Hallo
I am quite new to HPLC. So far I was using only isocratic methods, but recently I have started developing gradient method. I was surprised to find that number of theoretical plates calculated for my two analytes (according to USP, EP and JP) was: comp. A ~9000 and for comp. B ~67000.
RT_A = 5,27 min; W_A=0,22 min; W_A0,5 =0,13
RT_B = 20,07 min; W_B=0,31 min; W_B0,5 = 0,18

Is such a high number normal for gradient methods – 67 000 for 250 mm column.

I would appreciate if you could explain this to me. Thanks in advance.

Well, the way you're looking at this may be a bit skewed. Plate theory is only applicaible in isocratic mode. When comparing analytes they need to be influenced by the same parameters. so gradient doesn't allow for this. Many people still talk about N when using a gradient. a better way to compare things would be to use width at half height. Given all this, it isn't unusual to see a drastic decrease in width (or increase in "N") with a gradient. The gradient is used to optimize the elution conditions for each analyte. With the isocratic mode, you had a single condition for eluting both analytes. With the gradient, your analyte are eluted under different conditions (it also allows for you to preconcentrate, may be helpful if using low concentration and arent worried about time).

If you want to get a true measure of the separation power of your gradient separation, the best thing to do is to calcualte peak capacity. It is the gradient duration divided by the average peak width. It is very straighfroward to do.

Thank you Jim and Uwe.

Does the same apply to jump grsdient (or whatever you call it in english); 60:40 isocratic then quick change to 20:80 and again isocratic untill the end?

The same also applies to eluent step changes (I prefer not to use the term: step gradient since this is an oxymoron, a step is by definition not a gradient). You can easily see why the number of plates measured can be quite meaningless if you consider the following example. Start with an eluent in which the analyte has a very large k' (i.e. it binds to the stationary phase so tightly that essentially none of it is present in the mobile phase). Inject the sample, and wait one hour and then step to an eluent where the analyte has a k' of 0. You will get a rather impressive efficiency using this method. But it has relatively little to do with the performance of the column involved. Furthermore, if you repeat the experiment but wait 24 hours before making the change in eluent to the higher elution strength you will observe a dramatic increase in the "efficiency". So as you can see, the measured value in the case of eluent step changes and gradient separations is rather meaningless except as a basis for comparison such as in the case of comparing two different columns run under exactly the same gradient conditions on the same instrument (even instrument delay volume will effect the apparent efficiency of the separation).

In step gradients with isocratic plateaus, you can calculate the plates pretty much as you would do in standard isocratic chromatography. However, you need to consider the moment at which the step change reaches the column as the equivalent of your injection in isocratic chromatography. Thus the equivalent inejction time is the time at which the step gradient is executed + the time it takes for the step gradient to reach the column inlet. The latter is caled the gradient delay volume of your instrument and can be determined in an independent experiment, without a column.