GC/Ms Peak area drop

[chronomag]

Greetings folks!
I've come across the forum many times in searches for help, and finally joined up to ask about a confounding problem.

My lab has had a new 5977E with 7820 for about 6 months.
We mostly use it for PAH and OPAH analysis at low ppb levels (lowest standard = 2 ng/ml). Peak areas for these ~2000 with RTE integrator.

The sensitivity of the system has always been good, but there is a strange phenomenon I have noticed lately, with the peak areas dropping over time for the same standard/sample.

It's probable that this has been the case since the install, but only lately have we been really testing the %rsd over time.
The brief specs of our method are:
60m HP5 MS UI column
inlet 300
injection pulsed splitless 25 psi
2 ul injection (verified ok by liner volume calc)
solvent dcm
source 300
transfer line 280
oven 60-300 over 65 minutes


We just had a tech out here to check the system, clean the source (I had just done this as well), get the atune to verify, run checkout standards, etc.
The system can get good %rsd (1-3 %) over 24 hrs, no drift, with test compounds like azobenzene and octadecane.

But when we run a labile compound like malathion, or our 16 PAH mix, we get larger %rsd of 5-10 %, with a drift downward that is about -5% as a rolling average.
I haven't run enough injections to see how far the drift will go (down to zero?), but so far about 20-30% over 24 hours is typical.

What are the likely reasons for this type of peak area decrease?
The Agilent engineers are out of ideas for now...

Thanks in advance

[Rndirk]

Hello,

Could you add a chromatogram of a PAH mixture? Preferably a good and a bad one of the same mixture.

It sounds like loss of analyte or excessive tailing to me. PAHs are not labile compounds but they are prone to active sites in your system: dirt, leaks, damaged column,.... anything really negatively impacts the peak shape and area, especially of the heavier ones.

Did you lately:

- Check for leaks
- Replace the liner
- Clean the inlet
- Trimmed your column inlet

[chronomag]

Hello,

Could you add a chromatogram of a PAH mixture? Preferably a good and a bad one of the same mixture.

It sounds like loss of analyte or excessive tailing to me. PAHs are not labile compounds but they are prone to active sites in your system: dirt, leaks, damaged column,.... anything really negatively impacts the peak shape and area, especially of the heavier ones.

Did you lately:

- Check for leaks
- Replace the liner
- Clean the inlet
- Trimmed your column inlet

Thanks, all of the above maintenance was performed, and more.
I can post the chromatograms once I figure out how to do that.

I have read posts on the forum about problems with splitless injections...One thing, based on tests overnight, I am seeing no improvement after switching to a split injection. The split method was 4:1 with a 1 ul injection.

[Yama001]

It may be useful to solvent rinse the inlet lines (yes this can be a big old pita to do). There seems to be a lot of interaction with the upstream plumbing, even with a pulsed injection.

[Rndirk]

It may be useful to solvent rinse the inlet lines (yes this can be a big old pita to do). There seems to be a lot of interaction with the upstream plumbing, even with a pulsed injection.

Can you elaborate? Do you mean there is a stream of 'stuff' besides helium flowing trough the system?

Maybe it would be useful to check your MS in full scan, see what happens if you increase/reduce the inlet flow, cold/hot inlet etc..

[chronomag]

The engineers have a suggestion besides cleaning the inlet (it isn't that old or likely very dirty, but it might help a bit)

The plan is to do an exact mass test, since the target masses might be off just enough to miss the signal just a bit more each time there is a new injection. Slight deformities in the peak shape are a sign this might be happening as ions drop out randomly.

Start with the SIM ions in the table for the PAH method, say one group is just 272.0
Change that entry to include 272.0 +/- .3 amu
So enter 271.7, 271.8, up to 272.3 in that row, call this an amu group.

Do this for each SIM mass in the table.

Inject a standard mid-high range, and run in SIM.
Look at the result by extracting data for all ions in the amu group using the chemstation extract ion feature, overlay all of the ions for each SIM ion in the amu group (one row at a time of course).

Each ion will have an abundance, take the one that has the highest abundance for each and make that the new target ion for normal analysis. Say 272.1 is the largest, then it is now the target
Redo the table with only the optimized mass entries.

