If I understand it right, your channels were enclosed by the glass and Si
wafers. Only when you dice it, there appears an opening.

I don't know whether it is feasible in your design. But usually before
dicing, we put photoresist to protect the device. Could you fill the
channels with photoresist before dicing? Then you could wash the dust away
with Acetone. When the dusts sit there, my experience is that it is almost
impossible to get rid of them. You can try RCA or even Piranha if they are
compatible with the materials in your device.

Another thought might be to cleave the edges of samples. Maybe the dust
didn't go all the way into the channels. Or, next time you could perform
cleaving instead of dicing. Cleaving is generally much cleaner than dicing
if you don't need to control where to cut that accurately.

Jie

On Fri, May 7, 2010 at 11:38 AM, Miyakawa, Natsuki <
Natsuki.Miyakawa@eads.net> wrote:

> Dear all,
>
> After dicing my bonded wafers (glass with Si) I found out that my
> microchannels covered by glass wafer got partially stuffed up by dicing
> dust. The problem is that I have openings on the cut face so there is no
> possiblity to protect them from cooling wafer with dice dust.
>
> My question is, is there any method to removed the trapped dicing dust
> from microchannels covered by bonded chip? I tried ultrasonic-bath with
> alcohol, but it dind't help much.
>
> Thank you!
>
> Regards
>
> Natsuki

*  Zou Jie (Jay)
*  Department of Physics
*  University of Florida
*  Tel: +1-352-846-8018
*  Email: zoujiepku@gmail.com
*  Homepage: http://plaza.ufl.edu/zoujie/