Perovskite Solar Cells: Do We Know What We Do Not Know?

D 1−4, 2014 saw one of the largest gatherings of scientists and engineers on the topic of hybrid organic− inorganic perovskite solar cells, which dominated symposium W of the fall meeting of the Materials Research Society (MRS) in Boston on Perovskite-Based and Related Novel Material Solar Cells. The symposium included many contributions as well as invited talks, open discussion sessions (to highlight major themes and stimulate discussion of unpublished data), and a “rump session”. The latter was intended (and indeed mostly used) for last minute news in the field. Over the 4 days, well over 120 papers were presented, including some 50 posters, 3 of which were awarded the symposium’s poster prizes. Furthermore, on the first day of the symposium, Henry Snaith (University of Oxford) received the 2014 Outstanding Young Investigator award of the MRS and gave a meeting-wide lecture at Symposium X. In this Guest Commentary, we summarize the topics that, in our opinion, were the more salient ones of symposium W, with a focus on recent progress and future challenges in understanding the unique properties of hybrid perovskite solar cells. The salient issues that were discussed included hysteresis, which appears to be, at least in part, an interfacial phenomenon, primarily relevant for interfaces with oxides; ion migration, which remains a hypothesis because although it is supported by a variety of experimental results, it still lacks fully conclusive experimental evidence; materials preparation, where solvent annealing of the product of a 2-step process seems to be beneficial for cell efficiency, as is the presence of water; and electronic traps, where evidence for the presence of traps just below mid gap was presented, which is somewhat inconsistent with the high Voc of cells and may present a case of photoinduced effects. Furthermore, a variety of stability tests give promising short-term results, but the reported very low formation energy of hybrid perovskite materials (from the “binaries”) could be detrimental for long-term stability. For mixed I−Br perovskites that are relevant as “ideal bandgap” materials for spectral splitting and tandem cells, phase separation was reported, which, at least for the moment, cannot be explained well by thermodynamics or kinetics. An overall impression is that likely not everyone is, as yet, working with the same material as some effects of exposure to ambient could be explained by doping of the perovskite or holeconducting material. Other results remain puzzling, as even N2 was seen to change the material’s properties. Finally, not all ef f iciency measurements (and their results) appear to be comparable with each other. Subsequent to the first public report of hysteresis in the current−voltage curves of hybrid-perovskite-based solar cells by Hoke et al. at the MRS 2013 Fall meeting and its discussion in the 2013 “rump session” (see Figure 1), the phenomenon was reported in the literature and was the focus of a number of talks at this year’s symposium. Ideas that were brought up to explain it were analyzed in more detail in the discussion sessions of the symposium. Part of the interest stems from the need to arrive at a well-documented, easily reproducible measurement of cell efficiencies, part from the hope that revealing the origin of such “hysteretic behavior” may help to understand the mechanism behind the photovoltaic properties of hybrid perovskites. Furthermore, the hysteresis might be connected to the long-term stability of the devices. In several talks, layers of organic molecules (discussed in talk W4.06) and device geometry (see, e.g., talk W2.05) were reported to play an important, if not essential, role in the magnitude of the hysteresis effects. These findings suggest that the hysteresis is at least partly an interface phenomenon and highlight the importance of (inter-) sur-face passivation. On the other hand, several talks throughout the symposium discussed the complementary possibility that reorientation of the polar organic molecules, contained inside the inorganic cage, due to an external electric field could occur. One hypothesis is that such concerted dipole alignment results in ferroelectric domains, which could then be responsible for the observed hysteresis in the current−voltage curves (see talk W1.01). Hysteresis was shown to be independent of light intensity, implying that it is not arising entirely from stored chargesone of the other possible explanations (talk W1.06). While not stated explicitly, we got the impression that without TiO2 or other electron-conducting hole-blocking oxide layers, there is no or very little hysteresis, suggesting that interface phenomena play a dominant role. Although there is as of now no fully comprehensive explanation for the hysteretic behavior of hybrid perovskites, it was generally agreed in a number of discussion sessions that at the moment, for generating reliable physical insight in terms of I−V curves and solar cell efficiencies, it is highly important to report a number of other experimental parameters. These include a steady-state maximum power point, the scan rate, the scan direction, and the preconditioning of the device prior to measuring (e.g., was it electrically biased, illuminated, or measured immediately upon completion of cell fabrication).