© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1 wileyonlinelibrary.com COMMUNICATION Iodine Migration and its Effect on Hysteresis in Perovskite Solar Cells Cheng Li, Steffen Tscheuschner, Fabian Paulus, Paul E. Hopkinson, Johannes Kießling, Anna Köhler, Yana Vaynzof, and Sven Huettner* DOI: 10.1002/adma.201503832 while the ferroelectric domain polarization/depolarization pro- cess is much faster, in the picosecond range. [23,24] In addition, piezoresponse force microscopy [13,25] results also suggested that polarization switching in CH 3 NH 3 PbI 3 is unlikely to be the dominant reason for the hysteresis behavior. (2) By tran- sient photovoltage and capacitance measurements, O’Regan et al. [26] proved that J–V curve hysteresis is not associated with the change of recombination rate and charge separation effi- ciency. (3) Simple trapping and detrapping of charges at defects in the bulk or at the interface is also unlikely to be the main mechanism due to the long duration and large magnitude of the current decay. [19] However, the trapping/detrapping pro- cess associated with migrating defects, such as iodine intersti- tials, [15,27] which can be driven by an electrical field, may play an important role in the hysteresis. In brief, we used three different methods to reveal the under- lying processes relating to hysteresis. First we employed electroab- sorption (EA) spectroscopy, a noninvasive in situ characterization approach [28,29] to determine the built-in potential ( V BI ) in solar cell devices. In a second experiment we applied a staircase voltage profile to these devices and measured the time dependent current at a series of temperatures in order to study the activation energy of the migrating species in perovskite based devices. In the last experiment we used X-ray photoemission spectroscopy (XPS) measurements to study the redistribution of elements within laterally configured devices after long-term electrical biasing. Our results suggest that the hysteresis in J–V curves originates from the interfacial barrier associated with the drift of iodide ions or the respective interstitial under an electrical field. Figure 1a shows the scheme of a FTO/compact TiO 2 / CH 3 NH 3 PbI 3-x Cl x /Spiro-OMeTAD/Ag planar perovskite solar cell. As a p-i-n structure, [30,31] photogenerated electrons are collected by a FTO (fluorine doped tin oxide) electrode through a compact TiO 2 layer, and the holes by an Ag electrode through a layer of Spiro-OMeTAD (2,2′,7,7′-Tetrakis-( N, N -di-4- methoxyphenylamino)-9,9′-spirobifluorene), acting respectively as electron and hole transport layers (ETL/HTL). The ionization energy and the electron affinity of materials were estimated according to values reported in literature [1,14] and measure- ment of the optical bandgap (Figure S1, Supporting Informa- tion). The J–V curves of a CH 3 NH 3 PbI 3-x Cl x perovskite solar cell under AM 1.5G illumination in forward and reverse voltage sweeps are shown in Figure 1b. As shown in Table 1 , there is a shift in V oc of 0.22 V and a significant difference in the per- formance between the two sweeping directions. Note that the red dashed line depicts the extension of the J–V curve during The past few years have witnessed a dramatic development of inorganic–organic halide organometal perovskite solar cells, which have captured the attention of the scientific community due to their high power conversion efficiencies [1] and simple fabrication processes. [2,3] Besides further improving the con- version efficiency [4] and stability, [5] important device related questions still need to be answered: [6,7] the hysteresis behavior, i.e., the discrepancy of the performance between two voltage- sweeping directions when performing a current–voltage ( J–V) measurement. The possible origins include ferroelectricity, [8,9] low frequency capacitance, [10] trap states in the bulk, [11] inter- facial dipoles, [12] as well as mobile ion screening. [13–15] Several groups have fabricated devices without significant hysteresis, for example, by using mesoporous TiO 2 layers instead of a com- pact TiO 2 layer, [16] solvent annealing to obtain large crystalline grains, [17] as well as a fullerene passivation method. [18] Although there are several proposed mechanisms, recently more and more attention is focused on the role of ions/vacancies migration under an electrical field. [15,19–21] (1) The character- ized decay time scale in perovskite solar cells is on a second timescale [22] (also see Figure S6, Supporting Information), Dr. C. Li, J. Kießling, Prof. S. Huettner Macromolecular Chemistry I University of Bayreuth Universitätstr. 30, 95447 Bayreuth, Germany E-mail: sven.huettner@uni-bayreuth.de S. Tscheuschner, Prof. A. Köhler Experimental Physics II University of Bayreuth Universitätstr. 30, 95447 Bayreuth, Germany F. Paulus Organic Chemistry Institute Im Neuenheimer Feld 270 Heidelberg University 69120 Heidelberg, Germany Dr. P. E. Hopkinson, Prof. Y. Vaynzof Kirchhof Institute for Physics Im Neuenheimer Feld 227 Heidelberg University 69120 Heidelberg, Germany Dr. P. E. Hopkinson, Prof. Y. Vaynzof Centre for Advanced Materials Im Neuenheimer Feld 225 Heidelberg University 69120 Heidelberg, Germany Adv. Mater. 2016, DOI: 10.1002/adma.201503832 www.advmat.de www.MaterialsViews.com