Partitioning of fluorine and chlorine between biotite and granitic melt: experimental calibration at 200 MPa H2O

Experiments from 640 to 680?°C, 200 MPa H₂O at?f _O2?≈?NNO, employing a natural?F-rich?vitrophyric rhyolite from Spor Mountain, Utah, assessed the effect of variable Mg′ 100Mg/(Mg?+?Mn?+?Fe)] on the partitioning of fluorine and chlorine between biotite (Bt) and melt. Over this temperature interval, Bt (?±?fluorite, ?±?quartz) is the sole liquidus phase. Partition coefficients for fluorine between biotite and glass (D_F ^Bt/melt) show a strong dependence on the Mg′ of Bt.?D_F ^Bt/melt varies from???1.5 to 7.2 over the range of Mg′ from 21 to 76. A strong linear correlation between?D_F ^Bt/melt?and Mg′ has a slope of 9.4 and extrapolates through the origin (i.e., D_F ^Bt/melt?≈?0 at Mg′?=?0, an annite-siderophyllite solid solution in these experiments). D_Cl ^Bt/melt values (???1 to 6) in the same experiments vary inversely with Mg′. The Al-content of biotite does not vary with the aluminum saturation index (ASI?=?molar Al₂O₃/Σ alkali and alkaline earth oxides) of melt, but two exchange mechanisms involving Al appear to operate in these micas: (1) Al^vi?+?Al^iv?? Si^iv?+?Mg^iv, and Mg^iv?+?2Al^iv? 2Si^iv?+?□^iv. The effects of other components such as Li or other intensive parameters including f _O2 have yet to be evaluated?systematically. At comparable Mg′ of Bt, however, the Spor Mountain rhyolite yields higher D_F ^Bt/melt values than an Li-rich, strongly peraluminous melt previously investigated. The results indicate that the Mg′ of Bt exerts the principal control on halogen partitioning, with ASI and T as second-order variables. The experimental partition coefficients compare well with other experimental results but not with most volcanic rocks. Magmatic Bt from most rhyolites records higher D_F ^Bt/melt due to reequilibration with degassed (H₂O-depleted) magma and perhaps with F₂O_?1 exchange that may accompany oxidation (Fe³⁺O] Fe²⁺OH]_?1). This behavior is evident in magmatic biotite from a zoned peraluminous rhyolite complex near Morococala, Bolivia: Bt is sharply zoned with F-rich rims, but Bt(core)-melt inclusion pairs fall on our experimental curve for D_F ^Bt/melt. These experimental data can be used in part to assess the preservation of magmatic volatile contents in plutonic or volcanic silicic rocks. For plutonic rocks, the actual F-content of melt, not a relative activity ratio involving HF species, can be reasonably estimated if the mica has not undergone subsolidus reequilibration. This information is potentially useful for some shallow-level Ca-poor magmas that are thought to be rich in F (e.g., A- and S-type granites) but do not conserve F well as rocks.