Communication Perovskite Polarization Dynamics
www.advmat.de
Exploring Anomalous Polarization Dynamics in Organometallic Halide Perovskites Mahshid Ahmadi, Liam Collins, Alexander Puretzky, Jia Zhang, Jong Kahk Keum, Wei Lu, Ilia Ivanov, Sergei V. Kalinin,* and Bin Hu* between electronic, optical, and transport functionalities in this class of materials. The wealth of studies of optoelectronic, photovoltaic, and transport properties in these materials have revealed a broad spectrum of anomalous physical behaviors, including multiple relaxation time scales, hysteresis in transport, optoelectronic characteristics, along with many other phenomena.[17–19] To explain these, a number of groups suggested that ferroelectricity[19] and ion migration[20] can play a significant role in these systems. Notably, the coupling between ferroelectricity, optoelectronic, and transport phenomena is necessarily mutual, whereas polarization-induced fields will result in internal field which is believed to increase charge separation and suppress charge recombination,[21] the external field and photoexcitation stimuli can result in ferroelectric switching.[22,23] Ferroelectricity is an electrical ordering for developing multiferroic properties. Given that the high levels of disorder in of these materials can precipitate transition from classical ferroelectric to relaxor-like behaviors, separation of ferroelectric and ionic phenomena becomes complicated. Importantly, for external measurements such as polarization voltage or macroscopic electromechanical coupling, the two are fundamentally inseparable. It should be mentioned that in regular ferroelectrics, polarization is defined at the level of individual unit cell and is constant on the larger length scales (with the possible exception of surface and interface behaviors). In relaxor ferroelectrics, the atomic disorder leads to the formation of polar nanoregions, somewhat similar to nanoscale ferroelectric domains. These nanoregions can be either frozen (nonergodic relaxors) or fluctuate (ergodic relaxors).[24,25] Macroscopically, relaxor ferroelectrics can be distinguished by diffuse phase transitions, anomalously large frequency dispersion of dielectric constant, and often large effective responses such as dielectric constant or electromechanical coupling. In hybrid ferroelectrics, relaxor behavior can originate from the phase separation on anionic or cationic lattice and the rotational instability of organic cations. Alternatively, somewhat similar macroscopic responses can be associated with the coupling between ferroelectric and ionic responses. For example, ion migration can locally break the crystal symmetry by hopping ions to the nearest neighbors inducing nanosize polar regions (nanodomain) which can cause ferroelectric instability and can be further enhanced under photoexcitation or bias stimuli.
Organometallic halide perovskites (OMHPs) have attracted broad attention as prospective materials for optoelectronic applications. Among the many anomalous properties of these materials, of special interest are the ferroelectric properties including both classical and relaxor-like components, as a potential origin of slow dynamics, field enhancement, and anomalous mobilities. Here, ferroelectric properties of the three representative OMHPs are explored, including FAPbxSn1–xI3 (x = 0, x = 0.85) and FA0.85MA0.15PbI3 using band excitation piezoresponse force microscopy and contact mode Kelvin probe force microscopy, providing insight into long- and short-range dipole and charge dynamics in these materials and probing ferroelectric density of states. Furthermore, second-harmonic generation in thin films of OMHPs is observed, providing a direct information on the noncentrosymmetric polarization in such materials. Overall, the data provide strong evidence for the presence of ferroelectric domains in these systems; however, the domain dynamics is suppressed by fast ion dynamics. These materials hence present the limit of ferroelectric materials with spontaneous polarization dynamically screened by ionic and electronic carriers.
