# Inorganic–organic halide perovskites for new photovoltaic technology

Inorganic–organic halide perovskites for new photovoltaic technology Abstract Perovskite solar cells (PSCs) have received worldwide attention due to excellent power-to-electricity conversion efficiency (PCE). Currently, a PCE record of as high as 22.1% has been achieved, comparative to those of copper indium gallium selenide and CdTe solar cells. However, there are still some critical issues to solve, especially the stability issue at the current state, in order to promote PSC commercialization. In this respect, various device architectures and the working principle are briefly introduced. Then, key features of inorganic–organic hybrid perovskites and related film deposition methods are discussed. Typically, experimental and theoretical efforts to reveal the physical issues of the device are emphasized as well as the cell stability. Finally, a short summary, challenge and prospect are given. inorganic–organic hybrid perovskite materials, perovskite solar cells, heterojunction, photophysics, stability INTRODUCTION Thin-film solar cells are a promising type of cost-competitive solar power via cost-effective materials and fabrication technology, comparable with commercial silicon-based photovoltaics [1]. The bandgaps of CdTe (1.45 eV) and copper indium gallium selenide (CIGS, 1.04–1.68 eV) semiconductors can well meet the spectrum of sunlight, guaranteeing a power-to-electricity conversion efficiency (PCE) of over 22% for CdTe and CIGS thin-film solar cells; however, vacuum deposition and/or high-temperature processes are still needed [2–3]. Solution-processed technology represents the lowest-cost production method for thin-film solar cells, including spin-coating, blade-coating, spraying, inkjet printing, gravure printing, slot-dye coating and so on [4–5]. Recently, solution-processed perovskite solar cells (PSCs) have attracted worldwide attention due to their extremely high efficiency exceeding 22%, comparative to those of CIGS and CdTe solar cells and far exceeding dye-sensitized solar cells (DSCs) and organic photovoltaics (OPVs) [6]. It is very likely that a PCE of 25% may be achieved in the future [7]. Surprisingly, this progress has been achieved in only a few years along with rapid development of different device architectures, processing techniques and novel materials. The excellent cell performance of PSCs mainly benefits from inorganic–organic hybrid perovskite materials with advantages of ease of fabrication, panchromatic sunlight absorption, high carrier mobility, superior carrier diffusion length, long carrier lifetimes and so on [8]. The first PSC based on CH3NH3PbX3 (X = Cl and Br) as sensitizers was reported by Kojima et al. in 2009, exhibiting 3.8% PCE, which was further improved to 6.54% by Im et al. in 2011; however, a rapid degradation of the cell performance perplexed scientists [9–10]. The breakthrough came in 2012 when Kim et al. fabricated all solid-state mesoscopic heterojunction PSCs by using spiro-OMeTAD, presenting 9.7% PCE and relatively good stability [11]. Since then, some milestones in PCEs have emerged successively, which further stimulated the development of PSCs. In 2012, Lee et al. fabricated a ‘meso-superstructured’ PSC with 10.9% PCE by replacing the mesoporous TiO2 with Al2O3 [12]. The successful application of this meso-structured insulating scaffold demonstrated that the CH3NH3PbI3-xClx perovskite material is extraordinarily effective, and can transport both electrons and holes between cell terminals and function well in a simplified device architecture even without scaffolds. In 2013, Burschka et al. reported a sequential two-step dipping method to fabricate PSCs with 15% and a certified 14.14% PCEs [13]. Subsequently, by the aid of compositional engineering and the ‘anti-solvent one step deposition method’, Saliba and Shin's groups independently achieved over 21% PCE [14–15]. In the meantime, the stability of the PSCs was improved by introducing Cs+ and Rb+ cations, adopting 2D perovskite films or using La–BaSnO3 to replace conventional TiO2 [15–17]. In addition, a certified 19.7% PCE for PSCs with an aperture area exceeding 1 cm2 was also obtained [18]. Although the high efficiencies of PSCs have been frequently refreshed, the perovskite layers are subject to compositional degradation under high humidity and thermal treatment. Recent investigation reveals that the PSC may suffer from a change in the semiconductor electric properties while the cell operates. Therefore, long-term stability becomes a primary concern of the community [19]. On one hand, thermal stability is particularly important for solar cell operation; on the other hand, solar modules have to operate continuously between temperatures of −40 and 85°C as well as high solar irradiance over