REAL-TIME, IN-SITE METROLOGY TO DRIVE REAL-TIME APC

摘要

Real-time, in-situ process sensing already claims wide-spread applications for fault detection as part of advancedrnprocess control (APC) in semiconductor manufacturing. However, it is yet to be implemented to drive real-time courserncorrection. Instead, to date, course correction has been mainly focused on run-to-run control based on in-line, ex-siturnmetrologies. We have developed accurate real-time, in-situ metrologies based on three different sensing techniques (massrnspectrometry, acoustic sensing and FTIR) to drive real-time course correction in W CVD processes. Concurrent use of allrnthree sensing techniques for the same batch of deposition processes at 10 Torr has all resulted in film thicknessrnmetrologies with accuracies of the order 1% or better (in terms of average uncertainty). In addition, mass spec basedrnmetrology was further implemented to drive real-time end point control of film thickness in 0.1 Torr SiH4/WF6 process. Thernsuccessful demonstrations of accurate sensor-based metrologies and real-time end point control provide for an excellentrnprognosis in adopting real-time, in-situ sensing technology to support both the fault management as well as courserncorrection aspects of APC in semiconductor manufacturing.