Agilent 5890 peak area doubling?

[bisnettrj2]

Good evening GC people:

A coworker of mine is having an odd problem with her Agilent 5890 GC/FID. She noticed over the past few sequences that occasionaly an injection would occur that would result in a 200% recovery of her surrogate (in an extracted sample) or standard (calibration check), and subsequent re-injection would result in the expected amount. This problem has now become the norm - almost every injection is now almost perfectly double what her original calibration of the instrument indicates it should be. She changed the inlet liner and septum; the syringe; the injection tower; and finally the controller box. The problem is still occurring, and has not been resolved by any of the changes she made. She is running a splitless injection of diesel and motor oil standards and samples. She called Agilent, and they suspect her detector, but I would expect a problem in the detector to be more variable (e.g. not always double).

I know the post does not supply enough info, a cardinal sin of the Forum. However, I'm not well versed in GC, so I'm not sure what is pertinent in this post and what is not. So, if anyone could give me some ideas as to what might be causing this, or what information is needed to make more informed guesses, I would appreciate it.

[Peter Apps]

Are all the peaks on the chromatogram doubled, or only the surrogate and standard ? If all the peaks are double the problem lies somewhere in the GC, if only the surrogate then it is probably a preparation issue (although then it is difficult to explain how the problem goes away from run to run).

How close to double is the change, and how repeatable ?. On some older instruments there was an attentuation setting that changed signal by powers of 2.

Check the position of the column in the inlet - not likely to be a problem with a splitless injection, but worth a look.

How repeatable were areas before the doubling started ? It is not beyond the bounds of possibility that there used to be a leak (at the infamous gold seal for instance) that is now fixed.

Check the septum purge - if this is crudded up you will have less flow during the "splitless" phase, and more sample going to the column.

How does the solvent peak look - is it wider on doubled injections ?

Peter

[larkl]

As soon as I saw "motor oil", my first thought was plugged gas line. I'd check your split vent and septa purge lines.

[bisnettrj2]

Peter/larkl:

Thank you for your responses.

My coworker showed me an example of what she was talking about in terms of doubling - the solvent peak is larger, and the peaks were almost exactly double. She was seeing 150% recovery of a surrogate, where she would normally see 75%, and a calibration check would come out at 100% high, but a reinject of the same vial would come out "normal". So, it seems like double the sample is making it to the column, at first infrequently, but now almost every time she injects.

My thoughts went to a flow problem as well, and I inquired about the split/splitless line yesterday. She has a new split/splitless line that she intends to install, but she claims she removed the old one and ran solvent through the line to clean it out recently, just in case that was the problem. I'm not sure how easy it is to clean one of those lines, or how effective her cleaning job was,.

I'm not sure how well the GC performed before this started occurring. She is running the same method when the doubling occurs and when the injections are normal, so I'm not sure how an attenuation setting would change in between runs, and the problem is occurring at an increasing rate - happening more often as time passes.

Peter, it's funny that you mentioned a fixed problem being the reason for the doubling, as I suggested the same thing to her yesterday. I couldn't explain, how it would be fixed one injection, and then bad the next though. Although, if it was a partially plugged split line, maybe she got it mostly cleaned out, and after re-installing, the increased rate of "doubled" injections means the line is now completely clear (or now completely blocked)? Or she had a leak somewhere that is now plugged?

I'll inquire about the condition of the split vent and septa purge lines today, and I will inquire about her standard prep, and see if she can trim and re-install her column, just to be certain it is installed properly. I'll also ask about the gold seal condition. Thanks for the input.

[aldehyde]

Is she injecting 1 uL? Have her open the door on the autosampler and look at the white plastic "steps" above the syringe. When the plunger is raised there is a piece that fits against the steps and that is how the ALS knows whether it is injecting 1 uL or 2 uL or 5 uL etc. I wonder if it is incorrectly drawing up 2 uL of sample?

[Bigbear]

Don't forget the split solenoid. If there is no trap installed between the injection port and the soleniod it could be contaminated. Try cycling it to see if it opens and closes consistantly. As I remember you can address the split valve on the gc keypad. Then cycle from on to off and listen for the click (under the left side cover).

Don't ask how I know!

[aldehyde]

Yeah my first suspect is also the split vent line. I wouldn't worry as much about the solenoid, since its splitless if it was stuck open or closed they would probably notice (contamination/ghost peaks if stuck closed, reduced sensitivity if stuck open). HOWEVER, I have replaced so so many partially and completely clogged split vent lines and almost every time the symptom is different. As a result of clogged split vent I have seen:

1. extremely high baseline MS baseline / column bleed / sensitivity reduction with each injection.

2. incorrect split flow, this was caused by the user's boss injection 1 mL instead of 1 uL while she was on vacation.

3. reduced sensitivity for a particular analyte but no other obvious symptoms.

4. cycling pressure which would never settle on the setpoint ONLY AFTER the first run would finish and the GC would try to return to the method's start conditions. if you turned it off and let it sit for awhile when you turn it on it would looked fixed until you would make the run. we also saw higher than normal column bleed on this one (another MS)

in most of these cases the customer was analyzing oily plant extracts, or pestacides. very common with high boiling/oily compounds that can back flash into the line and then get trapped since it is not heated.

[Peter Apps]

A block in the split vent line is not likely to double the peak areas unless the splitless time is set so short that it was originally venting half of the sample, which would not have been repeatable.

That leaves the septum purge, which (AFAIK) on a 5890 (and other GCs of that vintage) stays open during a splitless injection - effectively giving a split ratio of about 1 or 2:1, depending on the flow. It is easy to check for septum purge flow from the relevent vent.

Peter

[Bigbear]

Another thought is the injection port. Have you tried injecting solvent only? If youu get the peaks of interest they have been traped in the inlet. You can then either solvent scrub the inlet or change it. There is an instruction paper on cleaning on the Agilent website. I have cleaned my injectors many times, but they are consumable and have to be replaced at some point.

[bisnettrj2]

OK, first a MAJOR correction on my part - it's a 6890, not a 5890. Shows what I know. I'm not sure if that leads to any other ideas, or discounts anything pitched so far, but I apologize for the confusion.

Update so far:

Changed the split line; peaks are still 'double'. Swapped the trap filter from the split line with one from another GC. Still 'doubling'. When the problem first occurred, the gold split seal was changed, along with all the aformentioned fixes.

She has injected both solvent and air, and does not get any peaks in the chromatogram.

She has watched the injection to make sure the 2uL she is injecting is in fact what the syringe is picking up - and it is.

If the GC people had a flow meter (which they don't, and I find that odd), I could measure the flows for all relative gasses. However, I can't.

Here's the kicker - my coworker gave up on fixing the issue, and has accepted it as the 'new normal' for this instrument, and has recalibrated to take into account the problem. Also, the analyst who previously used the instrument told me this morning that the problem happens on and off, and has for the last couple of years. Ugh.

As my coworker has given up on trying to fix it and it's not my instrument, I'll have to drop the issue. Thanks to everyone who pitched an answer - I'm frustrated to leave the question unresolved, but unresolved it will be - for now. Until the doubling returns to half at some point in the future...

[Peter Apps]

As long as she is using the internal standard to correct for whatever is going (drastically) wrong I suppose that the erratic behaviour will not influence the results. It would be tricky justifying it to a QC auditor though.

Peter

[bisnettrj2]

Unfortunately, she's not using an IS (although I suggested the same thing to her - that it would correct for the issue). And I agree, if/when she gets audited, it will be very hard to explain. Sigh... thank goodness it's Friday.