Random Peak Area Reductions

[mcris]

Greetings. We're working on a method for the determination of ethylene oxide in aqueous solution. We're using a water stable PEG column, ZB-WAXPlus, 30m x 0.53mm and FID detection. Everything was going along great, good RSD's, good recoveries, good chromatography when all of a sudden we begin to get dips in peak area/concentration over multiple injections from the same sample vial. The first injection will be right on, then the next two will decrease until it bottoms out, then begins to rise again and by the end of the run, from that vial, it's back to where it's expected to be. We know EO is very volatile and must be kept chilled to avoid evaporation, but we've done the studies and found it's stable over the number of injections we're performing and if it were a stability issue, it makes no sense that the concentration would go back up. Has anyone else seen something like this happen? Thanks.

[HPLC chemist]

EO is very reactive. Be sure your solvent is only HPLC water. It seems you have something else reacting with the EO which is reversible. Perhaps by adding a pH buffer you may be able to hinder this reaction.

[James_Ball]

How well is the temperature control of the samples?

If the autosampler is not chilled and the room temperature fluctuates it could be shifting equilibrium in the headspace above the water as the temperature fluctuates.

[mcris]

Both interesting answers, thank you. We're also analyzing for 2-chloroethanol in the same chromatogram and that seems to get even lower than EO. I couldn't find any likely reactions that 2-chloroethanol might undergo and the room temperature is monitored. It fluctuates less than 1 degree C during the day.

[Chemiger]

I agree with the comment about the change in equilibrium due to ambient surroundings, this is most likely why you are seeing the dip then recovery in analyte.
I do have one thing to add, are your inlet liners inert ?
non inert liners can have adsorption sites where analytes can stick to the liner.
This can affect results especially if looking for trace amounts of analyte.
all GC consumable companies sell an inert line of consumables.
Agilient use a range called UI ( ultra inert)

Thanks,

Chemiger

[mcris]

Thanks. I found the comment about the change in equilibrium intriguing. I had checked the room temperature and found it to be very consistent. It's a Perkin Elmer GC and the oven vents upward, so I thought that maybe someone walking by or an air current in the room my have directed some warm air toward the sample. I changed the equilibration time from 1 minute back to the default value of 2 minutes to, hopefully, allow the original equilibrium to be re-established.
In the meantime we discovered that the moisture trap in the hydrogen line was saturated. So, we are theorizing that random releases of water vapor may have cooled the FID flame at just the right time leading to lower peak areas. We replaced that trap and so far, so good, but the phenomenon was random so we don't really know if it will show up again.