The rapid pace of device scaling in recent years has outrun the ability of today's generation of semiconductor manufacturing equipment control systems to keep up. New device architectures and the materials and processes used to realize them create sources of variability that require better techniques for real-time sensing, filtering, detection, and response. This is not simply a question of faster data collection of the existing equipment variables. In many cases, the data needed to accurately determine the real-time process conditions are not even available in the equipment, so the traditional time-based techniques and endpoint detection methods for controlling recipe execution are not sufficient… rather, these advanced processes are effectively "flying blind." GLOBALFOUNDRIES, a long-time leader in equipment automation and process control, has addressed these issues by integrating specialty sensors into the broader control environment, effectively closing this gap through a variety of advanced sensorization initiatives that supplement the capabilities of the equipment as delivered by the OEMs. The broad expectations regarding the capabilities of manufacturing equipment and the embedded control systems have recently been described in [1]. Specific examples of advanced sensorization use cases and a standards-based implementation approach have been described in [2]. Now that the measurement gaps for a number of key processes have thus been closed (see Tables 1 and 2), new application opportunities exist for leveraging this enhanced visibility into equipment and process behavior. In particular, the ability to achieve "first-time-right response" to a wide variety of process conditions is now a practical reality. This paper highlights a number of these applications possibilities enabled by "unambiguous signals" on the equipment, and touches on the integration solution used to present these signals seamlessly to the application environment.

Requirements for first-time-right response in advanced manufacturing

The rapid pace of device scaling in recent years has outrun the ability of today's generation of equipment control systems to keep up. New device architectures and the materials and processes used to realize them create sources of variability that require better techniques for real-time sensing, filtering, detection, and response. This is not simply a question of faster data collection of the existing equipment variables. In many cases, the data needed to accurately determine the real-time process conditions are not even available in the equipment, so the traditional time-based techniques and endpoint detection methods for controlling recipe execution are not sufficients rather, these advanced processes are effectively “flying blind.” GLOBALFOUNDRIES, a long-time leader in equipment automation and process control, has addressed these issues by integrating specialty sensors into the broader control environment, and shares highlights of their experience for a number of specific process areas in this presentation. It also covers techniques for leveraging today's data collection standards to streamline the integration process, and concludes with a list of suggestions for future equipment control design that deliver the required capabilities more effectively to support the next stages of process evolution.

Brian Conerny

Papers

Device scaling vs. process control scaling: Advanced sensorization closes the gap — Alan Weber