a long period of time [14]. In this review, we will briefly introduce the device architectures and the working principle. The key features of perovskite materials and film deposition techniques will also be reviewed. Typically, experimental and theoretical efforts to reveal the physical issues of the device are emphasized as well as cell stability. Finally, a short conclusion and prospect will be given. DEVICE ARCHITECTURES The huge success of PSCs not only lies in their high efficiencies but also their diversity in device architectures, all of which are derived from the distinctive optoelectric properties of the perovskite materials. It is also found that the evolution of device structures is also accompanied by the development history of the devices. In this section, we will briefly introduce different PSC structures and the development history of the devices. The first PSC employed a mesoporous TiO2 layer as a scaffold, which supported hybrid perovskite materials as sensitizers and transported photo-generated electrons [9]. Due to the fine carrier transportation ability of inorganic–organic hybrid perovskites, the device architecture has evolved and different PSC structures have emerged [20]. Mesoscopic structure employs a semiconductor oxide scaffold, which is fully infiltrated with the perovskite and can extend over the nanostructure to form a capping layer, so-called ‘mesoporous structure’, see Fig. 1a. A specific mesoporous PSC structure worth mentioning is based on a mesoporous Al2O3 insulating scaffold [12]. PSCs can also be fabricated without a mesoporous scaffold, so-called ‘planar PSCs’ [21]. Considering the position of the electron- or hole-transporting layer in front of the incident light, planar PSCs can be classified into n-i-p and p-i-n (inverted) structures; see Fig. 1b. Regardless of the light irradiation direction, the PSC generally consists of a conductive substrate (fluorine-doped tin oxide (FTO), indium-tin-oxide (ITO) or flexible ITO-polyethylene terephthalate (PET)), an electron-transporting layer (ETL), a perovskite layer, a hole-transporting layer (HTL) and a back electrode (Au or Ag). When PSCs are defined by whether they use an HTL or not, the HTL-free PSC is also an important branch; the HTL-free full-printable PSCs in particular exhibit good stability [22]. At present, mesoscopic systems exhibit a certified PCE of 22.1% on the National Renewable Energy Laboratory's best efficiency chart, slightly superior to planar embodiments with a PCE of 20.8% [6,23]. Figure 1. View largeDownload slide Schemes of basic PSC structures: (a) mesoporous PSCs; (b) planar PSCs with n-i-p and p-i-n cell architectures; (c) and (d) energy-level diagrams (eV) of routine PSC materials. Figure 1. View largeDownload slide Schemes of basic PSC structures: (a) mesoporous PSCs; (b) planar PSCs with n-i-p and p-i-n cell architectures; (c) and (d) energy-level diagrams (eV) of routine PSC materials. PSCs have been proved to be a free carrier-based device, in which the charge separation and transportation processes are more like a heterojunction solar cell [24–25]. To achieve efficient PSCs, the energy levels of these different components need to be well matched (see Fig. 1c, d). In addition, the key materials, fabrication techniques and interface engineering have to be optimized. In the meantime, a better understanding the dynamic processes of the device is also needed. Detailed discussion will be presented in the following sections. PEROVSKITE MATERIALS AND FILM FABRICATION TECHNIQUES Generally, optoelectrical materials with different chemical compositions exhibit different trends in their application. The hybrid organic–inorganic perovskite structures closely bridge organic and inorganic materials through a well-defined molecular assembly. In this section, we will review the basic properties of perovskite materials and the recent efforts in understanding the relationship between the composition, properties and fabrication techniques. Engineering perovskite materials as light absorbers Brief introduction to perovskites The perovskite structure was first recognized as mineral calcium titanate (CaTiO3). Halide perovskite, as a new branch of perovskite materials, exhibits attractive photovoltaic characteristics [26]. So far, all inorganic CsBX3 (B = Sn, Pb; X = Cl, Br, I) and inorganic–organic hybrid perovskites have been successfully used in photovoltaic devices. Stoichiometric inorganic–organic hybrid perovskites, ABX3 (A = organic cation, widely used methylammonium (MA) or formamidinium (FA); B = Pb, Sn; X = Cl, Br and I), are generally used to describe this kind of perovskite, which has a 3D structure with corner-sharing anionic BX6 octahedra with the X atoms located in the corners and the B atom in the middle of the octahedra, and A cations located at interstices