Test solution for co-eluted analytes

[avitan]

In-house method for GC-FID analysis of 10 alcohol volatile congeners on WAX column.

Active amyl alcohol and Isoamyl alcohol are co-eluted on WAX.

I have following dilemma in test solution preparation:

Should one add both co-eluted analytes into the test solution or it's enough to perform determination for both analytes as their sum vs. one of them?

I know that regulator allow to determine this pair as a sum. It seems not reasonable to me to take both analytes for calibration: peak will have much higher response than others, it complicates calculations. But in official method they put them both into the test solution. Which way is preferable and why?

Thanks in advance.

[chromatographer1]

I would follow the official method.

Why?

Justification is not required. Anything you say may be held against you [even if you are correct] if you go against the official method.

Argue less. Do more work.

best wishes,

Rod

[Peter Apps]

If calibrating a peak area against the sum of the amount of two analytes that co-elute is making your calculations so complicated that it is worth changing an official method then you probably need to look at how you are doing the calculation.

Peter

[vballest]

Do you see a valley on the chromatogram? If you do, you can decrease the sample size, decrease carrier pressure, or temperature. If you don't, you may want to increase sample (as long as you don't saturate the preamp), spike with an internal standard that would elute between the two peaks and make those peaks "riders", or try adding a union and piece of restrictor tubing (<0.10" ID) to the column to reduce flow.

[chromatographer1]

Unless the coeluting peak causes an integration problem why is it a problem.

If you want to compromise, half the concentration of each analyte so the concentration of the two peaks summed equals about the concentration of the other peaks in the mix.

Be sure you have your argument ready and fully justified before you do this however.

best wishes,

ROd

[avitan]

Thank you all for your answers. It will be probably easier to stick with the official method, indeed.

[Peter Apps]

Do you see a valley on the chromatogram? If you do, you can decrease the sample size, decrease carrier pressure, or temperature. If you don't, you may want to increase sample (as long as you don't saturate the preamp), spike with an internal standard that would elute between the two peaks and make those peaks "riders", or try adding a union and piece of restrictor tubing (<0.10" ID) to the column to reduce flow.

That is certainly a novel approach - what makes you think that any of those changes will increase the resolution betweenthe two peaks ?

Peter

[vballest]

Peter,
It's something I've been playing with lately...

[Peter Apps]

Peter,
It's something I've been playing with lately...

I am intrigued - how do the results look ?

Peter

[vballest]

For each of the troubleshooting methods listed the results are different, but to summarize they all have an impact on response. The degree of impact is certainly application dependent but in most applications I find varying the sample size (injected) is the easiest way of manipulating response. Of course the preamplifier must be able to tolerate the response so the concentrations can not be too high and the reliability depends on how continous the GC runs without calibrations in between. I found that as the sample concentration increases, linearity to the minimum detectable limit may decrease. Which is why I prefer decreasing sample sizes to separate peaks.

[Peter Apps]

I would certinaly expect that injecting different amounts of sample, and adding an extra overlapping peak would change the peak areas (=response I presume), but what I am really interested in is the effect on the resolution between two peaks that are already overlapping - which is the situation in the original post. The only effect that I would expect is that if the quantity in a peak gets large enough to concentration overload the column, then the resolution would get worse. Apart from that I cannot think how changing the quantity in a pair of peaks can affect the resolution between them.

Peter

[vballest]

It is difficult to explain on a forum. If you can, take a known standard and inject 50% and then try 300% of what you would normally inject. In some cases there is too much sample and resolution occurs “offscale” on the high side of the preamplifier, in other cases there is not enough sample and resolution may occur “offscale” on the lower end (usually isomers such as 2,2dimethylbutane and n-hexane) . For co-eluting peaks the baseline will not completely resolve but, you can separate the maxima of each.. which will allow you to drop a valley between the two and integrate that way. Some amines co-elute on wax columns, partial resolution can be achieved by manipulating the response ( response factor = peak area/mole%) of the detector: mole % (concentration) would remain constant but Δpeak area (peak area directly reflects the amount of sample injected) would manipulate the overall Δresponse. If complete separation is desired the method (time, temp and pressure) must be adjusted or you may need to look for a more efficient column.

[Peter Apps]

So, am I right in understanding that by changing the size of the peaks you can change whether or not you see the valley between them, even though the retention times of the peaks, and the distance between the peak maximums do not change ?

What kind of quantities are you putting onto the column ? - and are you using ordinary capillaries (0.25 - 0.53mm i.d, 0.25 - 1 um films) ?

Peter

[HW Mueller]

vballast, Peter has tried to tell you that unless the column is overloaded at on injection amount the resolution does not change. By going to a lower injection amount you may get a chromatogram that fools you into thinking that the valley between two peaks has dropped to the baseline. If your software allows normalization to the same peak hight you can easily proof to yourself that amt of injection did not change overlapping.