John Innes Centre site search

You are here - Home ::

Separation and HPLC

Imagine a chromatogram long enough that you can resolve 1000 peaks. I cannot draw such a chromatogram on this page because the chances are that your screen has nowhere near enough resolution to display a peak that narrow. This is an excellent chromatogram.

Now scatter a few peaks along it. If we want to be 90% certain that a peak is pure, how many peaks can we scatter at random along our baseline?

The answer turns out to be 14, which is not many. And of course, if we do separate 14 chemicals this way, then although we are 90% confident of the purity of each of them, that leaves us expecting that on average more than one of the peaks is coeluting with another.

Coeluting compounds are a disaster for quantification. There is no point in having a really accurate calibration curve in your method if 10% of the peak was actually something else.

And think about it: would you like to take a pharmaceutical where the analyst was only 90% certain...

For more details, have a look in:
Mass Spectrometry, Principles and Applications, by Edmond de Hoffmann and Vincent Stroobant
Relevant details on page 158 of 2nd edition.