Artificial intelligence technique in solving nano-process parameter optimization problem / Norlina Mohd Sabri...[et al.]

• Main Authors: Mohd Sabri, Norlina (Author), Puteh, Mazidah (Author), Md Sin, Nor Diyana (Author)
• Format: Book
• Published: Faculty of Mechanical Engineering Universiti Teknologi MARA (UiTM), 2017.
• Subjects: Article; NonPeerReviewed
• Online Access: Link Metadata
• Notes: https://ir.uitm.edu.my/id/eprint/36804/1/35265.pdf
• Keywords: TJ Mechanical engineering and machinery

Nanotechnology has brought huge impacts to our modern life in many ways. The technology has been adapted in various domains such as biotechnology, chemicals, computing, electronics, metals, materials, renewable energy and also telecommunications. This research is focusing on the RF magnetron sputtering process, one of the nanotechnology processes which are widely used in the thin film technology. The conventional method that is currently practiced in the optimization of the RF magnetron sputtering process parameters is trial and error method. However, the practice of repetitively conducting experiments has consumed a lot of time and cost in the thin film construction. This research is proposing artificial intelligence (AI) technique as the alternative technique to overcome the sputtering process parameter optimization problem. Three artificial intelligence techniques have been selected to solve this parameter optimization problem. The techniques are Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Gravitational Search Algorithm (GSA). In the process parameter optimization, automated process has never been implemented to optimize the six deposition parameters of the RF magnetron sputtering. Based on the results, GA has shown the best performance compared to PSO and GSA in terms of the fitness value and processing time. Laboratory results have shown that GA has produced the highest values of electrical and optical properties of thin film. It is expected that AI techniques could complement the conventional method of trial and error in obtaining the optimized process parameter combination.