Towards an electronic model for CuIn1 − xGaxSe2 solar cells

We model some aspects of highly efficient CuIn1 − xGaxSe2 solar cells with x ≈ 0.3 as well as wide band gap cells with x = 1 and ask for the dominant recombination mechanism which limits the Voc of these devices. For CuIn1 − xGaxSe2 solar cells with x ≈ 0.3, interface recombination combined with Fermi-level pinning is a possible but unlikely recombination mechanism. We argue that these cells are rather limited by recombination in the quasi-neutral region (QNR) including the back contact. Using the expression for the QNR recombination rate we calculate the derivative of the collection function in the absorber at the space charge region edge which is in reasonable agreement with the experiment. It turns out that the diffusion length must approximate the absorber thickness. Based on this information, we draw a band diagram for a CuIn1 − xGaxSe2 solar cells with x ≈ 0.3 and plot the simulated collection function. For cells with x = 1 (Cu-poor CuGaSe2), the experimental activation energy of the recombination rate mostly equals the absorber band gap, i.e. Ea ≈ Eg,a = 1.67 eV. As the experimental interface band gap is smaller than Ea, interface recombination must be ruled out. Thus, the carrier lifetime in the Cu-poor CuGaSe2 absorber should be so small that bulk recombination is more efficient than interface recombination. From this consideration, we postulate an electron lifetime value of 10−12 s for CuGaSe2.