregation in ferropicrites from the
PechengaComplex, Kola Peninsula. Russia J. Petrol. 41,
1721–1742.
Campbell I. H. and Naldrett A. J. (1979) The influence
ofsilicate:sulfide ratios on the geochemistry of magmatic
sulfides.Econ. Geol. 74, 1503–1506.
Campbell I. H., Naldrett A. J. and Barnes S. J. (1983) A model
forthe origin of the platinum-rich sulfide horizons in the
Bushveldand Stillwater complexes. J. Petrol. 24, 133–165.
Capobianco C. J. and Drake M. J. (1990) Partitioning
ofruthenium, rhodium, and palladium between spinel and silicatemelt
and implications for platinum group element fractionationtrends.
Geochim. Cosmochim. Acta 54, 869–874.
Chai G. and Naldrett A. J. (1992) The Jinchuan
ultramaficintrusion: cumulate of a high-Mg basaltic magma. J.
Petrol. 33,277–303.
Chazey W. J. and Neal C. R. (2005) Platinum-group
elementconstraints on source composition and magma evolution of
theKerguelen Plateau using basalts from ODP Leg 183.
Geochim.Cosmochim. Acta 69, 4685–4701.
Chu Z. Y., Wu F. Y., Walker R. J., Rudnick R. L., Pitcher
L.,Puchtel I. S., Yang Y. H. and Wilde S. A. (2009)
Temporalevolution of lithospheric mantle beneath the eastern
NorthChina Craton. J. Petrol. 50, 1857–1898.
Chung S. L. and Jahn B. M. (1995) Plume–lithosphere
interactionin generation of the Emeishan flood basalts at the
Permian–Triassic boundary. Geology 23, 889–892.
Cohen A. S. and Waters F. G. (1996) Separation of osmium
fromgeological materials by solvent extraction for analysis by
TIMS.Anal. Chim. Acta 332, 269–275.
Dale C. W., Gannoun A., Burton K. W., Argles T. W. andParkinson
I. J. (2007) Rhenium–osmium isotope and elementalbehaviour during
subduction of oceanic crust and the implica-tions for mantle
recycling. Earth Planet. Sci. Lett. 253, 211–225.
Dale C. W., Pearson D. G., Starkey N. A., Stuart F. M., Ellam
R.M., Larsen L. M., Fitton J. G. and Macpherson C. G. (2009)Osmium
isotopes in Baffin Island and West Greenland picrites:implications
for the 187Os/188Os composition of the convectingmantle and the
nature of high 3He/4He mantle. Earth Planet.Sci. Lett. 278,
267–277.
Day J. M. D., Pearson D. G. and Hulbert L. J. (2008)
Rhenium–osmium isotope and platinum-group element constraints on
theorigin and evolution of the 1.27 Ga Muskox layered intrusion.J.
Petrol. 49, 1255–1295.
Eales H. V. and Cawthorn R. G. (1996) The Bushveld Complex.
InLayered Intrusions (ed. R. G. Cawthorn). Elsevier, Amsterdam,pp.
181–230.
Ernst R. E., Buchan K. L. and Campbell I. H. (2005) Frontiers
inlarge igneous province research. Lithos 79, 271–297.
Evans-Lamswood D. M., Butt D. P., Jackson R. S., Lee D.
V.,Muggridge M. G. and Wheeler R. I. (2000) Physical
controlsassociated with the distribution of sulfides in the
Voisey’s BayNi–Cu–Co deposits, Labrador. Econ. Geol. 95,
749–769.
Fleet M. E., Crocket J. H., Liu M. and Stone W. E.
(1999)Laboratory partitioning of platinum-group elements (PGE)
andgold with application to magmatic-PGE deposits. Lithos
47,127–142.
Foster J. G., Lambert D. D., Frick L. R. and Maas R. (1996)
Re–Os isotopic evidence for genesis of Archean nickel ores
fromuncontaminated komatiites. Nature 382, 703–706.
Francis R. D. (1990) Sulfide globules in mid-ocean ridge
basalts(MORB), and the effect of oxygen abundance in Fe–S–O
liquidson the ability of those liquids to partition metals from
MORBand komatiite magmas. Chem. Geol. 85, 199–213.
