What describe ( divot under the bond bad) is quiet a common problem.
Your process is crushing the silicon under the pad. Ultrasonic wire
bonders see this problem if not set up correctly. Try reducing the force
at contact. Alternatively you may be seeing a problem caused by
differential expansion / contraction problems. Have you though of using
a ribbon based wirebonder? A gold metallic wirebond can tolerate 500 deg
C.

Regards,

Steve Adamson
Market Manager
Asymtek
760-930-7274 office
760-802-1641 Cell


-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Steven Hubbard
Sent: Wednesday, April 12, 2006 8:47 AM
To: [email protected]
Subject: [mems-talk] microbrazing on Si

Hello,  I am developing a MEMS sensor for harsh environments --
specifically temperatures above 500C.  I'm looking for some resources to
assist with our electrical connections.  Currently, we are using
microbrazing to attach a Au-kovar ribbon to a pad.  The pad is 2 microns
Ni with a half micron of Au (with a Ti adhesion layer).  These pads are
sputter-deposited.  The microbrazing process bonds the wire well to the
pad, but the pad separates from the die often leaving a divot in the
underlying silicon.  I am curious if anyone has had similar situations
and could suggest some alternate metallurgies or processes that we could
explore.

We're investigating electroplated pads now as a way to increase the
thickness (and hopefully remove some heat from our Si).

Steven