Assuming that the polymer hetero junction diode follows the diode shockley equation, I = Is ( exp [ V- I Rs]/nVt + V/ Rsh,then all you need to plot I versus V on a semilogarithmic scale. That is Log10 I versus V.
You will observe a straight line in the middle current ranges, the slope of this straight line Log10 I2/I1 = Delta V/ nVt * log10 exp, with Delta V corresponds to the voltage change from I2 to I1. At low currents. the current through the shunting resistance will be dominant and Rsh= I / slope of the I-V characteristics near the zero diode voltage.
In order to determine Rs you extrapolate the straight line at the middle currents to the higher current range and see how much the I-V curve deviates from the extrapolated straight line . At certain current I, the voltage deviation the extrapolated line is Delta V, then Rs= Delta V / I.
Is is determined by extrapolating the straight line to the Zero voltage. Tfrom the intercept current is Is.
For more information and to see an example please follow the Link:https://www.researchgate.net/publication/3062547_A_distributed_SPICE-model_of_a_solar_cell
There are other methods based on curve fitting and numerical techniques.
I agree. You can directly deduce the ideality factor from the "slope" of the exponential region of the jV curve measured in the dark. A common problem for OPV is that this region is hard to determine due to low shunt and high series resistance. What you can try is to take the differential "slope" to judge whether you can take the plateau value: http://link.aip.org/link/APPLAB/v99/i15/p153506/s1&Agg=doi To get more reliable data, you should do this analysis at different temperatures.