Inconsistency in Preparative HPLC

[pancaker]

Hello,

I’m looking for some advice to improve my chromatography. As a warning statement, I’m fairly new to HPLC, and I’m working with a compound that is difficult to separate on prep scale.

I’m separating a compound from a mixture very similar derivatives/impurities – appox 80% pure originally -- using a normal-phase prep column (Rx-SIL). Sample is dissolved in DMSO. Usually manually inject 1mL of concentrated sample in a 2mL loop. Mobile phase is mostly ethanol, first 100% then some gradient action.

I’ve done a lot of method development, and the chromatography is pretty ugly. Very poor resolution. Still, analyzing fractions by analytical HPLC (which has great resolution), I am usually able to collect a few fractions that are 95% pure, which is fine for my purposes.

I need a high volume of high purity compound and am continuously running performing prep runs. Unfortunately, the chromatography is not consistent. I will get runs with numerous high purity fractions (95%+), and other runs where the highest purity fraction is 92%. Shape and retention time varies slightly.
I equilibrate at 100% EtOH for 20 or 30 mins. To wash the column, I inject 2mL DMSO and run 70:30 ethanol:water w/ 0.1% acetic acid for about 60 mins, then back to 100% ethanol. I equilibrate and go to the next run.

As someone relatively new to HPLC, just looking for some brainstorming for what could be causing the inconsistency or anything I can do to encourage consistency.

One thought I had is precipitation in the column due to high concentration of sample injection / low solubility in ethanol. On top of this, my wash method is probably not ideal, so advice there would be great.

[entropiclabs]

Highly recommend you lyophilize your sample first to remove dmso before injecting, then dissolving in pure etoh. Injecting dmso on a normal or reversed phase column on a prep scale can definitely cause problems.

[pancaker]

Highly recommend you lyophilize your sample first to remove dmso before injecting, then dissolving in pure etoh. Injecting dmso on a normal or reversed phase column on a prep scale can definitely cause problems.

I totally agree. Unfortunately the sample has very low solubility in EtOH -- less than 1 mg/ml. This is why I suspect a precipitation issue. During method development, I tried dissolving in a DMSO/Etoh mixture. Had to use a larger injection as well. Overall resolution was reduced, though results did seem more consistent. The impurities elute so closely to the desired peak that a larger injection volume doesn't seem like an option.

[HPLCaddict]

You're doing "old school" normal phase chromatography, right? No HILIC stuff? In normal phase, especially on plain silica, you normally shy away from using water as eluent, as it's strongly adsorbed by the silica and hard to remove completely afterwards. Furthermore I wonder what your gradient looks like. Your starting eluent pure ethanol is already quite a strong eluent in NP. If your component shows so much retention, maybe RP would be a better option?

[pancaker]

You're doing "old school" normal phase chromatography, right? No HILIC stuff? In normal phase, especially on plain silica, you normally shy away from using water as eluent, as it's strongly adsorbed by the silica and hard to remove completely afterwards. Furthermore I wonder what your gradient looks like. Your starting eluent pure ethanol is already quite a strong eluent in NP. If your component shows so much retention, maybe RP would be a better option?

I don't really know the details of HILIC. I was originally doing true "old school" normal phase with an isocratic EtOH method but found a very slight gradient including water was helping to pull out one of the impurities.

I am washing the column with a water/etoh mixture as mentioned in my first post. In the past I've always done reverse-phase so I didn't even consider water binding to the silica. Is washing with water a bad idea? What's the best way to flush everything out of the column?