Praveen,
I would suggest a look at the web pages for various column vendors. (like
http://www.restek.com/guide_gccolsel_sect2.asp or
http://www.sge.com/support/training/col ... tion-guide, for example.) Or, check the catalogs they send out. There are a couple of things to be gained: 1) They will often have a discussion of column chemistry, but more useful is: 2) Each vendor will have sample chromatograms run on various columns. For column selection, it is often helpful to look through chromatograms of similar mixtures and find a column that does the kind of separation that you want. This removes many theoretical assumptions - the separation is what you see.
Also, while a 50% phenyl column will have a strong pi-pi interatction, there will still be a boiling point component to the separation. There are no separations that are purely based on polarity, boiling point, induced dipole interactions or such.
If you are separating a series of known compounds from each other, careful analysis of these interactions can be used to compute the most optimal separation - and there is software available that will do this for you. If you are attempting to do an an analysis in a complex matrix, like a soil extract or a biological extract, the problem comes, not from the molecules you know about and want to find, but all those unknown compunds in the extract.
I have taught classes in the use of GC/MS systems and have been asked how to select a column - and my answer is one the column vendors do not like: My favorite column is the one already in the instrument. (Vendors like - and with good reason - to sell columns that are optimized to various applications.) Then I do not have to change it and the fellow who was using the instrument does not have to change back to continue his work. A large number of analytical separations will work on many columns. in GC/MS work, for instance, I have done most of my work over the past fifteen or twenty years with a methyl silicone or 5% phenyl methyl silicone column - not because of superior separation, but because of the lower bleed into the mass spectrometer, particularly if I need temperatures above 250 degrees C. Before I worked with GC/MS, I was quite happy to run everything on a carbowax column, unless I was forced away from it.
There are columns with particular characteristics - and there a particular problems that are best solved with picking just the right column. And the best way to pick one of the specialized columns is by looking at the vendor's literature. Sometimes the cross linking or end capping technique used by a particular manufacturer seems to make just the right difference for a particular separation, even though the stationary phase may be polyethyleneglycol or 5% cyanopropyl or some other common stationary phase. And, the vendors do not give a lot of detals on some of these finer points - otherwise their competitors will know how to make and sell the same (or better) columns.
If your work is on the nature of column analyte interactions, I have wandered off into the unnecessary. On the other hand, if you are using chromatography as a tool for other research, spend only as much time on selecting a column as you need to get a robust separation. Look at sample chromatograms or pick up the telephone and talk with a technical specialist at your favorite column company.
