Dear Kai Hiltmann,

a solution for feedthrough from the glass to the Pyrex wafer with andic
bonding is described in the Ph.D. thesis of Rob van Kampen, Technical
University of Delft, 1995 (ISBN 90 - 407 - 1175 - 5), on pages 181-183. He
uses press-on contact of aluminium. One 0.6 µm contact is placed in a 1
µm-deep recess in the Pyrex, whereas the other 0.6 µm aluminium conductor is
placed on the oxide surface of the silicon wafer. During bonding, the 1.2 µm
aluminium "srack" is compressed to 1 µm thickness, and a contact resistance
of 0.5-23 Ohm is obtained, depending on the contact size.

The other problem is getting through a hermetically sealed region (lateral
direction). You can consider buried conductors in the silicon wafer, so
there is a planar region on the wafer for bonding the Pyrex glass. For a
description of this technique see: W. Henrion et al., Proc. Solid-State
Sensor and Actuator Workshop, Hilton Head Island, S.C., 1990, pp. 153-157
(note figure 2).

Yours Sincerely,
Patrick Scheeper