Column bleed or real peaks

[PEfan]

I am detecting compounds derived from the Silica in the GC column. I am getting these in varying amounts and species from sample to sample (and not increasing over time).
For instance I am getting hexadecanoic acid, trimethylsilyl ester and I can see that it is significanly different in different samples. I am not sure if compounds like this one are "real" or if I should treat those as column bleed only.
Could anyone help on this matter, please?
Thanks

[Answer Man]

Capillary columns do not generate their own peaks (unless the phase contains contamination), they only generate column bleed (baseline). One way to confirm this is to do an analytical run but do not inject anything - you should not see any peaks (please repeat if you do see peaks, as these are probably just carry-over).

The peaks are probably caused by contamination in the injection port, the loss of the deactivation layer on your injection port liners, and/or the septa (either from the sample vial or the injection port septa).

The peaks are 'real', but are probably not from your sample.

[Consumer Products Guy]

In the consumer products business, we oftentimes assay products containing soap, even soap bars. A small bit of soap (fatty acid salts) will dissolve in organic solvents, so the fatty acid salts will deposit in the GC liner/head of column, so in subsequent injections involving trimethylsilyl derivatizing agents a small amount of the deposited salts will form fatty acid trimethyl silyl esters, elute, and have spectra of hexadecanoic, lauric, stearic acid, etc. This is tough to avoid in our business, as assaying soap products by GC for stuff like glycerin and sorbitol are so easy. If we want to establish that the peaks really do come from the sample, we do reagent blank injections first and compare the peak sizes to those. So, as Answer Man replied: cleanliness can help here.