Virtual Metrology Modeling for CVD Film Thickness

Abstract

The semiconductor industry is continuously facing four main challenges in film characterization techniques: accuracy, speed, throughput and flexibility. Virtual Metrology (VM), defined as the prediction of metrology variables using process and wafer state information, is able to successfully address these four challenges. VM is understood as definition and application of predictive and corrective mathematical models to specify metrology outputs (physical measurements). These statistical models are based on metrology data and equipment parameters. The objective of this study is to develop a model predicting the CVD oxide thickness (average) for an IMD (Inter Metal Dielectric) deposition process using FDC data (Fault Detection and Classification) and metrology data. In this paper, two VM models are studied: one based on Partial Least Squares Regression (PLS) and one based on Tree ensembles. We will demonstrate that both models show good predictive strength. Finally, we will highlight that model update is key for ensuring a good model robustness over time and that an indicator of confidence of the predicted values is necessary too if the VM model has to be use on-line in a production environment.

Integration of scheduling and advanced process control in semiconductor manufacturing: review and outlook

“…For example, Lynn et al [36] propose a windowed VM scheme to model the etch depth measurements right after etch process based on Gaussian process regression. Besnard et al [28] use FDC data to predict the Chemical Vapor Deposition (CVD) oxide thickness based on Partial Least Squares Regression (PLSR) and a tree ensemble method. However, a reliable model is very critical and virtual metrology should be trustworthy so that the real measuring activity can be skipped.…”

Classification algorithms for virtual metrology

“…The model was updated as new output measurement was available. [3] studied a virtual metrology model using partial least squares. This model predicts chemical vapor deposition oxide thickness for an Inter Metal Dielectric deposition process.…”

“…This model produces observations that only x 1 , x 3 , and x 4 are useful for classification, and the other variables are noise. Second, we simulated a binary output using a quadratic function of x 1 , x 2 , and x 4 is generated…”

Growing Structure Multiple Model System for Quality Estimation in Manufacturing Processes

“…In the area of CVD, Olsen et al [18] describe updating PLS models with a moving window to predict wafer thicknesses in six chambers, which yielded better accuracy than what was achieved without windowing and updating the model. More recently, Bernard et al [19] compared PLS with a decision tree method and showed comparable accuracy in film thickness VM on a plasma enhanced CVD (PECVD) process. As for VM on etch tools, Lee et al [20] [24] and was used for the same purpose.…”