18 
BSH Report No. 56 
Performance and results 
GC-FID data 
The first object of data analysis was the bell shape comparison, meaning the 
form of the distribution of n-alkenes among the homologues. As described 
above, sample 1 (Set 1) and sample 10 (Set 4) were duplicates and because 
of the duplicate analysis of each samples, this resulted in at least four Individ 
ual measurements of the same sample. 
FID-Paraffin-Bell-Shape-Comparison 
Mean of sample_l and 10 (n=4), normalized on maximum 
Fig. 7: Mean of peak heights from GC-FID measuring of every participant 
For the data of each participant, mean calculation and normalisation from 
those four analyses were performed. The plotted results (Fig. 7) show In gen 
eral a fair accordance between the data sets. 
However, especially on the descending side of the bell shape, larger differ 
ences between the different data sets are found. One reason might be an 
effect of mass discrimination as higher boiling n-alkanes are known to cause 
such difficulties. 
Another reason for these differences might be the quality of the chromato 
graphic separation. If peak broadening appears, absolute peak height might 
be decreased which Influences the results accordingly. Therefore, an area- 
based analysis of the peaks might be appropriate. 
The laboratory performance is expressed as the relative standard deviation 
(repeatability) based on the raw data of the four analyses per participant. Fig. 8 
shows that repeatability decreases with the lower concentrated n-alkanes.