surrounded by eight octahedra in the cuboctahedral gap; see Fig. 2a. These hybrid perovskite materials present fascinating characteristics for high-performance photovoltaic devices, i.e. adjustable bandgap (optimal 1.4– 1.5 eV), high absorption coefficient (104–105 cm−1), low exciton binding energy (<50 meV), ambipolar charge transport capability and long carrier diffusion length (∼175 μm) [8,27,28]. In addition, relatively good crystallization and easy film formation derived from a solution-processed method at low temperature make them highly competitive, promising low-cost photovoltaic materials. Figure 2. View largeDownload slide (a) The basic perovskite structure ABX3 and their extended network structure connected by the corner-shared octahedral; (b) basic structures of 2D organic–inorganic perovskite with bilayer intercalated organic molecules; (c) calculated t and μ factors for 12 halide perovskites; EA represents CH3CH2NH3 (Note: redrawn figure). Adapted with permission from [26]. Figure 2. View largeDownload slide (a) The basic perovskite structure ABX3 and their extended network structure connected by the corner-shared octahedral; (b) basic structures of 2D organic–inorganic perovskite with bilayer intercalated organic molecules; (c) calculated t and μ factors for 12 halide perovskites; EA represents CH3CH2NH3 (Note: redrawn figure). Adapted with permission from [26]. Generally speaking, the crystallographic stability and possible structure of perovskite ABX3, even selecting A, B and X ions, have to consider an octahedral factor μ and a tolerance factor t:   $$\mu \ = \frac{{{R_{\rm{B}}}}}{{{R_{\rm{X}}}}}\$$ (1)  $$t\ = \frac{{{R_{\rm{A}}} + {R_{\rm{X}}}}}{{\sqrt 2 \left( {{R_{\rm{B}}} + {R_{\rm{X}}}} \right)}}\$$ (2)where RA, RB and RX represent the ionic radii of A, B and X ions. If t is in the range of 0.89–1.0, the perovskite material usually adopts a cubic structure, as shown in Fig. 2c; however, a lower t value will result in less symmetrical tetragonal or orthorhombic structures. When t is close to unity, the ideal cubic perovskite structure can be obtained. Calculated t and μ factors for a range of halide perovskites are shown in Fig. 2c [26]. Despite these constraints, transitions between such structures while heating are common for any given halide perovskites, such as MAPbI3 (Fig. 3a) and FAPbI3 (Fig. 3b), and the high-temperature phase is generally cubic [29]. In fact, a cation is generally considered to be the charge compensation in the lattice, and does not contribute to the band structure. However, variation of the A cation's size will strongly influence the optical property due to its deformation of the MX64− octahedral network. Namely, smaller A cations (i.e. MA or FA) can maintain a 3D structure, whereas larger A cations (long chain alkyl or aryl ammonium) will result in lower-dimensional layered or confined perovskites, as seen in Fig. 2b. Figure 3. View largeDownload slide Graphical schemes of the structural transformation of (a) the MAPbI3 perovskite and (b) the FAPbI3 perovskite. Adapted with permission from [29]. Figure 3. View largeDownload slide Graphical schemes of the structural transformation of (a) the MAPbI3 perovskite and (b) the FAPbI3 perovskite. Adapted with permission from [29]. The 3D halide perovskites have a high extinction coefficient (α) up to 105 cm−1 [30–31]. For example, the light absorption capability of MAPbI3 is even higher than that of GaAs, which contributes to the best-performing single-junction solar cell, and two orders higher than that of crystal silicon [6]. According to the first-principles calculation, the electronic structure of MAPbI3 is dominated by characteristics of the Pb–I bond, the valence-band maximum of which is characterized as an anti-bonding hybrid state of Pb 6s and I 5p orbitals, while the conduction band minimum is characterized as a non-bonding hybrid state of Pb 6p and I 5p orbitals [32]. The high α is attributed to the high density of states in the conduction band minimum (CBM) of the perovskite, and direct p–p transition may be the origin. Due to the high α and strongly reflective back electrode, a 500-nm-thick perovskite film is enough to capture most incident photons with energies lower than the energy gap (Eg). Perovskite film fabrication techniques The morphology and crystallization of perovskite films will strongly influence the cell performance. In the meantime, high-quality perovskite film is susceptible to film deposition conditions and techniques. For the moment, three main film deposition methods are usually adopted: one-step precursor solution deposition, two-step sequential deposition (including the vapor-assisted solution process) and vapor deposition, as shown in Fig. 4. Figure 4. View largeDownload slide Three general perovskite film fabrication methods: (a) one-step precursor deposition method (CBZ: chlorobenzene); (b) sequential deposition method; (c) dual-source vapor deposition method. Adapted with permission from [33,35,43]. Figure 4. View largeDownload slide Three general perovskite film fabrication methods: (a) one-step precursor deposition method (CBZ: chlorobenzene); (b) sequential deposition method; (c) dual-source vapor deposition method. Adapted with permission from [33,35,43]. One-step precursor solution deposition At the initial development stage of PSCs, the one-step precursor solution deposition method was used most often; it was carried out by directly spin-coating the mixed solution on the substrate, and the precursor solution was obtained by dissolving all the precursor salts in high-boiling-point polar solvent (i.e. dimethylformamide (DMF), dimethylsulfoxide (DMSO), γ-butyrolactone (GBL)), as shown in Fig. 4a [11]. Aiming at uncovered pinhole areas derived from large MAPbI3 grains, Zhao et al. developed a one-step, solvent-induced, fast crystallization method (the so-called ‘anti-solvent method’) involving spin-coating a DMF solution of MAPbI3 followed immediately by exposure to chlorobenzene to induce perovskite crystallization for planar PSCs [33]. So far, by using the anti-solvent method combined with compositional engineering, over 21% PCE has been achieved [14–17,34]. In addition, a simple alternative vacuum flash-assisted deposition method was reported by Li et al., which enabled a certified record PCE of 19.6% for a 1-cm2-sized device and a maximum efficiency of 20.5% without hysteresis [35]. Overall, the anti-solvent one-step method can clearly induce rapid nucleation and slow crystal growth processes during the perovskite film formation, thus providing an effective way to achieve high-quality perovskite crystals. In the meantime, this method is suitable for all PSC structures (planar vs. mesoporous), and it may be one possible way to realize the large-area manufacturing. Two-step deposition method A sequential two-step dipping technique was first reported by Mitzi in 1998, then Burschka et al. in 2013 [13]. In a typical two-step method, PbI2 is firstly spin-coated onto the substrate, followed by exposure to the organic ions either in solution or in vapor to afford the perovskite film (Fig. 4b). This method presented the first 15% PCE with a certified 14.14% PCE [13]. Subsequently, aiming at solving incomplete conversion of PbI2 and uncontrollable crystal size and surface morphology, different post-treatment methods have also been developed; e.g., Zhu et al. and Xiao et al. reported a pressure-assisted morphology reconstruction method and temperature-assisted sequential deposition method to control film morphology and interfacial charge transport property for high-efficiency PSCs, respectively [36–37]. The vapor-assisted solution process (VASP) can be considered as another kind of two-step method, as shown in Fig. 4b [38]. This method is different from the conventional two-step solution process and vacuum deposition, and takes advantage of the kinetic reactivity of methylamine iodide (MAI) and the thermodynamic stability of perovskites during the film growth process and affords well-defined grain structures with sizes up to micron-scale, full surface coverage and smooth surface without pinholes. Currently, various VASP film growth techniques have been developed, such as hybrid chemical vapor deposition (HCVD), hybrid physical chemical vapor deposition (HPCVD), low-pressure chemical vapor deposition (LPCVD), in situ tubular chemical vapor deposition (ITCVD) and modified chemical vapor transport (mCVT) [39]. 18.9% PCE was obtained under ambient atmosphere by a modified VASP method [40]. VASP presents a simple, controllable and versatile method for high-quality perovskite films and high-efficiency PV devices. Vapor deposition method Vapor deposition techniques are widely used in the semiconductor industry for largescale production in optoelectronic applications. Perovskite films derived from the vapor deposition process under vacuum were first reported by Salau et al. [41], then by Mitzi et al. [42]. In 2013, Liu et al. modified this technique as a dual-source vapor deposition method for pinhole-free MAPbI1-xClx perovskite films with hundreds of nanometer-scale crystals for planar PSCs; see Fig. 4c [43]. Malinkiewicz et al. fabricated an inverted p-i-n planar PSC with sublimated MAPbI3 perovskite layer [44]. Recently, the chemical vapor deposition (CVD) method was also reported by precisely controlling the crystallization process [45]. Vapor deposition techniques are suitable for multilayered thin-film structures and a variety of substrates; however, this technique demands high vacuum. Compositional engineering As we know, tailoring the bandgap of the light absorber is important for minimizing