Change in GC (IRMS) chromatogram after maintenance

[UBC Forestry SIF]

Hello all,

I am new to this forum and relatively new to GC operation. Apologies for cross-postings, and to posting this under the wrong heading!

I have an Agilent 6890N GC, with an injector autosampler and combustion furnace, connected to an IRMS. The system is about 6 years old, but has only seen action for about the last 3 years. I have used the GC for only 3 months.

I have been analyzing fatty acid samples (FAME) dissolved in hexane. I did about 2 weeks of analyses before I saw some peak tailing and stopped to perform some maintenance.

After fixing the vacuum on the IRMS, checking the FID, checking the autosampler syringe, cutting bits off the end of the column (~4 years old), putting in a new septum and checking the liner (clean), and going through a software upgrade (MassLynx to IonVantage), the chromatogram I get on a BAME standard is not the same as before maintenance (or on a GC-MS), despite using the same set-up parameters. I have not changed the furnace tube (~4 years old) or changed the Nafion tube (~3 years old).

Differences I've noticed in the BAME chromatogram:

1. Retention times of peaks are slower despite the slightly shorter column. The slow-down is also not linear between peaks between the new analysis and previous analyses.

2. "Profile" of the chromatogram has changed. Compounds that used to produce the tallest peaks are now shorter and vice versa. This presents a problem since my student wishes to use the peak area to calculate absolute concentrations of various C compounds in her unknown samples. If the peaks now have different heights or areas than what we used to get, we can assume that the calculated concentrations will be different too. It's too cost-prohibitive for her to run a second set of samples on a GC-MS to get peak height data, if we could get both peak height and isotope data from the same run for her (these are natural isotopic abundance samples).

The good thing is that, despite the change in peak heights, the C isotope values of the equivalent peaks of the new and old analyses are relatively the same, as are the isotopic trends between peaks.

The problem is not with my IRMS as samples analyzed via another connected peripheral device yields satisfactory results. I don't think this is a software issue either, so it must be a hardware issue with the GC or GC combustion furnace.

I would appreciate any ideas or suggestions or similar experiences anyone may have. Thanks.

Alice

[bhuvfe]

Alice,

Longer retention times are generally indication of a leak. Since you did some maintenance at the inlet of your instrument that's the first place to check (septum, connections, etc..).
I suppose you cannot see an air leak with your IRMS. You can use either a leak detector or run a inlet pressure drop test described in the manual of your 6890.

No clue about question number 2 ( I have no experience in IRMS..). An inversion in peak heights might be due to some problem in your split injection (if it is split injection) or if you are in splitless with a too long splitless time.


''The problem is not with my IRMS as samples analyzed via another connected peripheral device yields satisfactory results. I don't think this is a software issue either, so it must be a hardware issue with the GC or GC combustion furnace''

Which peripheral device?

Good luck.

bhuvfe

[CE Instruments]

You have a four year old column for which you analysed how many samples in two weeks before you saw tailing peaks ? Although you chopped of the ends of your column are you sure it does actually still work ? Although not an expert on Fames analysis I do remember that column life can be short, very short if the sample has not been prepared perfectly as excess derivatisation agent will eat your column. If it is not a leak , try swapping the column for a new one.

[Peter Apps]

Does your problem number two relate to retention time ?, in other words is it early peaks that have got shorter and fatter and later ones taller and thinnner or vice versa, or are the height and width changes independent of retention ?

Is there any way with an IRMS that you would detect an injection of light hydrocarbon gas - methane propane or butane. If there is it would be useful to check the gas flow through the column by measuring the retention time for one of these gasses - a cigarette lighter is a convenient source.

Does you inlet liner have glass wool in it ?

Peter

[skunked_once]

checking the liner (clean),

Just because the liner looks clean doesn't mean that it is clean enough for you analysis. When I start to see tailing peaks on my FAME (with FID) a fresh liner with glass wool almost always sharpens up the peaks. I have also found that the gold seal may require replacement if the new liner doesn't help. Occasionally I do have to cut off a meter from the inlet end of the column also.

