Tetrachloroethylene

[wraymogg]

Hello,

I have a problem with quantification of tetrachloroethylene by GC-MS. This particular method is for dichloromethane, 1,2 dichloroethane, chloroform and tetrachloroethylene. For the other 3 I have no problem, they work perfectly, but tetrachloroethylene does not have repetability, and a linear response to concentration.

The most important thing is I DO NOT use a headspace, but a L-L extraction with pentane, so the standards injected are also liquid.

The most important parameters:

Injector: 200 Celsius grades
Pre-dwell time: 0
Post-dwell time: 0
Injection volume: 2 ul
Transfer line and source: 230 Celsius grades
Quantization ion: 166
GC program: 25 Celsius grades 6 minutes, then 15 grades per minute to 120

There is not a separation problem, the peak looks awesome, but it's height and of course, area varies a lot.

I injected a Toluene-D8 as ISTD (not the perfect match) just to see if it levels the injection defects, etc. They close retention times, but the thing is Toluene-D8 is very repetable, tetrachloroethylene is not in the same injections series. Toluen-D8 will stay constant in 6 repeated injections, tetracholoroethylene will vary A LOT.

The column is Thermo TG-5SILMS which is a 5% phenyl methylpolysiloxane, 30m with 0,25 um film. I know is not the best column for VOC separation, but again, not the separation is a issue.

Why chloroform has a perfect and repetable calibration curve and tetrachloroethylene hasn't ?

I tried this:
- varied the GC program
- varied the pre and post needle dwell time
- varied split ratio (10:1, 50:1, 100:1)
- varied injector temperature (200-250 Celsius grades)

At higher or equal with 40 Celsius grades starting temperature, they will not separate. I tried to start the temperature curve from beggining, from 25 to 120 with 15 grades per min, the retention time dropped, still a good separation, but the same problem. Peak will vary a lot.

Tried to switch to ion 164, same, they are 100% proportional.

If it reacts with something to sub-speciate in column, why the other clorurated VOCs will not ? Somehow this problem is related ONLY to tetrachloroethylene.

Any ideas ? I am in a dead end.

Thank you
Vlad

[Yama001]

PCE will degrade to other chlorinated ethylenes while your other analytes are better behaved. Do you see extraneous peaks (Trichloro, etc?).

[bor9455]

What kind of liner are you using on your inlet?

[wraymogg]

Hello,

I don't see any degradation. There is no peak for trichloro or other isomers. The column is quite inert and also the injector bottom seal.

Do you say tetrachloroethylene is more susceptible to degradation than chloroform ?

The liner I am using is the same I use for PAHs, actually is a splitless liner with single taper and deactivated wool. All other VOCs are fine with it.

I can try to replace it with a straight split liner with no wool, to see if anything changes.

Also I can try to inject it splitless...

Regards,
Vlad

[Steve Reimer]

The main degradation problem I have with tetrachloroethylene is the degradation of pentachloroethane into tetrachloroethylene in neutral pH water. Only a problem if you have a mix with extra compounds in it.

Is pentane your injection solvent?

[wraymogg]

Hello,

yes, pentane is my extraction and injection solvent. At this moment I try to fit a calibration curve, so I am only injecting pure standards. The standards are in-house prepared from pure substances so inside the injected solution there are only 4 compounds: DCM, 1,2-dichloroethane, chloroform, tetrachloroethylene.

I can change the solvent. It has to be the same for extraction and injection (i cannot solvent change due to high volatility and I don't have the cryo-concentrator). Just now i am experimenting with isooctane mainly for DCM (because pentane is overlapping DCM on this column) but I still get poor results.

Today I will try a splitless injection and if is still bad, I will change the liner and try again split.

Do you have a recommendation for other proper solvent for extracting tetrachloroethylene and minimize the risk to react ?

Thank you,
Vlad