Gangopadhyay A. and Walker R. J. (2003) Re–Os systematics ofthe
ca. 2.7-Ga komatiites from Alexo, Ontario, Canada. Chem.Geol. 196,
147–162.
Ghiorso M. S. and Sack R. O. (1995) Chemical mass transfer
inmagmatic processes. IV. A revised and internally
consistentthermodynamic model for the interpolation and
extrapolationof liquid–solid equilibria in magmatic systems at
elevatedtemperatures and pressures. Contrib. Mineral. Petrol. 119,
197–212.
Guo F., Fan W. M., Wang Y. J. and Li C. W. (2004) When did
theEmeishan mantle plume activity start? Geochronological and
http://dx.doi.org/10.1016/j.gca.2011.01.009http://dx.doi.org/10.1016/j.gca.2011.01.009
-
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1639
geochemical evidence from ultrmafic–mafic dikes in southwest-ern
China. Int. Geol. Rev. 46, 226–234.
Hannah J. L. and Stein H. J. (2002) Re–Os model for the origin
ofsulfide deposits in anorthosite-associated intrusive
complexes.Econ. Geol. 97, 371–383.
Hanski E., Huhma H., Rastas P. and Kamenetsky V. S. (2001)
ThePalaeoproterozoic komatiite-picrite association of
FinnishLapland. J. Petrol. 42, 855–876.
Hanski E., Walker R. J., Huhma H., Polyakov G. V., Balykin P.A.,
Hoa T. T. and Phuong N. T. (2004) Origin of the Permian–Triassic
komatiites, northwestern Vietnam. Contrib. Mineral.Petrol. 147,
453–469.
Haughton D. R., Roedder P. L. and Skinner B. J. (1974)
Solubilityof sulfur in mafic magmas. Econ. Geol. 69, 451–467.
He B., Xu Y. G., Chung S. L., Xiao L. and Wang Y.
(2003)Sedimentary evidence for a rapid crustal doming prior to
theeruption of the Emeishan flood basalts. Earth Planet. Sci.
Lett.213, 389–405.
He B., Xu Y. G., Huang X. L., Luo Z. Y., Shi Y. R., Yang Q.
J.and Yu S. Y. (2007) Age and duration of the Emeishan
floodvolcanism, SW China: geochemistry and SHRIMP zircon U–Pbdating
of silicic ignimbrites, post-volcanic Xuanwei Formationand clay
tuff at the Chaotian section. Earth Planet. Sci. Lett.255,
306–323.
Hellebrand E., Snow J. E., Hoppe P. and Hofmann A. W.
(2002)Garnet-field melting and late-stage refertilization in
‘residual’abyssal peridotites from the central Indian ridge. J.
Petrol. 43,2305–2338.
Herzberg C. and O’Hara M. J. (1998) Phase equilibrium
con-straints on the origin of basalts, picrites, and komatiites.
EarthSci. Rev. 44, 39–79.
Horan M. F., Walker R. J., Fedorenko V. A. and Czamanske G.K.
(1995) Os and Nd isotopic constraints on the temporal andspatial
evolution of Siberian flood basalt sources. Geochim.Cosmochim. Acta
59, 5159–5168.
Horan M. F., Morgan J. W., Walker R. J. and Cooper R. W.(2001)
Re–Os isotopic constraints on magma mixing in theperidotite zone of
the Stillwater complex, Montana, USA.Contrib. Mineral. Petrol. 141,
446–457.
Huang K. N. and Opdyke N. D. (1998)
Magnetostratigraphicinvestigations on an Emeishan basalt section in
westernGuizhou province, China. Earth Planet. Sci. Lett.
163,1–14.
Irvine T. N. (1977) Origin of chromitite layers in the
Muskoxintrusion and other stratiform intrusions: a new
interpretation.Geology 5, 273–277.
Keays R. R. (1995) The role of komatiitic and picritic
magmatismand S-saturation in the formation of the ore deposits.
Lithos 34,1–18.
Kerr A. and Leitch A. M. (2005) Self-destructive sulfide
segrega-tion systems and the formation of high-grade magmatic
oredeposits. Econ. Geol. 100, 311–332.
Kerr A. C., Saunders A. D., Tarney J., Berry N. H. and Hards
V.L. (1995) Depleted mantle-plume geochemical signatures: noparadox
for plume theories. Geology 23, 843–846.
