If I remember my thesis work correctly, you can get fracture strength
(of Si) bonding between the wafers using argon or nitrogen plasmas, or
oxygen plasma + post plasma RCA1 treatment. Annealing at temperatures
between ~150C and 300C in air for an hour or so maximizes the wafer to
wafer bond strength. I believe it is easier to get high bonding
strength if you bond a thermally oxidized wafer to a bare (HF dipped)
Si wafer. You can  look up papers by S.S. Lau, Nastasi, M. Cai, K.
Henttinen, T. Suni, or  F. Lu  (not to mention U. Gosele) in the time
period between ~1998-2006 for more process details. I can also send
you my thesis if you would like more details.

Felix


On Jun 25, 2009, at 12:20 PM, Chilcott, Dan - NV wrote:

> Hafizah,
>
> I agree with the plasma treatment approach. Oxygen plasma can also
> be used. Two things to consider. First, the bond area needs to be
> flat so you will need to CMP if you have metal runners under the
> bond surface. Second, the optimum plasma process window for oxygen
> plasma SiO2 bonding is more narrow than for bare silicon. You may
> need to experiment a little with process power and time.
>
> Dan Chilcott

Felix Lu
[email protected]