the energy loss and improving the cell performance. A material engineering in terms of AMX3 compositional control is necessary; namely, the absorption range and the Eg value of these perovskites can be adjusted by ion substitutions [28,47]. In addition, the Eg value of a semiconductor is related to the lattice size and the atomic interaction according to the tight-binding approximation theory. MAPbI3 can absorb the widest range of visible light, showing an Eg value of about 1.5 eV; however, the Eg value of MAPbBr3 is about 2.3 eV. Noh et al. reported (MAPbI3)1-x(MAPbBr3)x mixed perovskite by material engineering, the bandgap (Eg) of which can be changed from 1.57 to 2.29 eV by the following equation Eg(x) = 1.57 + 0.039 x + 0.33 x2. This mixed perovskite with x ≥ 0.2 exhibits higher stability toward the humidity than those with x < 0.2 [48]. In the chlorine substitution of MAPbI3, the main role of the chloride is not to regulate the bandgap. Instead, the introduction of chloride into MAPbI3 mainly improves the film morphology, preferential orientation, crystallinity, grain size and so on, which can further improve carrier transport across the heterojunction interface, passivate the grain boundary and result in higher cell performance [49]. Partial replacement of halides with the pseudohalide SCN−, MAPbI3-x(SCN)x, has also been reported; this exhibited good humidity-resistance characteristics, thus leading to enhanced moisture stability of the PSCs [50–51]. In fact, based on the investigation of single crystals grown under the same conditions, 2D structured (MA)2Pb(SCN)2I2 is suggested instead [52]. Xiao et al. further investigated this 2D (MA)2Pb(SCN)2I2 by the aid of experimental measurement and DFT calculation, and found that its bandgap is about 2.0 eV, larger than that of MAPbI3; in the meantime, the carriers are confined within the 2D layers [53]. PCEs of 8.3% were reported for mixed (MA)2Pb(SCN)2I2/MAPbI3 devices [54]. When larger-sized FA (HC(NH2)2+) occupies the A site of the perovskite lattice, the Eg value can be reduced to about 1.47 eV, closer to the desired Eg of 1.4 eV. In addition, FAPbI3 has other advantages over MAPbI3, such as a smaller polarization dipole, stronger p-type characteristics and higher thermal stability. For FAPbI3-based PSCs, a higher photocurrent can be obtained but with a lower photovoltage, and a non-perovskite phase is found to restrict the cell performance [55]. Therefore, the Eg needs to be optimized for a balance. (FAPbI3)0.85(MAPbBr3)0.15 mixed perovskite derived from FAPbI3, FAPbBr3, MAPbI3 and MAPbBr3 can afford a complete perovskite phase, in which the PCE was increased to 20.8% with a negligible hysteresis [15]. Very recently, PSCs based on triple and quadruple cation perovskites (CsRbMAFAPb(I/Br)3) have been fabricated; which exhibited a PCE of 21.6% with an average PCE of 20.2% and stabilized 19.0% on a 0.5 cm2 cell. In addition, polymer-coated cells can maintain 95% of their initial performance at 85°C after 500 aging hours under full illumination and maximum power point tracking [14,16]. Obviously, compositional engineering is feasible to fabricate highly efficient PSCs with better stability and reproducibility. Replacing the metal cation is another method of narrowing the bandgap; in the meantime, developing lead-free perovskite photovoltaics is also in line with the environmental regulations. Jain et al. used high-throughput density functional theory calculations to screen lead-free perovskite-like materials and found that about 10 compounds (A = Na, K, Rb, Cs, Cu and Ag; B = Ga, In and Sb; X = Cl, Br and I) have bandgaps in the range of 1.5–2.5 eV and exhibit potential application in lead-free photovoltaic devices [56]. Pure Sn halide perovskite itself only exhibited a PCE of over 6% and poor stability [57–58]. However, according to the first-principles calculation, when Pb atoms are partially substituted by Sn atoms, the Eg decreases instead, probably due to weaker atomic interactions between Sn and halides [59]. Eg can be further reduced to 1.16 eV when mixed Sn/Pb atoms occupy the B site, and a steady decrease in Eg is also verified by theoretical calculation and experimental measurement [60]. Although a higher photocurrent could be obtained by decreasing the Eg with Sn substitution, a much lower device performance was presented instead due to high hole doping or defects, and the existence of Sn atoms also resulted in low stability [57]. Bismuth (Bi3+) can be used as a nontoxic replacement for Pb-based perovskite due to its similar electronic structure to Pb2+. Recently, some efforts have been made to incorporate Bi3+ as the B-site cation in the ABX3 perovskite framework. (MA)3Bi2I9 perovskite possesses a bandgap of 2.1 eV and cannot realize the high cell performance; however, the bandgap was significantly