Lambert D. D., Frick L. R., Foster J. G., Li C. S. and Naldrett
A.J. (2000) Re–Os isotope systematics of the Voisey’s Bay Ni–Cu–Co
magmatic sulfide system, Labrador, Canada: II. Implica-tions for
parental magma chemistry, ore genesis, and metalredistribution.
Econ. Geol. 95, 867–888.
Lassiter J. C. and Hauri E. H. (1998) Osmium-isotope variations
inHawaiian lavas: evidence for recycled oceanic lithosphere in
theHawaiian plume. Earth Planet. Sci. Lett. 164, 483–496.
Li C. and Naldrett A. J. (1999) Geology and petrology of
theVoisey’s Bay intrusion: reaction of olivine with sulfide
andsilicate liquids. Lithos 47, 1–31.
Li C., Lightfoot P. C., Amelin Y. and Naldrett A. J.
(2000)Contrasting petrological and geochemical relationships in
theVoisey’s Bay and Mushuau intrusions, Labrador,
Canada:implications for ore genesis. Econ. Geol. 95, 771–799.
Li Z. X., Li X. H., Kinny P. D., Wang J., Zhang S. and Zhou
H.(2003) Geochronology of Neoproterozoic syn-rift magmatismin the
Yangtze Craton, South China and correlations with othercontinents:
evidence for a mantle superplume that broke upRodinia. Precamb.
Res. 122, 85–109.
Li X. H., Li Z. X., Sinclair J. A., Li W. X. and Carter G.
(2006)Revisiting the “Yanbian Terrane”: implications for
Neoprote-rozoic tectonic evolution of the western Yangtze Block,
SouthChina. Precamb. Res. 151, 14–30.
Li C., Ripley E. M. and Naldrett A. J. (2009) A new genetic
modelfor the giant Ni–Cu–PGE sulfide deposits associated with
theSiberian flood basalts. Econ. Geol. 104, 291–301.
Ludwig K. R. (2003) Isoplot 3.00 User’s Manual: A
Geochrono-logical Toolkit for Microsoft Excel, Berkley
GeochronologicalCenter, Special Publication No. 4, Rev. May 30,
70p.
Luo Y. N. (1981) The characteristics of Ti-chromite
mineralizationin Xinjie layered ultramafic–mafic intrusion in
Panzhihua area,China. Geochimica 10(1), 66–74 (in Chinese).
Maier W. D. and Barnes S.-J. (1999) Platinum-group elements
insilicate rocks of the Lower, Critical and Main Zones at
Unionsection, western Bushveld Complex. J. Petrol. 40,
1647–1671.
Mao Y. S. and Sun S. H. (1981) Petrological characteristics
andorigin of layered basic–ultrabasic intrusion in Xinjie,
Miyi,Sichuan Province. Mineral Rocks (6), 29–40 (in Chinese
withEnglish abstract).
McCandless T. E., Ruiz J., Adair B. I. and Freydier C. (1999)
Re–Os isotope and Pd/Ru variations in chromitites from theCritical
Zone, Bushveld Complex, South Africa. Geochim.Cosmochim. Acta 63,
911–923.
Meisel T. and Moser J. (2004) Reference materials for
geochemicalPGE analysis: new analytical data for Ru, Rh, Pd, Os,
Ir, Ptand Re by isotope dilution ICP-MS in 11 geological
referencematerials. Chem. Geol. 208, 319–338.
Meisel T., Moser J. and Wegscheider W. (2001)
Recognizingheterogeneous distribution of platinum group elements
(PGE)in geological materials by means of the Re–Os isotope
system.Fresen. J. Anal. Chem. 370, 566–572.
Morgan J. W., Stein H. J., Hannah J. L., Markey R. J.
andWiszniewska J. (2000) Re–Os study of Fe–Ti–V oxide and Fe–Cu–Ni
sulfide deposits, Suwalki anorthosite massif, northeastPoland.
Miner. Deposita 35, 391–401.
Naldrett A. J. (2010) Secular variation of magmatic sulfide
depositsand their source magmas. Econ. Geol. 105, 669–688.
