I have been in a discussion with some of my colleagues and they are convinced that reducing the internal diameter of the HPLC column will increase the efficiency of the separation. I have been trying to rationalize this and have not been able to. I thought the main contributions to efficiency were particle size, stationary phase depth and surface area. My belief was that reducing the internal diameter was mainly to address samples of reduced volumes, and interfacing of the separation with MS or NMR. Assuming the system is setup to handle reduced flow (10 uL/min) the linear velocity of the separation is matched for the two column diameters and there are no extracolumn effects shouldn't the number of plates generated be the same for the 4.6 and 0.5 mm ID columns? The particle size is unchanged at 5 um.

Andrei,

Indeed you are correct in that there is no connection between column diameter and chromatographic efficiency assuming that linear velocity is held constant. It is, nonetheless a commonly held misconception that smaller diameter will result in higher efficiency. In reality there are three main "advantages" of smaller diameter columns: 1) lower volumes of eluent are required for given separation, 2) higher mass sensitivity is achieved if one does not scale injection volume with column diameter (this can be advantageous with limited amounts of sample) and 3) generally smaller diameter columns have higher permeability (i.e. the column back pressure is lower for a given column length when the internal diameter is reduced). Generally, the reason people think that higher efficiencies are achieved was smaller diameter columns is that higher total plate counts are readily achieved was smaller diameter columns because it is simpler and more convenient to use columns that are longer than 0.25 m. In essence, the misconception is based on comparing apples and oranges (i.e. a short 4.6 mm ID column is compared with a 0.5 m long 0.5 mm ID capillary).

I agree with Chris on everything except for the pressure effect. As long as the internal diameter of the column is larger than about 30 particle diameters, there is no effect of column diameter on pressure. If effects of this nature have been observed, they are due to the pressure in extra-column tubing.

The formulars which are related to resolution do not have diameter as a variable. Conceptually, nevertheless, it is hard to imagine that application of the sample in larger diameter columns doesn´t result in a broader initial band (thusly a broader final band). Any thoughts on this?

Actually, cm for cm (length)you generally lose a few theoretical plates as you go down in id because of the relatively greater contribution to band broadening of the "wall effect."

I agree with Uwe that theoretically there shouldn't be a major difference in operating pressure between small and large diameter columns. However, for practical reasons it is commonly observed that smaller diameter columns have a somewhat higher permeability (i.e. they are more poorly packed). This isn't to say that this is always the case but it is often the case. I have spoken to several researchers to have tried to transform this problem into an advantage by focusing on this "improved permeability".

As I said, this is true in capillary columns, say at diameters of 320 micron and less, packed with 10 micron particles. For larger diameters, it is hogwash.

Cohesively speaking, it seems to be dangerous these days to think about permeability, monolithics, turbo flows, turbulences, cyclones in the column, perfusions. Are we being pulled over the table?? Is chemistry changing to an innovation of words, of playing with the PC, of fleeing into super special specialties (one molecule anal., etc.)?
More seriously: I burned my fingers, long ago, on a micro column, havn´t tried smallies since, but nevertheless, keeping an eye on this micro - nano development.

Thank you all very much for your in-depth replies to my query. I hope that others have learned as much as I from these responses.

Andre,
One bit more, post column band broadening destroys resolution. The detector itself often contributes a large amount of broadening.
In GC to a FID make-up air alleviates the broadening. You can do it in LC also with a tee and another pump. Peak shapes will sharpen and resolution improve. Without a single change to the column, plates become smaller, etc. Magic.
To get a good comparison, use a smaller column with a smaller tubing, with a smaller detector volume, etc. Or compare with make-up flow, otherwise the comparisons do not transfer from instrument to instrument and you will get discussions taking all sides of the argument with equal vigor.
Jim