From: putnam@msc.cornell.edu (Albert M. Putnam)
Newsgroups: sci.materials
Subject: Re: Smooth thin films?
Date: 26 Oct 1994 05:51:59 GMT
Organization: Cornell Materials Science Center

In 1988 we were trying to make smooth gold films 
(as were requested in the original post)
for use as substates in the STM imaging of biological specimens.

We wanted surface roughness of below 10 angstoms over areas 1000x1000A.
The whole business is VERY tricky.
You have to stike a happy medium between "epitaxy" and "random laydown".
We embarked on a program of plating gold on mica.
The coordination of gold and mica is fairly good and we hoped
we could build up mono-layers by SLOW hot evaporation.
That does not work. Islands "bead up".

Likewise fast evaporation on a cold surface does not work
you get a simple square-root of N variation in surface height where
N is the number of monolayers you put down (where they hit they stick).

Gold films that are too thin (<100A?) will often not conduct.

To make months of research a short story, you have to balance
evaporation speed (sputtering is thought to be hopeless - due to the energy/momentum 
distribution of the arriving particles - might work on another system?), 
the substrate (mica) temperature (as a function of time - 
you need to consider annealling time),
vacuum, surface preparation of the mica, and so on.
Phase space for this problem is large and not well understood.
You have a competition between - smoothing and island formation.
both due to restructuring on the surface during and after deposition.

BUT we published our the formula for a smooth surface:
A.M.Putnam, B.L.Blackford, M.H.Jericho and M.O.Watanabe
Surface Science 217 (1989) p276-288

It's pretty empirical, but it might serve needs.

Albert Putnam
putnam@msc.cornell.edu
Cornell Materials Science Center