By Rachel Thierry on Wednesday, June 30, 2004 - 12:32 am:

Hi all!
I have to separate two peaks which are only partially coeluting.
Right now I work with a reverse-phase isocratic system with 25% acetonitrile, and the peaks have a retention time of about 2.2 minutes.
I tried to lower the acetonitrile percentage (to 20 and to 15%), however the peaks become broader and resolution doesn't become higher.
So, I was thinking about gradient elution.
Would it be reasonable to have a low, constant content of acetonitrile (i.e., 15%) at the beginning of the run (for about 2.5 min) and then raise it (let's say, to 30 %) over the time needed for the compounds to elute?
Initially I thought that a constant, linear gradient of AN (e.g. from 10 to 50%) from the beginning until the end of the run would be OK, then I thought that this would give me longer retention times and the compounds would be eluted when the AN content would be high, thus probably coeluting...
Any opinion?

By Ann on Wednesday, June 30, 2004 - 02:55 am:

Have you tried altering the pH of your mobile phase? If you haven't, it is well worth trying this before you resort to a gradient method (especially as changing the percentage of organic modifier is seemingly having little effect). A RT of 2.2 mins sounds very quick to me, as though your compounds are hardly being retained by the column at all - what is the RT of the solvent front? A pH change may make all the difference, try it! Good luck :o)

By Rachel Thierry on Wednesday, June 30, 2004 - 05:07 am:

Well, I'm already at pH 6.5 and I guess that I can't go much higher with a silica-based column. I think that going lower wouldn't help me because my analytes are basic, they would be less retained.
And I'm doing LC/MS analysis on a short column (100x2 mm), so an RT 2.2 min isn't so unusual, I think... RT of the solvent is about 0.9 min.

By MG on Wednesday, June 30, 2004 - 07:03 am:

Gradient elution will give you sharper peaks. Based on what you've told us, I'd try 5% to 30% over 10 min, and adjust from there.

"Initially I thought that a constant, linear gradient of AN (e.g. from 10 to 50%) from the beginning until the end of the run would be OK, then I thought that this would give me longer retention times and the compounds would be eluted when the AN content would be high, thus probably coeluting... "

Eluting in higher ACN is better for LC/MS sensitivity. What you describe has never been a problem for me with gradients. Since the compounds were migrating more slowly in the early part of the gradient, and were better separated, having them elute in higher organic at the end does not automatically cause them to become unseparated.

Another easy thing to try is switch your organic modifier to methanol.

Although, I'm surprised you don't get better resolution at lower isocratic ACN conc. As is, your k' is only 1.4.

By Dave on Wednesday, June 30, 2004 - 11:36 am:


Before trying a gradient, have you tried a 150mm column? 250mm? Granted, you will increase your retention time, but since your peaks are eluting so quickly, the increase in RT would not be that drastic. And since they are "only partially coeluting", perhaps a greater column length will be all that is needed to achieve the resolution you desire.

Just out of curiosity, what is your resolution as it stands now?


By HW Mueller on Wednesday, June 30, 2004 - 11:41 pm:

How can a gradient increase resolution in a case like this? (I am assuming there are no macromolecules involved here).

By MG on Thursday, July 1, 2004 - 12:49 pm:

k' is only ~1.4 for those compounds. Chromatographic theory says you can increase k' to improve resolution, although there is little benefit beyond k' = 10. But simply lowering the organic content and keeping it isocratic will result in broader peaks, and sensitivity loss in ESI-MS. So a gradient allows you to have your cake and eat it too, as the saying goes, by changing k' from high to low over the course of the run. Or is there an error in my thinking?

By tom jupille on Thursday, July 1, 2004 - 01:38 pm:

You can *never* have your cake and Edith too. Two issues have cropped up in this thread: resolution and sensitivity.

For a given pair of peaks, a gradient will usually not improve resolution compared to an isocratic separation; the primary purpose of gradients is to accomodate a range of k' values too large to fit into a single isocratic run. Exceptions can occur if selectivity changes as a function of k' (e.g., as with a mixture of small and large molecules), because gradients allow access to a wider range of k' values.

For linear gradients in reversed-phase, the results can be characterized by an "average k'" (usually symbolized by k*): the k' of a peak when it is at the midpoint of the column. This varies with the steepness of the gradient and can be plugged into the resolution equation in place of isocratic k'. A shallower gradient will increase k*, and the effect will be pretty much the same as increasing isocratic k': increased resolution (if selectivity doesn't change!) and broader, lower peaks.

As far as sensitivity goes, the same arguments apply. A steep gradient will generate peaks with lower k* values. They will be sharper and come out at effectively higher acetonitrile concentration, both of which will improve sensitivity. However, the effect for a given pair of peaks will be no different than the effect of the same isocratic k'

The sharpening effect of gradients is most notable (and useful) when late-eluting peaks are involved. Because all peaks in a gradient separation have approximately the same k*, the last peaks come out with the same sharpness as the first ones. Because the k' values in this particular question are low, this doesn't apply.

By MG on Thursday, July 1, 2004 - 03:09 pm:

OK, that makes sense. Thanks Tom!

By Alexander on Thursday, July 1, 2004 - 06:44 pm:


Are your nasty compounds isomers? Normally you wouldn't get much from changing to gradient. Changing the column could be more efficient. Try to play on differences in nature of the compounds. Depends on that vary pH, phase type (C4-C18 , CN or Ph), counter ion, chiral column, etc.
I know, you may loose sensitivity, but going to C4 column and lower MeCN concentration may give you higher resolution.
Good luck!

By HW Mueller on Thursday, July 1, 2004 - 11:30 pm:

Amen, Tom....... that Edith must be some female, though, not allowing a cake next to her!!

By Anonymous on Sunday, July 4, 2004 - 10:45 pm:

It may be late to jump in discussion, but I have a suggesstion from my experience. The better resolution of peak, as it is a problem not having in low ACN, I found sometime the injection module may play an important role. I mean, usually in lc/ms/ms, I use microlitre pick-up mode, in which transfer solvent is required, this solvent can be altered to improve resolution of peak. It partly behaves like gradient, even your run is isocratic. You can use more aquous phase in transfer solvent and as your initial 25% ACN in MP, it can improve your resolution and separation. I hope you understand me.....
I am not sure which autosampler you are using but I thing in most there is micro-litre pick-up mode which requires transfer solvent.
I think you may solved your problem.....otherwise it may be helpful.