Hi -

Can anyone give me a reference that describes the conversion of LC methods from conventional HPLC columns (4.6 mm i.d.) to 3.0 or 2.1 mm id columns?

This will get you started....

http://www.waters.com/waters_website/pdfs/wpp207.pdf

For isocratic methods, I've used half the original flow rate for 3 mm i.d. columns and one-fourth the original flow rate for 2.2 mm i.d. columns. For gradients, trickier because dwell volume plays an effect. Sometimes I've also decresed the injection volume in half.

Assuming you are keeping the same column length:

1. Scale the flow rate to the square of the column diameter (e.g., going from 4.6 to 2 mm, drop the flow rate by a factor of 2.3^2 = 5.3)

2. Ideally, decrease all of your other volumes (injection volume, extra-column tubing volume, gradient dwell volume, detector cell volume, etc.) the same factor.

If you are also decreasing column length, then you may also have to adjust gradient time and/or flow so as to keep k* (and therefore peak spacing) constant.

-- Tom Jupille / LC Resources Inc.

A couple of references, both from John Dolan's "LC Troubleshooting" column in LC-GC:

LC-GC, 8(2) 98 (1990)
LC-GC 18(10) 1034 (2000)

I'll sneak in a plug here: we have John's columns from 1983-2000 available as Adobe pdf files with an index on CD-ROM:
http://www.lcresources.com/Resources/restscd.html

-- Tom Jupille / LC Resources Inc.

I have just set up a fluorescence system using 2.1mm columns instead of our 4.6mm id columns. The tubing from the column outlet should be 7thou id. I was surprised just how much difference this made.
As previously stated you must be aware of your dwell volumes if you are using gradients. I would go as far as to say that many older instruments are just not usable with this size of column!

For you trivia buffs: "standard" 0.010" id connecting tubing has about 0.5 microliters of volume per cm of tubing length.

Cutting the diameter to 0.007" cuts that in half (volume is a function of the square of the diameter), so you also have to be careful to minimize tubing lengths and downright paranoic about how you assemble fittings.

While I "almost" agree with the dwell volume comment above, I would amend it to add "without significantly modifying the method". It *is* possible to work around excessive dwell volume (e.g., by using a delay-to-injection, or by re-optimizing the gradient profile). In general, the result is a significant decrease in throughput.

-- Tom Jupille / LC Resources Inc.

Tom - If you use the delayed injection principle fully, there should be no reason for a decrease in throughput. The next injection can be prepared while the instrument is going through the re-equilibration cycle. Think about it in the following way: you create a sequence of gradients and reequilibration cycles as you used to run them on the larger diameter column with the higher flow rate. Now you time the injections (by delaying them) to happen at the same relative time in the cycle. Since the cycle times remained constant, your analysis times remained the same.
Best regards
Uwe

Oops! You're quite right with respect to what I said, but not with what I *meant* to say! :-)

You are correct in saying that the delay-to-injection technique by itself does not hurt throughput. Scaling down column diameter and flow rate without changing the dwell volume, however, *does* kill throughput because you still have to pump at least 1X the dwell volume for each injection. At 1/5 the flow rate, this takes 5X as long.

-- Tom