Hi Florian,

If the layer is relatively defect-free, try the SC1 cleaning step. You
could even also try Piranha as a pre-step, as the H2SO4 is unlikely to
etch the oxide. If the steps are to 'frothy', just increase the
proportion of DI in the mixtures.

Regards,

Michael Larsson


> ---------- Forwarded message ----------
> From: Florian Felderer 
> To: General MEMS discussion 
> Date: Wed, 09 Apr 2008 14:48:14 +0200
> Subject: Re: [mems-talk] Pirahna attack of TiO2
> Hello Kagan
>
> The TiO2 layer is made by Atomic Layer Deposition on a PE-CVD-Silicondioxide
> using an aluminum oxide buffer. So there shouldn´t be a lot of defects
>
> Regards
> Florian
>
> Michael Larsson schrieb:
> >  Hi Kagan,
> >
> > My suggestion was to use SC1, rather than Piranha to wet-clean the TiO2
> > layer. I have no practical experience to back-up my comments unfortunately,
> > but I am unaware of any wet etchants capable of high etch rates in Ti that
> > don't either include HF or -Cl groups. With regard to TiO2, this is
> > the product formed when Ti oxidises under galvanic attack. Once grown, the
> > oxide acts as a barrier, inhibiting further chemical attack. The ability to
> > protect underlying Ti from oxidation can obviously be reduced when
> > considering a deposited layer (with defects) relative to pure, bulk
> > material. Nevertheless, theory would suggest that an SC1 clean
> > (NH4OH:H2O2:H20) should not attack Ti, and certainly not it's chemically
> > inert oxide; even if this is indeed formed during the SC1 clean.
> >
> > Florian: is the layer sputtered or evaporated?
> >