The C-V curve of a junction can be used to characterize the junction where one can get the width of the depletion region, the doping concentration in the lightly doped side of the junction and the built in voltage. There are models relating the the junction capacitance with its voltage, doping concentration and surface state density. These models are obtained by solving the Poisson eqaution in the transition region of the junctions.
As for the dielectric constant it is measured directly by preparing a homogeneous semicondcutor slice and supplying it with two metallic electrodes having ohmic contacts. This forms capacitor structure where its capacitance can be measured by a capacitance meter.
I would like that you refer to the paper in the link where the C-V of heterojunctions are treated in details: Article Capacitance and conductance of ZnxCd1-xS/ZnTe heterojunctions
For the C-V in general you can visit the site: https://www.researchgate.net/publication/236003006_Electronic_Devices
Principially you can use it, provided the carrier concentration is homogeniously. It could be, the carrier concentration has a gradient, especially in surface near region. But it is a good approximation. To get informations concerning the donor profile, I would carry out SIMS (secondary ion mass spectrometry).
Sir Rüdiger Mitdank is there any difference between free carrier concentration n/p calculated from hall effect and Donor/ acceptor concentration ND/NA?
there is a difference. If the temperature is sufficient high (kT > EI, EI energy of ionization), you can have n = ND - NA or p = NA - N (depletion region). Here is presupposed that the thermal energy is smaller than the gap energy, e.g. the intrinsic carrier concentration can be neglected.
Generally, for only one donor (here we discuss only n-type) ND and a compensating acceptor concentration NA we find (impurity reserve):
n*(n + NA)/[ND - NA - n] = NC*exp(-EI/kT)
with NC the effective density of states. This implicit equation must be solved for n.