Optimal Design and Operation of a Spatially Distributed Multiscale Process, with Regard to Layered Heterostructure Growth

Abstract

Determining the optimal operation of processes modeled by multiscale systems presents many challenges, because of complexities in modeling and the extensive computational requirements needed to solve them. Application to layered heterostructure deposition is considered here, where an additional challenge of interface characterization is encountered. A solution methodology is applied to the fabrication of a device consisting of alternating layers of gallium arsenide and aluminum arsenide, with the intent of minimizing thickness nonuniformity and interfacial step density. Using a finite-element solver and kinetic Monte Carlo simulations, we were able to balance reductions in thickness nonuniformity and interfacial step density with maintaining favorable operating conditions in a computationally efficient manner.

Current challenges in the design and control of multiscale systems

“…Gallivan and Atwater, 2004;Varshney and Armaou, 2006;Hu et al, 2009;Zhang et al, 2010). Comparisons between the simulation results and experimental tests have not been (Lou and Christofides, 2003) Microelectronic process (Rusli et al, 2006) Reaction-diffusion processes (Armaou et al, 2005) Molecular design approach Design a microscopic system that can achieve the desired performance in the presence of external perturbations Catalytic process (Prasad et al, 2010) Metal organic Chemical Vapour Deposition (Behrens and Armaou, 2010) reported yet for model-based control approaches in multiscale modelling. Another alternative is to use intensive properties as adjustable variables to manipulate the behaviour of the system at the small length scales and select coarse (macroscopic) process variables to measure the performance of the multiscale system.…”

“…Nevertheless, the key contribution of that work was that the resulting design of experiments was used to identify the desirable properties of the catalyst that may improve the quality of the ammonia decomposition process for a given range of operating conditions. Likewise, Behrens and Armaou (2010) determined the optimal design and the best operating conditions for a CVD that considers a heterogeneous structure for deposition. In that work, an optimisation formulation that included a spatially distributed multiscale process model was used to determine temperature of the wafer, the annealing time and the inlet velocities of the materials being deposited in the wafer that minimises thickness non-uniformity and the interfacial layers of the heterostructure, which was characterised using the step density method.…”