首页    期刊浏览 2024年11月24日 星期日
登录注册

文章基本信息

  • 标题:Machine learning-based analysis of the physio-chemical properties for the predictive thickness control of atomic layer deposition
  • 本地全文:下载
  • 作者:Changsu Kim ; Thai Ngan Do ; Jiyong Kim
  • 期刊名称:IFAC PapersOnLine
  • 印刷版ISSN:2405-8963
  • 出版年度:2022
  • 卷号:55
  • 期号:7
  • 页码:626-631
  • DOI:10.1016/j.ifacol.2022.07.513
  • 语种:English
  • 出版社:Elsevier
  • 摘要:AbstractAtomic layer deposition (ALD) is an outstanding thin film deposition technique based on the surface chemical reaction. As the conventional device fabrication method is reported to be ineffective under the 5nm process, ALD drew attention to its ability to control the film growth. To develop an efficient ALD process, controlling film thickness is one of the most important factors. The key requirement for the film thickness control is to understand their chemical reactivity upon adsorption on different substrate surfaces. However, the current research on viable ALD development have remained inefficient because of the expensive and time-consuming experiments. In this study, we aim to analyze and suggest a strategy for identifying the contribution of ALD process features on the device film thickness based on principal component analysis (PCA) method. First, the features of ALD experiment (Precursor properties, reactant type, substrate properties, operating conditions) are compiled to define the chemical dimension of ALD process. Then, the contributions of ALD experiment features on the thickness growth are analyzed by projecting ALD chemical dimension into the information space using PCA technique. As a result, we could identify highly sensitive features for ALD film growth and suggest the strategy for controlling film thickness. From providing the analysis for the film growth, we elucidate the importance of high dimension analysis of ALD process and improve the understanding on film growth control.
  • 关键词:KeywordsAtomic layer processingMachine learningArtificial intelligence
国家哲学社会科学文献中心版权所有