Chromatography Forum

There are well defined definitions of turbulence, and there is no turbulence in HPLC columns.

There is probably also no turbulence in knitted tubes, but I have to admit that from a macroscopic point of view, it is not possible to tell the difference between pure secondary flow (=radial mixing) and turbulence. The performance (HETP) of a knitted tube is as good as a straight tube would be under turbulent conditions, and - contrary to straight tubes - one can not see a boundary between laminar flow and turbulent flow in knitted tubes, neither from the standpoint of performance nor from the standpoint of pressure.

In straight tubes, the onset of turbulence is marked by a sharp increase in pressure as the flow rate is increased. This does not happen in packed columns, where the increase in flow rate results in a directly proporational increase in pressure. No change in permeability = no change in flow character = no evidence of turbulence. Hans, I hope that this answers your question.

In knitted tubes, the change in the permeability is rather gradual, and it happens at MUCH slower velocities than those that cause turbulence. The ratio of the onset of secondary flow in well designed knitted tubes to the onset of turbulence in straight tubes is in the order of 200+. That is about the difference between a jogger and an airplane...

Hi Hans,

The stirring thing:
I gave this example, because it illustrates the contribution of motion generally.
But the following is a more detailed description:
Molecules of a solute will diffuse in the surrounding environment (the solvent) creating a somewhat saturated layer, which will slow down the diffusion of more molecules until the partially saturated layer “loosesâ€

One needs to distinguish between a patent and science. The original patent on "turbulent flow" in chromatography is based on a misinterpretation of the experimental results obtained. But it is still a patented chromatographic technique...

Hi

I don’t know how valid this point is but here goes, - does this mean that the flow pattern in a column having 80 Ǻ and 1000 Ǻ column will be same??

At 1000 A you are getting close to the grey area, but for a 5 micron packing, the ratio of particle size to pore size is still around 50, and I do not think that you can get any measurable perfusion effects yet.

I did some experiments, maybe two years ago, on peak broadening, etc., due to laminar flow. I tried to get peak sharpening (due to Uwe´s explanation I am not sure whether I can say that I tried to get turbular flow) without shortening the distance a tube can cover (that is, no knitting). This was reasonably successful (maybe I should write this up?). If I remember correctly there are publications by Uwe which showed that flow formulars (have to look those up again) didn´t yield parameters which indicate turbular flow in twisted, and so forth, tubes. I just wondered at the time whether these parameters are indicative of peak shape changes in modified tubes. (It would take too long to review all of this for a better presentation, but maybe I should restate what I did: I twisted and dented tubes with very little change in distance that could be bridged with this tube, but got a strong improvement in peak sharpness over unmodified tubes. I don´t know whether flow formulars and/or pressure measurement would have indicated a change from laminar to turbulent flow).

Danko, I am not sure of your argument anymore, but the initiation of bulk movement has long since been used to speed mixing, which can be extremely slow if diffusion is the only operating mixing mechanism. Are you trying to say that diffusion into pores creates bulk movement (flow)?

Hans, I believe you’re joking now.
If I’m wrong - then you’re contradicting your self.
I thought I stated a number of times, that any motion/movement, resulting from the interaction between the flowing mobile phase and the “stagnantâ€

I think there are no turbulences inside the column, it's all laminar (isn't that the credo of microfluidics?). Turbulent flow occurs at high Reynolds numbers.

Putting in some (generous) numbers:
* flow velocity v = 10 cm/s (= 6 m/min)
* characteristic length (diameter) l = 10 µm
* => Re = v l rho/eta = 1.1
(with density rho = 1000 kg/m^3, viscosity eta = 8.91e-4 Pa s)

This is several orders of magnitude below the limit of Re ~ 2000.


http://modular.mit.edu:8080/ramgen/iflu ... er_Flow.rm
..something for your morning coffee :-)

Danko, sorry I confused you, but the term "bulk movement" is not a synonym for turbulence. One can create all kinds of bulk movement: laminar, turbulent, chaotic....

I swear I’m of the same opinion
And I believe I wrote “any kindâ€

Just read Uwe´s Feb. 28 contribution again, + moino´s. That seems to explain it all, I think that I finally am beginning to understand this flow stuff. I was hooked on the idea that turbulent flow was necessary to get mixing...., obviously wrong.