Serpentine flow in post column derivatisation

[Ary]

Has anyone ever tried to 'wind' a stainless steel tube in order to reduce band broadening? I am stuck with a particularly difficult post column derivatisation method which requires a long analysis time. From what I understand using a serpentine tube improves thermal transfer and reduces band broadening significantly. However what info is out there and readily available on line is a little scant on actual practical details. Can anyone help?

[Mark Tracy]

Yes serpentine winding does help, especially with tubing diameters over 0.3 mm i.d. Almost all commercial reactor coils are convoluted in some way. You can buy knitted coils from Dionex (my current employer) or Pickering Laboratories (my previous employer) or a couple of other vendors. Unless there is some specific reason for using stainless steel, use PTFE, which is what almost all commercial products use. For obvious reasons, nobody knits steel tubing. For a quick hack using steel tubing, repeatedly fold the tube in half until it fits your heater, or becomes too difficult to bend any more.

If you don't mind me asking, what is the application?

[Ary]

Unfortunately I cant use anything that is gas permeable hence the stainless(I'm trying to measure CO2, like I said a somewhat complex method for LC). I've got some thin walled stainless which is easy to bend but I'm not sure how best to get the desired results. Theory seems to suggest the diameter of the bends needs to be less than 3x the internal tube diameter though some real results suggest I can get away with less than this. As the id is 0.25mm that means the bends need to be 0.75mm and as the od is 0.6mm its quite a tall order. Any ideas?

[Mark Tracy]

How long is your tubing? When I was at Pickering, we used 9 meters of 0.28mm tubing for a 500µL reactor without any knitting. The band-spreading was about an additional 15% in peak width at 1.3 mL/min. I believe they have since started knitting, but it was acceptable then.

I suspect that attempting to get a radius of 0.75 mm will collapse the tubing; 1-2 is probably all you can practically get. I have bent 0.5 mm id x 0.8 mm o.d. tubing, and it kinked very easily; a few kinks didn't cause trouble, but it was worrisome.

Try this: pound two parallel rows of nails into a thin strip of wood. Space the rows about 1 cm apart, and have about 1.5 cm of nail protruding. Fold your tubing in half. Hook it over the first nail. Wind the doubled tubing in an S-pattern over the first pair of nails. When it fills up, move to the next pair. When you are done, put the whole works in a forced-air oven.

316 stainless steel tubing is soft when you first get it, but it gets stiff as you bend it, so be patient and have plenty spare tubing.

Best of luck.

[ravenwork]

Analytical Scientific Instruments (ASI)
El Sobrante, CA
800-344-4340

makes some excellent SS reactors that produce very minor band broadening. I am using one, and I am very pleased with it.

I am not affliated with ASI other than using one for their reactors.

Evan L. Cooper, Ph.D.

[Ary]

The tubing I'm currently using is 1.5m long though I've inherited this and the initial experiments I've run suggest I am getting incomplete reaction which is a pain as I am trying to do quantitation, so I will probably need to go even longer(already maxed out on my heater at 150C). Thanks for the tip on winding the reactor and thanks also for the tip about where to get off the shelf reactors.

[HW Mueller]

Does anybody have some literature on this serpentine bending? I tried to reduce peak spreading with ~0.8 mm tubing. Knitting and knots every ~5 cm was best (plastic tubing). I had some success with alternating twists and dents in steell tubing (without much change in length, or better, reach of the tubing). Bending (radius too large?) or twisting alone were useless. There is an extensive literature on knitting, much of it by Uwe Neue and a group in Saarbrücken.
I also checked into 0.1 mm PEEK tubing which gave ~ no band spreading, so Mark´s results are not surprising to me. Right now I don´t remember: Is there too little mixing in 0.1 mm tubing for reactions?
Ary, did I miss something, or is it possible to slow your flow?

[Ary]

PEEK may or may not be a problem for me to use. Its something I need to check out purely from a chemistry point of view. The guy I've picked this up from "thought" it would be a problem but never actually checked so that's another experiment I have still to try. Unfortunately I can't give out exact details of what it is I'm doing so apologies to all if I appear a little vague or evasive at times. The flow rate is 0.2-0.5 ml/min depending on the exact application though again from what I've learned so far I cant see any reason why it shouldn't be fixed at 0.2ml/min. All in all I've picked up a bit of a dogs breakfast, all I have to do now is turn it into champagne and caviar. All in a days work eh!

[Einar Ponten]

One relatively recent paper deals with some of these things.

"Extra-Column Band Spreading Concerns in Post-Column Photolysis Reactors for Microbore Liquid Chromatography"
Current Separations 17:1 (1998), Arlen D. Kaufman, Peter T. Kissinger

We produce KOT (Knitted Open Tubular) reactor hardware based on the compiled conclusions made in the literature taking practical considerations into account. As a rule of thumb for PTFE tubing a linear flow rate (cm/s) > 10 is sufficient to get efficient mixing essentially without any band spreading. Based on this the suitable KOT length can then be selected from the desired residence (reaction) time. Often the required reaction time is less than initially expected and this can probably be attributed to the very efficient mixing within the KOT.

In practice one may also need to adjust for the type of solvent (viscosity) and temperature which influence the flow profile within the tube.

http://www.sequant.com/products/kot/kot.php

[charlier]

You can also read a short discussion of serpentine tubing at http://www.chromatography-online.org/EC ... /rs27.html


The orginal reference for the serpentine tubing is:
E. D. Katz and R. P. W. Scott, J. Chromatogr., 268(1983)169.

I think there is a method for making it in the experimental section, as I recall it was a fairly simple bending device.

[Uwe Neue]

I am the original inventor of these three-dimensionally bent tubes. Here is the simplest and best approach to make such tubes from steel. It involves two steps:

Step 1. Bend your steel tubing in tight bends that oppose each other. The final shape should look like a sine wave.

Step 2. Take this tube and bend now every upward or downward turn in the third dimension by 90 degrees away from the previous turn. Two right turns followed by two left turns. You will end up with a shape that looks like a meander from the top.

The result is a tube with a 3-D wave pattern and optimal bandspreading properties. Plus, it is fun...

[Ary]

Thanks everyone and especially Uwe for your time. I'll let you know how I get on. There's actually a craze for kids in the UK for making these things from thin bits of plastic at the moment so I've been getting tips from my daughter too.