Chromatography Forum

Can someone please explain how to calculate the detection limit for an analyte.
I understand that the detection limit is 3 times the signal to noise ratio, but I don't know what that ratio is and how to get it!

Thanks in advance!

berettagtz,

There are several ways to measure/determine the noise. You can measure the actual analyte peak height and noise amplitude on the chromatogram, you can use standard deviation of the signals and you can use statistics from the linear regression analysis of your calibration curve. There are alos instrumental ways to measure noise.

Rather than attempt to describe it all here, I recommend that you check out these references:

LCGC article by John Dolan (an expert in LC):

http://www.lcgceurope.com/lcgceurope/ar ... ?id=284453

Discussions on this forum:

http://www.sepsci.com/chromforum/viewtopic.php?t=8607
http://www.sepsci.com/chromforum/viewtopic.php?t=7290
http://www.sepsci.com/chromforum/viewtopic.php?t=7146


The article by John Dolan includes pictures and that is probably the best way for you to see and understand the S/N measurement. So, check out those references. If you have further questions, ask and the forum members will galdly help.

Regards,
Dan

I am still not sure about calculating the DL manually. What if the peaks are broad and they are not always proportional to the height?
Calculating the area and then finding the DL makes more sense to me, however I don't know how to calculate the area of noise.
How would you find the area of noise?

Thanks once again!

The ICH method validation document shows alternate methods of LOD/LOQ calculations

PDF can be dowloaded from here http://www.ich.org/cache/compo/276-254-1.html

berettagtz,

Manually measuring the peak and noise heights is a commonly used method for the detection and quantitation limit determinations, please see the ICH guideline referenced by JGK and the John Dolan article that I referenced before.

You cannot find the area of the noise. If you look at the noise in the chromatogram you will see that the noise consists of sharp peaks that are random, of various amplitudes, too narrow for proper area integration, both negative and positive in direction from the baseline, drifting negative or positive, etc. The measurement of the noise must take into account all of these aspects of the noise. There is no one place within the noise where you can decide that a specific peak is representative of all of the noise and then determine the area of just that one noise peak. It is doubtful (impossible or very difficult) that you could actually integrate the noise peak to get an accurate area measurement.

If you are uncomfortable in measuring the peak height and noise amplitude, then maybe one of the other noise determination techniques in the ICH guideline will be suitable to your interests.

Regards,
Dan