18 
Oil Spill Identification - Round Robin 20045 
peaks. This might be concentration related: in combination with a small 
(ID 0.18 mm) column overloading might be a reason for this behavior. 
The MS chromatogram of Source II is also strange. Resolution is far less 
compared to the other chromatograms. As a result Source II has been 
excluded. Comparing Fig 3 and Fig 4 of the BMM report indicates that 
the injection concentration between source and spill samples is a factor 
of 10 or more. It is advised to optimize the injection concentration. 
The evaluation of the chromatograms is laborious and susceptible to 
failures. Patrick has indicated that he intends to improve data handling 
Differences between the ratios of small peaks can be larger than normal 
because of the difference in the injection concentration. This is e.g. 
visible in the m/z 191 and m/z 216 chromatograms. 
GC-screening may be used to adjust the concentration of samples, as it 
is done, for example, by SKL: 
After a first preliminary GC-FID run we eventually dilute or concentrate the extracts so 
they all will have approximately the same concentration. Most of our samples that are 
collected from the water are taken with a teflon cloth, so the concentration can vary a 
lot. 
2.2 Centre de Documentation de Recherche et 
d'Experimentation sur les pollutions accidentelles des 
eaux (Cedre) 
Contact: Julien Guyomarch 
Cedre (http://www.le-cedre.fr/1 was created in 1978 within measures 
taken after the wreckage of the "Amoco Cadiz", to improve 
preparedness against accidental water pollution and strengthen the 
national response organization. Its expertise encompasses both marine 
and inland waters. 
The RR-samples have been analyzed with GC-MS in SIM mode on 
alkanes, PAH's and biomarkers. 
=> The comparison of the samples Source I and II, and Extract I and II led to 
the following conclusions: 
- n-alkanes show differences in the region n-Cl 1 to n-C19, but these 
variations are not significant due to possible evaporation processes affecting 
these distributions. On the other hand, the relative abundances of compounds 
ranking from n-C20 to n-C30 did not show differences. 
- The biomarkers (m/z=191, 217, 218, 231) did not allow to differentiate the 4 
samples due to high variability. 
- Finally, the diagnostic ratios calculated from PAHs analyses present 
variations that could explain differences of origins for Source I and II. These 
significant differences showed for fragment 216 and 234. 
=> The conclusion of this oil spill identification is that the samples collected 
at the water surface do not seem to come from the two boats. Moreover, the 
two bilge samples and the two surface samples have different origins. 
Notes/remarks: