Matching columns for tandem HPLC

[Mark Tracy]

Dear readers, please bear with me while I do a bit of informal market research.

I have a question about column matching. For tandem columns in a high throughput LC system, what are your expectations for the degree of column matching? Obviously, you want the retention times of the columns to be the "same" but what does that mean? Exactly how much mismatch is tolerable? What about peak shapes? Are there any official guidelines?

I would appreciate your opinions on this subject.

[unmgvar]

Hey Marck
your question actually touches many parameters. I hope that i will answer most of it's facet, especially the important ones.
1.the USP has created a special specification for columns. in order to create some orderin the chaos.
this set of specifications permits you to change column manufacturers if you stay within the system suitability criteria of your method.

2. both USP and EP have come out with a set of rules in order to be able to change columns if for exemple a certain type is no longer manufactured (this is for the common columns not the chiral). you can change column lenght, particle size and particle diameter if necessary, but only according to the rules that they have set.

3. EP has a chapter concerning what it regards to be correct peak shape and other parameters. USP has also a chapter like this but from what i know there are trying to implement a new set of rules that currently can be found in the forum and that serves has a guidelines for the industry.

for example according to EP, unless stated otherwise in the method, a correct peak shape should be between 0.8 to 1.5.

4. All and all ot os in the end up to you. you have to find, check, prove and validate your methods in order to prove robustness, repeatability, accuracy, precision and so forth. in the future you will only be able to accept mismatch that will be in accordance to your system suitability parameters. if you were flexible in the method validation then you will easily be able to switch column from different manufacturers.
if you have all the "goodies" in you system suitability RT, RRT, k', resolution, tailing, plate counts, S/N, and you can't achieve one of the list then you must stick to the column that will get you all those parameters.

hope this helps

[Mark Tracy]

Actually, I deliberately left the question vague. The situation I'm interested in is a little different that the answer you gave. You spoke of the situation of substituting columns of different make. I am talking about the situation of matching between columns of the same make. If a vendor sells you a "matched pair" of columns, what do you expect that to mean?

[tom jupille]

Mark, I'm probably not a valid data point for you (I come from the vendor side of the desk), but my $0.02 worth would be two columns from the same lot of stationary phase.

[Mark Tracy]

Well, since I too come from the vendor side, that would be my natural answer also. I certainly don't expect other vendors to share their hard-won market research. I was just hoping that some of our readers are doing or contemplating tandem LC for high throughput work.

[Victor]

Mark,

I think users might be interested in the resolution of a critical pair of analytes in the mixture they were using the columns for, using their normal mobile phase and other conditions. This parameter would encompass some aspects of retention time and peak shape. Presumably in a high throughput situation the resolution of such a critical pair is not generally vast, otherwise it would not be a high throughput situation....