Al/Au contact

Arif Sanli Ergun sanli at stanford.edu
Fri May 13 12:12:39 PDT 2005


Hi all,
Thanks a lot for the responses. I didn't know this was
such a popular subject. Here is a summary of the responses:

1. It is not a good idea to contact Au and Al directly, because
Al/Au alloy, also known as purple plague, is nasty, and high
resistive. I was aware of this and therefore was putting
Ti/Pt on top of Al before putting Au.

2. It is possible to wire-bond to Al pads both with Au and Al wires,
Al being the preferred choice. To do this you have to scratch the Al
with the tip of the wire bonder. This works for small volumes, but I have
about 400 bonds per device. In addition to the time consumed in making
400 scratches and wire bonds, I need the wire bonds to be all working
and reliable. That's why I'm trying to put Au on the pads.

3. The Ti/Pt/Au layer peels of while trying to wire bond. That is, Ti is
not sticking to Al. In fact, I'm depositing Al in gryphon, and evaporating
Ti/Pt/Au in Ginzton which means there is actualy Aluminum oxide on the
surface, not Al. So, the problem seems to be sticktion between Aluminum
oxide and Ti. Monty says Ti sticks by reducing native oxide, which happens
at an elevated temperature, for example at 400 C. Because I have gold and
because of stress issues I cannot do a heat treatment. Then, that is
probably why Ti is not sticking to Aluminum oxide. The solution I thought
is to deposit a thin layer of Ti on top of Al in gryphon without breaking
the vacuum, so that I don't have Ti/Alumina interface. When I break the
vacuum, Ti will immediately oxidize. Then, I will evaporate Ti/Pt/Au as
before. Assuming that Ti/TiOx sticktion is good (which shouldn't be a wrong
assumption) I won't have the peeling problem while wire-bonding.

Thanks again for the help,

Sanli




More information about the labmembers mailing list