Dear all,

I do strongly second Marc's explanation. Those of you who've been to
the MRC sputter school will have seen SEM pictures of this effect.

I personally tried to sputter 5000A of gold recently onto thermal oxide
wafers of controlled temperature. Even with a film this thick,
coalescence (de-wetting) starts as low as 200 C and the film is
non-conductive at 300 C, or peels off right away

The only way is to use a wetting/adhesion layer. I never tried
annealing at 700C , but Au diffusion into oxide definitely is
a concern, so your wetting layer should also be a good diffusion
barrier.

hope this helps
klaus


             (TEEL)        Tokyo Electron Europe Limited
                     PVD Process Support  (ex MRC)
Klaus Beschorner                   Tel +49-7033-45683
Drosselweg 6                       Fax +49-7033-45631
71120 Grafenau, Germany                  Mobile +49-174 315 7754



 >Allow me to add to my original answer.  I am quite sure this is what is
 >happening, though I am a little surprised that it would occur for a >1000
 >angstrom film.  Gurvinderjit- are you sure of this thickness?
 >
 >I used to make a form of tunneling sensor just this way- depositing
 >300-500
 >angstrom gold films on a glass slide, then heating the suface until the
 >uniform gold film broke apart and migrated to form discrete but closely
 >spaced "islands".  I then filled the space in between with one of >several
 >organic semiconducting films to enhance the tunneling efficiency as >they
 >were exposed to various VOC's.
 >
 >Hope this helps.
 >
 >marc
 >
 >
 > -----Original Message-----
 > From: Marc Straub [mailto:marc.straub@coventor.com]
 > Sent: Monday, May 20, 2002 10:17 AM
 > To: mems-talk@memsnet.org
 > Subject: RE: mems-talk digest, Vol 1 #291 - 7 msgs
 >
 >
 > Gurvinderjit,
 >
 > I suspect that your film is "beading" like water on a waxed
 > surface and forming into discrete gold islands.  Gold does not
 > wet or adhere to oxides well, particularly in very thin layers.
 > Heating the surface gives the Au atoms the energy they need to migrate.
 >
 > Depending upon your application, you should try a thicker layer
 > and/or flashes (thin layers) of Cr/Cu, Ti or Ti/W under the Au
 > for better adhesion.  Good luck.
 >
 > --
 > Marc Straub
 > Staff Engineer, Product Development Services
 > Coventor, Inc.  (Formerly Microcosm Technologies)
 > 7 Corporate Park, Suite 260
 > Irvine, CA  92606
 > Ph: 949-756-0033  Fax: 949-756-0070
 > marc.straub@XXXXcoventor.com [remove the "XXXX"]
 >
 >

 > > -----Original Message-----
 > > Date: Mon, 20 May 2002 11:20:15 +0530 (IST)
 > > From: Gurvinderjit Singh 
 > > To: mems-talk@memsnet.org
 > > Subject: [mems-talk] Gold film
 > > Reply-To: mems-talk@memsnet.org
 > >
 > > Dear friend i have one query.
 > > I am depositing gold film (thickness 0.1 micron)by thermal

 > evaporation on

 > > oxides,
 > > but when  i am annealing my sample to 700K this gold film become non
 > > -conductive may be  duo to formation of gold oxide. But from
 > > literature i found
 > > that gold should not become oxide till 1200K. I also read people using
 > > gold electrodes (thin film by thermal evaporation) on their samples and
 > > work till 900k kindly let me know where i am doing mistake,

 > wether to take

 > > more thick film or something else.