Comparative Study for Modern LC Columns

[SIELC_Tech]

I would like to set up comparative study for few modern RP columns. My objective is to study loadability, peak shape and retention control for hydrophilic analytes (Something like David McCalley is doing)

I have Sorbax SB AQ, Synergi Polar RP, Atlantis T3, and Gemini which I would like to compare with our columns (SIELC). I don't want to use conditions which people will call "unfavorable" for these columns. So my question is what are "good" conditions for modern RP HPLC columns for the following analytes: dopamine, phenylephrine, ephedrines, and propanolol (just starting with these 4). This is not a tricky question; I just need opinion of people on conditions to run this comparison. Some people on Chromforum stated that it is not hard to produce poor chromatography on competitors' columns, so I want to do a fair study.

Do I need to use 0-5% organic and low pH to produce retention and good peak shape or go to hiigher pH? I need unified conditions for all columns to compare "apples with apples".
So far I tried ACN/water/Ammonium acetate=5/95/20 mmol pH-5 and my peak shape and retention are not very good on these reverse phase columns.

Regards,

Vlad


P.S. Dear Uwe may be you can recommend conditions for Atlantis T3 for the analytes I mentioned above.

[Hollow]

Dear Vlad,

I'm just writting my first ideas which came in mind when I read your text.

1) Did you look for a applications on your competetors columns and your molecule of interests? Think for the catecholamines there are a few.

2) Do you know about the ACD-Labs column selector?
It's a add-on for ChemSketch (drawing prog) which let's you compare many columns against each others. They were tested on a standardized protocoll and everybody can submit their datas according to the guidlines. Maybe you also want to get your columns listed there. See guidlines and infos:
http://www.acdlabs.com/download/colsel_contribute.html
http://www.acdlabs.com/download/docs/co ... sketch.pdf

3) For your catecholamines I would do a gradient starting with no or just few % organic and raise it up.
Cause your analytes are bases, I would recommend to also raise the pH to 8-10 to get the free bases of your molecule. With low pH the molecules are ionized and not or little retained.
If you use the salts I would switch to normal phase HPLC or maybe HILIC-columns (?is it the same? I'm not familiar with HILIC)

4) for direct comparability use same diameters of your columns otherwise beware of the differents linear velocities

[SIELC_Tech]

Thanks Hollow,

I would like to compare individual compounds (not a mixture) at isocratic conditions on each column. These conditions might be different for each compound (and may be slightly different for each column). My plan is to use low organic (0-5%) for traditional RP columns, retain them with K'=4-8 and see loadability for prep separations, effect of pH on peak shape, retention control, etc.
Some of these columns "especially designed to retain polar compounds".
I am not sure that a lot of columns can handle pH-10 (Gemini only on my list).

Vlad

[Uwe Neue]

If you want to get the best retention, the best loadability and the best peak shape for bases, especially polar bases, you should run under alkaline conditions on a packing that is truly stable, such as the XBridge packing.

Here is a reference on the loadability of basic and acidic analytes as a function of pH (especially for preparative chromatography):

U. D. Neue, T. E. Wheat, J. R. Mazzeo, C. B. Mazza, J. Y. Cavanaugh, F. Xia, D. M. Diehl, “Differences in the preparative loadability between the charged and uncharged form of ionizable compoundsâ€

[SIELC_Tech]

I have plans to study four columns I have and I am sure that these columns will not tolerate high pH (9-10). Some of the basic compounds with pKa above 10 are going to be still partially ionized, plus silanols might give you a bigger problem.
At this point only two columns on the market (xBridge and Gemini) claim to be stable at high pH. If you have a mixture with both acidic and basic compounds going to high pH will make acidic compounds fully ionized and not retained. We are trying to develop universal platform without going to extreme: mild pH ranges (2-7), manageable in terms of solubility organic content (5-50%) and reasonable buffer concentration (5-50 mmol). This is good for both UV. ELSD, LC/MS and prep application, plus allows you to separate various compounds in the mixture without having some of them unretained.

[HW Mueller]

Sielc_Tech, if you tell us why you are really doing this, it should be possible to tell you the results without wasting time with experiments. As a matter of fact, I would venture a stab at the results already: Your columns are the best, the problem having been so chosen as to obviate this.
I hate to regurgitate the discussions on optimum perormance, etc.

[koen_shimadzu]

Hello Vlad,

I like to bench mark columns too. It’s fun and interesting work.

But before you start with this you have to ask you’re self the question; why am I doing this?
If the question is to demonstrate that you’re column is better than the Atlantis and XBridge, you waste you’re time. It you want to prove that you’re columns are different, than you may can find nice results.

I would suggest that you take some strong acid, basic and neutral compounds. Run all these compounds at three different pH levels for each column. A PCA c plot can give you a nice result for which condition which column to use. Or even more important, under which condition a column not should be used.

The ACD software is also a simple, but effective, tool for users.

Good luck

[SIELC_Tech]

Koen,

Do you think the following is a good set of compounds: DOPA, paraquat, benzensulphonic acid, maleic acid, dexthromethorphan, pseudoephedrine and naphthalene? I represents major classes of small molecules (neutral, zwitterionic, basic hydrophilic, basic hydrophobic and acidic).
I am screening pH range from 2 to 7 (may be will go to 10 on Gemini).

Vlad

[koen_shimadzu]

Hello Vlad,

To be honest, I cannot tell you that. There are groups that invest how to characterise columns which as less as possible experiments/compounds.
The compounds you’re using are looking fine to me.

I also would include some compound that everyone is using for selection of columns. Like e.g. amitriptyline, nitrobenzoic acid and tricarboxy benzoic acid.