Abstract:

The development is still prosperous in semiconductor manufacturing industry in recent years. The wafer size enlarges from 200 mm to 300 mm , and the factory automation is getting more and more important. Since machines are vital resources in the factory automation, the effective monitoring of the machine statuses and the good diagnosis and maintenance analysis are helpful to make the operation processes stable. Through internet, machine statuses in the clean room of IC foundry can be monitored by users or engineers remotely.

The purpose of this paper is aimed at the architecture and development for remote diagnosis and maintenance system of a cluster tool. A statistical process control (SPC) and run-by-run module are used to detect and eliminate the process variations. A diagnosis module uses a neural network and Internet Interactive Case-Based Reasoning (IICBR) to predict and diagnose a cluster tool separately. A maintenance module is supplied to forecast maintenance time and choose an adequate policy. All significant information can be notified and interacted via GMPP and web server. Hence, a three-tiered architecture is developed for a cluster tool. All modules are integrated to construct the remote diagnosis and maintenance system.