Peak integration Question

[Samba_Zero]

Hi,

I have integrated the peaks in my chromatogram and wanted to ask here if I have done this correctly.

Here are the chromatograms of three different measurements. Each measurement was retrieved three times for reproducibility.

Now my question, what is the correct method for integrating the peaks?

Unfortunately, I also have a relatively strong noise floor, so I am not quite sure how to proceed with image 2 and 3, but I think that image 2 should be integrated correctly.

For image 1 I had it done automatically, I'm not sure if it can be done that way and whether I should proceed in this way?

The analysis is to quantify and qualify diverse carboxylic acids.

[rb6banjo]

To me, the key to integration is doing it the same way every time. There's really no other secret or formula.

Carboxylic acids are tough because they don't always chromatograph really well/cleanly and as the column ages, how they interact with the stationary phase changes as well.

Good luck.

[Samba_Zero]

Thanks, indeed, carboxylic acids are not quite trivial in chromatography.

Currently, I only have an older WAX elite column at my disposal.

What do you say about the separation performance, how could this plateau be interpreted?

[rb6banjo]

It looks like you have an analog detector. Do you think it is your analyte? If you inject a sample/standard with larger concentration, does the plateau move with the peak?

If yes, you could try to integrate it all as one thing. If no (the peak gets larger and the plateau sort of disappears), then you need to work on your separation.

[Samba_Zero]

Yes, the detector is a FID.

Move along, you mean that the plateau increases in the direction of the peak?

When the concentration of my standards becomes larger, I think the plateau is not moving along in the height with the peak.

Here I have pictures of different concentrations:

First pic is a concentration of 10 mg/ml and second is about 1 mg/ml ... the zoom is a bit different.







And I wouldn't say that the plateau stretches over time in the chromatogram as the concentration gets larger either.

But at the moment I can't say that with 100% probability.


Ok, so if I have to improve my seperation, which possibilities should i consider? I mean temperature, split or other column, what would have the best effect for carboxylic acids?

Seems like I will have to do some research on that as well.

[sbashkyrtsev]

Since your peaks are not fully resolved, you can only guess how where you to put your baseline and peak boundaries. It's possible that there's yet another, third peak which is wide and overlaps with both of them. So depending on how important the precise number is, you can either just accept your guess or keep improving your method until peaks are fully resolved.

But notice that your peak is cut at the top:
This means that the concentration is too high. The detector can't measure concentrations that large, so you have to dilute your sample/split the flow to get any meaningful area.

[Samba_Zero]

Since your peaks are not fully resolved, you can only guess how where you to put your baseline and peak boundaries. It's possible that there's yet another, third peak which is wide and overlaps with both of them. So depending on how important the precise number is, you can either just accept your guess or keep improving your method until peaks are fully resolved.

Thanks for the reply, an overlay from another peak would not be excluded.

However, the chromatograms in Fig. 2 and 3 are only the standards, that is, a pure substance that we had ordered for the measurement. I would almost rule out contamination because this measurement was repeated several times and always gave the same result.

Could it be that the steep rise of the baseline is a mass or temperature gradient?
(I mean the rise at the point just before the big peak comes from the standard.)

But notice that your peak is cut at the top:
This means that the concentration is too high. The detector can't measure concentrations that large, so you have to dilute your sample/split the flow to get any meaningful area.

The peaks are not cut off, this is due to the zoom and the display of the software.

[AICMM]

Samba_zero,

I see three issues with your chromatogram and do not speak directly to your integration question (my apologies.)

First, for an FID, you seem to have a lot of noise. The pattern in front of the large peak suggests an electronic noise problem of some sort since it repeats (although I would expect it after the peak as well....?) If you are watching the FID signal, how many pA's of swing are you getting when sitting idle?

Second, it appears to me you have a peak width threshold (or slope sensitivity ) that is inappropriate since you are calling out lots and lots of peaks that are really just baseline noise.

Finally, the fronting of your peak is not really an integration issue, it is a chromatographic issue. IF this were a valve injection (I am betting it is not...) I would say you have a leaky rotor but something is certainly getting on column when it should not be...

Fix the chromatography issues first, then tackle the integration.

Best regards,

AICMM

[James_Ball]

Curious on that plateau. Is this a manual injection and if so is there a pause between when the needle is inserted and when the plunger is pressed?

[Peter Apps]

The plateau in front of the peak with a little spike at its leading edge is a classic solvent effect - what are you injecting, what volume and under what conditions (inlet split ratio, inlet and column temperatures)?

Don't worry about the flat top on the big peak - it is just the plot going off the top of the chart.

Peter

[Samba_Zero]

Samba_zero,

I see three issues with your chromatogram and do not speak directly to your integration question (my apologies.)

First, for an FID, you seem to have a lot of noise. The pattern in front of the large peak suggests an electronic noise problem of some sort since it repeats (although I would expect it after the peak as well....?) If you are watching the FID signal, how many pA's of swing are you getting when sitting idle?

Second, it appears to me you have a peak width threshold (or slope sensitivity ) that is inappropriate since you are calling out lots and lots of peaks that are really just baseline noise.

Finally, the fronting of your peak is not really an integration issue, it is a chromatographic issue. IF this were a valve injection (I am betting it is not...) I would say you have a leaky rotor but something is certainly getting on column when it should not be...

Fix the chromatography issues first, then tackle the integration.

Best regards,

AICMM

Thank you for the suggestions, there may be something wrong with the instrument, unfortunately I can't change or even repair much on the instrument myself, I would have to discuss this with the lab manager first.

The autosampler makes the injection method via the injection port. I can't say more about this at the moment, I can't just look inside the device to check for leaks.

No manual injection!

With standard



rinse


however, I had made several flushes, I can no longer say exactly between which measurement this flush was made?


Is it possible that some of the carboxylic acids absorb onto the column and gradually detach from it again on the next measurement?

The plateau in front of the peak with a little spike at its leading edge is a classic solvent effect - what are you injecting, what volume and under what conditions (inlet split ratio, inlet and column temperatures)?

Don't worry about the flat top on the big peak - it is just the plot going off the top of the chart.

Peter

Thank you, its a WAX Elite column at 240 °C, split and volume I must check don't know the data at the moment.

What is a solvent effect, can tell me please more about this?


Ok I iwill first try to find someone to help me with the hardware. should i also use another column?

Best Regards

[Peter Apps]

Solvent effects are what happens when there is enough solvent introduced to the column to affect the retention of the analytes. Usually it messes up peak shapes, but it can be exploited to focus peaks into very sharp bands, and for analysing large, dilute samples. In the 1980s it was a hot topic - I did my doctorate on it.

[AICMM]

Samba_zero,

The scale on your chromatogram, e5? FID's are very, very low baseline current detectors, and e5 makes me think your FID is running a very high background.

Possible to post what the raw signal is at the GC in idle state along with the units. An FID should run 0.6 to maybe 15 pA's depending on the quality of the gases. Much higher than that and you have an FID issue.

Best regards,

AICMM

[Samba_Zero]

Samba_zero,

The scale on your chromatogram, e5? FID's are very, very low baseline current detectors, and e5 makes me think your FID is running a very high background.

Possible to post what the raw signal is at the GC in idle state along with the units. An FID should run 0.6 to maybe 15 pA's depending on the quality of the gases. Much higher than that and you have an FID issue.

Best regards,

AICMM

Thanks again, I have a chromatogram here that corresponds to an idle where only heated out, on this one, no analyt was abandoned.

The background that leads to the increase of the baseline. What could it be, I have very little practical experience with GC, please excuse my stupid questions.