Increasing peak area in GC

[Fdollet]

Hello ! I am having a bad time with a GC analysis and hope someone will help me.

By injecting several times a solution containing THF (internal standard) and Propylene Glycol (the analyte), the area of the propylene glycol increases whereas the area of the internal standard is stable.

To try to solve the problem, first I thought I needed to prime the column (packed column with chomosorb FID - solvent: N,N-dimethylacetamde). So I inject a concentrated solution of propylene glycol, then I inject 5 times the "normal" solution. It did not solve the problem.

Then I thought it was a carry over issue. A blank after the injections of my solution did not show any and the solvent used to clean is the solvent of the method (N,N-dimethylacetamide).

Is there something else I should try ?

I was thinking about the FID ? Could it be a detector issue ?

You might need more information about the method. Feel free to ask

Merci


Thank you

[Don_Hilton]

Glycols are tricky. And you may need many injections to prime the column - and then there is the problem of keeping the column "primed" during a sequence of injections. Can you derivatize your samples before injection (such as adding a silylating agent)?

[Consumer Products Guy]

We've assayed glycols for over three decades by dissolving samples in DMF or pyridine and simply mixing with BSTFA reagent. Propylene glycol will elute early, just after the excess reagents elute, so do a blank, and start at about 80C on a non-polar capillary.

These derivatives are stable, easy to use, and we've published their use.

When we've investigated without derivatization, we use the conditions for trace ethylene glycol and diethylene glycol in USP glycerin monograph.

[Fdollet]

Thank you for your help !

Unfortunately I am not allowed to change the method. I am trying to validate an old already filed method under GMP/ICH guidelines.

The only change I can make are the one allowed by USP.

[AICMM]

fdollet,

Rarely do I see the FID as the problem. Much more common to see the inlet as the problem. So tell us more about the inlet (on column, packed port, liner, temperature, flow rates, etc, etc....(in my best Yul Brynner voice).)

Best regards,

AICMM

[Charles A. Burger]

Glycols are tricky.

I've heard that before but never actually heard what makes them so. Would it be possible to share some of the more common problems.

I also have to run underivitized Ethylene and Diethylene glycols. My first thought is that the glass wool, while helping disperse the injection, may also be a source of interation and bias in the injector. Would that be fair to say?

Thanks

[Don_Hilton]

glycols are very polar and do not disolve well in non-polar phases, thus poor peak shapes are common on methyl silicone columns or methyl silicone with some phenyl present. So, for good chromatography, you wind up looking at something more like a carbowax phase.

This does not solve problems wiht the inlet. Any sources of activity in the inlet will result in losses of glycols or bad peak shape starting in the inlet.

One solution to dealing with active sites is to mask them by tying them up with other compounds from the sample matrix or with glycols. The active sites retain the glycols well, removing them from the desired chromatographic peak, But the glycols slowly desorb from the active sites and eventually move off the column (or decompose, either becoming a non-issue or contributers to column or inlet activity).

Glass wool can be a problem, as glass can have lots of active sites. You may be able to make a few "conditioning" injections to cover up the activity.

[Charles A. Burger]

I'm using the an Rxi 624Sil-MS that is a G43 phase that's been modified to make it more inert. Started using it for the USP <467> analysis and when I saw the separation between Acetonitrile and Methylene Chloride, I was very happy. Peak shapes in general are really good.

The peak shapes are usually good, but my reproducibility is not where I think it should be considering I have an internal standard.

Have you ever tried the carbofrits from Restek? With out having read much about them, I would thing that active sites would be reduced.

Unfortunately, the USP ID test for Ethylene and Diethylene Glycol doesn't help you out on the backflash side of things either. I'll have to try and scale down the injection size to address that.

[mckrause]

Carbofrits are too active for glycols - you'd have to deactivate them extensively prior to use, and you'd still have the issue of selective sorption to deal with.

Glycols are so polar that they cause problems no matter what your chromatographic conditions are. We use FFAP or Carbowax (capillary equivalent), we do on-column work, and we deactivate the syringes. We still get RSDs in the 5-15% range on a FID - for our work that is acceptable, but if you doing QC work it's probably excessive.

If you have enough sensitivity go to a split injection, unpacked liner and 2-5 uL injections. Also, make sure your injector is hot enough - you probably want to run in the 300-320 range, as you want to flash the glycols and get them down into the column.

Good luck!