"Multi-Technique Problem Solving in an Industrial
Environment"
Shell, Thornton Laboratories, Chester-
Wednesday 6th of January 1999
Parallel Workshop Sessions
"Dirty Samples:- How Do We Clean Them To
Extract Useful data? "
Leader:- Dr. S. J. Harris
Raporteur:- Dr. K. R. Hallam
Steven started out by first raising the question: "Do we allow
dirty samples to enter our instruments?" Of course, in an ideal World, we wouldn't.
However, the World (apart from NPL) is far from ideal, so we do have to analyse dirty
samples. Assuming that it is not the actual contamination itself that we are seeking to
identify, we must remove the contamination in some way and this is, perhaps, an area in
which the group might like to pool its ideas. A list of techniques was produced (solvents,
ion beams, chemicals, plasmas, lasers, CO2, etc.) and the discussion then proceeded to try
and identify the best techniques to use for a variety of sample types. Firstly, though, we
quickly dismissed the use of all techniques when preparing samples for SIMS analysis - we
would be left to only analyse the cleaning method, rather than the sample. We also
reminded ourselves of the importance of being aware of the chemical state effects of our
chosen method, most especially if in advance of any XPS analysis.
An addition was made to the list to cover the use of gas streams to
remove surface particulate contamination. Blowing (from the mouth) was a sure-fire way to
introduce additional contamination onto the sample surface and "Dust-Off"-type
sprays were well capable of introducing carbon layers onto the sample. Instead, it was far
better to use either dry N2 from a gas cylinder or a simple puffer.
Solvents have been shown to have various effects on sample surfaces.
Contamination left behind or introduced, or surface oxidation has been shown to depend on
the particular type of organic solvent used and on the grade (GPR, Analar, etc.) selected.
Acetone was agreed to be on e of the worst organic solvents with regards to residues left
behind. If acetone had to be used, then it should be followed by a further cleaning with,
say, methanol. Environmental concerns were also now playing their part, limiting our
choice of solvents. Returning to our ideal World, we would carry out studies to select the
most appropriate solvent to remove any particular contamination and have minimal effects
on the results we obtain. This, though, is usually precluded by a lack of time and money.
Words such as "dangerous" and "aggressive" were
used throughout the discussion on chemical cleaning, and it was generally felt that such
techniques were more for the removal of surface layers of samples rather than for simple
cleaning prior to analysis.
Low power lasers had been shown capable of selectively removing organic
layers from both metals and polymeric materials, though there was considerable time and
effort required to calibrate any such system. Of more interest to those present in the
workshop was the use of CO2-based cleaning systems. Though there was limited experience
among the workshop participants of these tools, there were two variants discussed: a)
supercritical CO2, which acts like a solvent - an excellent, if rather expensive, method
for removing oxides, e.g. from the inside of pipework, and much used by the aerospace
industry; and b) DRICE - which could be described as sand blasting using CO2 pellets,
where solvent, impact and cooling effects combine to be effective in, for instance, paint
removal, though the effects of condensation should be taken into consideration and hot air
lines used where necessary.
Other methods covered included heating within UHV, obviously the
standard method for the preparation of single crystal specimens, cellulose acetate
replicas, for pulling away contaminant particles, uv/ozone, ex-situ plasmas (good for
removing organic layers) and mechanical abrasion (e.g. corks!).
A table was completed to indicate where we believed particular methods
could (within all the limitations discussed) be applied for metallic, polymeric and
semiconductor samples. We concluded by reminding ourselves that all cleaning methods have
an effect on the specimen, and it is most important to understand the possible effects of
any cleaning on the results we subsequently obtain and to pass on this understanding to
the "customer" who brings along the dirty sample in the first place.
|
