Abstract

Abstract This paper deals with missing and non measured properties for a Chemical Mechanical Planarization (CMP) process in semiconductor manufacturing. A virtual metrology (VM) module is built to estimate non measured properties using a new modified Just-In-Time Learning approach (JITL). The estimated data are integrated thereafter to ensure the smooth functioning of the Run-to-Run (R2R) control loop in order to improve product quality. In collaboration with our industrial partner ST Microelectronics Rousset, the efficiency of the elaborated approach is illustrated using data-set from a CMP process. Then, a comparison is made between estimation quality obtained with the classical and the modified version of JITL approach. The use of estimated data in control context is also presented. For this, another comparison is made for industrial R2R performances using real and estimated data.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call