Hello Jingru, There are many things that can cause the cracking. What do the cracks look like? Are they round cracks or half moon cracks? What type of bonder are you using, homemade hotplate and power supply or a commercial chamber bonder? Are you bonding in vacuum or atmosphere? If you are seeing half moon cracks or round chunks of glass popping out then you have air (or gas) getting trapped between two bond fronts. The gas will be compressed until it cracks the glass and vents. If you have long cracks the follow your channels then the trapped gas in the channels may be cracking the glass, I've never seen this case but it could be an issue. Try this: 1. If you are using a chamber bonder make sure it is setup properly and the upper and lower electrodes are clean and flat. Metal electrodes tend to warp and pit over time. If this is the case then it will cause poor contact with the bond stack and allow air or gas to become trapped. SiC electrodes will become contaminated with Na after several bonds. Give them a good clean in RCA over night and the conductivity should be back to new and uniform across the chuck. 2. Try bonding in a vacuum or at a lower pressure. If you bond at a low pressure make sure it is below 1e-2mbar or you risk the chance of generating a plasma in the chamber. You'll know if you have a plasma because the current will stop dropping and the voltage will stop increasing and stay and a constant state. 3. If you cannot bond in a vacuum and trapped gas is the issue the try bonding at a lower current. This will slow the bond front and may help to allow the gas to vent before it gets trapped in the bond. 4. If the cracking is general cracks across the wafer check your ramping and cooling times. Let the wafer stack slowly heat up then let them set at the bonding temp to come to a nice uniform temp before bonding. Any time I set up a first time bond I let the wafers ramp up to 350C over a 30 minute time. Then I will heat to 400 and hold for 15 minutes to allow the tool and the wafer stack to stabilize. After bonding I remove any force on the stack and cool the wafers slowly at about 100C over a 30 minute time. If you have the ability to apply force do not heat or cool the stack with force, this can not only cause a shift in the wafer stack but also cause cracking.. This sounds like a very long recipe but once you have a working process you can optimize it for throughput. Good Luck Brad Johnson Application Engineer DJK Global US Distributor, Semiconductor Inspection Systems 2447 W. 12th St. - Suite 6, Tempe, AZ 85281 480-968-3343 Ext 112 office 602-501-4413 cell bjohnson@djksemi.com http://www.djksemi.com -----Original Message----- From: jingru zhang [mailto:ruharvard@gmail.com] Sent: Thursday, February 26, 2009 9:12 AM To: mems-talk@memsnet.org Subject: [mems-talk] glass cracking during anodic bonding Dear all, I have deep channels (around a hundred microns) on the silicon substrate. and during the anodic bonding of silicon and glass, they bonded well but the glass cracked. Maybe it is the stress around the channels walls which caused cracking. The voltage I am using now is 500 w, and the temperature goes to 400 deg c. Does anyone have similar experience? How to avoid this? Thank you very much. Best Regards, Jingru