We present a systematic approach for converting a legacy wafer fab from manual wafer handling to fully automatic wafer handling. Our strategy began by quantifying the need for automation in terms of impact on die yield, identifying a seven percent die loss associated with scratches from wafer handling. We then addressed the fundamental changes in production equipment and processes as well as overall fab goals and attitudes that are required to achieve full wafer handling automation. After considering several approaches to staged fab automation, we selected an approach which eliminated all manual handling within specific fab modules, completing the automation within one group of modules before embarking on another module set. In this way, we limited both the initial scope and cost of the project while preparing to leverage its initial successes. This paper summarizes the methodology and metrics found useful for preparing the fab for change, executing the change, and successfully managing the overall project.

No abstract

No abstract

The earliest attempts by human inspectors to classify defects found during in-line inspection of integrated circuits were fraught with difficulties in clarifying defect definitions and in training a diverse and changing inspector staff. These deficiencies were exacerbated by the challenges of expanding classification categories as new defects were discovered. Our diversified product mix had accumulated a knowledge base of approximately seventy defect types, posing a formidable learning challenge for even the most knowledgeable inspector. Not surprisingly, the average accuracy of the group in classifying defects was -55%, and even the best inspector scored around 70%. To address these issues, we developed a comprehensive methodology for classifying defects. This methodology includes both word descriptions of the physical appearance of defects and a hierachical questionnaire leading to precise defect classification. After adopting this methodology and implementing strong training programs, our team significantly improved its defect review process, ultimately reaching 8O% classification accuracy. With this degree of accuracy, we were able to implement defect specific statistical process control (SPC) charts, together with formalized "decision tree" procedures for correcting defect excursions. These formalisms then became an effective part of the fab's yield improvement program. Today, as technology advances into the realm of automatic defect classification (ADC), the lessons learned from human defect inspection form a strong foundation by establishing a comprehensive set of defect categories uniquely related to causality and supporting defect identification standards that can be used by the entire community ofADC training engineers.

When yield analysis revealed extensive die losses associated with wafer scratches, our fab management commissioned a comprehensive program to completely eliminate manual handling of wafers during manufacturing. This experience constituted a true cultural change for our legacy fab, which throughout a 15-year history had excelled at low cost, low cycle time manufacturing but had neglected fundamental improvements in wafer handing automation. This paper describes a team approach to quantifying the components of scratch associated yield loss, planning for remedial actions, and completing the task. Key ingredients for success included achieving buy-in at all levels of fab operations, effectively mobilizing fab resources and developing a group of indices to measure progress. As a result of these actions, our phase-1 program eliminated all manual wafer handling in our photolithography and contiguous plasma etch areas within a seven month period, reducing scratches on resist coated wafers by 83% and setting the stage for beginning additional phases of work to extend the automation to the rest of the fab.