45035

Poor Titanium Tungsten adhesion on Alumina after Nickel electroplate [California] 

May 7, 2007

I have an issue regarding adhesion lost after a Nickel Chloride Sulfamate plate.

Our substrate is 99.6% Alumina with TiW seed layer. We sputter and plate gold on the TiW which has good adhesion.

However, I would try to deposit Ni on top of this stack and the adhesion is lost at the Alumina / TiW interface. The Ni to Au adhesion is still good.
Nickel plating occurs after a HCl clean/rinse. Woods Ni strike at 50ASF and Ni sulfamate plate at 25ASF.

Adhesion lost can be noticed right after plating with a scratch / tape test.

To me this seems as a stress related issue.
Does anyone have any experience as to what might be causing this adhesion lost.
Does it seem reasonable that a Ni plate induces a stressful layer that would cause the entire metal stack to peel off of the Alumina?

Thanks,

Peter Tran
Process Engineer - San Jose, CA

My first bet is the chlorides are finding their way thur the gold and into the alumina which is very sensitive to them.

Guillermo Marrufo
Monterrey, NL, Mexico

Trying to electroplate any thing to Si/Ti/W is very difficult, so a poor sputter adhesion is not shocking.
Normally, in plating , it would indicate that the surface has not been properly activated. It is probably something similar for sputtering.

James Watts
- FL

I also think the problem is that the chloride ions are getting through the gold layer and attacking the substrate. However, I do not know how it is doing so from a corrosion aspect. I can only surmise that the chloride is attacking the TiW in some way, because to attack alumina would be very difficult, especially in the concentrations normally found in hydrochloric acid cleaners. I would suggest a few things; firstly try using a chloride-free acid cleaner.
Secondly, try using a sulphamate nickel strike as opposed to Woods nickel - this should further reduce the risk of chloride attack during plating.
Thirdly, increase the thickness of the gold to a couple of microns - this will be expensive, but will minimise its porosity and hence minimise the migration of chloride through the layer.
Finally, try putting down an electroless nickel instead of using Woods nickel. Then you will get your nickel layer without the chloride problem.

Trevor Crichton R&D practical scientist - UK

Thanks for the response.

I have done some trials and have noticed that the TiW adhesion is lost only with the nickel plating on top. Only the selected areas with the Ni would peel from the TiW/alumina interface.

I have tried a NiBr sulfamate bath and noticed similar issues.

One thing to note is that if I strip off the Ni, the underlying TiW interface does not peel off.

This to me proposes the idea of a mechanical issue where the hardness and thickness of the Ni makes it easy to peel off the entire metal stack.

I might try increasing the TiW thickness to see if this helps or not. Anyone else with any other suggestion would help.

Thanks,

Peter Tran
- San Jose, CA, USA

Have you try sputtering gold at a high bias voltage, about 70 angstron, then your TiW, then your gold. If you have reactive capability I would try the NITRIDES of Ti,W.

good luck,

Hamilton M. Solidum
- Mays Landing, NJ

I would guess you have a mechanical problem, too much compressive stress in the coating. The shape of the parts could tell you. If it peels more readily from concave surfaces than from convex, it is most likely a stress problem.

Jim Treglio
- Gardena, California