LC separation of compounds with wide range of polarities

[karen j]

Does anyone have suggestions for developing an impurity method for a pharmaceutical (MW~600) that has no measureable water solubility but has known impurities ranging from very water soluble to completely insoluble? Any column suggestions for this wide polarity range? I’m currently using reversed phase HPLC with a 2-stage gradient, using formic acid/methanol to elute the early eluting peaks and then adding a stronger solvent to elute the later peaks. I’m having problems getting adequate peak shape for the earliest impurity since I need to dissolve my sample in 100% methanol and inject at least 5 mcL.

[Consumer Products Guy]

I think that you're on the right track with your gradient.

I’m having problems getting adequate peak shape for the earliest impurity since I need to dissolve my sample in 100% methanol and inject at least 5 mcL.

I believe that your issue is related to your 5 µl injection solvent being methanol (strong solvent), and your initial mobile phase being high on aqueous, low in methanol. This typically affects early eluting peaks much more than late-eluting ones; actually, your issue seems like the "classic".

Question 1. Are you locked into 5 µl as your injection volume? We have several methods that use 3.0 µl injection sizes to overcome such issues. I'd at least experiment with 2 µl and 3 µl injection sizes.

Question 2. Will the material stay in solution if you first dissolve in some 100% methanol and then add some water or aqueous buffer?

Question 3. Sometimes diluting a sample with aqueous and then injecting a larger volume (like 10 µl in your instance) actually makes the peak sharper as the injection solvent is weaker than with 100% methanol. Care to try that?

If you're still searching for better chromatography, you sure are not at the validating stage yet...consider, or at least try, some changes.

[karen j]

I'm using an old HP1100 LC, and haven't historically had very good luck with injection volumes below 5 mcL; that's why I set 5 mcL as my initial parameter - however, it's been difficult to find conditions for which I can inject 5 mcL of 100% organic onto the column. Or perhaps it's more accurate to say that any conditions that produce adequate retention and separation at the front end of the chromatogram produce other more difficult resolution problems at the back end of the chromatogram. I guess I could revisit the idea of using 2-3 mcL injection volumes, and confirm what the practical minimum is with my specific method conditions.

Unfortunately any dilution with, or addition of aqueous buffers in the diluent is not an option. This material is completely insoluble in aqueous solutions at any measureable concentration. I've looked at the other solvent options for the diluent, and methanol is definitely the best for reversed phase LC.

[JMB]

If CPG's suggestions do not work out, then a few thoughts are,

1) there are/were E(xtended P(hase) RP columns, made to handle very polar compounds.

2) Check column manufacturers applications notes vs. your active for similar structural types.

3) If all else fails, and depending on your time-frame, it might be easier/quicker to develop two separate methods---one for the polars and one for the non-polars. Check with your regulatory people t osee if the FDA will allow this. As long as it is scientifically defensible, you should be OK.

Regards.

[Consumer Products Guy]

I'm using an old HP1100 LC, and haven't historically had very good luck with injection volumes below 5 mcL; that's why I set 5 mcL as my initial parameter.

I am using "an old HP1100" HPLC as well. But I also used 3 mcL with my older 1050 models.

Agilent lists RSD specification in the 1100 Autosampler manual for injection sizes below 5 mcL. I'm at home now, I'll leave that for you to look up. Let's just say that our RSDs are significantly lower than Agilent's spec limit. When's the last time you changed the autosampler metering seal, the needle, and the needle seat (and check the metering piston); these are typically involved with autosampler injection repeatability.

[mckrause]

Try switching to an aqueous type phase and/or HILIC. HILIC may be too polar for your insoluble compound. We've had good luck with aqueous-type phases, running 100% aqueous eluants and small injection volumes of methanol. I typically run 2.1 mm columns, so my flow rates normally run in the 0.3 to 0.4 mL/minute range.

BTW, I am working on a 1050.

[KM-USA]

Nothing wrong with 1050, that was my favorite, like it better than 1100 or 1200 series, liked the pushbuttons in front as option to the computer Chemstation for some stuff.

Anyway, if you ever need parts for that 1050 not available new, you can contact me through this board, in my Announcements topic of vintage parts for sale, quite reasonable.