Quality improvement at a semiconductor equipment manufacturing facility through error re-categorization and proper inventory management

Abstract

This thesis addresses the improvement of first pass yield on the assembly floor of Varian Semiconductor Equipment and Associates (Varian) - a low volume complex semiconductor capital equipment manufacturing facility. First pass yield refers to the proportion of fully built modules that pass testing without the need for additional rework. The first pass yield (FPY) project began in 2011 and showed steady improvement for its first three years. But over the following two years, the primary yield metric has stayed level. The goal of the work presented here is to analyze the reasons for the leveling and propose novel ways of improving the metric once more. Two major quality improvement recommendations are presented in this thesis. The first is a new way of categorizing error reports that will lead to targeted corrective action to reduce the recurrence of pre-identified failure modes. A multi-step methodology is presented on how to determine an efficient corrective action, along with a case study based on data acquired from Varian. The second quality improvement recommendation involves reducing the number of part and sub-assembly shortages on the assembly floor that were shown to be correlated with poor first pass yield. A new management system for a subset of the total inventory at Varian is presented, along with a discussion on how to execute the recommendation.