this may be an odd question, but if you inject on a reverse phase silica c18 column, (prep scale=large column) and you overload it. When it comes out, will you lose compound somewhere else. It does not come out in the wash (assume the wash is stong enough to remove everything) and minimal loss in the solvent front. This has now become a debate between a colleague and I and i would like a second opinion. I got a low yield on a purification and we are wondering where it went. Any insight. I can think that when you load it will eventually come out. the loss was very large. If you need more info let me know

thanks to all who would like to contribute to this debate

Losses in prep are often associated with a difference in time between the peak showing up in the detector and the time the material shows up at the end of the collection tube. Don't forget that the outlet tubing of the detecor has a large diameter and therefore a large volume. This is where I would check first.

thanks uwe. The distance between the detector and my collection flask is not long and i do manual collection. I think that the crude i was purifying contained more salt than it should have. the crude was really 96% pure anyways. I started out with 1g and ended up with 550mg. I was purifying really for safe measures. I always like to pass anything through the hplc to minimise any future problems. a quick pass never hurts unless you loose too much of course. The crude was passed through a gel column to desalt before given to me. Are there any chances of there still containing any salt in this crude after gel. Keep in mind the gel columns used are used frequently with different compounds, over and over and over again.

thanks

Low recovery in prep LC is common - the problem usually is not loss on column but impure crude. How do you know it is 96%? HPLC at higher wavelengths? Then you might miss most of the impurities. What about water content? Alkanes?
When I did natural product isolation from plants 10-50% recovery was usual. The "purified samples" were pure by HPLC but when given to NMR we often found alkane signals (column bleeding?) and of course water and methanol.
A colleague once recieved a sample for LC-NMR together with a chromatogram with only one peak. However, he wasn't even able to see that peak in the LC-NMR, but the first chromatogramm was recorded at 450 nm. Guess what he saw when he looked at 220 nm.