Naldrett A. J. and Lehmann J. (1988) Spinel non-stoichiometry
asthe explanation for Ni-, Cu-, and PGE-enriched sulphides
inchromitites. In Geo-platinum 87 (eds. H. M. Prichard, P. J.Potts,
J. F. W. Bowles and S. J. Cribb). Elsevier, pp, pp. 93–110.
Naldrett A. J., Gasparini E. C., Barnes S. J., von Gruenewaldt
G.and Sharpe M. R. (1986) The upper critical zone of theBushveld
Complex and the origin of Merensky-type ores. Econ.Geol. 81,
1105–1117.
Naldrett A. J., Wilson A. H., Kinnaird J. and Chunnett G.
(2009)PGE tenor and metal ratios within and below the MerenskyReef,
Bushveld Complex: implications for its genesis. J. Petrol.50,
625–659.
Niu Y. and O’Hara M. J. (2003) Origin of ocean island basalts:
anew perspective from petrology, geochemistry, and mineralphysics
considerations. J. Geophys. Res. 108, 2209.
doi:10.1029/2002JB002048.
O’Driscoll B., Day J. M. D., Daly J. S., Walker R. J.
andMcDonough W. F. (2009) Rhenium–osmium isotopes and
http://dx.doi.org/10.1029/2002JB002048http://dx.doi.org/10.1029/2002JB002048
-
1640 H. Zhong et al. / Geochimica et Cosmochimica Acta 75 (2011)
1621–1641
platinum-group elements in the Rum Layered Suite,
Scotland:implications for Cr-spinel seam formation and the
compositionof the Iceland mantle anomaly. Earth Planet. Sci. Lett.
286, 41–51.
Pang K. N., Zhou M. F., Lindsley D., Zhao D. G. and Malpas
J.(2008) Origin of Fe–Ti oxide ores in mafic intrusion:
evidencefrom the Panzhihua intrusion, SW China. J. Petrol. 49,
295–313.
Peach C. L., Mathez E. A., Keays R. R. and Reeves S. J.
(1994)Experimentally determined sulfide–silicate melt partition
coef-ficients for iridium and palladium. Chem. Geol. 117,
361–377.
Philipp H., Eckhardt J. D. and Puchelt H. (2001)
Platinum-groupelements (PGE) in basalts of the seaward-dipping
reflectorsequence, SE Greenland coast. J. Petrol. 42, 407–432.
Prytulak J. and Elliott T. (2007) TiO2 enrichment in ocean
islandbasalts. Earth Planet. Sci. Lett. 263, 388–403.
Puchtel I. S. and Humayun M. (2001) Platinum group
elementfractionation in a komatiitic basalt lava lake.
Geochim.Cosmochim. Acta 65, 2979–2993.
Puchtel I. S., Brügmann G. E. and Hofmann A. W. (2001a)
187Os-enriched domain in an Archean mantle plume: evidence from2.8
Ga komatiites of the Kostomuksha greenstone belt, NWBaltic Shield.
Earth Planet. Sci. Lett. 186, 513–526.
Puchtel I. S., Brügmann G. E., Hofmann A. W., Kulikov V. S.
andKulikova V. V. (2001b) Os isotopic systematics of
komatiiticbasalts from the Vetreny belt, Baltic Shield: evidence
for achondritic source of the 2.45 Ga plume. Contrib.
Mineral.Petrol. 140, 588–599.
Puchtel I. S., Brandon A. D. and Humayun M. (2004) Precise
Pt–Re–Os isotope systematics of the mantle from 2.7-Ga komat-iites.
Earth Planet. Sci. Lett. 224, 157–174.
Qi L., Hu J. and Grégoire D. C. (2000) Determination of
traceelements in granites by inductively coupled plasma
massspectrometry. Talanta 51, 507–513.
Qi L., Zhou M. F. and Wang C. Y. (2004) Determination of
lowconcentrations of platinum group elements in geologicalsamples
by ID-ICP-MS. J. Anal. At. Spetrom. 19, 1335–1339.
Qi L., Zhou M. F., Wang C. Y. and Sun M. (2007) Evaluation ofthe
determination of Re and PGEs abundances of geologicalsamples by
ICP-MS coupled with a modified Carius tubedigestion at different
temperatures. Geochem. J. 41, 407–414.
Qi L., Wang C. Y. and Zhou M. F. (2008) Controls on the
PGEdistribution of Permian Emeishan alkaline and
peralkalinevolcanic rocks in Longzhoushan, Sichuan Province, SW
China.Lithos 106, 222–236.
