A transferable force field for gallium nitride crystal growth from the melt using on-the-fly active learning

Atomic-scale simulations of reactive processes have been stymied by two factors: the general lack of a suitable semi-empirical force field on the one hand, and the impractically large computational burden of using ab initio molecular dynamics on the other. In this paper, we use an “on-the-fly” active learning technique to develop a non-parameterized force field that, in essence, exhibits the accuracy of density functional theory and the speed of a classical molecular dynamics simulation. We developed a force field suitable to capture the crystallization of gallium nitride (GaN) using a novel additive manufacturing route and a combination of liquid Ga and ammonia gas precursors to grow GaN thin films. We show that this machine learning model is capable of producing a transferable force field that can model all three phases, solid, liquid and gas, involved in this additive manufacturing process. We verified our computational results against a range of experimental measurements and ab initio molecular dynamics simulation, showing that this non-parametric force field shows excellent accuracy as well as a computationally tractable efficiency. The development of this transferable force field opens the opportunity to simulate liquid phase epitaxial growth more accurately than before, analyze reaction and diffusion processes, and ultimately establish a growth model of the additive manufacturing process to create gallium nitride thin films.
In this archive, we included the mapped Gaussian Process force field parameters of gallium and gallium nitride for LAMMPS simulations. Users can download these force field parameters to test and recreate similar Molecular Dynamic simulation discussed in the paper.

原子尺度反应过程的模拟受到两大因素的制约：一方面是缺乏合适的半经验力场，另一方面则是使用从头算分子动力学计算所带来难以承受的计算负担。在本文中，我们运用一种“即时”的主动学习方法，开发了一种非参数化力场，其本质上兼具密度泛函理论的精度和经典分子动力学模拟的速度。我们采用一种新型的增材制造途径，结合液态镓和氨气前驱体，开发了一种适用于捕获氮化镓（GaN）结晶的力场。我们展示了该机器学习模型能够产生可迁移的力场，能够模拟增材制造过程中涉及的固态、液态和气态三种相态。我们通过一系列实验测量和从头算分子动力学模拟验证了我们的计算结果，表明这一非参数化力场不仅展现了卓越的精度，还具有易于处理的计算效率。该可迁移力场的发展为以前更精确地模拟液相外延生长、分析反应和扩散过程以及最终建立增材制造过程以制备氮化镓薄膜的生长模型提供了机遇。在本档案中，我们包含了适用于 LAMMPS 模拟的镓和氮化镓的映射高斯过程力场参数。用户可以下载这些力场参数以测试和重现论文中讨论的类似分子动力学模拟。