Pergamon Geochimicaet Cosmochimica Acta, Vol. 60, No. 14, pp. 2595-2608, 1996 Copyright© 1996 Elsevier ScienceLtd Printed in the USA. All rightsreserved 0016-7037/96 $15.00 + .00 PIIS0016-7037(96)00112-3 Oxygen isotope fractionations in muscovite, phlogopite, and rutile THOMAS CHACKO,I XIANGSHENGHU, 2 TOSHIKO K. MAYEDA, 3 ROBERT N. CLAYTON, 4 and JULIAN R. GOLDSMITH 5 Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta T6G 2E3, Canada 2 Department of the Geophysical Sciences, University of Chicago, Chicago, IL 60637, USA 3 Enrico Fermi Institute, University of Chicago, Chicago, IL 60637, USA 4 Enrico Fermi Institute, Department of Chemistry, and Department of the Geophysical Sciences, University of Chicago, Chicago, IL 60637, USA 5 Department of the Geophysical Sciences, University of Chicago, Chicago, IL 60637, USA (Received June 9, 1995; accepted in revised form March 26, 1996) Abstract--Oxygen isotope fractionations in laboratory systems have been determined for each of the following minerals relative to calcite: muscovite, phlogopite, fluorophlogopite, and rutile. Statistical mechanical calculations following the method of Kieffer (1982) were fit to the experimental data and then used to extrapolate the experimental results to higher and lower temperatures. The calculations are represented by a series of equations which allow the reduced partition function ratios (~ factors) for each of these minerals to be calculated at T > 400 K. These equations can be combined with correspond- ing equations for calcite, quartz, albite, anorthite, diopside, forsterite, and magnetite (Clayton and Kieffer, 1991 ) to give a large number of mineral-pair fractionations for use as isotopic thermometers. It was found that the high-frequency vibrations of OH bonds contribute such a small amount of the fractionation factors that they do not introduce significant nonlinearity to plots of A vs. T -2. The commonly used calibrations of quartz-muscovite and quartz-biotite fractionations are not in good agreement with the present experimental measurements. This probably refects disturbance of the rock assemblages on which those calibrations were based, as a consequence of the high diffusivity of oxygen in micas. The experimen- tal quartz-rutile fractionations are in good agreement with some earlier hydrothermal experiments and with an empirically determined calibration. The calculated rutile partition functions of Kieffer (1982) are not consistent with the experimental results, probably due in part to the neglect of the effect of cation mass on the vibrational energies. The large number of mineral systems with measured fractionation factors allows a test of various empirical relationships based on oxygen bond strengths. In general, these relationships are successful for anhydrous silicates, but do not adequately account for the behavior of hydrous minerals or metal oxides. 1. INTRODUCTION Accurate determinations of equilibrium oxygen isotope frac- tionations between minerals are central to the correct inter- pretation of oxygen isotope systematics in natural samples. As a result, numerous attempts have been made to obtain this information using a variety of methods (see reviews by Clayton, 1981; O'Neil, 1986; Kyser, 1987). One of the most successful of these methods to date has been an experimental technique that uses calcium carbonate (rather than water) as the oxygen exchange medium (Clayton et al., 1989; Chiba et al., 1989). Fractionations obtained for a number of minerals using this technique are in excellent agreement with those derived from statistical mechanical calculations (Kieffer, 1982; Clayton et al., 1989; Chiba et al., 1989). This result is very encouraging because it suggests that the approximations inherent to the theoretical calculations are generally reason- able, and, consequently, that theoretical models provide reli- able guidelines for extrapolating the fractionation data out- side the experimentally investigated temperature range (Clayton and Kieffer, 1991). Until recently, detailed comparisons of fractionation fac- tors obtained from experiments and statistical mechanical theory had been made only for anhydrous silicates and car- bonates (Bottinga, 1968; Chacko et al., 1991; Clayton and Kieffer, 1991). It is important that hydrous silicates and oxides be included in this discussion as these phases have also been used widely for isotopic thermometry. Experimen- 2595 tal data on exchange between calcite and phlogopite and between calcite and tremolite have been presented (Fortier et al., 1994; Zheng et al., 1994b). Hydrous silicates are interesting in that the high vibrational frequencies of the hydroxyl units are thought to result in a more complicated temperature dependence of fractions than is observed in frac- tionations between anhydrous silicates (Bottinga and Javoy, 1973 ). Oxide minerals may also present some complications as these minerals have different bonding characteristics from silicates and commonly involve cations with significantly greater mass than those that determine the structure of sili- cates. Thus, the approximations made in the calculations for silicates (Kieffer, 1982) may not be directly applicable to the oxides and must be evaluated in light of experimental data. Our objectives in this paper are fourfold. First, we present experimental fractionation data for muscovite, phlogopite, fluorophlogopite, and rutile obtained using the carbonate- exchange technique. Second, we compare fractionation fac- tors obtained from our experiments with those calculated from statistical mechanical theory, in particular the calcula- tions of Kieffer (1982) for the muscovite-calcite and rutile- calcite systems. Third, following the approach of Clayton and Kieffer (1991), we link the theoretical calculations to the experimental data by adjusting the calculated partition function ratios for the silicate or oxide mineral so as to fit the experimental data. The advantage of this combined