Natsuki,

Small scale surface roughness can be a tricky problem. The best way to
characterize a surface is through three parameter roughness (Correlation
length, fractal dimension, and RMS). The affects that you are seeing are
most likely the result of measurement artifacts from profilometry. If you
have an alternative technique (XRR, AFM, etc.) you would be able to get more
insight into the problem that you are having.

There are SEMI and ANSI standards for surface roughness measurement, but
they cost money.

Thanks,
Zak

-----Original Message-----
From: antwi nimo [mailto:[email protected]]
Sent: Friday, July 31, 2009 7:58 AM
To: General MEMS discussion
Subject: Re: [mems-talk] Surface roughness measurement of Si

There are tons of Si surfaces u can measure...was it a polished silicon
surface...anodized or what...and there are a number of ways u can use the
find the average surface roughness..
AFM is one such example....

With your profilometer, this should not be a problem at all...it is just
specification problem....and for you to know very well what machine you are
using.