Decrease in peak areas

[saana]

Hi,

I'm relatively new to GCs and would greatly appreciate some help..

We have been analysing some metal extractant reagent samples with our Agilent 5975C GC/MS. Column: Agilent HP-5ms, column flow 1 mL/min, injection: 1 uL, split 75:1, syringe washes with acetone and hexane. The sample contains kerosene as a diluent, the reagent (organic molecule), the degraded reagent and some modifiers. The samples have been stripped of metals.
The aparatus is quite new (got it in spring 2011) and we have no one here with real expertise regarding it.

We look at the ratio of the reagent and the degraded reagent (both come out as peak groups, which are then integrated). The results expressed as ratios are quite consistent between repeat samples. However, I have noticed a decrease in peak areas. I would not be too worried if the area fluctuated slightly between the samples, but now there is a clear decreasing trend. I have run a series of about 10 analyses with just a std sample and have also run the std sample between the actual samples. I noticed that especially if there have been other samples between the std samples the decrease in area is more noticable.
Could this be a sign that there is something wrong with the GC/MS?
Similar analysis has been done at another facility, but they had a GC/FID. Apparently the area of the peaks stayed quite constant in their analyses.

I have changed the syringe (thought that there was a blockage build up) and tried a different column. Neither helped.
Could it be that the MS is getting dirty and the sensitivity of it is decreasing? And so, it is just a characteristic of the MSD. If so, the FID would not have this problem, since I don't think it get dirty as quickly as the MSD.

As I mentioned before our results are expressed as ratios, and this problem does not seem to affect them. I am just concerded about our system and that is it possibly being harmed in some way and this problem a symptom of something and could there be worse problems up ahead.

Any ideas?

Thanks in advance,
Saana

[Peter Apps]

Hi Saana

Welcome to the forum.

Your suspicion that the MSD response is decreasing is a very good one. Routine maintenance such as source cleaning on the MS and inlet liner and septum changes on the GC are an absolute requirement to maintain instrument performance.

Peter

[saana]

Thanks for the quick reply.

We have been changing the liner and septum about every 150 injections. And they haven't looked too damaged or dirty at that point.
I'm still wondering about the frequency of cleaning the MSD. We were advised that it should be done yearly or if something is the matter. We don't run very many samples (less than 80/week and sometimes 0/week). Does that sound about correct?

-Saana

[Peter Apps]

A lot depends on the nature of the samples - if there is a lot of heavy muck or reactive compounds then the source gets dirty more quickly. At 80 samples a week I would think that a year between source cleans is rather long.

How often are you tuning the MS - besides a dirty source (which generates high noise as well as low signals) a decline is response can be due to a deteriorating ion multiplier, it is normal for the voltage on the multiplier to be increased with each tuning to compensate for this.

Peter

[saana]

We check the tune weekly (Agilent Quick tune- function) and do an actual tune about every 3-4 weeks and after maintenance. This is what was recomended to us. I still don't really understand the difference between the functions- is it that the quick tune does not save the checked parameters?
I've been wondering why even do the check if you can just tune it everytime? Especially since it does not take any longer to tune than to check.

I haven't noticed anything alarming in the tunes and everything seems in be order with them so far. The EM voltage has been increased slightly.

The samples we are running at the moment are slightly mucky and it is possible that they are making the source dirty.. However cleaning the source too often seems like a waste of recources and time, especially if the results (ratio of compounds) we get are constant. The source was cleaned at the start of November and the system was not in use full time at the end of the year.

-Saana

[Peter Apps]

If the source clean did not solve the problem, then a dirty source was not the problem.

I am not quite clear whether you are seeing a long term decline in peak areas over periods of days or weeks, during which multiple batches of samples and standards are run, or whether the peaks at the end of a batch are smaller than those at the beginning.

A deteriorating column can also make peaks smaller, especially if the analytes suffer adsorption, has the tailing or the peak width increased as the peaks have got smaller ?. Are you measuring peak height or area ?

Peter

[nvalentine]

Forgive me if I am stating the obvious, but to be clear, peak areas will gradually decrease over time on a GC/MS. This is part of why using internal standards is so important. The area may change, but, relative to the internal standard, the concentrations will not. This is under normal conditions (valid tune, clean source, etc.).

If your areas are dropping dramatically run to run, then you probably have an issue. The fact that your ratios are still fine would seem to point to dirtying of the source. You can retune all you want, and the system will "work", but until you clean everything you won't get that response back. It would make sense that the FID would be more stable as far as the peak areas go, since you are simply burning everything that comes through and not worrying about fragmenting/ionizing and the associated matrix issues.

[aldehyde]

It sounds like you're doing a good job with the way you're tuning. The software keeps a historical record of your tune values (in the tune and vacuum control window go to file > view tunes and a series of plots will open. Also a spreadsheet version of the values will pop up.)

