Let me be blunt here: that method will *never* work well in its present form.
Here's why:
Your first peak is essentially unretained (i.e., it has very little chemical interaction with your column). That's why I asked about t0 (the "dead time", which is the time required for compound that does not interact with the column to wash through from the injector to the detector). If you think about it, t0 is the time required to pump one column volume of solvent through the system, which means that we can estimate it as t0 = Vm/F where Vm is the volume of mobile phase in the column and F is the flow rate. You know the flow rate is 1 mL/min, so if we can estimate Vm, we can give a reasonable estimate for t0. It turns out that for typical reversed-phase columns of 4.6 mm internal diameter, Vm can be estimated by 0.1 x L(cm) where L is the column length in cm (I know the units don't cancel; this is simply a rule of thumb, good to about +/- 15%).
You have a 25-cm column, so your column volume is about 2.5 mL. At 1 mL/min, that makes t0 about 2.5 minutes. Your first peak elutes at 3 minutes, which is a negligible difference. To put this in perspective, the US FDA in their Guidance for HPLC methods suggest that any peak being quantitated should have a retention time greater than or equal to 3 times t0. In fact, without knowing how many individual standards you ran, I can speculate that it may not be a real peak at all, but simply the equilibrium upset from your injection (this is commonly referred to as "t0 noise").
That's for starters.
Next, 300 mM buffer is *extremely* high for HPLC. 10 - 50 mM would be more typical.
Finally, you did not specify the A/B ratio in your mobile phase. If you ran a gradient, this is yet another problem; as previous posters have indicated, equilibration times with ion-pair separations are notoriously long, which means that gradients are generally a bad idea.
Here's a summary of the approach that we recommend in our
Advanced HPLC Method Development course (shameless plug!):
1. Ignore your basic compound for the moment and begin working with the acid. pH 3.5 is a bit high for a phosphate buffer (phosphate has a pKa of 2.1, so you're on the ragged edge), but if the pKa of your acid is above 5 or so it shouldn't matter. I'd be tempted to drop down to 3.0. Triethylamine is fine as a counterion. Do *not* use any octanesulfonate at this point.
2. Adjust the %B as necessary to make your acid elute at 6 - 10 times t0 (anything in the 15-25 minute range would be about right, given your column configuration).
3. Now start working with your basic compound. It will probably be unretained under the conditions of step 2 (i.e., it will elute somewhere around 2-3 minutes). Do a series of experiments in which you hold %B constant but gradually increase the octanesulfonate concentration, going in geometric progression (0, 1 mM, 2 mM, 4mM, . . . ) to "walk" the base peak out to elute at 3 - 5 times t0 (anything from about 7 to 15 minutes). Make sure to allow plenty of equilibration time with each change (20 to 50 times the column volume is a safe rule of thumb to start with). The retention of your acid will probably decrease somewhat as you add the octanesulfonate.
4. Fine-tune the separation by "tweaking" %B and the octanesulfonate concentration.
Assuming you only have the two peaks, you should be able to get a great deal of selectivity, which means that you could speed up the separation considerably by increasing the flow rate and/or shortening the column.
To answer your other questions:
- expiration date: TEAP (triethylammonium phosphate) is a great nutrient medium for microbes: it contains nitrogen, phosphorus, oxygen, and carbon. I would make it up fresh daily just to be safe. The expiration date for your samples should be set based on the stability of your compounds.
- For isocratic separations, the triple-distilled water should be OK (HPLC grade would be better). 0.45 micron filter is OK for the column geometry you are using (I'm assuming the column is packed withe 5 micron particles).
- washing the column may not be a good idea with ion-pair separations because of the equilibration time issue. The column will have a longer lifetime if you flush it with buffer-free mobile phase and then store it in acetonitrile, but you will waste a good proportion of that lifetime re-equilibrating the column. I would definitely *not* use the column for any non-ion-pair method because there is always a chance that the surface chemistry has been permanently altered.