Poor %RSD for a peak eluting at the solvent front

[gtma]

I'm running a method that elute diazolidinyl urea compound at the solvent front. The sample solvent baseline is flat with no significant interference. However, multiple injections of the std or sample shows significant %RSD. Any suggestions why?

The peak area response trend is random ie. no general decrease within vial or between vial for the same (std or sample) solution. Therefore, degradation is not suspected.

1. An internal std compound eluted later in the chromatogram shows very good RSD, hence ruling out instrument issue. Also, high %RSD (>5%) was observed for multiple inst.
2. The response generally decrease with multiple injection from the same vial (60% methanol:40%water as sample solvent & mobile phase) but there are some exceptions where the %RSD is very good.
3. Tried different LC columns, insts, LC vial/caps, etc.....still high %RSD.
4. Diazolidinyl urea exist a complex mixture (apparently 2 peaks) but the UV spectra of the peaks are very similar therefore the response should be similar.
5. UV spectrum is very steep at the selected wavelength (for optimal response) but I don't suspect it will cause high % RSD within a run.
6. I don't believe it's related to the compd pKa.....does anyone know it's pKa. The sample solvent is appprox pH 8.4. Could it be the mp is close to the compd's pKa? I plan to use a pH 5 buffer solution to check.

I tried HILIC but I'm not getting the chromatography outlined in a literature paper. The diazolidinyl urea compound elutes at the solvent from for all the different columns/mp combinations tried.

[tom jupille]

You cannot quantitate a peak that elutes at the dead time ("solvent front"). In fact, the USFDA suggests that any peak quantitated should have a k' value greater than 2 (i.e., more than 3x t0). This is good science, because the "solvent front" is generated by the equilibrium upset from the injection. It represents a "chaotic" state of the system.

Without knowing anything about the HILIC column/mp combinations you have tried, it's hard to make specific suggestions, but very often a gradient is a good starting point (for either reversed-phase or HILIC) for exploration.

[gtma]

Thanks Tom. Yes, I understand k>2 is ideal. However, the compound is very polar and just wouldn't budge in RP-LC even with 100% aqueous/SDS ion-pair reagent/different ligands. I came across a paper (see link below) that uses Cosmosil column (HILIC mode) with 95%ACN:5%water to 85%ACN:15%water. I just can't replicate the DU chromatograms outlined in the paper. I even tried ammonium acetate buffer as the aqueous phase. In general, the baseline is poor and the DU retention time is not reproducible even for isocratic run......makes me think I'm not achieving HILIC mode.

http://www.sciencedirect.com/science?_o ... archtype=a

[Gerhard Kratz]

Don't give up with HILIC. If you go to the product information for your Cosmosil HILIC column many other mobile phases are listed. And please look for other HILIC columns as well. Novell NH2 columns with endcapping are available. HILIC is never the same. Diol columns are also listed as HILIC columns. For your very polar compound a new generation NH2 column could do the job. Just ask different manufacturers for a free of charge test column. Give another 2 or 3 columns a chance. Good luck.

[tom jupille]

In HILIC mode water, of course, is the strong solvent and acetonitrile is the weak solvent. A reasonable general-purpose HILIC screening run is a gradient from 50% buffer to 95% buffer. You don't worry too much about the baseline at that point; all you want to do is to see a peak and then estimate a reasonable % buffer further development.

[carls]

A reasonable HILIC screening gradient starts at 90% ACN 10% aqueous buffer and ends at 50% ACN. I recommend ammonium formate pH 3.5 as a good starting point. You'll want the buffer concentration in the mobile phase to be ~5mM, i.e. 10% of 50mM = 5mM. You can run this screen using a binary system with A = 90/10 ACN/buffer and B = 50/40/10 ACN/water/buffer.

Prior to running a gradient flush the column with 10 column volumes of 100% B (50% ACN) then re-eqilibrate with 20 column volumes of 100% A (90% ACN). 15-20 column volumes is required for re-equilibrating after each gradient run

A good column format allowing for fast screening is 100x2.0mm with 3um sorbent. Due to the high ACN content you can likely run at 0.5mL/min which gives a void time of ~0.5min. A 5 minute gradient to 100%B with 1 min initial hold and a 2 minute final hold should be reasonable.

Hope this helps