We present experimental results constraining the partitioning of fluorine (F) between titanite and metaluminous silicate melt under H2O-saturated conditions at 50–200 MPa, 875–925 °C. Resulting experimental melts are metaluminous with ASI values [aluminium saturation index, calculated as molar Al2O3/(CaO+Na2O+K2O)] falling in the range 0.68–0.97. Titanite crystals and coexisting melts have F contents that vary from 0.2 to 2.6 wt% and from 0.2 to 3.2 wt%, respectively. The calculated proportion of the F–O substitution is within 0.02–0.28 per formula unit, which is predominantly compensated by the substitution of Al and Fe3+ for Ti, as well as a subordinate amount of Ti–Mg2+ substitution. The mass-ratio partition coefficient DTtnmeltF is within 0.5–1.2, and there is little influence of pressure under the conditions investigated. However, the potential effect of temperature needs further investigation. DTtnmeltF is positively correlated with melt ASI, and can be predicted using the following relation: DTtnmeltF=2.26ASI1.05.

This equation is valid only for metaluminous melts (ASI < 1). The effect of melt ASI on DTtnmeltF is related to structural properties of F-bearing melts, which can incorporate higher F concentrations at lower ASI. Assuming that the high-Ti titanite investigated here is subject to ideal mixing, we calculated a Margules parameter (WmeltF) of 48.6 ± 4.3 kJ/mol for F in silicate melt with a symmetric binary mixing model. The high value of WmeltF confirms that O–F mixing in silicate melt is far from ideal. The reliability of using of DTtnmeltF to calculate melt F contents has been checked by investigating plutonic rocks of the Liujiawa pluton in the Dabie orogen of eastern China. Our new estimates of melt F content are consistent with those derived from amphibole, biotite and apatite compositions.

You do not have access to this content, please speak to your institutional administrator if you feel you should have access.