Inconsistent responses in early eluting analytes

[tangaloomaflyer]

We are experiencing variation in responses that only seems to affect the earlier eluting compounds of a method based on EPA8270. The analytes are mostly phenols but a dichlorobenzene internal standard is also affected so I don't think it is a prep issue. The later eluting analytes have consistent responses and the internal standards co-vary with the analyte responses.

Instrumentation: Agilent 6890/5973
Injection/solvent: 1µl DCM with 10% toluene
Inject speed: 30µl/sec
inlet: pulsed splitless 280oC, with 50ml/min purge flow
Column flow: ramped, 1.7 to 2ml/min
Oven ramp: 50oC to 330oC at 25oC/min
Column: J&W DB5.625 [20m x 0.18mm x 0.36µm)
1m pre-column installed

[gurlwithgcms]

Hi tangaloomaflyer,

The only thing I can think of besides changing the septa and liner, and making sure the column is not old, is to change some of the settings in the run method, focusing on the ramp settings. We run method 525 in our lab, which is similar, but we use dual columns; 30m length, 0.25 mm I.D., 0.25 µm film, like the Restek Rxi®-5MS. For something to start with, the ramps set up for our runs are: 1: 20 rate, 205 final temp. 2: 2 rate, 230 final temp. 3: 25 rate 320 final temp, 1 final time. 4: 0 off. When I was running volatiles (524) I had some of the same problems. I changed the ramp settings, so there was more of a ramp, (5 rates instead of 2). I was using a 20m column, and switched to a 30m, and both of those changes helped a lot. If you have a used 30m column available, it might be worth switching out the column to see if that makes a difference. Your method should allow for the column and ramp changes, but it would be a good idea to double check just in case. I hope those suggestions help because that's all I can come up with since I don't have that much experience with the semi-volatile methods. Also if you go to Restek's website (www.restek.com: go to the Chromatograms tab, then GC Chromatograms), you can filter through chromatograms that pertain to your method. They list the instrument parameters along with the chromatogram, and that was very helpful when I was editing the 524 parameters.

[Peter Apps]

This dichloromethane - toluene mix has come up before as a source of problems.

This has to be some form of inlet discrimination - what is the splitless time, and for how much of that is the pressure increased ?

What kind of liner are you using and does it have any packing in it ?

Peter

[Steve Reimer]

Is the inconsistency in your standards or only in your samples/QC?
Your oven ramp starts at 50 degrees and the boiling point of toluene is 110. I can see where it may be a little touchy about inlet conditions.

[tangaloomaflyer]

Thanks for the detailed replies.

The method has been fine up until recently and has been reproduced on other systems without any problems. That is why I think this is more of a malfunction somewhere, rather than a quirk of the method. Of course, it could be an aspect of the method exacerbating a problem.

Peter: splitless time is pulsed for 1 minute, pressure raised from 31psi to 50psi. The liner is a single taper (at bottom) packed with silanised glass wool at the bottom.

It is one of our oldest instruments, and the problems started after unintentionally running a large volume injection (10µl on the same method) for a while (several months)

[Peter Apps]

Thanks for the detailed replies.

The method has been fine up until recently and has been reproduced on other systems without any problems. That is why I think this is more of a malfunction somewhere, rather than a quirk of the method. Of course, it could be an aspect of the method exacerbating a problem.

Peter: splitless time is pulsed for 1 minute, pressure raised from 31psi to 50psi. The liner is a single taper (at bottom) packed with silanised glass wool at the bottom.

It is one of our oldest instruments, and the problems started after unintentionally running a large volume injection (10µl on the same method) for a while (several months)

10 ul injections will have put liquid solvent onto the column. When you say pre-column I assume this is uncoated and deactivated ?. Replace it anyway and trim a metre off the column.

How are the peak shapes - can you post a chromatogram ?

