Online information source for semiconductor professionals

Product Briefing Outline: A real-time, visual predictive and preventive maintenance (PPM) system that could save semiconductor fabs thousands of dollars annually per tool has been developed by Quality Wise Knowledge Solutions (QWiKS) in cooperation with ATDF, the Austin-based R&D foundry. The PPM system, called QWiKS CBM/PHM, combines condition-based maintenance (CBM) with prognostic health monitoring (PHM) to monitor and maintain semiconductor process equipment more effectively.

Problem: According to market research by International SEMATECH Manufacturing Initiative (ISMI), 17 percent of a typical fab's tool asset value must be allocated to maintenance to support round-the-clock operations – so that a factory with $4 billion in equipment assets spends about $680 million per year on regular tool maintenance. The price goes even higher for unscheduled downtime, which costs four to seven times more than routine maintenance.

Solution: PPM technology helps a production fab reduce unscheduled downtime by
providing early warning of equipment failure. For example, ATDF uses the CBM portion of the system to monitor furnace throttle valve positions and alert manufacturing technicians when it is time to clean the furnace pump exhaust trap. This usage avoids unscheduled downtime due to process accumulation. The QWiKS system also includes tools for managing and improving tools’ performance tracking, which include
real-time tool utilization charts, a preventive maintenance module to link schedules, procedures and history to a single module, and EKB (equipment knowledge base)/tool repair knowledge base to archive history and knowledge for future repairs. An escalation module is available to automate the escalation process when tool downs become critical to the production process. Also included is a searchable tool maintenance comments facility that allows users to find repair data as needed.  The system also includes a failure mode effects analysis (FMEA) module to allow both process and equipment FMEAs to be implemented to ensure that the root causes of a failure are understood, addressed and corrected. A powerful search function within the module allows users to search for archive FMEA studies using keywords or a risk priority number (RPN). Also, an FMEA pattern signature function can capture the tool and process parameters associated with the root causes of a failure. Equipment or process diagnostic rules can be generated for detecting early sign of tool or process failures, or pinpointing the root causes of tool failure to reduce equipment repair time.

Applications: R&D to large-scale fabs.

Platform: Patented technology uses an open architecture, based on ETEST (Wafer Electrical Test) applications and uses an ‘Excel’-style spreadsheet environment that allows domain experts to integrate their knowledge into the system without having to write code.

Availability: November 2007 onwards.