durusmail: mems-talk: Bonding and CMOS
Bonding and CMOS
1995-08-03
Bonding and CMOS
Wood
1995-08-03
This list seems a bit quiet compared to many others, so how about a few
opinions on the following subject?

There's been a discussion in our lab recently about the bonding process and
CMOS compatibility. The question is: given that some MEMS processes use wafer
bonding, can you then make CMOS circuits in the same process flow? Can you have
CMOS circuits and a MEMS device on the same wafer if the latter requires
bonding, particularly if you are using a wafer thinning procedure after
bonding?

There are three ways you could approach the CMOS/bonding technology. The
insertion of the MEMS manufacture could be before/after bonding:

1. Make the CMOS circuits, then protect them and bond the wafers. Of course,
this would not allow you to thin the wafers.

2. As in 1, but thin the wafers after CMOS (as already happens to, say, GaAs
MMIC devices) and before bonding.

2. Bond the wafers, thin as necessary and then make the circuits.

This latter approach seems to be available via BESOI wafers. What about any
insurmountable problems arising the other processes? WE can produce successful
fusion bonds at as low as 200 C, so I don't think temperature is a big issue.
Anything else?

Are bonding and CMOS compatible? All opinions, with appropriate technical
arguments, gratefully received!

Cheers,

David Wood.


reply