Thanks!

It's been a lot of help from all the replys I've got.

I think I'll be trying on diamond bits for a start.

Best,
TJ

> Date: Mon, 4 May 2009 11:23:48 -0700
> From: robm@shearwaterscientific.com
> To: mems-talk@memsnet.org
> Subject: [mems-talk] Bearing holes in pyrex/glass wafers
>
> I spent a long time trying to make .5mm holes through a .5mm pyrex wafer
> as an access port for a fluidic device. At the end of the process we
> found that microabrasive powder blasting could be made to work. They
> have a tool for this at the WTC which is dedicated for this purpose.
> The advantage of this is that you can do patterns other than holes.
>
> In my opinion this is a difficult process which is also very common in
> micromachining and in need of a good textbook solution.
>
> Here are some things I tried which did not work:
>
> 1. PECVD silicon etch mask for 49% HF. The PECVD silicon did not etch,
> but it did flake off and contaminate the bath. Not fun to clean up.
>
> 2. Cr/Au etch mask for HF etch: same as above. Material flaked off.
> Also, this material had lots of pin holes.
>
> 3. UV laser cutting. This was just slow. I don't remember the wattage or
> the wavelength, but we didn't have success.
>
> 4. CO2 laser. This caused cracks in the wafer.
>
> 5. Drilling with a dremel and a diamond bit. This worked, but the
> success rate was low, and we lost wafers in the process. At the end of
> the drilling, the bit would break through the backside at times and
> cause fractures in the glass.
>
> 6. Sand blasting with sand that was too large: this caused
> micro-fractures in the glass.
>
> I would recommend micro-abrasive powder blasting.
>
> I believe that with the proper tooling diamond drilling can be made to
> work. I would also like to hear about peoples experience with water-jet.
>
> WTC recently got a new etch tool just for glass etching. My
> understanding is that it will make verticle side walls at fine pitch,
> but the masks only hold up for about 30um of etching. I have not worked
> with this tool myself.
>
> Thanks,
> Rob MacDonald
>