The maximum open-circuit voltage of a solar cell can be evaluated in terms of its ability to emit light. We herein verify the reciprocity relation between the electroluminescence spectrum and subband-gap quantum efficiency spectrum for several photovoltaic technologies at different stages of commercial development, including inorganic, organic, and a type of methyl-ammonium lead-halide CH 3 NH 3 PbI 3−x Cl x perovskite solar cells. Based on the detailed balance theory and reciprocity relations between light emission and light absorption, voltage losses at open circuit are quantified and assigned to specific mechanisms, namely, absorption edge broadening and nonradiative recombination. The voltage loss due to nonradiative recombination is low for inorganic solar cells (0.04-0.21 V), while for organic solar cell devices it is larger but surprisingly uniform, with values of 0.34-0.44 V for a range of material combinations. We show that, in CH 3 NH 3 PbI 3−x Cl x perovskite solar cells that exhibit hysteresis, the loss to nonradiative recombination varies substantially with voltage scan conditions. We then show that for different solar cell technologies there is a roughly linear relation between the power conversion efficiency and the voltage loss due to nonradiative recombination.

We study the appearance and energy of the charge transfer (CT) state using measurements of Electroluminescence (EL) and Photoluminescence (PL) in blend films of high-performance polymers with fullerene acceptors. EL spectroscopy provides a direct probe of the energy of the interfacial states without the need to rely on the LUMO and HOMO energies as estimated in pristine materials. For each polymer, we use different fullerenes with varying LUMO levels as electron acceptors, in order to vary the energy of the CT state relative to the blend with [6,6]-phenyl C61-butyric acid methyl ester (PCBM). As the energy of the CT state emission approaches the absorption onset of the blend component with the smaller optical bandgap, , we observe a transition in the EL spectrum from CT emission to singlet emission from the component with the smaller bandgap. The appearance of component singlet emission coincides with reduced photocurrent and fill factor. We conclude that the open circuit voltage is limited by the smaller bandgap of the two blend components. From the losses of the studied materials, we derive an empirical limit for the open circuit voltage: 2

We investigate the reasons for the dependence of photovoltaic performance on the absorber thickness of organic solar cells using experiments and drift-diffusion simulations. The main trend in photocurrent and fill factor versus thickness is determined by mobility and lifetime of the charge carriers. In addition, space charge becomes more and more important the thicker the device is because it creates field free regions with low collection efficiency. The two main sources of space-charge effects are doping and asymmetric mobilities. We show that for our experimental results on Si-PCPDTBT:PC71BM (poly[(4,40-bis(2-ethylhexyl)dithieno[3,2-b:20,30-d]silole)-2,6-diyl-alt-(4,7-bis(2-thienyl)-2,1,3-benzothiadiazole)-5,50-diyl]:[6,6]-phenyl C71-butyric acid methyl ester) solar cells, the influence of doping is most likely the dominant influence on the space charge and has an important effect on the thickness dependence of performance.

The application of Mott−Schottky analysis to capacitance−voltage measurements of polymer:fullerene solar cells is a frequently used method to determine doping densities and built-in voltages, which have important implications for understanding the device physics of these cells. Here we compare drift-diffusion simulations with experiments to explore the influence and the detection limit of doping in situations where device thickness and doping density are too low for the depletion approximation to be valid. The results of our simulations suggest that the typically measured values on the order of 5 × 10 16 cmfor doping density in thin films of 100 nm or lower may not be reliably determined from capacitance measurements and could originate from a completely intrinsic active layer. In addition, we explain how the violation of the depletion approximation leads to a strong underestimation of the actual built-in voltage by the built-in voltage V MS determined by Mott−Schottky analysis.