Hello everyone,
I'm getting familiar with photodiodes modeling since it is the topic of my internship.
I guess you want an accurate model at inverse bias voltage...
Well, in this case it has been demonstrated lately that the reverse current depends on the voltage (thus creating a dynamic parallel resistance). The current can be due to several macanisms : diffusion, band to band tuneling, trap assisted tuneling.
It has been shown that the increase of the current with the reverse bias voltage can be model using a current gain coeficient (M = 1 + 2^(2(V-Vth)/Vth) with Vth = 6,8*Eg). At very low reverse bias votlage a parameter should be added to take into acount the field dependance on bias voltage.
Then the dynamic resitance can be easily dedued calculating the slope (R(V)=(Vd(i)-Vd(I+1))/((Id(i)-Id(I+1)).
Last, it's worth knowing that the M coefficient accounts for avalanche phenomenon, consequently according to the quality of the photodiode M accounts for the reverse current for a variable range of bias voltage. In case the diode is badly desined and fabricated the leaking current becomes prevaling at certain bias voltage.