When I first started doing column chromatography, I had a very particular idea of what a “good column” should look like. I preferred to use a long and thin column, with dry loading, and I liked to run it by gravity. To be fair, the separation was usually pretty good. The problem was the time. A single column could easily take me an entire day, and sometimes even two days from packing, running, checking fractions and finally collecting the product.
At that time, I thought this was simply the price you had to pay for good separation. I thought that a longer column would give the product and impurities a longer distance to travel through the silica, so they would have more opportunity to separate from each other. I also thought that running the column by gravity was somehow “gentler”. In my mind, using pressure might force the compounds through the silica too quickly, while gravity would reduce unnecessary interaction to give cleaner separation. And because the results were often acceptable, I did not question it very much.
That changed after I discussed my column technique with my supervisor. When I described how I normally ran a column, he was quite surprised by how long it was taking me. His approach was almost the opposite of mine. Instead of using a very long and thin column, he would normally use a much shorter column with a larger diameter.
And instead of relying on gravity, he would run it as a flash column, applying pressure using a pressure bulb. His target was to collect all the fractions within around 15 minutes for a straightforward separation. That sounded extremely fast to me at the time.
“Slower” does not automatically mean “better”. As compounds move through the column, their bands gradually broaden. If you run the column unnecessarily slowly, the product band and impurity bands can become wider, which may make it easier for them to overlap.
The other important thing I learned was how to think about reaction scale. When the amount of crude material increases, you do not necessarily keep making the column longer.
In many cases, you increase the diameter of the column so that the silica can handle the larger sample loading, while keeping a sensible bed height. That was quite different from the way I originally thought about column design. For a straightforward flash column, I now try to prepare everything in advance — solvent, collection vessels, expected column volume and fraction size — and then collect the fractions continuously before analysing them together by TLC.