Paper chromatography

Why do the colours in a spot of ink separate as solvent creeps past them?

A dot of black ink on filter paper, the bottom edge in solvent, and twenty minutes later there is a ladder of colours strung up the strip. The classroom explanation is that lighter molecules travel faster and heavier ones lag, which is intuitive and mostly wrong. Change the solvent and the order can rearrange itself, sometimes reverse. Whatever sets a dye's position, it is not a property the dye carries around with it.

So here is the better question. If the same molecule can end up near the top with one solvent and near the bottom with another, then position is not being read off the molecule at all. It is being produced by a relationship. Between what and what?

Notice first that the paper is not simply a road. Cellulose is covered in hydroxyl groups and holds a layer of water bound to them, which air-dried paper carries whether or not you added any. So there are two liquids in play: that stationary bound water, immobile in the fibres, and the developing solvent climbing past it by capillary action. Every dye molecule moves constantly between the two.

Now ask what a molecule's speed actually is. When it is dissolved in the advancing solvent, it moves at the solvent's speed. When it is in the bound water, it is stationary — the paper is not going anywhere. Those are the only two options, and neither is slow. There is no "moving slowly" in this system at all.

Which means the average speed of a molecule is the solvent's speed multiplied by the fraction of time it spends in the mobile phase. And that fraction is set by an equilibrium: how the molecule distributes itself between two solvents it finds differently comfortable. One that strongly prefers the bound water sits out most of the run; one that prefers the developer rides most of it.

This is exactly what the retardation factor measures. Divide the distance a spot travelled by the distance the solvent front travelled and you get a number between zero and one — and that number is the fraction of time spent in the mobile phase. Not a speed, not a mass: a partition, expressed as a ratio of distances.

Now a consequence worth pausing on. Each molecule crosses between phases thousands of times over a run, and each crossing is a matter of chance, so two identical molecules take different numbers of steps and end up in different places. The spot spreads. But note the shapes of the two effects: separation between two dyes grows in proportion to the distance run, because it is a difference in average speed, while spreading grows only as the square root of it, because it is a random walk. Run the plate longer and the gaps outgrow the blur — which is why chromatography works at all, and also why it yields diminishing returns rather than perfection.

There is a second reason the thousands of crossings matter. Because exchange is fast compared with the run, every molecule of a given dye samples roughly the same balance of the two phases, so they stay together as a band. If exchange were sluggish, molecules caught in the stationary phase at the wrong moment would fall behind, and the spot would smear into a tail.

The relevant contrast is polarity-like affinity, not size. Because the stationary phase on cellulose is aqueous and the developer usually less polar, more polar solutes generally sit lower on the strip. But that is a statement about a pair — swap in a more polar developer and the same solutes rise, which is why a retardation factor is only ever quoted alongside its solvent system. Mass enters weakly and indirectly.

The factor is also less portable than it looks. It depends on the paper, the temperature, how saturated the tank's vapour is, and how far the front was allowed to run — and the front decelerates as it climbs, since capillary rise advances roughly as the square root of time. This is why identification is done by running a known standard on the same strip rather than trusting a published value.

One honest uncertainty: whether the mechanism on cellulose is really partition into bound water or adsorption directly onto the cellulose hydroxyls has been argued for decades, and the answer is almost certainly both, in proportions that shift with how much water the solvent carries. The reasoning above holds either way — what matters is that the molecule alternates between a phase that moves and one that does not — but the tidy two-liquid picture is a model, not a photograph.

Finally, the ink is not being decomposed. Black inks are blends of separately manufactured dyes; separation reveals a mixture that was always there.

• Why running a second solvent at right angles to the first separates things a single run cannot.
• How the same partition logic scales into gas and liquid chromatography, where "distance run" becomes "time to elute".

Every term the collection defines is gathered in the glossary.