Equipment Productivity Enhancement by SMC (Statistical Machine Control) Application on Furnace Area

摘要

In this work, we clearly address the strategies of monitor and via Statistic Machine Control (SMC) on furnace process at ProMOS technology, DRAM manufactory, including low pressure and atmosphere pressure process. Especially for LPCVD Nitride deposition process, particle and condense in the piping which was formed by by-product, NH{sub}4Cl, is an undesired factor to affect process stability. This work have demonstrated four strategies of Statistic Machine Control (SMC) models, (A) VG2 clog (B) SV clog (C) VV1 clog (D) FCV angle control, to predict the timing and location of by-product clog at LPCVD furnace nitride processes and warm us when the by-produce condense had generated and the clog was accrued. Further more, to predict the schedule to clean the powder by short corrective action. It was found that the counts of leak check fail were obviously getting improvement for different nitride process equipments, the overall improvement of leak check fail average count is 15.8 time/month before May to below 4 time/month. The impacted lots also were reduced 79.3% which equals to 2.6 K wafers move gaining per month. With the reduction of tool down time and maintenance, the productivity can be increased and manpower and cost was saved.