41 
Oil Spill identification - Round Robin 20045 
The black line shows mass 231 of sample sourcel together with the 
aromatic steranes of a crude oil (Brent-crude, red line). Obviously, 
aromatic steranes were definitely not present in the samples. 
If in this case the oiltype found by GC-screening (mixture of light fuel 
with lubricating oil) is not considered and a fixed set of parameters is 
used for all oiltypes, then the analyst will suddenly be in a situation, 
where highly overlapping peaks on a hump are found instead of 4 to 5 
normally very well resolved peaks. 
He may, nevertheless, find peaks at the proper retention-times and 
even integrate them. This might not be a bigger failure, if he does the 
same with all 4 samples, since all peak relations must match, if samples 
are regarded as identical. But the integration of highly overlapping 
peaks on a hump is very crucial and should be avoided because those 
peaks cannot be measured precisely. 
It may even have been tested, whether peaks can be measured 
precisely or not by determining the corresponding peak ratios in two or 
three consecutive injections of the same sample. 
But such a test cannot be regarded as sufficient for excluding ratios 
because it is possible that those ratios can be measured precisely in this 
case by chance (because the peaks are measured here definitely in the 
same sample). But there would be the high risk of a non-math, when a 
comparison is made afterwards between the spill sample and the 
possible source sample although these samples actually originate from 
the same source because these samples may have a totally different 
"history" (The spill sample may even have been extracted, whereas the 
possible source sample may have been weighted out. The two samples 
may thus even have a quite different oil concentration). 
Generally, the analyst must use the knowledge here, that aromatic 
steranes are not present in lubricating oil. He should not be astonished, 
when he finds overlapping peaks on a hump instead of 4 to 5 well 
resolved peaks. 
In addition, he must avoid to integrate peaks in a highly overlapping 
cluster (bad chromatographic practice). 
This means that the analyst must be able to decide, which ratios are 
useful in a distinct oil spill situation and which are not. 
Using a fixed, broad set of parameters (compound ratios) for all 
oiltypes and in all oil spill situations from the beginning, and letting the 
system “decide", may thus easily lead to wrong conclusions. 
In the CEN guideline therefore a contribution of Gerhard Dahlmann can 
be found (annex H) describing the peculiarities of the different oil 
types. It has been finished already a few years ago and has been 
advised to all participants of the Round Robin last two years: 
The problems associated with the identification of bilge oil samples are 
especially well described in 
http://www.bsh.de/de/Produkte/Buecher/Berichte/Bericht31/Bericht31 .pdf. 
Hints, tips and precautions are given here concerning their analysis and result 
interpretation.