Physics-separating artificial neural networks for predicting sputtering and thin film deposition of AlN in Ar/N$_2$ discharges on experimental timescales
This work addresses a multi-scale, multi-physics problem in plasma-surface interactions for thin film deposition, offering a practical solution for experimental applications, though it appears incremental as it builds on prior scale-bridging machine learning models.
The authors tackled the challenge of predicting sputtering and thin film deposition of AlN in Ar/N2 discharges on experimental timescales, which was previously hindered by high computational costs, and achieved predictions for process times up to 45 minutes by developing a physics-separating artificial neural network.
Understanding and modeling plasma-surface interactions frame a multi-scale as well as multi-physics problem. Scale-bridging machine learning surface surrogate models have been demonstrated to perceive the fundamental atomic fidelity for the physical vapor deposition of pure metals. However, the immense computational cost of the data-generating simulations render a practical application with predictions on relevant timescales impracticable. This issue is resolved in this work for the sputter deposition of AlN in Ar/N$_2$ discharges by developing a scheme that populates the parameter spaces effectively. Hybrid reactive molecular dynamics / time-stamped force-bias Monte Carlo simulations of randomized plasma-surface interactions / diffusion processes are used to setup a physics-separating artificial neural network. The application of this generic machine learning model to a specific experimental reference case study enables the systematic analysis of the particle flux emission as well as underlying system state (e.g., composition, mass density, stress, point defect structure) evolution within process times of up to 45 minutes.