Ab initio phonon models of lattice thermal conductivity of lower mantle minerals

地球内部的热驱动着地表的板块构造，地幔内部的对流，以及核心的磁流体动力学。然而，作为全球热收支的重要组成部分，穿越核-地幔边界(CMB)的总热流仍然受到限制，这是因为实验室热传输实验还不能直接获得相关的高温/压力条件，而较低温度/压力下测量数据的长期外推导致先前对CMB上方热边界层导热系数的估计存在很大的不确定性。基于第一性原理量子理论的原子尺度计算和模拟为确定含铁下地幔矿物固溶体的晶格热导率提供了一种补充方法。大型并行超级计算机的出现，以及研究组成功的算法/数据结构并行化和优化，使得这项研究成为可能。新的第一性原理计算机模型不仅将改善对CMB热流估计的约束，而且计算结果还将有助于我们理解复杂材料系统中的微观热传导过程，从而揭示极端条件下热传输的基本物理。本研究所发展的计算方法可用于其他复杂材料体系的研究，如新型热电材料。对研究生在固体物理、地球物理和高性能计算方面的培训将为下一代计算矿物物理学家提供全面的跨学科教育。具体地说，将提出地球下地幔热导率的理论模型，该模型以温度、压力和铁浓度/自旋态与四种主要下地幔矿物组的晶格热导率的稳健第一原理模型为基础，包括铁方镁矿、含铁钙钛矿和后钙钛矿镁硅、立方钙钛矿，利用新实施的结合密度泛函理论和动力学声子输运理论的第一原理计算方法。第一原理理论数据将以地球动力学模拟和/或其他需要热传输数据的地球科学应用程序容易采用的格式进行表述和制表。此外，新的理论模型将以最新的实验数据为基准，(1)验证和进一步改进热输运性质的经验温度-压力外推模型，(2)探索极高温下新的热传导机理。

The Earth's internal heat drives plate tectonics at the surface, convection within the mantle, and the magneto hydrodynamics in the core. However, the total heat flow across the core-mantle boundary (CMB), a crucial component of the global heat budget, remains poorly constrained, because laboratory thermal transport experiment can not yet directly access the relevant high temperature/pressure conditions, and long extrapolations of measured data at lower temperatures/pressures results in large uncertainties in previous estimates of thermal conductivity of the thermal boundary layer right above the CMB. Atomic scale calculations and simulations based on the first-principles quantum theories provide a complimentary approach to determine the lattice thermal conductivity of iron bearing lower mantle mineral solid solutions. The proposed study is made possible by the availability of large parallel supercomputers and the successful algorithm/data-structure parallelization and optimization by the investigator's group. Not only the new first-principles computer models will improve the constraints on the estimated CMB heat flow, the calculated results will also help us to understand microscopic heat conduction processes in the complex materials systems and therefore reveal the basic physics of heat transport at extreme conditions. The computational methodology development in this study can be adopted to study other complex materials systems, such as novel thermoelectric materials. The training of graduate students in solid-state physics, geophysics, and high-performance computing will provide a well-rounded interdisciplinary education for the next-generation computational mineral physicists.Specifically, theoretical models of thermal conductivity of the Earth's lower mantle will be proposed, anchored on robust first-principles models of temperature, pressure, and iron concentration/spin-state dependences of lattice thermal conductivity in four major lower mantle mineral groups, including ferropericlase, Fe-bearing perovskite and post-perovskite MgSiO3, cubic CaSiO3 perovskite, derived using the newly implemented first-principles computational method that combines density functional theory and the kinetic phonon transport theory. The first-principles theoretical data will be formulated and tabulated in the formats that can be easily adopted by geodynamical simulations and/or other Earth Sciences applications that require thermal transport data. In addition, the new theoretical models will be benchmarked with latest experimental data (1) to validate and further improve empirical temperature-pressure extrapolation models for thermal transport properties, and (2) to explore new heat conduction mechanisms at extremely high temperatures.