Hello all
I am new to LC world, but I am enjoying it a lot. In my current project, I am separating five impurities and two active ingredients. Out of the five impurities, two of them are isomers (position of the OH and NH2 groups are different). One of them elutes early (like 6 mins) and the other elutes late (almost 30 mins). Thsi pattern remains the same no matter if you change the pH of the mobile phase or the column functionality. Both of them are probably weak bases and should have comparable Pka values. Why should they separate so differently? This might be a dumb question since I am new to this area.

You should be happy that they have such a good resolution! What kind of column are you using? If you are using a 4.6 x 150mm column, one of your impurities elutes at K=~19. That's kind of large for an impurity (I guess <0.1%) because peak widens as K increases for an isocratic method and this will have an adverse impact on your S/N ratio.

I understand that the resolution is good here. I am using Lichrosorb-RP-18 column that is 4.6 X 250 mm. I am using amine modifiers and so the peak shapes are still good at even large K's. I need to know what makes their elution pattern so different. I would think that the Pka values should be comaprable and so the elution times shouldn't be that much different. In this product, I am working with two drug product and 5 impurities (a total of seven peaks). Most of them have pyrimidine and pteridine rings (heterocyclics). One particular impurity elutes so early that it co-elutes with solvent. I have changed the pH of the mobile phase and also tried several solvent combination with little success. Any suggestions?

Are you disolving the sample in mobile phase or 100% organic?

Bonding between the compound and column may be a factor here since you've tried to change mobile phase.

No, I am not dissolving the sample in the mobile phase. Two of the impurities (isomers) are not completely soluble in the mobile phase (even 50 microgm/ml). I tried several organic solvents with no success. In DMSO both dissolve completely. But there is a huge solvent peak from DMSO which I think is more due to the refractive index change from DMSO than any absorbance in the detector. I found this out by changing the detection wavelength. So, the problem is multi-fold. Recently, I even tried adding ClO4- ions to move that first peak away from the solvent with little success. I know that this first eluent has three amino group and a Nitro group in the heterocyclic ring. I guess, it is the most basic component. Any suggestions on how to move this peak selectively. I have tried cyano column (slightly polar) to retain this species, and it does seem to work, but other peaks co-elute.

You said you changed the pH... From where to where?

I changed it all the way from 2.3 to 7.0. No change in the elution time for the first peak (2,4,6 triamino-5-nitrosopyrimidine). I tried CYANO column and the peak eluted little late, enough to circumvent the solvent problem peak:-) But why does the cyano column elute it a bit late. Is it becuase it is slightly polar stationary phase and my compound is also ionic? (protonated form?)

Aliphatic amines are protonated at pH <= 7 so that retention of the first peak cannot be improved in RP at least with conventional silica- based columns in RP mode. You could use an ion-pairing agent such as an alquilsulphonate to improve retention of amines. Moreover the greater is the number of NH2 the more will be the retention using ion pairing. Secondly, why do you need to use 250 mm columns? You can improve your method a lot (decreasing retention times by 2 or even 3) by using 100-75 mm and 3 um of particle diameter.
On the other hand, you could try NP mode with a CN column and a mixture of an hydrocarbon as non-polar solvent (hexane, heptane, ciclohexane), an alcohol as polar modifier (isopropanol, and if you need more polarity you can use ethanol) and a 0.1 % of TEA as basic modifier. Sometimes NP works pretty well for separtion of positional isomers.
It should be useful if you could say what the compounds are.
Hope this helps

The first peak is 2,4,6 Triamino -5-nitrosopyrimidine (heterocyclic) with three NH2 groups. I don't think this is aliphatic, but I am not sure if this one is aromatic either since I need to check the 4n+2 rule:-)
Second peak is 2,7 Diamino-4-hydroxy-6-phenylpteridine. Two pyrimidine rings fused together with two NH2 groups and one OH and a phenyl group.
Third peak is 4-Amino 6-chloro-1,3 Benzenedisulfonamide. Aromatic chloro sulfonamide (two SO2NH2 groups, one Cl and one NH2)
Fourth peak is Triamterene. Same as the second one, except it has NH2 instead of OH
Fifth peak is Hydrochlorothiazide
Sixth peak is isomer of second peak(2,4 Diamino 7-hydroxy 6-phenly peteridine). The position of phenyl and OH are changed.
WHY should these isomers elute so differently?

Have you tried a basic mobile phase on a base-stable column?

I was talking in general, but your right, they aren't aliphatic amines, however they are basic compounds. The triamino (1st peak) seems to be highly polar and because of that, it won't be retained by RP. Regarding the isomers you mention, these behavior can be very common, you should think in carbohydrates, for instance all aldoses are diasteromers but they can be easily separed, or cis-trans isomers such as fumaric and maleic acids or positional isomers like p-hidroxibenzoic and salicilic (o-HO isomer) which are easily resolved by chromatography. Interaction between compounds-stationary/mobile phases are no easy to describe, they depends not only the chemical structure but also the spatial orientation of molecules in solution and in the stationary phase. Sometimes they can be more or less easily explained and sometimes don't. Salicilic acid vs p- and m-HO-benzoic is one example, salicilic is less polar, its melting point is lower than its isomers, etc. because of its ability to form an intra-molecular hydrogen bond. I'm going to speculate now: 2,7 Diamino 4-hydroxy 6-phenly peteridine could be more polar because its HO (aromatic HO are naturally aicidic groups) can give intra-molecular protonation through a ring of 5 atoms (which is spatially favoured together with 6-members ring) with the N in 5. This would be increase the HO acidity and therefore polarity. But who knows!
Regards

Marcelo