The thinking is that taking the ions from SCAN mode originally and taking the larger abundance peak (s) is not exact enough, and finding the most abundant ion this way will improve the precision.

[Peter Apps]

If the mass calibration is drifting, and if you are doing the azobenzene and C18 in SIM mode (which you need to be if you want to compare their results with SIM runs of PAHs) then you should see a decrease in their peak areas also.

A quicker way to check whether mass drift is the problem is to run in full scan - then the masses do not matter, you are just counting total ions.

Peter

[Rndirk]

Correct me if i'm wrong, but isn't the autotune function supposed to correct that? Especially for ions around the fragments of the calibration gas.

It seems however a plausible explanation for your problem. Checking in full scan is a good idea. Make sure you also run a solvent blank in full scan, if you're not used to using your GC-MS in full scan you might see some weird stuff...

[chronomag]

I do run full scans, normally in method development when adding new compounds to the method, or when looking for interesting unknowns in samples.
Our normal sample method requires 2 ng/ml and up to a max in the hundreds of ppb, and full scans do not detect PAH along that range.

If I run a repeat test in full scan and see no area drop, that might indicate a mass drift, and if there is no area drop, then more likely we have an inlet problem.

is that right?

[Peter Apps]

I do run full scans, normally in method development when adding new compounds to the method, or when looking for interesting unknowns in samples.
Our normal sample method requires 2 ng/ml and up to a max in the hundreds of ppb, and full scans do not detect PAH along that range.

If I run a repeat test in full scan and see no area drop, that might indicate a mass drift, and if there is no area drop, then more likely we have an inlet problem.

is that right?

If it is ad/absorption in the inlet or at the head of the column that is reducing the peak area then it will have a proportionally greater effect with the small quantities that you need SIM for. So a lack of drift with larger quantities and full scan will not tell you whether it is an inlet or a detector problem. A mass calibration problem will affect all compounds run in SIM, not just PAHs and malathion. An inlet problem will affect only the intractable problems.

Peter

[chronomag]

A mass calibration problem will affect all compounds run in SIM, not just PAHs and malathion. An inlet problem will affect only the intractable problems.

Peter

Peter, if that is the case, then there isn't a mass calibration problem. The other two compounds, azobenzene and octadecane in the Agilent mix, did not have peak area drop over time.

So what would be the point of running in full scan?

[chronomag]

A mass calibration problem will affect all compounds run in SIM, not just PAHs and malathion. An inlet problem will affect only the intractable problems.

Peter

Peter, if that is the case, then there isn't a mass calibration problem. The other two compounds, azobenzene and octadecane in the Agilent mix, did not have peak area drop over time.

So what would be the point of running in full scan, or in doing an exact mass test?

[Peter Apps]

Peter, if that is the case, then there isn't a mass calibration problem. The other two compounds, azobenzene and octadecane in the Agilent mix, did not have peak area drop over time.

So what would be the point of running in full scan, or in doing an exact mass test?

Now that we have established that the test mix tractable components are not affected in SIM (which we did not know before) there would be no point in running full scan or doing an exact mass test - a classic case of the more you tell us and the sooner you tell us, the more we can help and the quicker we can hep.

Merry Christmas !

Peter

[tdgcuser]

Have you looked into Agilent's Jet Clean Source cleaning? We've had the same problems as you. In our case, a source cleaning always returned sensitivity to previous levels. Agilent approached us about trying the Jet Direct, mostly because we had a 77/75 with CI available for troubleshooting. I have to admit, we haven't gotten around to it yet. The paperwork is sitting on my desk. Give it a look, see what you think.

https://www.agilent.com/cs/library/appl ... 7254EN.pdf

[01310040231]

You do not mention how many injections you had made until you noticed this problem. Also, please state what type of cleaning you perform to prepare your samples and what is your gain set at compare to your tune. As you mentioned, the sensitivity of 77E is very good but you are making a 2ul injection. I am told by Agilent that the 77E are 10 times more sensitive than the regular 77s. I think you should be able to see 2ng/ml with a 0.5ul injection which if it works for you; it will reduce the amount of maintenance. Do you have another column to try to see if your column is still good? You may also consider using DB UI 8270D column.