Regelous M., Hofmann A. W., Abouchami W. and Galer S. J.
G.(2003) Geochemistry of lavas from the Emperor Seamounts,and the
geochemical evolution of Hawaiian magmatism from85 to 42 Ma. J.
Petrol. 44, 113–140.
Righter K. and Downs R. T. (2001) The crystal structures
ofsynthetic Re- and PGE-bearing magnesioferrite spinels:
impli-cations for impacts, accretion and the mantle. Geophys.
Res.Lett. 28, 619–622.
Roeder P. and Emslie R. F. (1970) Olivine-liquid
equilibrium.Contrib. Mineral. Petrol. 29, 275–282.
Rogers N. W., Davies M. K., Parkinson I. J. and Yirgu G.
(2010)Osmium isotopes and Fe/Mn ratios in Ti-rich picritic
basaltsfrom the Ethiopian flood basalt province. No evidence
fromcore contribution to the Afar plume. Earth Planet. Sci.
Lett.296, 413–422.
SBGMR (Sichuan Bureau of Geology and Mineral Resources)(1991)
Regional Geology of Sichuan province. Geological Pub-lishing House,
Beijing. 680pp (in Chinese).
Schaefer B. F., Parkinson I. J. and Hawkesworth C. J. (2000)
Deepmantle plume osmium isotope signature from West
GreenlandTertiary picrites. Earth Planet. Sci. Lett. 175,
105–118.
Shirey S. B. (1997) Re–Os compositions of mid-continent
riftsystem picrites: implications for plume–lithosphere
interactionand enriched mantle sources. Can. J. Earth Sci. 34,
489–503.
Shirey S. B. and Walker R. J. (1995) Carius tube digestions for
low-blank rhenium–osmium analysis. Anal. Chem. 67, 2136–2141.
Song X. Y., Zhou M. F., Cao Z. M., Sun M. and Wang Y. L.(2003)
Ni–Cu–(PGE) magmatic sulfide deposits in the Yan-gliuping area,
Permian Emeishan igneous province, SW China.Miner. Deposita 38,
831–843.
Tao Y., Li C., Hu R. Z., Ripley E. M., Du A. D. and Zhong
H.(2007) Petrogenesis of the Pt–Pd mineralized Jinbaoshanultramafic
intrusion in the Permian Emeishan large igneousprovince, SW China.
Contrib. Mineral. Petrol. 153, 321–337.
Tao Y., Li C., Song X. Y. and Ripley E. M. (2008)
Mineralogical,petrological and geochemical studies of the Limahe
mafic–ultramafic intrusion and associated Ni–Cu sulfide ores,
SWChina. Miner. Deposita 43, 849–872.
Tao Y., Li C., Hu R. Z., Qi L., Qu W. J. and Du A. D. (2010)
Re–Os isotopic constraints on the genesis of the Limahe
Ni–Cudeposit in the Emeishan large igneous province, SW
China.Lithos 119, 137–146.
Thériault R. D., Barnes S.-J. and Severson M. J. (2000) Origin
of Cu–Ni–PGE sulfide mineralization in the Partridge River
intrusion,Duluth Complex, Minnesota. Econ. Geol. 95, 929–943.
Vogel D. C., Keays R. R., James R. S. and Reeves S. J. (1999)
Thegeochemistry and petrogenesis of the Agnew intrusion, Canada:a
product of S-undersaturated, high-Al and low-Ti tholeiiticmagmas.
J. Petrol. 40, 423–450.
Walker R. J. and Nisbet E. (2002) 187Os isotopic constraints
onArchean mantle dynamics. Geochim. Cosmochim. Acta
66,3317–3325.
Walker R. J. and Stone W. R. (2001) Os isotope constraints on
theorigin of the 2.7 Ga Boston Creek flow, Ontario, Canada.Chem.
Geol. 175, 567–579.
Walker R. J., Morgan J. W., Horan M. F., Czamanske G.
K.,Krogstad E. J., Fedorenko V. A. and Kunilov V. E. (1994) Re–Os
isotopic evidence for an enriched-mantle source for
theNoril’sk-type, ore-bearing intrusions, Siberia. Geochim.
