Azide/Aniline resolution

[PJ8]

Hi everyone,

The main analyte in a reverse phase HPLC method is an R'N3R azide. The method itself is a water plus 0.1 % TFA MPA, MeOH MPB gradient method from 40-90% B on a Sunfire C18 150x4.6mm, 3.5um at 40degC.

We've started seeing the aniline impurity (R'NH2R) which currently is poorly resolved from the main peak. Can anyone give me a lead on what parameters might effect the resolution between an azide and an aniline. Do you think we can tweak this method to improve things, or will we need some more radical development from a different starting point?

Many thanks, Pat.

[tom jupille]

If your resolution is marginal, the "bigger hammer" is to simply couple two columns in series and double the gradient time. That will improve resolution by a factor of 1.4 (at the cost of a doubled run time)

for the rest, there are only six things you can do to change the selectivity (relative spacing of the peaks):
- gradient steepness
- temperature
- pH (TFA concentration)
- solvent type
- other additives
- column type

You could quickly screen the first 4 in a day (or overnight if you have an automated system with solvent switching):
- do one run with twice the usual gradient time
- do one run at a higher temperature (55 degrees ?)
- do one run with half the TFA concentration
- do one run with ACN instead of MeOH
If any of them show a siginficant shift in peak spacing, then you can "fine-tune". If not, then you're faced with a more involved method redevelopment problem.

[HW Mueller]

PJ8, you never tried to vary the pH, and I mean by much more than Tom suggested?
I am somewhat surprised that your azide stays intact under those conditions, of course, I don´t have too much experience with azides, and it was a long time ago . . . . .

[PJ8]

Thanks for your suggestions guys. There's been a number of things tried during the method development, but never with this impurity seen before so I had no handle on what method factors might most effect the resolution. pH might well be the most significant effect, but it was already optimised for 10 or so other imps so I was hoping not to move too far from the current set-points unless I had to. We do have a low and high pH screen in our method development tool kit.

I was planning either a screen (TFA conc/gradient and temp/gradient, modelled in DryLab) or perhaps an even more broad brush DoE approach, with flow as well.

I failed in my attempt to lit search separations of azides and anilines, so I thought a check on whether anyone had experience and knew what factors they were most sensitive to might narrow down my search and get me a quick win. Since there's not I'll use one of the broader brush approaches and let you know what the resolution is sensitive to.