# Predict polarity (TLC, column) using ChemSketch (-Draw)

I'm currently working on a project where I have to make and isolate often quite similar molecules. But, as I change the functional groups the overall electronic situation and polarity of the product and of the side-products often changes. I can just estimate what spot on my TLC might be my product and check later with the NMR but I'm still interested in a software-supported way. So I checked ChemSketch (I don't use ChemDraw) and found some parameters that might be of help:

The polarizability, the density, the molar mass, the index of refraction and (although often not displayed) the dielectric constant.

Now I know this is all just a very rough and simple calculation but sometimes I just need a very basic feeling if the molecule can be considered polar or not.

So I thought perhaps using the equations of Clausius-Mossotti and Lorentz-Lorenz I might be able to determine the dielectric constant (because the software is often unable to predict it).

To my understanding is the dielectric constant a very fundamental parameter here. Polar solvents have a much larger one than non-polar solvents and the dielectric constant is responsible for how well a salt is dissolved in a solvent. For TLC and column chromatography on a polar stationary phase, a more polar substance will move much slower.

So can I just break this down in comparison of two molecules, that the one with the higher dielectric constant will also have a smaller $\mathrm{R_f}$-value?

• A low tech alternative: Take some of the sample from a crude reaction and run a prep-TLC. Scrape off the bands, dissolve the compound off the silica with MeOH/DCM, filter and run NMRs. This way you should be able to ascertain which spot to aim for when running a column. Jan 24 '18 at 0:07