@ Dinesh Kulhari

A Digital LCRZ Bridge instrument is widely used

One can take dielectric studies with respect to frequency and temperature. Mostly the crystals are polished very thinly and are kept between two parallel metallic plates and an A.C frequency is applied between the plates also the temperature can be varied. The dielectric loss is then found out from the graph.

There are various papers describing this method. Try searching the dielectric studies of Nonlinear optical single crystals.

You have to develop a MIM structure i.e., two electrodes at two sides are required from where you can get the contacts.

The problem is, the irregular shape of the crystals. It makes very difficult to determine the effective area of the electrodes and the average thickness of the sample in-between.

I don't know, if it is possible for your sample, but for some alloys and ceramic materials, crystals are cut very nicely, to obtain a very regular or symmetric shape. If that's possible for your material, then you can do the dielectric study very accurately. Otherwise, there will be some error associated in the numerical values, for the conventional methods. Although the basic nature of the frequency-dispersion would be okay.