Solar cells based on Cu(In1−x, Gax)Se2 typically show time transient behavior of the open-circuit voltage Voc under illumination. In this work, we study both the red-light Voc(t) transient and the red-light capacitance transient at different temperatures of samples with different x. From the capacitance transient, we calculate a transient behavior of the Cu(In1−x,Gax)Se2 doping density NA,a(t). Then, using established models on the NA,a dependence of the dominant recombination mechanisms, we derive from Voc(t) that Cu(In1−x,Gax)Se2 samples with x = 0, 0.3 are dominated by bulk recombination and a sample with x = 1 is dominated by interface recombination—in agreement with the expectation. Further, the transients of NA,a(t) can be used to recalculate Voc(t) transients which are then compared with the measured Voc(t) transients. From the excellent agreement, we conclude that under red-light illumination Voc(t) indeed is dominated by NA,a(t) and other transient effects are of secondary importance. We further conclude that the sample with x = 1 can be described by an absorber/buffer/window energy band diagram with fully depleted buffer layer which here is CdS.

No abstract

No abstract

Suitable procedures for the determination of diode parameters of illuminated and dark current density-voltage (j-V ) curves of solar cells are investigated. Within these procedures we avoid misinterpretations by the commonly used shifting approximation which assumes a voltage-independent photo carrier collection. To this end, the voltage-dependent collection efficiency is determined by j-V measurements under incremental illumination. After that, illuminated curves can be converted into dark curves taking the voltage-dependent photo current density into account. The converted dark curve next is consigned into a program that numerically solves the implicit equation of the current density and derives the diode parameters by means of a weighted least-squares minimization. Only if fluctuations of the illumination are noticeably high this procedure cannot be used, since the converted dark curve also reveals a high noise level. Then, the determination of diode parameters is performed stepwise with the help of commonly used methods.

No abstract