Determination of water partition coefficients between upper mantle minerals and melts as a function of pressure, temperature, and CO2 content using a rapid quench cell in multi-anvil experiments

人们普遍认为，在不同的地区，地幔含有的微量水从10^(1)到10^(3)wt.ppm不等。然而，水在地球内部是如何分布和循环的问题仍然存在。尽管上地幔浅层水的行为已经受到自然采样的很好的制约，但由于缺乏自然采样，软流圈深层的水的行为仍然存在争议。了解软流圈条件下矿物和玄武岩熔体之间的水分配是了解地幔中水行为的关键。尽管已经报道了一系列关于主要上地幔矿物(橄榄石和辉石)与熔体之间水分配系数的实验研究，但所有这些数据都是基于在4 Gpa以下的压力下进行的，对应于岩石圈或最顶部的软流圈条件。深部软流圈(高达13 Gpa)中的水分配是未知的，因为它应该是压力依赖的，从橄榄石中水的溶解度的显著压力依赖性中推断出来。有限压力条件的原因是玄武岩熔体不能在高于4 Gpa的典型多砧或活塞柱实验中冷却成玻璃，因此无法直接测量熔体中的水含量，从而无法确定矿物与熔体之间的水分配系数。为了研究与深部软流圈高压相对应的高压下玄武岩熔体与上地幔矿物之间的水分配，本研究将设计一种冷却速度比典型多砧实验快得多的多砧槽组件(这里称为“快速猝灭槽”)来在高压下将玄武岩熔体冷却成玻璃。通过直接测量玻璃及其共存的镁橄榄岩和顽辉石中的水含量，得到了它们的水分配系数随压力1~13 GPa1300~1900K的函数。初步实验表明，与橄榄石共存的熔体在8 GPa1870K可以被淬灭，这暗示了以前的研究可能错误地解释了分配系数的压力和/或温度关系。此外，由于熔体中的二氧化碳含量可能会显著影响水的分配，因此也将研究水分配系数与二氧化碳含量的关系。

It is commonly accepted that the Earth’s mantle contains trace amounts of water varying from 10^(1) to 10^(3) wt. ppm in different regions. However, the question about how water is distributed and cycled in the earth interior remains. Although the behaviors of water in the shallow part of upper mantle are already well constrained by natural sampling, those in the deep asthenosphere are still under debate due to the lack of natural samples. Knowledge about water partitioning between minerals and basaltic melts under asthenosphere conditions is the key for understanding the water behavior in the mantle. In spite of that a series of experimental studies about water partition coefficients between the major upper mantle minerals (olivine and pyroxene) and melts have been reported, all of those data are based on experiments performed at pressures below 4 GPa, corresponding to lithosphere or topmost asthenosphere conditions. The water partitioning in the deep asthenosphere (up to 13 GPa) is unknown since it should be pressure-dependent inferred from the significant pressure dependence of water solubility in olivine. The reason for the limited pressure condition is that basaltic melts cannot be quenched to glass in a typical multi-anvil or piston-cylinder experiment at higher than 4 GPa, and therefore impossible to directly measure the water contents in melts, sequentially, unable to determine the water partition coefficients between minerals and melts.In order to investigate the water partitioning between upper mantle minerals and basaltic melts at high pressures corresponding to those in the deep asthenosphere, in this study, we will design a multi-anvil cell assembly with much faster quenching rate (here called “rapid quench cell”) than typical multi-anvil experiments to quench the basaltic melts to glass at high pressures. By directly measuring the water contents in the glass and coexisted forsterite and enstatite, their water partition coefficients will be obtained as a function of pressure from 1 to 13 GPa, temperature from 1300 to 1900 K. The preliminary experiments demonstrated that melt coexisting with olivine can be quenched from 1870 K at a pressure of 8 GPa, and implied a possibility that the previous study might misinterpreted the pressure and/or temperature dependence of the partition coefficients. Additionally, since CO2 content in melts may significantly affect the water partitioning, the CO2-content dependence of water partition coefficients will also be investigated.