Ethylene glycol RT shift issue

I am using GC6890 /MS5973 and COmbi PAL headspace to analyze ETHYLENE GLYCOL and other solvents. But always got RT shifts with concentration increase.

anybody could shine any lights for me to find the bug?


Here is detail infor:


Ethylene Glycol Peak RT shift issue


concentration of ethylene glycol = 2.5 ug/ul, pipette the following amount to 10 mL headspace vial for each calibrator


Calibrator 1, Ethylene Glycol Peak RT = 8.35 min, Amount = 0.8 uL

   Calibrator 2, Ethylene Glycol Peak RT = 8.38 min, Amount = 1.5 uL

   Calibrator 3, Ethylene Glycol Peak RT = 8.47 min, Amount = 2.5 uL

Calibrator 4, Ethylene Glycol Peak RT = 8.92 min, Amount = 5 uL

    Calibrator 5, Ethylene Glycol Peak RT = 9.22 min, Amount = 10 uL



1)other solvents RT are always consistent without any shifts

2)The calibrators were prepared by spiking pure analytes into triacetin.

3) 250 uL headspace gas was injected by CTC Combi PAL.

4) The peak width becomes wider with increasing concentration, peaks is showing tailing for calibrator 4 and 5

5) Peak area is linear for first three calibrators, then suddenly drop for calibrator 4, then back for calibrator 5.

6) RT for same calibrator (repeated injection) are consistent without any shift

7) All above obServed information could be duplicated by running another batch of 5 calibrators. We repeated several times, always got same results like above.

8) Inlet liner is Agilent ultra inert PN 5190-2293

9) GC conditions: Inlet 250 C, split 10:1, flow 4.0 ml/min, oven 35C hold 1.5 min then to 300 C at 30C/min

MS interface 280 C, Scan 15-125 amu.

  • Hi Wendy,


    I assume you use a wax-column for the EG?

    With increasing sample load you get, as you mention, wider peaks, and also more asymmetrical peaks.

    The RT is determined as the apex of the chromatographic peak, at that determination will shift with increased asymmetri. It is a well known "feature".

    Possible solutions could be to reduce the sample load or to increase column capacity so that the peaks don't overload. This will help the peaks stay symmetrical with more constant retention times.


    If identification/localization of the peak with the quantitation software is the problem, try to increase the expected retention time window - that is if your separation allows unequivocal identification within a wider RT-window of course.


    Hope this helps.


    Best regards,


  • Karina


    Really appreciated your kind reply!


    But it is not overloading issue as we generally experienced. 




    Sent from my iPhone

  • Hi Wendy,

    This is my first ever post here, but I have done a fair amount of work with loop & syringe injection of polar compounds by HS.


    I am assuming that the vial is empty and your CTC heats the vial to ~50C and equilibrates for a few minutes?

    If we assume that all of your Calibrators go into the vapour phase, the conditions in your column will vary depending on the amount of water you are adding. As you are starting at 35C water will tend to condense in the column with your ethylene glycol (EG).

    Assuming that all of the Calibrators are in the vapour phase, then for: Calibrator1 your column will see (very roughly) 5ng of EG and 2000ng of water.

    Calibrator4 would be 30ng EG and 12000ng of water, and Calibrator5 would be ~60ng EG and 25000ng of water.

    As you say on a >0.25µ polar phase those amounts of EG should not overload the column, but I would expect that the extra amount of very polar water condensing in the (polar?) column would give the sort of increases in RT that you have seen. If you have a low polarity column it will overload, but I would expect the peak to "front" rather than the tailing you describe. Your water may also tend to act as a mobile phase while it evaporates. On a polar column it could take a while before it all goes into the gaseous phase, as it will tend to stay dissolved in the stationary phase.


    The solution could be to inject the same amount of water for each level of Calibrator: Calibrator1 = 0.8 +9.2µLH20; Calibrator3 = 2.5 +7.5µLH20; etc.


    Alternatively, you could try to avoid condensing the water in the column by having the initial oven hold at 110C, but you may need a pre-column/guard column; otherwise you could damage the analytical column. The Agilent Polyethylene Glycol Deactivated Polar type should be OK, but if you are going to do a lot of samples it might be worthwhile looking at a speciality pre-column (Look for hydroguard water resistant types).







  • Hi wendywen,

    I just wanted to follow-up to see if this issue is resolved.  If so, please come back to the post and click the “Correct Answer” button on the response that help so it will make the solution more visible.  If you still need help, just let us know and we would be happy to continue working with you. Also if you found the solution other than what was posted, let us know and it will help the community. I have also moved your post to an area where it will be more visible.

  • Have you tried timstrutt suggestions? Let us know what has been done.

  • Sorry, it does not fully solved yet.




    Wendy Wen

  • Hi Wendy,

    I should have mentioned that if the problem is caused by water, increasing the split ratio should also improve the RT precision, as you are putting less water into the system. As indicated above, at least some of the separation of your EG is happening with injected water acting as a component of the mobile phase above 100C and as part of your stationary phase below 100C - Different amounts of water will change the partition ratio of your analyte/stationary phase - More water acting as a stationary phase will increase the retention time of your EG.

    As you say, you are not obviously overloading the system with EG, but increasing the split to 50:1 will still give you plenty of signal for your low level injection. Headspace injections at low split ratios are notoriously imprecise with differing polarity materials, so you should get better results by decreasing the amount of solvent (water) injected. A 50:1 split ratio is often a good place to start when doing method development - I would avoid <20:1 split with a polar compound and an apolar column (or a polar column and an apolar solvent).



  • Hi wendywen,

    Agilent does offer some application help as well. We have column chemists that can be reached in the US 1-800-227-9770 opt 3 , opt 3, opt 3 for GC. If outside of the US , please check with your local support numbers found on Agilent | Contact Us

  • Thanks for the info Wendy,

    Depending on the polarity of your column I would expect DMSO to elute near EG, and triacitine (glycerin triacetate?) after EG. I have not used triacitine, but it was an old GC/MS trick to determine volatiles by split injection by dissolving them in DMSO - Generally, even with similar amounts of DMSO injected for each concentration, the retention times could change.

    Cheers, Tim

  • Small variations in concentration, different speeds of cooling/evaporation/sorption/desorption in injector, particularly at low split ratios and lowish injection temperatures...

    Expanding my post: In your case, if you are still using the same concentrations:-

    Calibrator 1, Ethylene Glycol Peak RT = 8.35 min, Amount = 0.8 uL

    Calibrator 5, Ethylene Glycol Peak RT = 9.22 min, Amount = 10 uL...

    If most of the solvent is in the vapour phase you would be injecting roughly .03mg of solvent for Calibrator 5 and .002mg for Calibrator 1 (This is where it gets tricky, and my mental arithmetic could be wrong - It is late at night here and I am definitely old enough to need my beauty sleep) 0.03mg is also very roughly the same amount of stationary phase that you could expect coating ~1 metre of capillary column. If your peak is ~3 seconds wide and the carrier is at ~30cm/sec, your analyte peak also occupies ~1 metre of column - So your analyte will also be retained by very roughly about as much solvent as there is stationary phase material in 1 metre of your column for your Calibrator 5 and less than 1 tenth of that for Calibrator 1. You would expect more solvent to increase the retention time of your analyte as the polarity of your solvent is similar to your analyte...


Was this helpful?