Chromatography Forum

Hello,

I'm looking to pinpoint the source of a contaminant that's been popping up in small amounts on a few of my samples. I'd wager it could be from filters/septa/vials moreso than my MS, as repeat runs of tainted samples show near-identical counts, but you never know. Shows up in Hexane,DCM, and Ethanol solvents.

Analysing essential oils and ethanolic plant extracts under the following conditions (some temperature programs may deviate from this slightly)

TP: 40 degrees for 2 minutes, followed by 5 degrees a minute to 230 degrees, hold for 6 minutes. Transfer and source at 250 (max for this instrument)
Column: Varian VF-5ms, 30 x 0.25mm x 250um
GC: Agilent 7890a
Detector: SAI Kronus ToF-MS
Inlet: 270 Degrees, BTO Septa, Low pressure drop liner

Structure of contaminant (from NIST, at least) http://i.imgur.com/v3zh3.jpg


Any help is much appreciated....cheers!

The presence of m/z 105 in (and absence of m/z 115 from) your spectrum would indicate that the proposed NIST match is not the true structure.

1. can you run the GC/MS under CI conditions (NH3 or CH4) to give [M+H]+ ??

2. if not, can you repeat the EI analysis under reduced filament eV to reduce fragmentation and possibly enable M+. to be observed ??

3. if either (1) or (2) is OK, can you also get a clean accurate mass measurement on the molecular ion ??

This should enable the molecular weight and possible elemental composition(s) to be determined. First find the true structure, then pinpoint the origin.

If your are seeing the same contaminant in three different solvents (or is it only in extracts rather than solvent blanks ?) then you need to be looking at what is common to al three when they are injected into the GC - the obvious culprit is the vials and caps, presuming that you transfer direct from the bottle to the vial.

Peter

Don't overlook the injection port, split vent line and filter. then there is the syringe.
Do you get the peak with a no injection run?

The 207 m/z from the column bleed is making the NIST searching the wrong molecule, scan from 30 to 200 m/z and try the library search again.
With that 77, 91, 105 (120 too but i don't see it, maybe it's just hided) combination i would guess it could be an aromatic hydrocarbon like trimethylbenzene or an higher molecular weight one of the same type.
Of course i'm not sure, that spectrum needs to be cleaned.

I agree the NIST result is not correct. I also suspect that the 207 ion is not part of a correct spectrum, but...

Your spectrum (searched in NIST) shows a ??time?? of 21:44.796. Is that a retention time? If so, then this compound is eluting at 137 C temperature - I am surprised you have column bleed at that low of a temperature. Or is the 21:44 not a retention time? The RT would give us a ballpark idea of how big a molecule this really is...

Can you average several spectra across the peak, then subtract an averaged background (from a few scans to the right of the peak)? Is the 207 now gone? Is there anything that seems it may now be a molecular ion? If you plot the 91, 129, 207 ions, do they rise and fall together, or are these two - or three - separate materials?

If you search this averaged/background subtracted spectrum, do any of the hits look better?

Is this in some but not all samples? Blanks? Standards? We need more info...

Parts of the GC may be contaminated from previous injections and the solvents may be dissolving this contaminant each time something is injected. The 207 peak could very well be from column bleed (though I don't see the 281 peak), the smaller MW peaks could be something much larger getting blown apart when it encounters heat. Clipping off the front end of the column and reconditioning it as well as replacing the gold seal took care of this problem for me when I had similar symptoms. You may also want to consider the injection volumes used and the liner used; there could be backflash problems and gas lines could be contaminated. All fun things to consider.

The presence of m/z 105 in (and absence of m/z 115 from) your spectrum would indicate that the proposed NIST match is not the true structure.

1. can you run the GC/MS under CI conditions (NH3 or CH4) to give [M+H]+ ??

2. if not, can you repeat the EI analysis under reduced filament eV to reduce fragmentation and possibly enable M+. to be observed ??

3. if either (1) or (2) is OK, can you also get a clean accurate mass measurement on the molecular ion ??

This should enable the molecular weight and possible elemental composition(s) to be determined. First find the true structure, then pinpoint the origin.

Replaced the gold seal, liner, and changed out the solvent wash bottles. Gave the needle a good clean and snipped the column, basically gave it a spring clean. Searched for the ions together in current runs and it seems to have all but vanished. Still no joy with the structure, running at 30eV gives the same spectrum albeit in negligible amounts, thankfully. Didn't even consider the liner or gold seal giving me problems


Cheers for all the help - if it starts rearing its ugly head again, I'll be back for more!