Chromatography Forum

We are trying to calculate LOD/LOQ for a chromatography sample using the method in Harris Quantitative Chemical Analysis: measure 7+ blanks and 7+ samples of low concentration and the signal detection limit is (average of the blanks) + 3*(standard deviation of samples).

This is simple enough, but I can't figure out how to get Clarity to allow me to integrate the blank signals. I can make a new peak and it will often allow me to do so, but it won't let me place the starting and ending points where I want them to be.

Anyone have any thoughts on how to get Clarity to do this? If not, is there free and/or open source software that will do so?

Thank you!
Mark

I think the integration in Clarity suffers from the lack of a click-drag-unclick option. There are buttons for setting the start and end of a peak, but I can't remember what they look like... I think there is also a setting for minimum peak size.

But shouldn't your blanks be clean enough that there's essentially just a flat line with nothing to integrate? Thus all the results would be non-detect and there'd be nothing to use. Blanks are useful for figuring out the MDL in methods where you always get some kind of number. But if the software and/or user is never going to integrate a little noise peak and call it a hit, then using those peaks to figure out MDL is not really useful. If you are getting peaks large enough to integrate, could you eliminate them somehow?

Hi Michael,

I have tried the buttons you speak of, but they don't always work when there is no peak. I can even specify an beginning and ending point, but the software "helps" me by changing my settings.

You are correct and my blanks are clean, flat lines. There is very little to integrate, which is the source of the issue. What I'm after is to get a handle on the noise. Speaking as an analytical chemist, there is always noise in any measurement and the amount of it matters. That's what this is trying to get a handle on: what's the contribution of noise to the signal.

I recognize, though, that I am principally a spectroscopist, not a chromatographer, and I am approaching this problem as such. The confusion in your answer tells me that perhaps I am approaching this in the wrong way. Is there a better/more accepted way of determining limit of detection and/or limit of quantitation in chromatography?

Thank you!
Mark

There are a lot of methods to chose from for determining the LOD/LOQ in chromatography. One approach from this paper (https://pubs.acs.org/doi/10.1021/ac981179n) and the Harris analytical chemistry textbook involves running a standard near the LOQ 7 times and using a calculation from the standard deviation of those 7 replicates and the slope of a calibration curve.

Thanks for the paper! That's a great resource and I appreciate it.

The Harris approach is the one I was trying to use, but it also requires measuring seven blanks. I have the data, but can't integrate the blank samples. This is source of the original question.

I'll scour the paper you sent. I have also found a couple of other approaches in my searches.

Thanks again!
Mark

The equation I've used is:

LOD = (3.148 * s) / (m)

where s is the standard deviation, m is the slope, and 3.148 comes from the degrees of freedom.

Good luck.

Hi Michael,

I have tried the buttons you speak of, but they don't always work when there is no peak. I can even specify an beginning and ending point, but the software "helps" me by changing my settings.

You are correct and my blanks are clean, flat lines. There is very little to integrate, which is the source of the issue. What I'm after is to get a handle on the noise. Speaking as an analytical chemist, there is always noise in any measurement and the amount of it matters. That's what this is trying to get a handle on: what's the contribution of noise to the signal.

I recognize, though, that I am principally a spectroscopist, not a chromatographer, and I am approaching this problem as such. The confusion in your answer tells me that perhaps I am approaching this in the wrong way. Is there a better/more accepted way of determining limit of detection and/or limit of quantitation in chromatography?

Thank you!
Mark

Maybe the confusion is because I'm thinking in terms of GC/MS which is both a qualitative and a quantitative measurement. You only quantitate a compound after you've concluded that it's in your sample (or, maybe, that you can't rule out that it's in your sample).

Suppose I have a peak at the retention time for an analyte. The ion associated with the analyte is definitely present, but none of the qualifier ions are present (so I see the molecular ion but none of the fragments). Usually that will happen because a different compound is co-eluting. The concentration may give me a reading if I do integrate the peak, but what I actually conclude is *not detected*. There's nothing to count. No matter how many apples show up, the number of oranges is still zero. The correct integration is no integration: if the software auto-integrates noise, you delete that integration.

If you are using a detector like an FID or ECD, I'm not sure the same thing is true. Can anyone else weigh in on that?

PS: it did come to mind that as a rough estimate you could use peak height on your blanks and see if those tiny numbers even make a difference. Or! If you can't integrate, you could do it the old fashioned way and print the chromatogram then cut out the peaks and weigh them to get an area. In fact I think you should just because it's charming.