Hi, everyone, Well, I think I found the cause, and the problem has almost gone. I followed Gareth's suggestion, and tried to reduce the edge bead by leveling SU-8 layer for longer time at room temperature and lower temperature. I got the flatter layer. Prolonger soft-bake time and a little higher temperature drove the solvent to a lower level, even inside the edge bead. The ambient humidity may be another factor causing stiction. I tried to expose the sample right after the sample cool down to near room temperature. I used to let the sample cool down through night after softbake, and the air-conditioning was turned off. The humidity will change between day and night. This time, the samples exposed for 300seconds still had slight stiction. But for the sample exposed for 180 seconds, no stiction was found. The omega UV filter was applied during exposure. So, higher solvent concentration inside edge bead is the cause of stiction. Higher exposure dose will also make the top surface of SU-8 layer melt as more heat accumulats on it. By the way, does the Tg of SU-8 depend on the solvent concentration? OK, thanks again for the suggestions and advices you gave me! 2008/10/24 Gareth Jenkins> You can reduce it by leaving your wafer to relax for about 1 hour > before starting the softbake. > > I have also experimented in the past with a low spin speed and > cleanroom wipe well soaked in acetone but this often went wrong and > made matters worse! > > Best Regards, 郑瑞麟 Ruilin Zheng Address: Pen-Tung Sah MEMS Research Center, Xiamen University, Xiamen,361005, P.R.China