{"title":"分类回归方法的材料去除率预测","authors":"Kart-Leong Lim, R. Dutta","doi":"10.1109/EPTC50525.2020.9315140","DOIUrl":null,"url":null,"abstract":"Chemical Mechanical Polishing (CMP) is one of the most critical process step in the fabrication of advanced packages, such as Fanout Wafer Level Packaging (FOWLP). CMP process requires tight and dynamic control of process parameters to achieve palnarization, high quality and reliability of organic or in-organic redistribution layer (RDL) surface morphology. Typically, physics based or data driven approaches are implied to predict material removal rate (MRR) and run time control. The former models a closed-form expression between domain knowledge and MRR. Often, the domain knowledge are based on kinetics and contact interaction between the wafer, and the polishing tool. While the latter use time series based training data and machine learning to predict MRR. In this paper, we demonstrate to incorporate wear knowledge as classification and show its effectiveness in predicting MRR. Our experiments shows better overall accuracy being achieved through the proposed classification and regression framework.","PeriodicalId":6790,"journal":{"name":"2020 IEEE 22nd Electronics Packaging Technology Conference (EPTC)","volume":"3385 1","pages":"172-175"},"PeriodicalIF":0.0000,"publicationDate":"2020-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Material Removal Rate Prediction using the Classification-Regression Approach\",\"authors\":\"Kart-Leong Lim, R. Dutta\",\"doi\":\"10.1109/EPTC50525.2020.9315140\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Chemical Mechanical Polishing (CMP) is one of the most critical process step in the fabrication of advanced packages, such as Fanout Wafer Level Packaging (FOWLP). CMP process requires tight and dynamic control of process parameters to achieve palnarization, high quality and reliability of organic or in-organic redistribution layer (RDL) surface morphology. Typically, physics based or data driven approaches are implied to predict material removal rate (MRR) and run time control. The former models a closed-form expression between domain knowledge and MRR. Often, the domain knowledge are based on kinetics and contact interaction between the wafer, and the polishing tool. While the latter use time series based training data and machine learning to predict MRR. In this paper, we demonstrate to incorporate wear knowledge as classification and show its effectiveness in predicting MRR. Our experiments shows better overall accuracy being achieved through the proposed classification and regression framework.\",\"PeriodicalId\":6790,\"journal\":{\"name\":\"2020 IEEE 22nd Electronics Packaging Technology Conference (EPTC)\",\"volume\":\"3385 1\",\"pages\":\"172-175\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2020-12-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2020 IEEE 22nd Electronics Packaging Technology Conference (EPTC)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/EPTC50525.2020.9315140\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 IEEE 22nd Electronics Packaging Technology Conference (EPTC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/EPTC50525.2020.9315140","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Material Removal Rate Prediction using the Classification-Regression Approach
Chemical Mechanical Polishing (CMP) is one of the most critical process step in the fabrication of advanced packages, such as Fanout Wafer Level Packaging (FOWLP). CMP process requires tight and dynamic control of process parameters to achieve palnarization, high quality and reliability of organic or in-organic redistribution layer (RDL) surface morphology. Typically, physics based or data driven approaches are implied to predict material removal rate (MRR) and run time control. The former models a closed-form expression between domain knowledge and MRR. Often, the domain knowledge are based on kinetics and contact interaction between the wafer, and the polishing tool. While the latter use time series based training data and machine learning to predict MRR. In this paper, we demonstrate to incorporate wear knowledge as classification and show its effectiveness in predicting MRR. Our experiments shows better overall accuracy being achieved through the proposed classification and regression framework.