Research Article
Optimization of the Deposition Condition for Improving the Ti Film Resistance of DRAM Products
@INPROCEEDINGS{10.1007/978-3-030-69514-9_40, author={Yun-Wei Lin and Chia-Ming Lin}, title={Optimization of the Deposition Condition for Improving the Ti Film Resistance of DRAM Products}, proceedings={Smart Grid and Internet of Things. 4th EAI International Conference, SGIoT 2020, TaiChung, Taiwan, December 5--6, 2020, Proceedings}, proceedings_a={SGIOT}, year={2021}, month={7}, keywords={DRAM Response surface method Neural networks Genetic algorithms}, doi={10.1007/978-3-030-69514-9_40} }- Yun-Wei Lin
Chia-Ming Lin
Year: 2021
Optimization of the Deposition Condition for Improving the Ti Film Resistance of DRAM Products
SGIOT
Springer
DOI: 10.1007/978-3-030-69514-9_40
Abstract
Dynamic random access memory (DRAM) products are the key parts in consumer products. To fulfill the current market’s strict specifications, various customers have asked DRAM manufacturers to continue improving the quality of DRAM products. The resistance of the Ti film directly affects the electrical quality of DRAM products. At present, the DRAM products developed by the case company have caused customer returns due to abnormal resistance value of Ti film. Process engineers always adjust the engineering parameters based on experience, which resulted in slow improvement and inability to determine the setting of engineering parameters. Consequently, shipments of DRAM products are delayed. This study adopts the Ti film resistance of DRAM products as the main research object for improvement and applies the response surface method, neural networks, and genetic algorithms to help process engineers analyze and improve DRAM products. This work assists the case company in achieving a significant improvement in Ti film resistance from 210.33 Ω (the origin made by the case company) to 185.28 Ω (the improvement made by this work) where the specified target value is 185 Ω. The results are effective in shortening the improvement time and reducing customer returns.
