With all due respect, it looks like you're "flailing": trying things at random and hoping that something will work.
My $0.02 worth would be to do it systematically:
1. How much resolution do you need? 2.0? 1.5? FWIW, the FDA "suggests" that Rs should be > 2, but you may need more if there is a big disparity in your peak sizes or if there is a lot of tailing.
2. The classic relationship for what controls resolution is Rs ≈ (0.25) * SQRT(N) * (alpha -1) * k'/(1+k')
3. Your column is 250 x 4.6mm, which means that it has a dead volume of around 2.5 mL. At 1 mL/min flow rate, that's a dead time of 2.5 minutes. Retention times of 14 and 13 minutes then work out to k' values of 4.6 and 4.2, respectively, and that, in turn means an alpha value of 1.1. You didn't say how wide your peaks are, so I can't estimate the plate number, but if we assume N = 5,000 (which would be mediocre for a 250 mm column packed with 5 micron material), that works out to a resolution of about 1.4.
4. So you have three possible approaches to improving resolution:
- increase k'
- increase alpha
- increase N
5. If you increase k' from 4.4 (your current average) to 10, that should increase Rs to about 1.6 (assuming nothing else changes).
6. If you can increase the plate number from 5,000 to 10,000 (you should be able to get 10k plates on that column geometry) that should increase Rs to about 1.9 (again, assuming nothing else changes).
7. alpha (selectivity) is where you have the most leverage.
In reversed-phase, there are only six things you can do to change alpha:
- change the % organic (which changes k', but can also sometimes change alpha)
- change the temperature
- change the organic solvent
- change the pH
- change the column
- put in additives (e.g., ion pair reagents).
So now you can be systematic, as suggested by the preceding posts. Take each of those parameters and try tweaking them.
8.
- %B. Change it by 5% (either way). Did the resolution get better or worse? If it got worse, go 5% the other way and see what happens.
- temperature. Increase it by 10 degrees. Did resolution get better or worse. If it got better, increase is some more. If it got worse, decrease it.
- organic solvent. Switch from methanol to acetonitrile, then tweak the acetonitrile 5% either way to see if selectivity changes. If you don't object to THF, repeat with that (a lot of labs don't like THF in part because of the high UV cutoff, but that shouldn't be an issue in your case).
- pH. Are your compounds acids or bases? If so, what are their pKa values (if you don't know exactly, estimate from the structure and functional groups). Start 2 pH units away from the pKa and then run a series with decreasing (or increasing) pH moving toward the pKa in half-unit intervals. [Since you were using plain water initially, my guess is that pH won't be very effective].
- column chemistry. Rather than guessing, go to the USP column selectivity database and look for the most different selectivity:
http://www.usp.org/app/USPNF/columnsDB.html (there are actually two databases there; the second one (PQRI) has a lot more columns). Then start over from the top with the new column.
- additives. Same general comment as pH; these are most effective when you have acids or bases.
What this basically amounts to is a "QbD" (Quality by Design) approach to developing your separation. Although this seems more time consuming than guessing right the first time, it pays off when you go to validate the method, because you have a lot of data about robustness with respect to the various parameters.
Sorry for the long-winded post!