胶粘剂
薄脆饼
图层(电子)
材料科学
晶圆规模集成
计算机科学
复合材料
光电子学
作者
Vasarla Nagendra Sekhar,Hongmiao Ji,Shinji Arimoto,Toru Okazawa,Takenori Fujiwara,Masaya Kawano
标识
DOI:10.1109/eptc.2017.8277548
摘要
Present study focuses on room temperature mechanical peeling temporary bonding and debonding (TBDB) method using single layer adhesive material, mainly to reduce overall cost of the process. This method is compatible with standard semiconductor equipment and it can be readily adopted into existing fabrication lines. In addition, this method doesn't requires any preparation for debonding and it occurs at room temperature only. Main objectives of this work is to qualify an adhesive material in glue coater, temporary bonder, debodner and cleaning tools. Whole TBDB process is developed using 300 mm wafer size and process lines. For coating process optimization, a spin curve has been generated and selected the 20um thickness coating with lower total thickness variation. For temporary bonding optimization extensive design of experiments (DOE) have been conducted and parameters considered for this evaluation include, pre-bonding curing, bonding temperature, bonding loads and post bond curing steps. Additional curing steps have been employed, mainly to improve the cross linking and stability of the bonded material. Bonding quality of the bonded pairs has been assessed by employing thru-scan analysis. After bonding process, wafers have been subjected to two step curing process at 220C and 250C, to improve the cross-linking of the adhesive material. In back grinding process evaluation, device wafer has been thinned down to 30um and overall integrity of the bonded pair is good. For fabrication process evaluation, chemical vapor deposition (CVD) step has been selected, as it is most sensitive process step for bonded wafers in terms of thermo-mechanical stresses. TEOS oxide deposition of 2um in thickness has been carried out on bonded pairs. Both thru-scan analysis and visual inspection has been carried out to verify the bonding and stability of the bonded wafers is good. So far this material has shown promising results and it will be implemented in full integration.
科研通智能强力驱动
Strongly Powered by AbleSci AI