On-Column - a possible misunderstanding of the basics

[Spuzzin]

Hi.

I am currently developing a method for the analysis of glycols (propylene glycol up to and including triethylene glycol) in aqueous samples.

Initally I went for split/splitless injection since we had an instrument available. After numerous problems and some research I've switched to on-column with some better and some poorer results.

The current set-up is cold on-column analysis with the inlet at 80 Deg C injecting onto a 2.5m polar deactivated pre-column which is attached to a 25m wax column. The start temperature for the oven is also 80 Deg C. Peak shape and calibration curves are good for the first five compounds but diethylene glycol and triethylene glycol both have very small peaks relative to the early eluters. The only way I have been able to improve this is by raising the temperature of the inlet to >200 Deg C. Other experiments have shown that the problem isn't with the syringe and the degree of carryover/ghost peaks confirms this. There is next to no ghosting of the first analyte and the degree of ghosting increases as the molecular weight increases.

My question is this:- by my understanding the sample should flow rapidly through the uncoated precolumn to 'trap' on the analytical column but the evidence seems to suggest that the heavier analytes are sitting on the pre-column in the inlet (which remains at 80 Deg C throughout the run) and being rinsed onto the analytical column when another injection is made.

Is my understanding of this concept flawed ?

Any comments appreciated.

Rich

[gcguy]

In my experience water does not make a good solvent for on-col injections.
The general rule of thumb is to have the oven and injector 15-20C lower than the boiling point of the solvent used. As for carryover, if you run up the oven temp to the coulmn max and hold it for a while, then do a blank injection and check for peaks.

[chromatographer1]

I think your understanding is sound.

The problem lies in that you allow your injector to remain at 80°C.

You should have your injector temperature track along with the column oven in temperature.

The Agilent GC I have does have this option.

I suggest you use it.

Your problems should go away.

best wishes,

Rod

[Spuzzin]

gcguy,

I switched to on-column because I was getting such poor results from the combination of aqueous samples, glycols and a split/splitless liner. I'm getting better results from the on-column but again the general response of avoid putting aqueous samples on GCs may be appropriate.

Rod,

Unfortunately the instrument I'm using doesn't support programmed temperature ramps for the inlet so I will have to think around the problem.

Thank you both for your replies.

Rich

[chromatographer1]

Rich

It will be time consuming but here is one way to accomplish your task. Inject with the injection port cold and then reset the injector temperature to 250°C before you start to program the column oven.

After the injector temperature increases to 200°C or so you might want to start the slow programming of the column oven to begin. This delay may take several minutes.

After the run is over you will have to wait for the injector to cool down before you perform the next injection. You may have to wait 10 minutes or more. Awkward, but you can make it work.

Good luck,

Rod

[CE Instruments]

Firstly you do not have a cold -on column inlet 80C is not cold. The Grob Cold on column is cold and used properly the injection occurs inside the oven inside the column so that there is no transfer step. You seem to have a copy of the Cool on-column that is fitted to HP/Agilent and these should be temperature tracked to transfer the sample from the injector into the column in the oven.
Have you tried a water liner with your split/splitless ? Water injections are bad news and water and on-column injection is even more difficult. If you inject at a temperature greater than 100C you run the risk of flash back of the water into the carrier line. If you stick with Cool on column the manual swich of injector temperature is the safest way to transfer the heavier fractions.

[AICMM]

Spuzzin,

How much separation do you have between IPA and your first analyte? I have a friend who dilutes all water samples 1:1 with IPA and he says that this dramatically improves the flash and reproducability. However, he also does a split injection and I don't know what his ultimate detection limit is with this scheme. Something to think about though.

Best regards.

[Spuzzin]

Rod,

As you say your suggestion with regard to resetting the inlet temperature is awkward and not practical from a high throughput point of view but will definately help as proof on concept. I'll give it a crack.

CE,

I agree that 80C isn't cold but I usually do injections at 300C+ so its relatively cold, for me at least.
The instrument I'm running is a 1041 constant temperature on-column inlet and a Varian 3900 GC.

AICMM,

I'm currently looking at something similar using methanol to dilute all samples, for the same reasons as your friend. Initial results are encouraging but changing priorities are pushing me away from this project.

Hopefully in a few weeks I'll be able to get back to this project and post some results or at least some progession.

Thanks

Rich

[chromatographer1]

Thanks for keeping us updated. We are all interested in helping you achieve success. Hopefully your employer will give you the tools you need to accomplish your assignments well.

When you share your experience with the Forum many others who come after you can profit from your trials.

It will be appreciated.

best wishes,

Rod

[Spuzzin]

Update:

After trying various injection techniques I am getting the best calibration curves and the least carry-over/ghost peaks by using direct injection. The only issue now is that the liner gets contaminated fairly quickly and replacement is a little bit more complicated since the column has to be cut.

Regards

Rich

[chromatographer1]

Thanks for taking the time to report back your results. Others will profit from your reply. But may I pose some questions for additional clarification?

Did you need to use a wash plug behind your sample ?

Are you using a fairly small sample volume of water?

Did you dilute with methanol as a final solution?

What components are giving you ghosts and how much of a problem are they to your measurement?

Thanks again for getting back to the Forum.

best wishes,

Rod

[Willy]

Dear Rich,

I just read your posting and all others and i was wondering if you ever have considered using glass column. This way you can inject more (3-4 µL. Ive done some analysis already on Propylene glycol and Ethylene glycol on glass column with direct injection. I dont remember the phase we were using in the glass column (probably Porapak or carbowax type). It is better since DB-Wax doesnt like water very much (although there is some new DB-Wax column that resist water now.)

Willy the GC

[Spuzzin]

Additional Information

Apologies for the delay in replying. As in normally the case things kept on getting in the way.

The on-column method I developed didn't make it through the validation. Ghost peaks were the biggest problem for all compounds from ethylene glycol through to triethylene glycol with the triethylene glycol being the worst culprit. I believe this is down to the fact that the on-column inlet I am using isn't able to ramp with the oven.

I have now moved onto an direct injection method which is proving much more effective, in general. I say in general because it works well on clean and 'fairly' clean waters but heavily contaminated effluents result in a drop in response for diethylene glycol and triethylene glycol.

The method is fairly standard although diluting 2:1 with methanol is required to prevent excessive ghosting.


Willy,
I hadn't thought of using glass columns but I don't think any of our systems will readily convert to packed columns.

Regards,
Rich

[eljoe]

hi,

I used EN13130-7 to determine ehylene glycol and diethylene glycol in aqueous extracts (with 1% methanol). I use "cold"-on-column (60 to 70ºC) injection on a CpWax 57 (25*0.53mmID;df=1um) with oven @ 100ºC initial then till 160ºC (maybe not enough for triethylene glycol). The injection volume is 0.5 ul. I had no problem so far with those two.
Hope I could help
Joel

[Spuzzin]

Hi Joel,

The problem I have had is that the on-column inlet I have access to is fixed temperature and doesn't ramp with the oven. Like you I started about 70 deg C. But the column inside the inlet, even using nothing more than a transfer line, seems to hold onto the analytes.

Rich