Typically you would use something simple like lecithin. A weak solution in IPA, perhaps 2wt%, is normally sufficient. You can dip coat, spin coat, or just brush it on to your clean substrates, depending on how uniform you need your director alignment. It can be baked at slightly elevated temperature, 40-50C for 15 minutes, to ensure no solvent remains.
Bear in mind that sometimes ultra clean glass can spontaneously align LC's such as cyanobiphenyls homeotropically without any additional surface treatment.
Other options to create homeotropic alignment surfaces are silane-based surface treatments and chrome complex surface treatments.
If your measurement is very sensitive and necessarily very accurate, you may want to use a material for which you accurately know the permittivity.
For dielectric measurements, one other thing you can do is to use a non treated cell, and instead use a sufficiently high magnetic field (e.g. 1T) to align the director vertical as you are measuring dielectric properties.
I agree with Ben Snow - I have used those methods successfully in the past. Doping the liquid crystal with a non-ionic surfactant usually works very well.