Chromatography Forum

Hello,

I’m having difficulties finding the source of possible contamination of water samples in mineral oil index analysis. I prepare blank samples from tap water and ultra-pure water by carrying out extraction to hexane. Extracts are purified on a Florisil column and analysed on a Clarus 680 GC.

A chromatogram of a contaminated sample looks like this:

First peak is hexane and then this sort of a comb appears.

Now, let me break down my process so you guys might be able to point out something I’m missing here:

Sample bottle is being washed in lab-grade washing machine with tap and ultra-pure water with dishwasher salt in 75°C. After that, it's rinsed with acetone manually.

Glassware used in extraction process is soaked in tap water and regular dish soap, then rinsed 3-4 times with tap water, then rinsed again with ultra-pure water and then put to oven at 400°C for at least 8 hours – all prior to extraction.

Extraction process is based on adding hexane to water sample and stirring it so the potential oil transfers to hexane. Then, using the fact that hexane is on top of the water, it’s separated from the water by being pushed from the bottom to the separatory funnel by adding tap water. Not all hexane is collected in the separatory funnel because I don’t want any water to get into the funnel. Next, the extract is purified on a Florisil column and collected in vials which were also cleaned as described earlier.

Extracts are analysed on a system described as follows:
Chromatograph: Perkin Elmer Clarus 680 GC
Column: Zebron ZB-5HT Inferno 30 m / ID 0,53 mm / film thickness: 0,15 μm (brand new)
Detector: FID
Inj. Method: On-column

The problem seems to be in the sample preparation step in our lab because I had samples prepared by me and analysed in different lab - also on a Clarus 680 GC equipped with FID detector - showing the same peaks so the issue doesn’t seem to origin from hardware in our lab. The other lab prepared thoeir own blank samples using the exact same procedure and they don’t have these peaks showing.

I want to ask you fellow chromatographers if you could help me with locating the source if these peaks or maybe you have faced something similar in your lab. I’ll be happy to provide more details if necessary.

Hi

Are there PEG or Tween or silicon oils around your lab, which also went to the dishwasher and oven?
I guess it's verified that it is not column bleed? (you said the column is brand new.)
I don't really understand how the transfer to the separation funnel is done?
How do you push it? -> contaminated air / N2 source?

I guess here, only a systematic rule-out-one troubleshooting strategy will help.

I would just inject everything (possible) after each step and also compare with fresh solvents and consumables.
Also, if you could ask for aliquots of solvents and consumables from the other (good performing) lab.

e.g.
- injection of air
- injection of pure hexane, acetone etc.
- different lot/fresh pack of vials
- different vial caps / if possible injections without any cap / or without septum /different /cleaned crimper (contaminated crimper?)
- inject sample pre and post florisil
- an aliquot of the hexane phase prior the transfer to the separating funnel
- using glassware that were not in the oven, just rinsed with good aceton/hexane (in case the oven is contaminated and just coats the glassware)
- use of higher oven temperature if possible; at least prior to the glassware treating
etc...

What about the glassware/consumables used to add/transfer the samples/solvents?

Just check everything, step-by-step until you find the source.
Don't think to rule out anything, until proven.

Replying to anyone who might find this useful.

I found the issue causing these peaks to appear. It turned out to be a bad Florisil/Na2SO4 purification column. The extract was migrating through the column way too quickly to be purified properly. We're now using columns from a different manufacturer and the peaks are gone.

Replying to anyone who might find this useful.

I found the issue causing these peaks to appear. It turned out to be a bad Florisil/Na2SO4 purification column. The extract was migrating through the column way too quickly to be purified properly. We're now using columns from a different manufacturer and the peaks are gone.

Useful information, thank you.

Peter