Organometallic halide perovskites (OMHPs) have garnered considerable interest for applications in advanced electronic technologies, such as solar cells,[1–3] light-emitting diodes,[2,4] photodetectors,[5–9] high energy radiation sensors,[10–12] and even as gas sensors.[13] The unique functional properties of these materials further enable applications for photoferroelectrics and photopiezoelectrics,[14] enabling new generations of light-addressable sensors and internet of things (IoT) devices.[15,16] Such applications necessitate fundamental understanding of the coupling Dr. M. Ahmadi, J. Zhang, Prof. B. Hu Joint Institute for Advanced Materials Department of Materials Science and Engineering University of Tennessee Knoxville, TN 37996, USA E-mail:
[email protected] Dr. L. Collins, Dr. A. Puretzky, Dr. J. K. Keum, Dr. I. Ivanov, Dr. S. V. Kalinin Center for Nanophase Materials Sciences Oak Ridge National Laboratory Oak Ridge, TN 37831, USA E-mail:
[email protected] Dr. W. Lu Department of Chemistry University of Tennessee Knoxville, TN 37996, USA
DOI: 10.1002/adma.201705298
Adv. Mater. 2018, 30, 1705298
1705298 (1 of 10)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
www.advmat.de
These considerations stimulated a number of studies using piezoresponse force microscopy (PFM) as a tool to visualize nanometer-scale domain patterns and establish polarization switching.[26–31] These studies provided certain insight into domain structures and switching behavior in polycrystalline and single crystal of methylammonium lead iodide (MAPbI3). However, PFM is known for being notoriously sensitive to surface charging, ionic, and electret-like phenomena,[32] as utilized in the electrochemical force microscopy.[33–35] Similarly, surface topography gives rise to strong topographic cross-talk,[36,37] which can result in identification of ferroelastic domains or step edges as ferroelectric domains.[38] Finally, the surfaces of these materials are especially unstable in ambient environment, potentially leading to the formation of the secondary phases such as PbI2, etc. that can further mask intrinsic materials properties.[39–41] In addition, dynamic PFM is known to be influenced by electrostatic contributions originating from nonlocal forces acting on the tip, particularly when using soft cantilevers (approximately few N m−1), or studying poor ferroelectric samples, meaning electrostatic contributions are often nonnegligible.[42,43] In particular, PFM switching experiments are prone to artefacts, where high dc-voltage pulses are applied to induce ferroelectric polarization switching, but can also cause charge injection or activate chemical processes. Often, evidence for ferroelectricity is assigned primarily due to the observation of hysteresis loops in PFM spectroscopy and/ or in conjunction with domain contrast in PFM. However, both observations can also be made on clearly nonferroelectric materials.[44] Although ferroelectricity has been recently observed in a molecular organic–inorganic perovskite,[45] the ferroelectric or antiferroelectric nature of MAPbI3 perovskites is strongly debated. For example, ferroelectric P–E behavior was revealed in single crystals of MAPbI3 only at low temperature by using the lossy part of the dielectric response similar to ferroelectric insulators.[46] It was proposed that at low temperatures the electrochemical effects, ionic conductivity, and dielectric relaxation which can screen the ferroelectric switching decrease and consequently the spontaneous polarization increases. Very recently, ferroelastic domains were observed in MAPbI3 single crystals and polycrystalline thin films in both pristine state and under applied stress.[47] Although ferroelectricity was not detected, it was demonstrated that strain engineering can tune ferroelastic domains in both MAPbI3 single crystals and polycrystalline films. It was also shown that the antiferroelectric phase in MAPbI3 can be switched to ferroelectric polarization by poling treatment or under photoexcitation.[22,48] Here, we seek to establish the presence, observability, and character of ferroelectricity in these materials in ambient conditions using the combination of band excitation PFM (BE-PFM), Kelvin probe force microscopy (KPFM), and contact Kelvin probe force microscopy (cKPFM) for the first time. Of these, BE-PFM provides the information on local electromechanical coupling, obviating topographic effects,[49–51] KPFM ascertains the presence of surface dipoles and charges on the pristine surface and surface modified by an SPM tip, and cKPFM provides the spectroscopic information into long- and short-range dipolar behaviors (ferroelectric density of states), allowing to differentiate ferroelectric and ionic/relaxor responses.[42,52–54]
Adv. Mater. 2018, 30, 1705298
It is well established that in inorganic perovskites, both octahedral rotations and cation site displacements compete with ferroelectricity. Therefore, those compounds with the highest tendency to octahedral rotations (and the smallest tolerance factors) have the largest ferroelectric instabilities. In this respect, among OMHP compounds, formamidinium (FA) cation disfavor octahedral rotation compared to methylammonium (MA) due to its hydrogen bonding.[55] In addition, FASnI3, FAPb0.85Sn0.15I3, and FA0.85MA0.15PbI3 have highest tolerance factors compared to MAPbI3, respectively.[56] Therefore, to explore ferroelectricity in OMHPs, we have chosen the family of FA lead/tin iodide polycrystalline thin films including FASnI3, FA0.85MA0.15PbI3, and FAPb0.85Sn0.15I3 for the first time. Although lead-halide perovskites provide many superior characteristics, the toxicity of lead is problematic for the practical implementation. Thus, recently the fabrication of lead-free alternatives and environmental friendly perovskite materials, and devices have been conducted very actively.[57–62] To reduce the use of lead, as well as to tune the bandgap of material, substitution (total or partial) of Pb2+ by Sn2+, and MA by FA has been suggested.[59,63–65] In addition, it has been demonstrated that FA-based perovskites are more chemically and thermally robust compare to MA-based perovskite.[63,66,67] We also note that they are more stable during scanning with SPM tip. Relevant to the goals of this study, the selected materials differ in crystallographic symmetries and hence it is expected to have different ferroelectric and ferroelastic domain morphologies. At the same time, bulk and especially surface electrochemical properties driven by the instability of metal-sublattice can be expected to be comparable, providing comparative base between observed phenomena. As a first system, we explored rhombohedral FA0.85MA0.15PbI3 thin film. It was found that at room temperature (RT), FAPbI3 tends to crystallize in a wide bandgap hexagonal symmetry (P63mc) or δ-phase instead of the desired rhombohedral (P3m1) α-phase formed at a higher temperature (130 °C). Therefore, it was suggested that partial substitution of FA (
