[Ingeniosuccinimide]

Is there some guideline or a rule of thumb when choosing a column size for flashing? Here are the sizes we have available.

Do I decide on the basis of the approx. mass of the product in my crude, or the mass of the whole crude, or Rf of the product from a TLC plate? Also, do I fill the column with the stationary phase up to half, or 3/4, or how do I decide on that?

[TheUnassuming]

I'm assuming these are for some manner of auto-column/isco system? 

The amount you need to use will depend on the quality of the silica/packing, but in general assume the same ratios of material:silica you use when running a manual flash column.

[pgk]

The length of the column in relation to the cross sectional diameter, depends on the desired number of theoretical plates. The higher the number of theoretical plates, the better the separation is. An even higher number of theoretical plates, is necessary for flash chromatography. Methods for calculation of the necessary theoretical plates in relation to the Rf, are described in specialized chemical engineering books. But this is too much for ordinary lab work.
However, the Rfs in TLC and in the column, are not the same (even the static phases are not the same). Besides, retention of the substance, is significant in longer columns.
In practice, you choose long columns with short diameter by respecting the mass ratio of substance per static phase to be lower than a few per cent (e.g. 2.5-3.5%) depending on the chemical structure of the substance and the absoptability of the static phase (but this completely empirical; might be more, might be less). Thus, if the column is half-filled or not, does not matter and in addition it creates dead space.

[kriggy]

I asked similar question on reddit and got pointed to this article
http://pubs.acs.org/doi/abs/10.1021/jo00408a041

[TheUnassuming]

^ classic paper.  Easy to forget that something so common in lab was actually "discovered" not that long ago.

[tomek]

There's another paper I would recommend reading:
http://www.sciencedirect.com/science/article/pii/S0021967308016683

[clarkstill]

Clark Still, what a dude. 

[Ingeniosuccinimide]

In practice, you choose long columns with short diameter by respecting the mass ratio of substance per static phase to be lower than a few per cent (e.g. 2.5-3.5%)

What do you mean by substance? Approximate/theoretical mass of the wanted product in the crude or the sum mass of crude with all the components included?

depending on the chemical structure of the substance and the absoptability of the static phase (but this completely empirical; might be more, might be less). Thus, if the column is half-filled or not, does not matter and in addition it creates dead space.

Yes, this does not help me a lot. Basically I'm curious which rules do people follow when choosing the diameter and length of columns for both flashing and classic column chromatography and the amount of silica for these also. You can write your own experience if you wish.

[Dan]

Everyone has their own system, but for me:

Generally speaking, the volume of silica (in mL) I use is between 200 and 20 times the mass of the crude (in g), depending on difficulty of the separation (more silica for a more difficult separation). I use 40-63 μm silica. I select a column of a diameter such that the depth of the silica is 10-20 cm.

For elution, I normally use a step gradient whereby the solvent system in which the target material runs at about R_f 0.5 on TLC is approached over about five column volumes (a column volume = volume of silica being used). For a difficult separation I might reduce the gradient and approach the final solvent system over about ten column volumes. I almost never use isocratic elution as I find it is inefficient in comparison.

Example: For a fairly straightforward (in terms of separation) purification of 1 g of crude, where the target material runs at R_f 0.5 in 25% EtOAc/hexane, I would load the compound onto 100 mL of silica and elute with:

5% EtOAc/hexane (100 mL)
10% EtOAc/hexane (100 mL)
15% EtOAc/hexane (100 mL)
20% EtOAc/hexane (100 mL)
25% EtOAc/hexane (100 mL)

Usually by this point the column is finished. If not, I would continue and run 30%, 35%... etc.

[pgk]

1). I mean the mass of crude product per silica. The higher the mass ratio the better the separation is. On the other hand, the higher the mass ratio the compound retention is. Therefore, compounds that form hydrogen bonds with silica, e.g. alcohols, phenols and primary and secondary amines require a lower mass ratio per silica, e.g.  5-6% w/w.
2). The theoretical analysis of separation of separation is based on the number of theoretical plates. But this is a rather complicated calculation that demands the knowledge of some physical properties of the compounds. Therefore, there is a limited use in separation of new compounds with unknown properties. In practice, you respect crude product/silica mass ratio, you choose columns of > 10/1 of height/diameter ratio and you follow Dan’s advices.
3). Instead of pressure, you can also use a slight vacuum. But take care: If the vacuum is high, all constituents will be eluted, together.