Geochemistry of garnet xenocrysts and garnet-bearing xenoliths in Ordovician bentonites

Within outcrops of the black Utica shales south of the Mohawk Valley narrow re-entrants parallel with stratification are visible, often accompanied -by a marked orange-staining of the underlying strata. These rentrant features indicate the presence of bentonites weathered back faster then the enclosing shales. These bentonites are presumably magmatic ashes related to extensive volcanism when an island arc collided with the North American continent in the late Ordovician about 450 million years ago (Taconic orogeny). Bentonites are good stratigraphic markers because of the short time-interval they represent and the large area they cover. Therefore they are commonly used as correlation tools. In this study an attempt was made to use the minerals (i.e. garnets) as a correlation tool in contrast to earlier attempts to correlate bentonites by chemical fingerprinting using bulk chemistry (Huff, 1986; Kolata et al., 1986, 1987). The mineral content of these bentonites was separated and analyzed. In this study special attention was paid to the garnets and garnet-bearing rock fragments. The garnets were analyzed using an electron microprobe for major elements and trace elements, i.e. Ti, Sc, Zn and Y. In the single crystals as well as in the rock fragments two groups of garnets are observed: (1) pink grains containing 0.4-5 mol.% grossular, 56-79 mol.% almandine, 12-40 mol.% pyrope, 1-5 mol.% spessartine, and \u3c2.2 mol.% andradite; and (2) dark grains containing 14-24 mol.% grossular, 48-76 mol.% almandine, 2-35 mol.% pyrope, 1-9 mol.% spessartine, and \u3c7 mol.% andradite. For the trace elements some general trends are observed, especially a positive correlation of TiO2 with CaO, with up to 1150 ppm TiO2 in the CaO-rich garnets. A negative correlation of Sc and CaO is observed, leading to higher amounts of Sc (100-200 ppm) in the CaO-poor garnets compared to the CaO-rich garnets which contain generally less than 100 ppm Sc. Yttrium does not show a distinct behavior with changing CaO content, but it appears to favor slightly garnets with higher CaO content. Geothermobarometry calculations (Berman and Perkins, 1987; Lindsley and Anderson, 1983; Ellis and Green, 1979; Fuhrman and Lindsley, 1988; Ghent, 1976; Kohn and Spears, 1990) were attempted on rock fragments with suitable mineral assemblages. Calculated temperatures and pressures included: (1) 600ºC on pyroxene pairs; (2) 605-625ºC and 6.5-7.5 kbar on a rock fragment containing garnet-hornblende-quartz-plagioclase; (3) 450-550ºC on feldspars; (4) 600-650ºC on clinopyroxene-garnet pairs; and (5) pressures of 4.5 kbar at 600ºC for the assemblage garnet- plagioclase-aluminosilicate and quartz, which support the view that the garnets and rock fragments are derived from a metamorphic source of about granulite-facies grade. The incorporation of xenocrysts and xenoliths into the bentonites can be explained by the disruption and integration of Grenville-like crust into the ash during eruption. This explanation is consistent with the ages derived from a K-feldspar aliquot of one of the bentonites using the Ar/Ar method of dating, which shows a thermal history similar to Grenville crust K-feldspars and an age of 550-750 million years. This is too old to be derived from metamorphism during the Taconic event, nor can these feldspars be explained as authigenic phases; therefore, the best explanation is a different source than the volcanism at 450 million years, and a minimum age for this source of 750 million years, and probably older. The subject of this investigation is split into two parts: (1) Garnets as means of correlation and (2) garnets as indicators for the source of the magmas creating the bentonites. Both lines of investigation were unsuccessful. The xenocrystic garnets in the bentonites cannot be used for correlation purposes and are also of very limited use to define a source area because of their metamorphic character