A new olivine-melt thermometer/hygrometer based on Ni partitioning

Abstract:

Temperature and H₂O content are key to understanding the origin and evolution of arc basalts. Here we propose that the strong temperature dependence on D_Ni (olivine-melt), quantified by Li & Ripley (2010), can be used as a thermometer that is largely independent of melt H2O concentration. Currently, the most widely used olivine-melt thermometers (based on partitioning of Mg) are strongly dependent on melt H2O content, and application to hydrous basalts requires that melt H2O contents already be known. If an H2O-independent olivine-melt thermometer could be developed, then application to hydrous basalts will lead to accurate temperatures at the time of olivine crystallization. These temperatures can then be combined with Mg-based olivine-melt thermometers to obtain the magnitude of ∆T, the depression of the olivine liquidus due to melt H2O concentration. In turn, hydrous phase equilibrium data from the literature allow melt H2O concentration to be derived from ∆T (e.g., Medard &Grove, 2008; Almeev et al., 2007). Thus, an H₂O-independent olivine-melt thermometer can be used to calculate melt H2O concentrations at the onset of olivine crystallization.

To test this approach in hydrous systems, we obtained new Ni concentration data for the H2O-saturated basaltic andesite phase equilibrium experiments of Moore and Carmichael (1998). The predicted temperatures from the inverted L&R’10 model are within 25° C of the measured experimental temperatures. This approach was also tested on 192 olivine-liquid pairs from 18 experimental studies that reported Ni content in both olivine and melt. The average residuals between the calculated and experimental T is -16 °C. We also calibrated a new model based on the same format above, but with an additional pressure term (average residual of -2 °C). Application of this approach to lavas for which H₂O in olivine-hosted melt inclusions have been analyzed in the literature shows excellent agreement.