EM volts will increase over time because the EM has a lead oxide coating that is used up over time, requiring an increase in voltage to have the same peak abundances. A big jump in voltage can indicate a leak or other problem.

Repeller voltage is also a good thing to keep an eye on, the repeller pushes ions through the quadrupoles and the tune may increase the voltage to keep the same peak abundances.

When you do inlet maintenance do you find that your liners are visibly dirty? Have you changed the gold seal and split vent trap? You may want to wet a cotton swab with solvent, remove the liner and rub the swab against the wall of the inlet. If it comes away dirty you are contaminating your inlet. As sample reacts with the metal of the inlet wall, or if the liner gets dirty enough, you'll create active sites that can react with your sample and decrease sensitivity.

If you find that it is dirty or if you change the split vent trap and find grime there is a short length of copper tubing that is connecting the inlet and split vent trap. I had an instrument once where the user would clean the source, do inlet maintenance etc and then they would try to run in CI mode and within 10-15 injections their peaks of interest were basically gone. The split vent line was clogged with black goo and even though they were doing splitless injection when the valve would open to purge the inlet some of the goo would make its way back into the inlet and contaminate it, rip up them column, and dirty up the source.

I doubt you're dealing with anything as severe as this, but its possible that over time some contamination is accumulating somewhere and you are starting to see the effects. The majority of problems that slowly present themselves are either in the inlet and possibly the syringe (I'd also say the source but you've checked that.) Try a new syringe and see if it makes a difference, it is an easy thing to check if you have a spare. I keep a known good syringe for troubleshooting and it has saved me a lot of time.

[saana]

nvaletine: sometimes it is good to state the obvious, especially when a newbie is listening

I measure the area of the peaks, not the heights. It is very difficult to see changes in the peak shapes because the compounds we are looking at come out as peak groups (over a minute or so) due to the characteristics of the compound (saturated carbon chain attached to the benzene ring). And so I can not see a visible change between the chromatograms, but there is that trend in decreasing peak area. But the decrease in area is not sudden, rapid or dramatic, so I am not too worried by it.

It seems to be a longterm decline in the peak areas.. which could indicate that the MS is just slowly getting dirty. The GC/MS was in very little use before the source was cleaned (it was cleaned as a part of a training session) and I have mostly been following the peak areas after this. However it is also possible that the inlet is getting dirty. So, in the next maintenance I will check the split vent trap, the gold plate and also check if the wall of the inlet is dirty.

This problem was mostly raised to my attention when the researcher who gets the results from our lab noticed that the areas do not stay stable. He had been used to looking at the results from the GC/FID analyses, where the peak areas are more compareable between the analyses. This raised questions of if the results are correct or not. The hunch that with GC/MS the analyses are not comparable in the same way as with GC/FID (ie. comparing the absolute areas of the peak) got some confirmation from you guys. I tried searching literature, but I was still a bit uncertain with my deduction.

I think I will be revising the cleaning schedule of the GC/MS, so that the GC/MS is checked more frequently... just in case..

Thanks for the insight into this matter

-Saana

[MMJ88]

This problem was mostly raised to my attention when the researcher who gets the results from our lab noticed that the areas do not stay stable. He had been used to looking at the results from the GC/FID analyses, where the peak areas are more compareable between the analyses. This raised questions of if the results are correct or not. The hunch that with GC/MS the analyses are not comparable in the same way as with GC/FID (ie. comparing the absolute areas of the peak) got some confirmation from you guys. I tried searching literature, but I was still a bit uncertain with my deduction.

Hi Saana,

Forgive me if I am explaining something you already know, but just to be clear, on GC/MS, you do not want to use the peak area numbers to compare between analyses. You should be running your samples with an internal standard. This is something that has similar chemical properties to your analyte(s) which you add to each sample in a known amount. You then take the ratio between your analyte peak area and your internal standard peak area (referred to as the "response ratio") and use that. You can't count on the MS response to be the same every time you run an analysis. That's why you should calibrate with every batch and use an internal standard.

-Matty

[saana]

No worries Matty. I initially didn't compare the areas of the peaks between analyses. For some reason the person who reveives the results from us wanted to. I had not even noticed the decline in the area before he mentioned this because I assumed that you can not compare the analyses in that way. But then this trend got me a bit worried and I decided to ask people who have expertise with these systems.

In this case we are looking at the fraction of the degraded reagent. So, our results are of the form:
fraction of degraded reagent= (area of degraded reagent)/(area of degraded reagent + area of reagent)
fraction of reagent = (area of reagent)/(area of degraded reagent + area of reagent)
and of course the sum of these two =1.

We are not looking at concentrations but the degree of degradation.
However, if in the future we do do analyses where we have to determine the concentration we will use an ISTD.

-Saana