Our compound is a N-BOC, tert-butylester-protected amino acid. It was not very soluble in our chromatographic solvent, which was 35/65 ethyl acetate/hexanes. We added some DCM to solubilize our compound, added about twice the mass of silica, and removed the solvent by rotary evaporation. We applied the impregnated silica to the top of the column and ran under flash conditions (we more typically use gravity). The material tailed, and the product was still visible (though dilute, relative to the most intensely staining fractions) even after we finished collecting fractions. We used permanganate to visualize the product on TLC. We noticed that the recovery by mass was low; therefore, we flushed the column with 100% ethyl acetate and recovered more permanganate-positive material that had about the same Rf as previous fractions. We will soon have dry masses on the pools.
In the past we have often dealt with poorly soluble compounds by dissolving in DCM and applying this solution to the column. On one occasion we used the dry-loading technique. I have not tried to correlate yield with poor solubility, but suffice it to say that I have not noticed a recovery problem previously. My admittedly simple model for why poorly soluble compounds usually elute without difficulty is this: when a compound moves through a column, it is bound to silica for much of the time. Putting it another way, the silica helps to solubilize such a compound. Maybe the model is wrong or is only correct some of the time. However, the most important question is whether or not to change our technique for this compound or other compounds in the future. Thoughts?