Peter

[tangaloomaflyer]

pre-column is uncoated, and has been changed. Column has been clipped. Peak shapes are good. We replaced the whole MSD analyser door (quad and ion source) with another instrument and the chromatography is a lot better than it used to be. This instrument always struggled with peak shape, I think it is slower than the other instruments to change to the required voltages in SIM mode, resulting in less points across a peak compared to another instrument with identical dwell times and ions per SIM group. Technicians are suggesting we buy a new weldment, and also to run in split because 50psi is too high.

[Peter Apps]

This would go a lot quicker if you supply all the relevant information at the outset rather than feeding it to us in rationed portions. In addition to the information in your original enquiry we now know that the problem followed the larger volume injections, you have already done maintenance on the guard column and a quite drastic change to the MS, the inlet and liner is pretty standard. Sorry to go on about this, but it really does waste time and drag things out trying to troubleshoot with less than half the story.

Either your tech support is talking nonsense or you are using pulsed splitless uneccessarily. If you can go to split and still obtain good signal to noise you should certainly do so, having the pressure in the inlet fluctuating cannot help with repeatability - the only function of the pulse is to drive for sample into the column. Even if you need to stay with splitless in order to get good peak size you should drop the pulse if you can.

With the very narrow column and high inlet temperature you might be overfilling the inlet with solvent vapour. Agilent has a vapour volume calculator on their web site somewhere.

What is the purpose of the 10% toluene ? mixing solvents just makes method less robust to inlet conditions.

Peter

[Yama001]

8270 standard stock mixes are often shipped with some toluene; I believe it helps with keeping some analytes in solution at high concentrations. It is likely not needed for working levels, though this laboratory may have a good reason.

The fluctuating pressure during injection does not look like something done on purpose, it sure looks like that may be a big part of the problem. It could easily be occuring from damage to the pneumatics.

[Peter Apps]

Hi Yama001

I am pretty sure that the pressure fluctuation is deliberate - there is a pressure pulse option on most (all ?) GCs, the rationale being that temporarily increasing the gas flow rate will flush more sample onto the column and increase peak areas. I have never been convinced that its advantages outweigh its drawbacks. There are already way too many pressure and flow fluctuations during a hot vaporizing injection, without adding yet more of them.

Peter

[tangaloomaflyer]

But Peter, if I were to give you all the information straight away that would be far too easy. I'm posting on behalf of a chemist and I didn't have all the facts to hand.

All our extracts have 10% toluene. It is as a keeper when the extract is used in a solvent venting method or to aid in solvent focussing for large volume injection splitless techniques (an onorous method we have to use due to low sample sizes). The internal standard we use is made up in toluene and all extracts get spiked with the same volume. When running the extract on low-volume injection splitless, it is superfluous, but we are a high throughput laboratory and it is more efficient for us to spike all extracts with the same standard.

The pulsed splitless technique has been the norm for us, and has been fine on this instrument up until our recent troubles.

[Steve Reimer]

Do I understand this correctly? You were running 10uL injections pulsed splitless with toluene/dcm as the solvent.
I can see why they want to replace the weldment.

[tangaloomaflyer]

Do I understand this correctly? You were running 10uL injections pulsed splitless with toluene/dcm as the solvent.
I can see why they want to replace the weldment.

Yeah.... don't ask

[Yama001]

I think the pulsed splitless injection is done for good reasons in this case (Peter is correct about the drawbacks). What looks odd to me is "pressure raised from 31psi to 50psi". I usually see the pulse implemented as a static pressure for the first 10 to 60 seconds, rather than a ramp (for example, 50 psi for .5min).

Is the 31 to 50 psi part of the method, or is it just what is observed from the instrument panel during injection? If the latter, it looks like something in the pneumatics is not functioning correctly. The actual pressure will fluctuate as the solvent flash vaporizes, but the inlet pressure reading is usually stable for a splitless injection of 1 or 2 uL. A pressure reading that increases above the pulse setpoint, then decays slowly looks like a problem.

[tangaloomaflyer]

50psi pulse pressure is a set parameter, not a pressure reading.