Page 1 of 1

Peak purity percentage criterion in specificity/selectivity

Posted: Sat Aug 20, 2022 3:01 pm
by snisyriou
Hello,

which is the acceptance criterion of the peak purity? Is a value of 0.95 acceptable? Is there any official guideline about this?

thank you

Re: Peak purity percentage criterion in specificity/selectivity

Posted: Sat Aug 20, 2022 4:12 pm
by vmu
"Peak purity" (spectral) is a software/algorithm-dependent quantity. No official limits exist.
See also
Empower Peak Purity and Absorbance
https://www.chromforum.org/viewtopic.ph ... ty#p413124

Re: Peak purity percentage criterion in specificity/selectivity

Posted: Sun Aug 21, 2022 10:42 pm
by Multidimensional
No official "specs" or guidelines exist because the results can be tweaked by the operator to reveal any "purity" value you want. If you understand what the software does and what is assumed in using it, then you will understand how any purity values obtained from the software are marketing nonsense.

Here is a link to an authoritative article on the topic.
  • Peak Purity Determination by HPLC Diode Array Chromatography Software: Limitations and Uses
; https://hplctips.blogspot.com/2016/12/p ... diode.html

Re: Peak purity percentage criterion in specificity/selectivity

Posted: Wed May 03, 2023 4:24 am
by Scuman
"Peak purity" (spectral) is a software/algorithm-dependent quantity. No official limits exist.
See also
Empower Peak Purity and Absorbance
https://www.chromforum.org/viewtopic.ph ... ak+purity/geometry dash bloodbath
An unreliable method of determining whether the separation method is sufficiently selective is to perform a spectral peak purity check. The absence of hidden peaks of impurities beneath the primary peak cannot be demonstrated by this purity check. The absence of any discernible spectral differences between the various components of the main peak (these differences are sought in the peak purity check) may be caused by low concentrations of impurities and the similarity of their spectra to the spectrum of the parent substance. I think this answer to the above question.