On a Discovery HS F5 (150 x 4 mm) I noticed some strange behaviour when analysing three structurally similar neutral compounds under isocratic conditions:
Resolution (Rs=(2(RT2-RT1))/(W2+W1) ) changes with the flow rate. At 1 ml per min the second and third peak are just shoulders on the first peak. At 0.75 mL/min all three peaks are separated (Rs 2/1 = 2.4, Rs 3/2 = 1.3) and at 0.5 mL there is nice baseline separation (Rs 2/1 = 3.6, Rs 3/2 = 1.9).
Theoretically the flow rate can influence only the number of theoretical plates via the van-Deemter equation. With our example for the first peak efficiency increases with decreasing flow rate: N=6000 for 1 mL/min, 9700 for 0.75 mL/min and 12000 for 0.5 mL/min.
However, that doesn’t explain solely the improved separation at low flow rates.
Also k' is changing (for peak 1 1.9 at 1 mL/min, 3.4 at 0.75 mL/min and 5.3 at 0.5 mL/min). The alpha-value increases slightly for the second peak (1.11, 1.13 and 1.16) and is stable for the third.
I do not understand the increasing k’ with decreasing flow rate. Has someone experienced something similar? Has some one a explanation for this? Are there any articles about retention behavior on PFP columns?

Alex

The change in retention factor is the only thing that is strange. What is the compound with which you measure t0? Is V0 (t0* flow rate) constant?

Maybe not strange, but still unexpected is the increase in efficiency at low flow rates. With a 4 mm column and 3 µm material I would expect the highest efficiency around 1 mL / min, slowly decreasing with higher flow rates and more pronounced decreasing at flow rates <0.4 mL /min.
t0 is estimated, assuming a constant V0. If i take a minor disturbance in the chromatogramm as marker V0 is decresing with lower flow rates (at 0.5 mL/min the disturbance is observed at 3.12 min, for calculations 3.9 min (2x t0 at 1 mL/min).
When using the disturbance as t0 marker I would get even higher k' values for the lower flow rates.
I should try uracil or acetone as t0 markers

Thanks for the comment,

Alex

There is nothing extraordinary with a platecount of 12000 for a 15 cm 3 micron column. It is actually not up to standard. A decent manufacturer should supply higher numbers on a routine basis. Why do you expect the high platecount at a higher flow rate? What is your mobile phase?

N= 12000 is sufficient and realistic, especially if you do not work with 2 µL injections and selected compounds. Solvent is methanol / water (50/50).
Again: are there any publications on retention behaviour on PFP columns?. Has anyone observed change of k' values with flow rate?

It is not very likely that you are really looking at change in retention factor. I suspect that you are looking at a change in t0. Of course you won't know until you use a real t0 compound to measure it. There could be leak or airbubble during injection, or there could be a delay when your machine starts the chromatogram. I have seen both.