Arjan-

I would pursue the ultrasonic techniques (e.g. Bullen Ultrasonics,
www.bullen-ultrasonics.com) to see if they can accomplish your specs, since
I don't think your 2 approaches will work.  My experience with laser
drilling (in silicon) is that it is not even as good as Deep RIE can
accomplish.  The edges of the laser machined holes were very rough - one
company to check would be Potomac (www.potomac-laser.com) who may have a
better answer.  As for the SOG (Spin-On-Glass), it would be very difficult
to get anything resembling a machined glass wafer from a DRIE Si mold as
thick as 200µm - during the curing phase, the glass will form lots of
cracks, and I doubt you will get your 0.1µm accuracy.  A former Ph.D.
student in Prof. Mehregany's group at CWRU (A. Azzam Yasseen) did thick
spin-on-glass experiments - a reference to his work is below:

Thick glass film technology for polysilicon surface micromachining
Azzam Yasseen, A.; Cawley, J.D.; Mehregany, M.
Microelectromechanical Systems, Journal of , Volume: 8 Issue: 2 , June 1999

Page(s): 172 -179.

Good luck!

-Kevin

Kevin C. Stark, Ph.D.
Senior Program Manager
NineSigma, Inc.
21945 Chagrin Blvd.
Cleveland, OH  44122
216-295-4806
216-295-4825 (fax)
[email protected]
www.ninesigma.com