LEAPS-MPS: Spin-lattice interaction in paramagnetic cubic iron at high pressure

该奖项全部由《2021年美国救援计划法案》（公法117-2）资助。这项leap - mps奖支持磁性对铁原子在高温和高压下排列方式的影响的计算和理论研究。作为地球的主要成分之一，对我们的技术文明至关重要，铁是研究最深入的材料之一，但量子力学模拟同时包括磁性和铁原子因温度而偏离平衡位置的位移模式，这在计算上是禁止的。然而，这些模拟是重要的，因为材料的许多物理和化学性质取决于原子的特定排列及其在平衡位置周围的振动，而准确预测这种模式的能力可以使材料的设计仅使用计算机模拟。研究小组将通过使用特定的计算机算法来有效地搜索导致与实验一致的预测的一般参数，并通过使用这些参数来指导量子力学和其他模拟，从而解决这一计算挑战。该团队还将率先使用一种机器学习算法，该算法能够学习材料的两个或多个特征，并透明地显示它们的相互作用。该奖项还支持教育资源的创建，这些资源将用于向社区大学生展示计算材料科学的概念，并将促进由德克萨斯大学埃尔帕索分校的教师和本科生研究人员为社区大学生提供正式和非正式的指导，并与加州大学伯克利分校的研究人员建立研究合作。这项leap - mps奖支持计算和理论研究，研究面心立方铁和体心立方铁顺磁状态下的磁自旋激发（磁振子）如何通过与声子的相互作用影响两相在熔点和压力高达15 GPa的温度下的相对热力学稳定性。从第一性原理模拟晶格动力学的计算成本很高，并且包含自旋自由度会使相互作用的直接计算变得难以实现。然而，自旋-晶格耦合对该体系的热力学和相稳定性的计算中不可忽视，估计其对自由能的贡献高达35 meV/原子。研究团队将通过(i)经验约束原子间力常数，通过进化计算可以计算声子色散关系，以及（ii）基于能够学习原子和自旋配置的数学图开发机器学习模型，用于动力学模拟。该奖项还支持教育资源的创建，这些资源将用于向社区大学生展示计算材料科学的概念，并将促进由德克萨斯大学埃尔帕索分校的教师和本科生研究人员为社区大学生提供正式和非正式的指导，并与加州大学伯克利分校的研究人员建立研究合作。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This award is funded in whole under the American Rescue Plan Act of 2021 (Public Law 117-2).NONTECHNICAL SUMMARYThis LEAPS-MPS award supports computational and theoretical research on the effect of magnetism on the way iron atoms arrange themselves at high temperatures and pressures. As one of the principal constituents of the Earth, and being of paramount importance to our technological civilization, iron is one of the most intensely studied materials, but quantum mechanical simulations that simultaneously include magnetism and the displacement patterns of iron atoms from their equilibrium positions due to temperature are computationally prohibitive. These simulations are nevertheless important because many physical and chemical properties of materials depend on the particular arrangement of atoms and their vibrations around equilibrium positions, and the ability to accurately predict such patterns can enable the design of materials using only computer simulations. The research team will approach this computational challenge by using a particular computer algorithm to efficiently search for general parameters that result in predictions that are consistent with experiments, and by using these parameters to guide quantum mechanical and other simulations. The team will also spearhead the use of a machine learning algorithm capable of learning two or more features of the material and transparently showing their interaction.This award also supports the creation of educational resources that will be used to expose community college students to concepts of computational materials science and will facilitate formal and informal mentorship for community college students by faculty and undergraduate researchers at the University of Texas El Paso, and establishing research collaborations with researchers at the University of California Berkeley.TECHNICAL SUMMARYThis LEAPS-MPS award supports computational and theoretical research on how magnetic spin excitations (magnons) in the paramagnetic states of face-centered cubic and body-centered cubic iron affect the relative thermodynamic stability of the two phases at temperatures up to the melting point and pressures up to 15 GPa through their interaction with phonons. Simulations of the lattice dynamics from first principles are computationally expensive, and the inclusion of spin degrees of freedom can make the direct computation of the interaction prohibitive. Nevertheless, with an estimated contribution to the free energy of up to 35 meV/atom, the spin-lattice coupling cannot be ignored in calculations of the thermodynamics and phase stability of this system.The research team will approach the challenge by (i) empirically constraining the interatomic force constants, from which phonon dispersion relations can be computed via evolutionary computation, and (ii) developing machine learning models based on mathematical graphs capable of learning both atomic and spin configurations to be used in dynamics simulations.This award also supports the creation of educational resources that will be used to expose community college students to concepts of computational materials science and will facilitate formal and informal mentorship for community college students by faculty and undergraduate researchers at the University of Texas El Paso, and establishing research collaborations with researcher at the University of California Berkeley.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.