What approach would you recommend for developing a method (e.g. related subs)quickly?

Approach 1:
Hook up 5 different columns (using a column switcher)-C18, C8, phenyl, cyano, amino

Use 2 solvents -buffer (pH 2.5 PO4), ACN

Run a steep and shallow gradient (e.g. 90% buffer to 30% buffer with ACN). Evaluate optimal condition using DryLab or other means. Check similar solvent strength MeOH and THF. Blend solvent perhaps.

Repeat at different pH (e.g.6)if the above did not work.

Approach 2:
Evaluate one column at a time using different organic and buffer (pH 2.5, 6). Then use different types of columns (c8, phenyl, cyano, amino).

Analyze the results in DryLab or other appropriate means.

Would you explore different columns (c8, phenyl, etc) and pH upfront or later? I trying to come up with a generic kit to quickly develop a method.

First of all I have been developing methods for years and have never found a simple quick method for the development of methods. But, research about your analyte prior to the actual analysis will speed you up, for trial and error takes a long time.

What is your main component like? I know you wanted a related substances method. But, as the name implies the impurities are typicaly related. If the analyte is polar the impurities most likely are. Check out all information on the main component. If a assay method is known and works well start there.

What are the known impurities? The intermediates used in the synthesis?

These types of questions can get you in the right area for pH, buffer strength, and column type. Now, having a column switcher would be a great asset for even if you know that a C18 column would be good there are so many to chose from.

I would definitly start with a gradient and have it end with a large amount of organic to ensure all impurities are off. I have seen many methods miss impurites by stopping to early.

If not familiar with what change is needed to make the desired change from the gradient then Drylab or something simular is quite helpful, it also works well as a teaching tool.

I hope this has been of some help, even though it wasn't a step by step method to do it. There are many books and programs dedicated to this subject. Reviewing some of them would help get a feel for what needs to be done.

CN and NH2 are no good in RP mode, definitely not amino. If your samples are maines or acids, you should also change the pH.

Focusing on reversed-phase LC, there are only a half-dozen or so parameters that you can use to move peaks around:
- mobile phase strength (%B or gradient time)
- temperature
- mobile phase type (MeOH, ACN, THF)
- pH
- additives (type and concentration)
- column type

The order in which those variables are explored depends a great deal on what you know about the chemistry of your analytes (as pointed out by M Emerick) as well as on your past experiences and preferences. For general-purpose use, I like to use the following (pretty much following Snyder: "Changing Reversed-Phase High Performance Liquid Chromatography Selectivity. Which Variables Should be Tried First?" L.R. Snyder, J. Chromatog. B, 689, 105 (1997).):

Start: wide-range gadients; A=low pH buffer, B= ACN. Reason: the low pH will tend to protonate residual silanols on the stationary phase (minimize tailing) as well as "ion suppressing" weak acid analytes. I prefer ACN over MeOH because it has a lower viscosity (hence lower pressure and slightly higher plates) and a lower UV cutoff.

First variable: solvent strength (% ACN or gradient time). This is first on my list because you have to get retention (k') in the right region in any case, so you may as well look at peak spacing while you're at it. And yes, I do use DryLab (not surprising).

Second variable: temperature. underappreciated, but surprisingly powerful, especially when combined with mobile phase strength changes. Easy to do if you have a temperature controlled system.

Third variable: solvent type. first to MeOH, then to THF. If the critical pairs change with solvent change, then I'll consider a ternary mobile phase.

Fourth variable: pH (assuming ionizable analytes). My general recommendation is to increase pH in increments of about 1.5 pH units (e.g, 2.5 - 4.0 - 5.5 - 7) and look for major selectivity changes. If I see a noticeable change in selectivity, interpolate (e.g., 4.75) and then use DryLab to map out the details. Ph ranks fairly low for me because of the potential for robustness problems in the vicinity of analyte pKa values.

Fifth variable: additives. This is a catch-all category that includes ion-pair, tailing suppression (e.g., TEA), etc. Unless there is severe tailing, (in which case I will add something like TEA early in the game), I prefer to not resort to these because of the possibility for system peaks and long equilibration times.

Sixth variable: column type. This can be very powerful, but it has the disadvantage of being a "stepwise" change. Any of the other parameters allow "interpolation" when major changes in selectivity occur. If I find different (but inadequate) selectivities with two columns, I have no convenient way of getting something "in between).

When all is said and done, however, you have to choose a strategy that fits in with your samples and experiences.

Hope this helps!

-- Tom Jupille / LC Resources

Thanks everyone. I very much appreciate your comments.

Emerick, I agree that CN and amino may not be a good choice. I prefer C18, C8, phenyl and maybe C3.

Tom, I agree with your suggestions. Your approach is systematic and logical. I have read Practical Method Development textbook and several articles in LCGC (e.g. Method Dev by John Dolan). I don't have as much experience in Method Dev as you but I'm confidence that there must be a fast way of developing a method (esp. difficult methods) using some basic tools (DryLab, Column switching).

Tom, here is my suggestion to your approach above:

Variable 1, 3, and 6 can be investigated upfront by attach several columns (C18, C8, C3, phenyl) and run three different organic (methanol, ACN, THF) with aqueous (PO4, pH2.5). Gradient runs are conducted and drylab analysis used for interpreting the data. The analysis may take a day and interpretion two days.

Variable 4 (low and medium pH, not close to pKa of the compds) can be conducted using several columns with aqueous/ACN or MeOH organic.

Variable 2 I see more as a optimization, however, I did experience a change in profile (relative movement of peaks) with change in temp.

Variable 5 may not be the last desperate option. For example, if you have polar and non-polar compd at a specified pH, you may want to look at SDS method much earlier (as I found out painfully in my last method development experience). Equilibration time is not a problem if your method is isocratic.

My objective is to develop a method fast using a systematic method development approach based on the compd and column chemistry, chrom profiles/maps. I'm confident that we can use a systematic approach in parallel rather than in series. However, it may require multi-dimensional thinking upfront.

I appreciate all your comments and suggestions and would be interested in your take on my proposal.
Thanks

My wife likes chocolate ice cream, but I prefer French vanilla. I think our relative rankings differ in the same way.

When I teach the "Method Development Using DryLab" course, I go to great pains to point out that what I present is *a* strategy, not *the* strategy. That's why I tried to be explicit about the rationale (some people might say "irrationale") behind my suggestions. That's also why we went to great pains to make sure that the DryLab software doesn't lock you in to a particular strategy.

To be specific, I agree with your comment about ion-pair reagents; I probably ought to amend my initial comments to make them seem less negative. I will turn to ion-pair primarily for samples which contain hydrophilic bases (hence weakly retained under the usual RP conditions). Having said that, as the availability of high-pH-stable reversed-phase columns increases, I might consider the following starting points:
- neutrals: water/ACN
- neutrals + acids: low-pH buffer/ACN
- neutrals + acids + bases: low-pH buffer+ion-pair/MeOH
- neutrals + bases: high-pH buffer/ACN

Regarding "parallel" vs "sequential" method development, the question really boils down to: "given the capacity to carry out a certain number of runs (for the sake of argument, let's say about thirty runs overnight), which variables do I want to explore?". My preference for temperature and yours for column type are a chocolate/vanilla choice. Either one is likely to be satisfying.

Bottom line: my take on your proposal is that it makes a lot of sense and should work just fine!

Regards,

-- Tom Jupille / LC Resources Inc.