Cos-mochim. Acta 58, 4179–4197.
Walker R. J., Morgan J. W. and Horan M. F. (1995)
187Osenrichment in some mantle plume sources: evidence for
core–mantle interaction? Science 269, 819–822.
Walker R. J., Morgan J. W., Hanski E. J. and Smolkin V.
(1997)Re–Os systematics of Early Proterozoic ferropicrites,
PechengaComplex, NW Russia: evidence for ancient
187Os-enrichedplumes. Geochim. Cosmochim. Acta 61, 3145–3160.
Walker R. J., Storey M., Kerr A., Tarney J. and Arndt N. T.
(1999)Implications of 187Os heterogeneities in mantle plumes:
evi-dence from Gorgona Island and Curacao. Geochim. Cosmo-chim.
Acta 63, 713–728.
Walker R. J., Prichard H. M., Ishiwatari A. and Pimentel
M.(2002) The osmium isotopic compositions of the convectingupper
mantle deduced from ophiolite chromitites. Geochim.Cosmochim. Acta
66, 329–345.
Wang C. Y., Zhou M. F. and Zhao D. G. (2005) Mineral chemistryof
chromite from the Permian Jinbaoshan Pt–Pd–sulphide-bearing
ultramafic intrusion in SW China with petrogeneticimplications.
Lithos 83, 47–66.
Wang C. Y., Zhou M. F. and Keays R. R. (2006)
Geochemicalconstraints on the origin of the Permian Baimazhai
mafic–ultramafic intrusion, SW China. Contrib. Mineral. Petrol.
152,309–321.
Wang C. Y., Zhou M. F. and Zhao D. G. (2008) Fe–Ti–Cr oxidesfrom
the Permian Xinjie mafic–ultramafic layered intrusion inthe
Emeishan large igneous province, SW China: crystallizationfrom Fe-
and Ti-rich basalt magmas. Lithos 102, 198–217.
-
Re–Os isotope and PGE in Xinjie layered intrusion, SW China
1641
Xiao L., Xu Y. G., Mei H. J., Zheng Y. F., He B. and Pirajno
F.(2004) Distinct mantle sources of low-Ti and high-Ti basaltsfrom
the western Emeishan large igneous province, SW China:implications
for plume–lithosphere interaction. Earth Planet.Sci. Lett. 228,
525–546.
Xu Y. G., Chung S. L., Jahn B. M. and Wu G. Y. (2001)
Petrologicand geochemical constraints on the petrogenesis of
Permian–Triassic Emeishan flood basalts in southwestern China.
Lithos58, 145–168.
Xu Y. G., He B., Chung S. L., Menzies M. A. and Frey F. A.(2004)
Geologic, geochemical, and geophysical consequences ofplume
involvement in the Emeishan flood-basalt province.Geology 32,
917–920.
Xu J. F., Suzuki K., Xu Y. G., Mei H. J. and Li J. (2007) Os,
Pb,and Nd isotope geochemistry of the Permian Emeishancontinental
flood basalts: insights into the source of a largeigneous province.
Geochim. Cosmochim. Acta 71, 2104–2119.
Yao P. H., Wang K. N., Du C. L., Lin Z. T. and Song X.
(1993)Records of China’s Iron Ore Deposits. Metallurgic
IndustryPress, Beijing. pp. 633–649 (in Chinese with English
abstract).
Zhang Z. C., Mao J. W., Mahoney J. J., Wang F. S. and Qu W.
J.(2005) Platinum group elements in the Emeishan large province,SW
China: implications for mantle source. Geochem. J. 39, 371–382.
Zhang Z. C., Mahoney J. J., Mao J. W. and Wang F. S.
(2006)Geochemistry of picritic and associated flows of the
westernEmeishan flood basalt province, China. J. Petrol. 47,
1997–2019.
Zhang Z. C., Zhi X. C., Chen L., Saunders A. D. and Reichow M.K.
(2008) Re–Os isotopic compositions of picrites from theEmeishan
flood basalt province, China. Earth Planet. Sci. Lett.276,
30–39.
Zhang Z. C., Mao J. W., Saunders A. D., Ai Y., Li Y. and Zhao
L.(2009) Petrogenetic modeling of three mafic–ultramafic
layeredintrusion in the Emeishan large igneous province, SW
China,based on isotopic and bulk chemical constraints. Lithos
113,369–392.
