> [mailto:[email protected]] On Behalf Of g.balsubra manian
> Sent: dinsdag 7 maart 2006 16:22

> generally in my case after etching in Bufferd HF.,it creates
> a surface roughness which is usually few Angstroms (100 - 120
> A) for etching 1500 A film.

Buffered HF is one of the candidates. However, its etchrate is rather
low,
about 0.6nm/min for stoichiometric nitride (Si3N4) and even lower for
SiN
films that contain more Si. If time is an issue, phosphoric acid might a

better choice because it etches faster (can't find the rate right now)
and
is more selective with respect to other SiOxNy compounds. However, the
liquid
needs to be heated to approx 155 degrees Centigrade, a setup that may
not be
readily available in every fab.

(more info on phosphoric acid:
http://snf.stanford.edu/Equipment/wbgeneral/NitrideEtchWBGen.html)

I'm not too sure about the 100-120A increase in roughtness (is that the
RMS value?)
In my experience, etching about 200nm Si3N4 in BHF increases the
roughness from
0.2nm RMS to 1.5nm RMS.

> So, i dont think etching by bufffered HF creates rough
> surface thereby the light exposure may get scattered.

Note that a roughness >1nm RMS is generally unsuited for waferbonding
(Si fusion
bonding). I doubt that this roughness will influence the lithography but
if bonding
is performed later on, probalby a polishing step of the surface is
needed.

> You can try with agitation to get more of a smooth surface.I
> tried this before and got a smoother surface

Oh, that's something I must try~ :)

>(u can read the smoothness by surface profilometer)

or AFM.

Cheers,
Geert.
--
Geert Altena  -<=>- Integrated Optical
MicroSystems
\nabla \times H = J + \frac{\partial D}{\partial t}\\ \nabla \cdot D =
\rho
\nabla \times E = - \frac{\partial B}{\partial t}  \\ \nabla \cdot B = 0