HYDROXYL DEFECTS IN GARNETS FROM MANTLE XENOLITHS IN KIMBERLITES OF THE SIBERIAN PLATFORM

A suite of more than 200 garnet single crystals, extracted from 150 xe noliths, covering the whole range of types of garnet parageneses in ma ntle xenoliths so far known from kimberlites of the Siberian platform and collected from nearly all the kimberlite pipes known in that tecto nic unit, as well as some garnets found as inclusions in diamonds and olivine megacrysts from such kimberlites, were studied by means of ele ctron microprobe analysis and single-crystal IR absorption spectroscop y in the v(OH) vibrational range in search of the occurrence, energy a nd intensity of the v(OH) bands of hydrosyl defects in such garnets an d its potential use in an elucidation of the nature of the fluid phase in the mantle beneath the Siberian platform. The v(OH) single-crystal spectra show either one or a combination of two or more of the follow ing major v(OH) bands, I 3645-3662 cm(-1). II 3561-3583 cm(-1), III 35 15-3527 cm(-1), and minor bands, Ia 3623-3631 cm(-1), IIa 3593-3607 cm (-1). The type of combination of such bands in the spectrum of a speci fic garnet depends on the type of the rock series of the host xenolith , Mg, Mg-Ca, Ca, Mg-Fe, or alkremite, on the xenolith type as well as on the chemical composition of the respective garnet. Nearly all garne ts contain band systems I and II. Band system III occurs in Ti-rich ga rnets, with wt% TiO2 > ca. 0.4, from xenoliths of the Mg-Ca and Mg-Fe series, only. The v(OH) spectra do not correspond to those of OH- defe cts in synthetic pyropes or natural ultra-high pressure garnets from d iamondiferous metamorphics. There were no indications of v(OH) from in clusions of other minerals within the selected 60 X 60 mu m measuring areas in the garnets. The v(OH) spectra of pyrope-knorringite- and pyr ope-knorringite-uvarovite-rich garnets included in diamonds do not sho w band systems I to III. Instead, they exhibit one weak, broad band (D elta v(OH) 200-460 cm(-1)) near 3570 cm(-1), a result that was also ob tained on pyrope-knorringite-rich garnets extracted from two olivine m egacrysts. The quantitative evaluation, on the basis of relevant exist ing calibrational data (Bell et al. 1995), of the sum of integral inte nsities of all v(OH) bonds of the garnets studied yielded a wide range of ''water'' concentrations within the set of the different garnets, between values below the detection limit of our single-crystal IR meth od, near 2 X 10(-4) wt%, up to 163 X 10(-4) wt%. The ''water'' content s vary in a complex manner in garnets from different xenolith types, o bviously depending on a large number of constraints, inherent in the c rystal chemistry as well as the formation conditions of the garnets du ring the crystallization of their mantle host rocks. Secondary alterat ion effects during uplift of the kimberlite, play, if any, only a mino r role. Despite the very complex pattern of the ''water'' contents of the garnets, preventing an evaluation of a straightforward correlation between ''water'' contents of the garnets and the composition of the mantle's fluid phase during garnet formation, at least two general con clusions could be drawn: (1) the wide variation of ''water'' contents in garnets is not indicative of regional or local differences in the c omposition of the mantle's fluid phase; (2) garnets formed in the high -pressure/high-temperature diamond-pyrope facies invariably contain si gnificantly lower amounts of ''water'' than garnets formed under the c onditions of the graphite-pyrope facies. This latter result (2) may po int to significantly lower fH(2)O and fO(2) in the former as compared to the latter facies.