Plate count
Comments

Hi Scurran
For a wellpacked column with fully porous particles (like the UPLC BEH Phenyl 1.7um) the plate count can be estimated to N = 500 x L/dp, where L is the column length (in mm) and dp is the particle diameter (in um).
This will give around 44000 plates on your column, so the 70000 you get sounds a bit high. Is this a gradient separation ?
best regards
Rune
0 
Yes, this is a gradient separation.
0 
Hi Scurran
I suspected this. According to Snyder and Dolan in their book "HighPerformance Gradient Elution", "an experimental value of N in a gradient elution cannot be calculated by the usual expression for isocratic separation". If the equation is used it will result in much too large values in N.
Regarding the question whether to use the USP or EP calculation of plate count : The two different methods use the width at different points on the peak. The EP uses the width at half height and the USP use the width where the tangents cross the baseline. If the peak has a perfect Gaussian peak shape (symmetrical) the two calculations should give the same value, but the USP will be more sensitive to tailing, which is why you typically see lower values than the EP.
In my opinion plate count is mostly usefull to monitor performance/degradation of the column with age/use, and not to say something of the quality of the separation. For this the resolution between all peak pairs is more relevant.
best regards
Rune
P.S.: In the UPLC console you can see what plate count the BEH phenyl column gave when it was tested at the factory. Look under column>column QC.
0 
Thank you for this information.
0