Optimization of Multilayer Wear-Resistant Thin Films Using Finite Element Analysis on Stiff and Compliant Substrates

Extensive research has been carried out by researchers on the growth and characterization of multilayer protective coatings, but the design of these coatings still remains largely empirical. In this regard, recent progress has been made in developing a design approach for optimizing a multilayer coating structure before deposition, which would help save time and material. In pursuit of an optimal design, finite element analysis using a plane strain Hertzian contact model was developed to investigate the stress/ strain behavior within the layers of the system. The present study looks to find the optimal thicknesses of individual layers in a multilayer coating/substrate system that can reduce stresses and/or strains in the system. Multilayer Cr/CrN thin films were modeled and optimized to have effective “load support” by the films on stiff A2 steel and compliant 2024-Al substrates. Optimization was carried out using both multiobjective and single-objective procedures for the models of eight-layer film on substrates. For the multilayer on A2 steel substrate, the first test case is a multiobjective optimization performed by minimizing the strain discontinuities at the coating/substrate interface and the stresses developed in the uppermost layer under combined normal and tangential load conditions. Another option is a single-objective optimization (minimizing the strain discontinuity) and constraining the stress to values below the yield stress. The same two test cases were employed on the 2024-Al model, but the stresses considered were those in the substrate in order to keep the model within the elastic regime. Efficiency of several optimization algorithms, such as genetic algorithms and gradient based routines are discussed and the preliminary results are compared to experimental pin-on-disk wear results of empirically designed coatings. Architectures were found with improvements in the elastic measures