[chemwipe]

Just because the liner looks clean doesn't mean that it is clean enough for you analysis. When I start to see tailing peaks on my FAME (with FID) a fresh liner with glass wool almost always sharpens up the peaks. I have also found that the gold seal may require replacement if the new liner doesn't help. Occasionally I do have to cut off a meter from the inlet end of the column also.

^^^^This.

I'm surprised you're using such an old column.

And when doing maintenance, you mentioned "cutting bits off the end of the column" - how much are you cutting off? Try trimming 2-3 feet off of the column.

Hope that helps.

John

[UBC Forestry SIF]

Thanks everyone for your responses. I am using the GC in splitless mode.

I will have to check for leaks in the inlet area. The septum is new. I will put in a new liner too. The one I have in there now looks new after ~100 injections, compared to the one I took out (hundreds of septum bits) used by a previous operator with apparently no problem during manual gas injections. The liners I use are constricted at the bottom and have a plug of glass wool.

I haven't thought of cutting of part of the column at the inlet end. By cutting off "bits" I meant centimeters (being very conservative with my trimming!). I do have a new column on standby; perhaps it's time to try that. I may end up installing a new combustion tube too.

Early peaks (C13:0, C14:0) tend to be slightly taller than before. The later peaks (C18:0 to C20:0) are much taller than before. The peaks in the middle seem to be the same as before, exccept the short ones are shorter than before. The peak widths have not changed between then and now.

The other peripheral that feeds into the IRMS is an elemental analyzer.

Alice

[AICMM]

Ms. Chang,

The response symptoms you describe could very easily be attributed to a difference in the inlet liners. More wool than before could have improved the vaporization of the heavy constituents and given them much more response than before. Column position is very important but would not really cause the problems that you relate here (in my opinion.)

Best regards,

AICMM

[Peter Apps]

Hi Alice

Can you give us some hard numbers on how much the peak shapes and retention times changed.

What are the dimensions of the column, and what is the stationary phase ?

It would be helpful if you could post an example of before and after chromatograms.

Peter

[UBC Forestry SIF]

Hi Peter and AICMM,

I have a splitless liner inserted (Agilent p/n 5062-3587 for splitless GC mode), which worked well before, so I have a feeling my problems are due to the way I maintained and re-fitted the column. My HP-5 capillary column was 30 m long, with 0.32 mm i.d. and 0.25 um film thickness. During maintenance, I cut off a metre length at least. The previous operator may have also cut off some length, so I don't know the true length of the column now.

Peter: How do I post attachments to the forum? Otherwise I can e-mail you the files.

Thanks,
Alice

[Peter Apps]

Hi Alice

Instructions on how to post chromatograms are in a sticky at the top of each sub-forum page.

Peter

[UBC Forestry SIF]

Hi Peter,

Unfortunately I don't have my image on the internet, as the instructions suggest. May I e-mail you the chromatograms instead (pdf)? I know others won't be able to see it however.... I also have Excel files for numerical results.

You can find me at: http://web.uvic.ca/~tfp/People.html if you want to send me a message first.

Thanks,
Alice

[Peter Apps]

Hi Alice

I would really rather not get into one on one consulting. You can set up an account with an image hoster quite easily and for free I believe.

Peter

[AICMM]

Ms. Chang,

How far up is your column installed in the inlet? With the liner you describe, it should be just above the ferrule by about 2-3 mm.

What happens if you just pull the wool? (Probably more of a diagnostic than a solution...)

Best regards,

AICMM

[UBC Forestry SIF]

I put about 5 mm of column above the ferrule going into the inlet (Agilent manual suggests 4-6 mm). I don't think I will take out the wool in the liner. I don't want any septum bits falling straight through or clogging.

I just installed a new HP-5 column and that seems to have solved the problem of the change in relative peak heights, but now I am getting severe peak tailing. The column was conditioned, and the GC-IRMS fully settled before I started analyzing BAMEs. Anyhow, another forum topic, I suppose.

Thanks all.