J. metamorphic Geol., 1998, 16, 625–639 Effects of cation substitutions in garnet and pyroxene on equilibrium oxygen isotope fractionations M. J. KOHN 1 AND J. W. VALLEY 2 1 Lawrence Livermore National Laboratory, PO Box 808, Livermore, CA 94550, USA 2 Department of Geology and Geophysics, University of Wisconsin–Madison, Madison, WI 53706, USA ABSTRACT High-grade metamorphic rocks were used to explore oxygen isotope fractionations between pyroxene and garnet, and to investigate the effects on fractionation factors of the cation substitutions Fe3+Al -1 and Ca(Fe,Mg) -1 . Recrystallized, granulite facies (725 °C) wollastonite ores from the northern Adirondack highlands contain essentially only the minerals clinopyroxene (a Di–Hd solid solution)+garnet (a Grs–Adr solid solution)±wollastonite, and exhibit a systematic dependence of meas- ured fractionations on the Fe3+ content of calcic garnet: D(Cpx–CaGrt)=(0.14±0.12)+(0.78±0.20)X Adr and D(Wo–CaGrt)=(0.15±0.22)+(0.57±0.33)X Adr . In eclogites formed at T ≤650 °C, measured com- positions of Ca-poor garnet and omphacite combined with experimental data indicate that Ca-poor, Fe-rich garnet is enriched in 18O compared to both diopside and grossular: extrapolating to 1000 K, D(Alm–Di)#c. 0.2‰ and D(Alm–Grs)#c. 0.5‰. Orthopyroxene and clinopyroxene from Gore Mountain, New York, show a constant fractionation that is independent of rock type, as expected if they have the same closure temperature. These data imply D(Opx-Cpx)#c. 0.7‰ at 1000 K. Measured fractionations among Ca-poor garnet, orthopyroxene, clinopyroxene and hornblende in the Gore Mountain rocks further indicate an 18O enrichment in Ca-poor garnet over Grs (#c. 0.5‰ at 1000 K). The new measure- ments are indistinguishable from expected equilibrium values based on experiments for the minerals enstatite, diopside, grossular, wollastonite and feldspar, but consistently indicate a significant isotope effect for the simple octahedral cation substitutions Fe3+Al -1 (Grs vs. Adr) and Ca(Fe,Mg) -1 (Ca-poor garnet vs. Grs; Opx vs. Cpx). Neither cation substitution has been directly investigated for its effect on 18O/16O fractionation with experiments in silicates. Chemical characterization of minerals is required prior to petrological interpretation of oxygen isotope trends. Key words: cation substitutions; fractionations; garnet; oxygen isotopes; pyroxene. achieved for analysing natural rocks and comparing INTRODUCTION measured fractionations to an extensive database of independent determinations, it has been clearly recog- An accurate knowledge of equilibrium isotope fraction- ations among geological materials is required before nized that most rocks should not maintain equilibrium fractionations. isotope compositions can be interpreted. Consequently, much recent effort has been focused on precise As described by Kohn & Valley (1998a), although most rocks are indeed inappropriate for empirical experimental determination of fractionation factors through carbonate exchange methods (e.g. Chiba et al., calibrations, careful selection criteria can be applied to screen for those few samples that do probably maintain 1989; Clayton et al., 1989; Rosenbaum et al., 1994; Rosenbaum & Mattey, 1995; Chacko et al., 1996), and equilibrium fractionations. Each sample must be characterized concerning the timing of mineral growth on theoretical models predicting isotope partitioning (e.g. Richter & Hoernes, 1988; Zheng, 1991, 1993a,b; (i.e. prograde and retrograde reaction histories), the abundances and diffusivities of the minerals, and the Hoff bauer et al., 1994). In contrast, despite substantial improvements in the reliability of isotope measure- cooling rate. Samples that satisfy the selection criteria contain pairs of minerals whose compositions reflect ments for ever smaller samples of silicates and oxides (e.g. Sharp, 1990, 1992; Mattey & Macpherson, 1993; equilibration at an identical temperature (a closure temperature), thus a determinable temperature signifi- Valley et al., 1995), the use of natural rocks for empirically calibrating mineral-pair fractionations has cance can be ascribed to the measured fractionation. Many such samples have a specific mineralogy that fallen into disfavour. This abandonment of natural rocks in preference for experiments and computers has allows minerals to retain peak metamorphic composi- tions, but others contain mineral pairs that are reset largely been driven by the recognition that diffusional resetting of isotopes is inevitable in natural samples to a known, common closure temperature. Samples that do not satisfy the criteria simply do not contain (e.g. Giletti, 1986; Eiler et al., 1992, 1993; Jenkin et al., 1994). Thus, just as greater capabilities have been pairs of minerals with the same closure temperature, 625 © Blackwell Science Inc., 0263-4929/98/$14.00 Journal of Metamorphic Geology, Volume 16, Number 5, 1998, 625–639