Chromatography Forum

Dr. Neue made the comment below:

viewtopic.php?t=11144&highlight=retention
PostPosted: Mon Aug 17, 2009 2:55 pm
Dancho is right: the retention is way too low. Add some more water to the mobile phase! You want a retention of at least around 10 minutes with the flow rate and the column that you have.
-note- flow rate = 1ml/min, column Vydac C18 218 TP (4.6 * 250 mm)

My question is, how does someone know what the retention of a compound should be based on column dimensions? Is it a calculation from a specific capacity factor?

For my newest method, I have a capacity factor of 1.3 so far and really I don't know if that is optimum. I can think of how it might be a problem when I enter into the sample prep development phase, but until then I want to make sure I have the basics down.

Thank you all in advance for your infinite HPLC wisdom!

Dear Kerri

for isocratic separations one preferably wants to tune your components into a retention window of about 1 < k < 10 (maybe up to 20).

In combination with the dead time t0, which depends on the colum dimension and the flow rate, one can calculate the corresponding retention time window: tR=t0*(k+1)

The dead time t0 can be estimated by dividing the free column volume by the flow rate:
t0=1/6*d^2*Pi*L/F
(d and L in [cm], Flow in [ml/min], t0 in [min]).
Note: the free column volume can be estimated as about 2/3 of the cylindrical volume of the column, this means that only 1/3 is occupied by the skeleton of the stationary phase.

In the given example, the t0 of the method would have been about 3 min, which turns into a retention factor of only 0.5 for a compound eluting at 4.4 min, whereas 10 min corresponds to a k = ca 2.3

Hollow,

Thank you very much!

Kerri

Hi,

I saw an article where the authors report a dead volume RT of 4.9 for the column

ZIC-HILIC 150 mm x 4.6 mm, 5 μm column (Merck Sequant), The flow rate was 300 μL/min

but with the formulaabove I get 5.53, m I missing some thing? Do I have to take the 5um particle in consideration?

You're not missing anything, but that forumula is only an approximation (give or take 15%). So the measured and predicted are in the same ballpark.