Resolution parameter during method Development

[Charles A. Burger]

One topic I've not run into while looking at articles on method develpment is when and how to set parameters such as Resolution, Tailing, k' and etc. I bring this up because of a method I recently ran where the resolution was NLT 5.0. It was a residual solvents method with 5 solvents and the critical pair was Methylene Chloride and n-Hexane.

With a brand new column, my best result was 5.2 which is not a lot of room to work with on a brand new column. I can understand having resolutions that appear to be excessive for related compound/impurity methods because of the possibility of unknowns being present, but these are fixed solvents. Would the method still be good if 4.0 was chosen instead of 5.0?

[zokitano]

The resolution factor is efficiency indicator of the column. It shows the efficacy of the column to separate two different peaks (compounds).
Your goal is to separate residual solvent by GC. As you previously said about the resolution needed when assaying related compounds/impurities, minimum resolution have to be achieved when running residual solvent analyses also (they are also one impurity type).

Best regards

[Dan]

Charles,

If you need to follow the USP or EP, then some limits for system suitability parameters are predetermined for you. The USP states you should evaluate resolution, repeatability and tailing.

With 5 injections, the RSD is NMT 2.0%. So, the limit is already set for you.

People like to see nice peaks and a tailing of NMT 2.0 makes for good peaks and good chromatography. Again, the limit is already set for you.

Resolution is another matter. Typically, you would want a minimum resolution of 2.0 for baseline resolved peaks. However, I have seen many methods (ones I worked on and also mongoraph methods) where the resolution limit was much higher, NLT 10.0 is not all that unusual. Where does such a value come from? It comes from the method development. However, the final limit should come from the method validation.

Where I have worked (pharmaceutical labs), we propose a limit for resolution (or theoretical plates) in the method validation protocol but state in the protocol that the final value will be determined from the results of the validation experiments. The proposed value comes from the development work. The final value will be based on the lowest value that we saw in the validation. We typically take 80% of that lowest value as our final limit. That gives a bit of leeway so that we don't end up shooting ourselves in the foot in the future for the approved method.

So, if I had a method that the development and validation showed a resolution of around 5.0 as the lowest value, then I would put a limit in the approved method of something like NLT 4.8 (maybe even 4.5). It depends a bit on how much resolution is enough for a good chromatogram.

I have seen methods where the resolution was set at a value that was typically seen during development/validation and then seen people go nuts trying to get that value when they ran the method.

Taking 80% of the lowest value seen in the validation has always been useful for me.

Regards,
Dan

[Bruce Hamilton]

One thing I've noticed is that different brands of columns that are supposedly the same stationary phase can offer differing resolution of critical pairs of compounds.

I noticed this especially with polar capillary columns, such as the 624 ( cyanopropylphenyl polysiloxane ), which is often used for residual solvents. In one case, changing the brand of column for a 624 phase greatly improved the separation of IPA and DCM.

Please keep having fun,

Bruce Hamilton