Page 1 of 1
selectivity
Posted: Mon Dec 24, 2007 1:18 pm
by olivia
Hi, can anyone give me an acceptable value for selectivity. Is it >1.5.
Is there a quantified number anywhere
Hope someone can help
Liv
Posted: Mon Dec 24, 2007 1:47 pm
by JGK
I attended a seminar recently on optimizing separations. Selectivity, was not assigned an optimum value but the as an effect on resolution.
The example given showed a slight increase in selectivity (1.1 to 1.2) producing a resolution change from 2.0 to 4.0.
The emphasis of the lecture was to gain maximum resolution and then to trade off some resolution modify the analysis to meet other targets.
Posted: Mon Dec 24, 2007 9:28 pm
by tom jupille
"Acceptable" selectivity is a function of the desired resolution, the plate number, and retention.
The general relationship for two close peaks is:
Rs ≈ 0.25 * N^0.5 * (α - 1) *[k'/(1+k')]
You can usually get sufficiently high retention to make [k'/(1+k')] approximately equal to 1, so that simplifies to:
Rs ≈ 0.25 * N^0.5 *(α -1). Solve for a and you get:
α ≈ 1 + 4 * Rs /(N^0.5)
How many plates do you have ? and how much resolution do you need? Plug in the values and find out what kind of selectivity (alpha) value is required.