Chromatography Forum

We are checking one method by using Waters alliance HPLC. The method was originally developed on Agilent HPLC .

Our problem is that system suitabilty is not passing but retention
time and elution order is matching.

Criteria for sytem suitabilty is

1.Resolution b/w peak A and B NLT 14.0 ( Obtained 10)
2.Column efficincy is NLT 3500 ( Obtained 1400)

Is it due to different makes of the HPLC or some other reason could be there ?

Can some one hlp us out to solv ethis problem.

If anyone had faced similer type of problem kindly share ?

Regards

Amit Kumar Jian

So,

probably you're getting high RSD for the Area under curve of your standard, that doesn't match with your method requirements. What kind of injection are you using (manual or by autosempler)?
If your method is reproducible (i suppose it is, when it is already developed) then applying it on different HPLC system should give reproducibe results (of course if your Waters system is qualified and verified, I mean if it is ok and suitable for that kind of analyses).
Is your sample easily degradated? Did you pay attention of its stability, the method of sample preparation, special conditions that should be met to ensure injection of intact chemical(s) in your system?
If you're sure that everything is done right, considering the sample preparation and sample handling, then the possible problem might be in your HPLC system

Best regards

zokitano,

I do agree with your suggestions.

One thing to add here,
Agilent uses std capillaries to connect all the HPLC modules, try mathcing the same with your waters system. probably this may be the problem.

regards

[quote="zokitano"]So,

Dear Zokitano

% RSD is below 1.0%. Molecule is stable . Our HPLC are qualified and calibrated. We have tried other HPLCs also( Waters only).
Can it be a column problem?

Regards

Amit Jain

Amit,

You can check the column first, alternatively you can try running same set of sequence on a validated Agilent system, if it reproduces the same result as earlier, probably optimizing the delay/dead volume on your waters system could solve the problem. if not chek the reagent & sample preparation as suggested by zokitano.

regards.

Amit,

You can try other suggestions then.
regards,

Dear Amit Kumar Jain

I suppose then that this is a column problem. You are able to reproduce the exact retention times of your analytes, but you lack column efficiency to separate those peaks according to the limits set in your method.

Try to change the column and use another equivalent column to ensure yourself that this is/isn't column problem. Also check your guard column. Maybe you should replace it (check the peak shape, if they are irregular<peak tailing or fronting> this may be an indicator to use new guard column)

Also you could make a slight change in your method (according to pharmacopoeia if you run pharmaceuticals) for example: small change in the ratio of the solvents that you're using in your mobile phase, small change in flow somethimes could give better results. If you're using ACN in your mobile phase, a little adjustment (increase) of the percent of ACN can give better shapes of your peaks and better resolution or number of plates.

The changes that are allowed for liquid chromatography method (by the European pharmacopoeia) are not more then +/-30% change of the minor component in your mobile phase, and +/- 2% absolute change of that component in your mobile phase. No other component in the mobile phase should be changed more than +/- 10% absolute. If you don't want to revalidate the method you should make the changes within these/or others Ph's limits

Hope that this will help you

Best regards

1) is it a gradient method?

2) Agilent - single or binary pumping (high pressure gradient mixing is a little different from low pressure gradient mixing)?

3) With that loss of column efficiency, my first thought is that the column you used may have been used for something else in the past and that it is no longer suitable for your method, or that there is some difference between the procedures as run by each lab.

If the retention times match closely, then your problem is entirely due to a loss of efficiency (theoretical plates). There are really only three possible causes that come to mind:

1. A bad column (as has been suggested by several people). Try a new column and (if you have access to the Agilent system, try swapping columns).

2. Excessive extra-column volume on the Waters system. Possible causes:
- transfer line is bigger diameter than it should be
- large sample loop with the injection valve plumbed backwards (it happens!)
- poorly assembled fitting(s)

3. Sample diluent is stronger than your initial mobile phase (in this case, extra-column volume helps because it allows mixing with the mobile phase before the column inlet).

Dear DR / TOM

Yes, method is a gradient method.

Column we used new only.

Diluent is mobile phase itself.

RETENTION TIMES OF THE PEAKS ARE 1 and 2 minutes respectively.

Regards

AKJ

RETENTION TIMES OF THE PEAKS ARE 1 and 2 minutes respectively.

Whoa!! what are the column dimensions and flow rate? If you are using a very small column, this greatly increases the probability that the problem is caused by extra-column volume.

[quote="tom jupille"][quote]

Column dimension is as folows

50*3.0mm, 2.5um

USP L1 Packing

OK - if we rule out bad column, too small a column, and strong solvent effects, and take into consideration that it is a gradient method, I would check into the dwell volume of each system. If there is a "large" difference between the dwell volumes, this will cause issues with reproducibility between the systems.

Please calculate the dwell volume (no column) for each system by seeing how long it takes to elute a trace amount of acetone.

But a dwell volume difference shouldn't affect the peak width!

I'm betting on extra-column volume. This is an out-take from a paper we gave at the Pittsburgh Conference in 2006. It plots predicted resolution as a function of extra-column volume for different size columns. It doesn't take much to kill resolution on a small column.



By the way, I'm sidestepping the whole issue of plate count. How is it being measured? The isocratic formula N = 16*(tR/wb)^2 is not meant for gradient conditions.