Hi Rick. Thanks a lot for your comments. The underside is gold in color indeed. The delamination occurs everywhere and starts from the edges. All the deposition was done in vacuum and I used the Filament Evaporation for the deposition. I will repeat my process very carefully and check the photoresist issue. But if it's the photoresist, I cannot explain why this didn't happen in the only Au or only Cr case. Do you think the oxide created by O2 plasma may hinder the adhesion although we have a Cr layer for adhesion? Thanks again. All the best, Jay On Sun, Jun 15, 2008 at 7:35 PM, Rick Willistonwrote: > Jie Zou > > You didn't mention where the delamination is occuring or details about your > deposition technique. If you break vacuum between the first Cr layer and the > Au deposition, the adhesion could be severely impaired. Gold sticks well to > most metals but not too well to oxides. Normally I would say look at the > remaining layer and tell us the colour, but with 3nm this amount of Cr could > be readily oxidized and be transparent, so I would look at the underside of > the peeled metal and see if it's gold in colour or not. > > Even if you don't break vacuum, if there is a combination of long time > between depositions, higher than normal pressure in vacuo or higher than > normal outgassing of your gold target you could cause some semblance of the > oxidation required (or other contamination). > A lot of these same issues exist if the delamination can be shown to be > under the bottom layer of Cr. But in this case the most common candidate is > the lithography process, it may be inadequate and may have to be > redeveloped. A thin organic layer of resist could be left if you have the > wrong combination of either expired resist, improperly stored resist, > uncalibrated or inadequate exposure, unclean or incorrect mask material, > expired or improperly stored developer, chemicals, DI water, etc. or the > part is left too long in air or inadequate storage after developing and > before deposition. > > As you can see, there are lots of ways to go wrong so you have to make sure > all of the logical candidates are ruled out and if you find any areas for > improvement, then correct them and try again. A good start would be to > consult the resist supplier and they can tell you what needs to be done > correctly for their material to work properly. They are particularly > sensitive to this as they frequently get blamed for a lot of process miscues > so they have to have the answers ready. > > Good Luck, > Rick Williston