Research in organic photovoltaics (OPVs) has recently brought about single-junction cells with 10% power conversion efficiency, made possible by high external quantum efficiencies and fill factors. However, the open-circuit voltage (Voc) remains relatively low, with most systems losing 0.6 ± .1 V compared to the energy of the charge transfer state, and rarely exceeds 1 V in high-performing systems. At Voc, the high rate of recombination in OPVs leads to charge densities on the order of 10^16 cm^-3. Consequently, only the tails of the donor HOMO and acceptor LUMO densities of states (DOS) are filled, and the voltage loss is therefore attributable to the width of the DOS and the degree of bimolecular recombination. Until recently, it has been difficult to obtain reliable measurements of the DOS within the gap, i.e., at densities below 10^19 cm-3, and measuring the energetics in BHJ blends remains particularly challenging. In polymer:fullerene cells, recent work has shown that changes in blend ratio and processing conditions[1,2] can significantly change donor and acceptor energetics. While small molecules and polymers are both capable of reaching high efficiencies, the relationship between bulk heterojunction morphology and energetic disorder has not yet been observed in small molecules. In this work, we determine the energetics—from the energy of the charge transfer state, through the range of densities under illumination, and down to states occupied at equilibrium—of six different solar cell blends based on a small-molecule donor and either fullerene or polymer acceptor. Using multiple methods to determine the shape and the relative disorder of the combined donor-acceptor DOS, we then gain insight into the factors limiting Voc in these systems. We find good agreement between estimates of energetic disorder based on band bending and on the density dependence of Voc, and that these distributions are consistent over a large range of the DOS. Interestingly, we find that disorder plays a nuanced role in the case of PCBM, where more energetic disorder is concomitant with increased phase separation, significantly higher charge density, and better overall performance. We find the 1-sun charge density at Voc to vary by over an order of magnitude in the systems studied here, and that these densities follow a roughly exponential dependence on the dielectric constant. This observation is consistent with recent work showing that synthetic modification to increase the dielectric constant can directly improve Voc[3]. Because of the relatively low energetic disorder observed in small molecule donor materials, we find that efforts to increase Voc are best focused on increasing the dielectric constant and further optimizing nanoscale morphology.
[1] Lange, I., Kniepert, J., Pingel, P., Dumsch, I., Allard, S., Janietz, S., Scherf, U., Neher, D., J. Phys. Chem. Lett. 2013, 4, 3865-3871. [2] Sweetnam, S., Graham, K.R., Ndjawa, G.O.N., Heumueller, T., Bartelt, J.A., Burke, T.M., Li, W., You, W., Amassian, A., McGehee, M.D., J. Amer. Chem. Soc. 2014, 136, 14078-14088. [3] Cho, N., Schlenker, C.W., Knesting, K.M., Koelsch, P., Yip, H., Ginger, D.S., Jen, A.K.-Y., Adv. Energy Mater. 2014, 4, 1301857