Osmium isotope disequilibrium between mantle minerals in a spinel-lherzolite
Introduction
The concentration of Os in mantle rocks is almost entirely controlled by sulphide 1, 2. Thus, the high compatibility of Os in the mantle is dependent upon the abundance and behaviour of this phase in mantle lithologies. Present understanding of sulphide solubility in silicate melts is consistent with available data for sulphur contents in oceanic basalts 3, 4, and the sulphur content of many 5, 6, if not all [7], mantle rocks is sufficiently high that melting is unlikely to remove sulphide as a residual phase [2]. These observations are consistent with the compatible behaviour of Os during mantle melting. However, it has long been recognised that if sulphide is residual during mantle melting then oceanic basalts should have extremely low Os contents (e.g. 2, 8). Many Mid-Ocean Ridge basalts (MORB) possess low Os concentrations [9], which suggests that sulphide may well be residual during melting. Nevertheless, some MORB and many Ocean Island basalts (OIB) have high Os concentrations 8, 9, 10, 11, 12, 13, 14, which have been taken to indicate partial dissolution or disequilibrium melting of sulphide 2, 11, where sulphide dissolution occurs faster than volume diffusion of Os, such that the interior of the sulphide is not in equilibrium with the melt. However, the diffusion rates for Fe and S in sulphide are at least as rapid as those for silicates [15]; moreover most sulphides melt at mantle temperatures [16]. Consequently, it is doubtful that any grains would not rapidly equilibrate with a melt.
Based on the covariation of Re and S in orogenic lherzolites and mantle peridotites 5, 17, and the high concentration of Re in sulphide inclusions in diamonds and magmatic sulphides 18, 19, it has been argued that sulphide is also a major host for Re in the mantle. However, recent direct measurement of sulphide in basalt glass indicates that very little Re enters the sulphide [20]. Similarly, parametrization of Re partitioning data suggests a minor role for residual sulphide during mantle melting [21], and it is difficult to explain the strong fractionation of Re from Os during mantle melting by sulphide control alone [21]. Experimental data for Re partitioning between garnet and a melt suggests that at least some silicate phases may exert a control on the distribution of Re in mantle rocks [22], but sulphide was not present in these experiments, and in natural systems the behaviour of Re remains poorly known.
Thus, at present, the interpretation of sulphide behaviour in the mantle source of many oceanic basalts is in conflict with understanding of phase equilibria and element diffusion in sulphide, and the distribution of Re and Os amongst coexisting sulphides and silicates remains poorly constrained. Until recently, the analysis of minerals from natural samples was limited by the low abundance of Re and Os in many phases. Recent work [20]provides Re and Os data for sulphides and their host basalt, but as yet there are no such data for both Re and Os for coexisting sulphide and silicates from mantle rocks. This study uses a low-blank solvent extraction technique for the Re–Os chemistry [23], and presents chemical and isotopic data for coexisting silicate and sulphide phases from a spinel-lherzolite from Kilbourne Hole, New Mexico. These results provide quantitative constraints on the distribution of Re and Os amongst mantle minerals and allow an assessment of the behaviour of sulphide during mantle melting.
Section snippets
Sample petrography
The sample studied is an anhydrous spinel-lherzolite (KH8312) from Kilbourne Hole, New Mexico. Silicate minerals have homogeneous major element compositions and the observed mineral modes can account for the major element composition of the bulk xenolith (Table 1). Both silicate and bulk xenolith chemistry, and mineral modes, are typical of those observed in other spinel-lherzolites from Kilbourne Hole (cf. 2, 24, 25). Olivine, ortho- and clinopyroxene occur as large crystals (up to 5 mm
Analytical techniques
Re and Os data have been obtained for the bulk xenolith and host basalt, all the major silicate phases, interstitial and included sulphide, and sulphide from the xenolith rind (Table 3). Silicate minerals were hand-picked and cleaned in ethanol, water and dilute HCl. Two interstitial sulphide grains were selected on the basis of their curvilinear faces. Four large isolate included sulphide grains were prised free of their host silicates, and six large sulphide grains were taken from the
Distribution of Re and Os amongst sulphide and silicates
The concentrations of Re and Os in the silicate phases are low, but the Re/Os ratios of all the silicates are higher than the bulk xenolith value, indicating that Re is preferentially incorporated into the silicate phases relative to Os. Using the mineral modes given in Table 1, silicate phases can only account for about 8% of the total Os, but account for about 35% of the Re in the bulk xenolith. It is difficult to estimate the proportion of sulphide present in the bulk xenolith. However,
Concluding remarks
These data confirm that sulphide dominates the Os budget of mantle rocks 1, 2but show that a significant proportion of the Re is located in the coexisting silicates. These results suggest that sulphide inclusions in mantle silicates may be shielded from reaction by their silicate hosts, in a similar way to sulphide inclusions in diamond [18], and in principle may preserve significantly older Re–Os age information than coexisting silicates. However, it is often the case that sulphides in mantle
Acknowledgements
We would like to extend special thanks to D.G. Pearson and an anonymous reviewer for their careful and constructive reviews. We are also grateful to Françoise Capmas, Mhamed Benbakkar, and Michelle Veschambre for assistance with chemistry and mass spectrometry, whole rock measurements, and electron microprobe analysis, respectively. KWB also thanks Alex Halliday for providing the time and support to finish this study during time spent at Michigan. [RV]
References (50)
- et al.
An evaluation of the solubility of sulfide sulfur in silicate melts from experimental data and natural samples
Chem. Geol.
(1990) - et al.
Sulfur in basaltic magmas
Geochim. Cosmochim. Acta
(1992) Abundance and distribution of Cu–Fe–Ni sulfides, sulphur, copper and platinum group elements in orogenic type spinel lherzolite massifs of Ariège (northeastern Pyrenees, France
Earth Planet. Sci. Lett.
(1989)- et al.
Content and isotopic composition of sulphur in ultramafic xenoliths from central Asia
Earth Planet. Sci. Lett.
(1992) - et al.
Osmium–strontium–neodymium–lead isotopic covariations in mid-ocean ridge basalt glasses and the heterogeneity of the upper mantle
Earth Planet. Sci. Lett.
(1997) - et al.
Re–Os isotope systematics of HIMU and EMII oceanic island basalts from the south Pacific Ocean
Earth Planet. Sci. Lett.
(1993) - et al.
187Os/186Os in oceanic island basalts: tracing oceanic crust recycling in the mantle
Earth Planet. Sci. Lett.
(1995) - et al.
Os isotope systematics in ocean Island basalts
Earth Planet. Sci. Lett.
(1993) - et al.
Os isotope systematics in the Azores: implications for mantle plume sources
Earth Planet. Sci. Lett.
(1996) - et al.
Sulphide inclusions in diamonds from the Koffiefontein kimberlite, S Africa: constraints on diamond ages and mantle Re–Os systematics
Earth Planet. Sci. Lett.
(1998)