Zhong H. and Zhu W. G. (2006) Geochronology of layered
maficintrusions from the Pan-Xi area in the Emeishan large
igneousprovince, SW China. Miner. Deposita 41, 599–606.
Zhong H., Zhou X. H., Zhou M. F., Sun M. and Liu B. G.
(2002)Platinum-group element geochemistry of the Hongge
Fe–V–Tideposit in the Pan-Xi area, southwestern China.
Miner.Deposita 37, 226–239.
Zhong H., Yao Y., Hu S. F., Zhou X. H., Liu B. G., Sun M.,
ZhouM. F. and Viljoen M. J. (2003) Trace-element and Sr–Nd
isotopic geochemistry of the PGE-bearing Hongge
layeredintrusion, southwestern China. Int. Geol. Rev. 45,
371–382.
Zhong H., Yao Y., Prevec S. A., Wilson A. H., Viljoen M.
J.,Viljoen R. P., Liu B. G. and Luo Y. N. (2004) Trace-elementand
Sr–Nd isotopic geochemistry of the PGE-bearing Xinjielayered
intrusion in SW China. Chem. Geol. 203, 237–252.
Zhong H., Hu R. Z., Wilson A. H. and Zhu W. G. (2005) Reviewof
the link between the Hongge layered intrusion and Emeishanflood
basalts, southwest China. Int. Geol. Rev. 47, 971–985.
Zhong H., Zhu W. G., Qi L., Zhou M. F., Song X. Y. and ZhangY.
(2006) Platinum-group element (PGE) geochemistry of theEmeishan
basalts in the Pan-Xi area, SW China. Chin. Sci. Bull.51,
845–854.
Zhou C. L. (1982) Petrological characteristics and genesis of
theXinjie basic/ultrabasic layered intrusion in Miyi. J.
ChengduColleg. Geol. (2), 5–12 (in Chinese).
Zhou M. F., Malpas J., Song X. Y., Robinson P. T., Sun
M.,Kennedy A. K., Lesher C. M. and Keays R. R. (2002) Atemporal
link between the Emeishan large igneous province(SW China) and the
end-Guadalupian mass extinction. EarthPlanet. Sci. Lett. 196,
113–122.
Zhou M. F., Robinson P. T., Lesher C. M., Keays R. R., Zhang
C.J. and Malpas J. (2005) Geochemistry, petrogenesis
andmetallogenesis of the Panzhihua gabbroic layered intrusionand
associated Fe–Ti–V oxide deposits, Sichuan province, SWChina. J.
Petrol. 46, 2253–2280.
Zhou M. F., Ma Y. X., Yan D. P., Xia X. P., Zhao J. H. and SunM.
(2006) The Yanbian terrane (southern Sichuan Province,SW China): a
Neoproterozoic arc assemblage in the westernmargin of the Yangtze
Block. Precamb. Res. 144, 19–38.
Zhou M. F., Arndt N. T., Malpas J., Wang C. Y. and Kennedy A.K.
(2008) Two magma series and associated ore deposit types inthe
Permian Emeishan large igneous province, SW China.Lithos 103,
352–368.
Zhu W. G., Zhong H., Hu R. Z., Liu B. G., He D. F., Song X.
Y.and Deng H. L. (2010) Platinum-group minerals and telluridesfrom
the PGE-bearing Xinjie layered intrusion in the EmeishanLarge
Igneous Province, SW China. Miner. Petrol. 98, 167–180.
Associate editor: Edward M. Ripley
Rhenium–osmium isotope and platinum-group elements in the Xinjie
layered intrusion, SW China: Implications for source mantle
composition, mantle evolution, PGE fractionation and
mineralizationIntroductionGeological backgroundPetrography of the
Xinjie intrusionAnalytical methodsResultsVariations in major and
trace elementsVariations in chalcophile elements and PGERe–Os
isotope
DiscussionEstimation of a Xinjie parental melt compositionPGE
behavior during the evolution of the Xinjie magmaModeling of the
role of cumulus sulfidesPetrogenesis of the PGE-rich sulfidesThe
Xinjie and ELIP mantle source in a global context
ConclusionsAcknowledgmentsSupplementary dataReferences