In this paper, an end-of-line quality control scheme based on wafer accepta
nce test (WAT) data is presented, Due to the multiple-stream and sequence-d
isorder effects typically present in the WAT data, an abnormal process shif
t caused by one machine at an in-line step may become vague for detection u
sing end-of-line WAT data. A methodology for generating robust design param
eters for the simultaneous application of Shewhart and EWMA control charts
to WAT data is proposed. This SHEWMA scheme is implemented in a foundry env
ironment and its detection and diagnosis-enhancing capabilities are validat
ed using both numerical derivations and fab data. Results show that the SHE
WMA scheme is superior to the current practices in detection speed. Its use
is complementary to the existing in-line SPC for process integration.