Optimization of plasma enhanced chemical vapor deposition process parameters for hardness improvement of diamond-like carbon coatings

Recently, there has been a surge in the use of metaheuristic algorithms to design materials with optimum performance. In this paper, the RPSOLC (Repulsive Particle Swarm Optimization with Local search and Chaotic perturbation) metaheuristic algorithm was used to design Diamond-Like Carbon (DLC) thin lms with improved hardness. Based on the Box-Behnken design, 15 independent DLC deposition experiments are performed in a PECVD (Plasma-Enhanced Chemical Vapor Deposition) setup by varying the CH4-Argon ow rate, hydrogen ow rate, and deposition temperature. The nano-hardness of the DLCs is evaluated using nano-indention tests. The hardness is then expressed as the function of the three process parameters using a polynomial regression metamodel. Finally, using RPSOLC, the metamodel is optimized and compared to the optimal predictions of a traditional Genetic Algorithm (GA). It is seen that RPSOLC has faster convergence and is more reliable than the GA. In general, a high H2 ow rate along with a low CH4-Ar ow rate and high temperature is found to be bene cial in improving the hardness.