Hi Praveen,
I would definitely take Vlad's suggestion more seriously-- separating isomers is a feature specific to specific columns. Separating some isomers will be possible on standard common columns like the CN or C18, but for better resolution a chiral selector is usually the way to go. I've had success separating diastereomers on a PFP(2) column from Phenomenex, but it could not resolve their enantiomeric pairs (4 diastereomers, 4 enantiomers) and would come out as the pairs of diastereomers. The PFP is marketed as a column that has some chiral separation, but clearly not enough in my case. We went through their screening process and found a chiral column that could split these pairs and we're now able to get all of the products efficiently. Positional isomers may be more difficult due to the subtle differences observed between the molecules, which may mean that you have to bump up your chromatography to that next-step. Alternatively, you could purchase (find, if you have some available) some chiral mobile phase additives which is a good alternative to purchasing a new column (depending on intended use of course). It sounds like you have them mostly separated, and depending on if you're running a gradient or not, you may be able to tweak the ratios of your mobile phases slightly (probably extending the overall run-time of your method) in order for your compounds to interact with the column more, and hopefully enhance your resolution. From what you said, it seems like you are able to separate them sometimes, and sometimes they come out as a single peak, this indicates to me that you could probably separate them but it depends on the gradient of your mobile phase.
As always, any conditions, mobile phase details, potential columns, etc. would be very welcome and help us assist you to reach your goals!
Thanks for reading