Opportunity for improving fab effectiveness by predictive overall equipment effectiveness (POEE)

The evolution of intelligent fab automation raises the opportunity of improving fab effectiveness such as product yield and machine efficiency by predicting their possible losses and taking proper actions in advance. With the huge amount of real-time data being collected in modern fab, the conventional standards and indexes defined for the yield and the productivity shall be inevitably altered in accordance with more predictable functions. Starting from the calculation for the tool productivity, this paper proposes the Predictive Overall Equipment Effectiveness (POEE) to evaluate and monitor the forthcoming effectiveness of a single tool. The POEE is composed of two parts: i) the deterministic effectiveness such as the preventive maintenance, engineering experiments, and job scheduling and dispatching, and ii) the predictive effectiveness in terms of the extra production time due to the abnormal tool condition and undesired product quality. In the diagnostic perspective, POEE serves as a consolidated index to be inspected and controlled. When the POEE is low, the practitioner can intuitively identify the critical causes beforehand in the three fundamental dimensions: tool availability, tool performance, and product quality. Through scheduling optimization, tool availability will be increased. By looking into the tool performance and product quality, abnormal tool conditions which impact on the process efficiency and product yield can be addressed and resolved. By monitoring the POEE over a set of tools in the same group, the high priority jobs can be dispatched to the tools with high POEE to avoid the yield loss and the long cycle time due to unscheduled machine breakdown. A Chemical Vapor Deposition (CVD) tool is taken as the study vehicle to demonstrate the proposed method in this research.