reduced to 1.45 eV after sulfur doping, which was supposed to be more suitable for application in photovoltaics [61]. Some work reported that, in the formation of 2D Bi3+-halide perovskites, Bi3+ has been incorporated into the lattice with vacancy formation on the B site and an organic cation as the template [62]. Slavney et al. reported a double-perovskite structure Cs2AgBiBr6 with an indirect bandgap of 1.95 eV, suitable for a tandem solar cell [63]. So far, compositional engineering with A and X site substitutions has brought about tremendous success. For a single-crystal MAPbI3 perovskite, the absorption edge is red-shifted with Eg lower than 1.5 eV. This may be due to the difference in lattice size and defect properties between the surface and the bulk of perovskite crystals [64]. Currently, the cell performance based on single-crystal perovskites is still unsatisfactory. Crystal growth engineering A high-quality perovskite film is the key to high-efficiency PSCs, and this kind of high-quality perovskite film refers to films with fewer defects and trap states derived from pinholes, crystal defects and grain boundaries. As we know, inorganic–organic halide perovskite films can be easily obtained in mild conditions due to a low lattice energy. Therefore, whatever fabrication method is adopted, it is necessary to effectively control the perovskite growth process. In fact, when the perovskite formation rate is high, it is hard to control crystal growth. The Lewis acid–base adduct method has been recognized as a viable way to regulate perovskite crystal growth. A Lewis acid is defined as an electron-pair acceptor and a base as an electron-pair donor. The Lewis acid–base reaction will lead to either a redox reaction or adduct formation, and adducts are generally a kind of coordination compounds. For inorganic–organic halide perovskites, Pb(II) halides are typical Lewis acids, which can afford adducts with monodentate or bidentate ligands in 1: 1, 1: 2 or 2: 1. Monodentate or bidentate ligands (Lewis bases) usually contain O-donors (e.g. DMSO, N-methyl-2-pyrrolidone (NMP)), S-donors (e.g. thiourea, thioacetamide) and N-donors (e.g. pyridine). Initially, this adduct-induced perovskite film growth method is supposed to strongly retard the rapid crystallization of PbI2, which is based on the strong interaction between Pb2+ and S = O from DMSO. Jeon et al. first prepared the PbI2(DMSO) precursor, which was further applied to fabricate the FAPbI3/MAPbBr3 by direct intramolecular exchange with FAI/MABr, presenting a PCE of 20.2%, as shown in Fig. 5 [46]. Then, the mixture of DMSO and DMF was directly adopted to give an adduct of MAI⋅PbI2⋅DMSO, since PbI2 prefers to coordinating with DMSO relative to DMF due to its stronger Lewis basicity [65]. This exchange does not induce volume expansion or change the film thickness due to the similar molecular volumes of DMSO and FAI, and the perovskite film is extremely uniform and flat as well. Recently, Jo et al. reported that a PbI2(NMP) complex (NMP = N-methyl-2-pyrrolidone) as an intermediate can also convert into high-quality perovskites. Larger perovskite grain size was obtained from the PbI2(NMP) film than from the PbI2(DMSO), which can reduce bulk defects and improve charge carrier mobility, thus leading to a PCE of 19.5% [66–67]. When the FAPbI3 film was fabricated by this adduct method, however, thiourea as an S-donor was preferred rather than DMSO as an O-donor due to its stronger coordination ability, and a 1:1:1 adduct of FAI⋅PbI2⋅thiourea is suggested. The resulting FAPbI3 films showed better cell performance because of faster charge transport and the reduced trap states [68]. In addition, this kind of Lewis acid–base intermediate adducts plays a crucial role in controlling the morphology of perovskites. Crystal growth engineering via Lewis acid–base adducts is expected to afford homogeneous films with large grain sizes since the removal of the Lewis base in the adduct can also reduce perovskite crystallization rate. Figure 5. View largeDownload slide Scheme for perovskite formation via the MAI–PbI2–DMSO intermediate: (i) PbI2 with edge-sharing [PbI6]4− octahedral layers; (ii) the formation of flat MAI–PbI2–DMSO intermediate phase by MAI and DMSO molecules intercalated between the layers while toluene is introduced to the wet film with PbI2, MAI and DMSO; (iii) conversion of MAI–PbI2–DMSO into the perovskite phase with corner-sharing octahedra by removing DMSO molecules while annealing. Adapted with permission from [46]. Figure 5. View largeDownload slide Scheme for perovskite formation via the MAI–PbI2–DMSO intermediate: (i) PbI2 with edge-sharing [PbI6]4− octahedral layers; (ii) the formation of flat MAI–PbI2–DMSO intermediate phase by MAI and DMSO molecules intercalated between the layers while toluene is introduced to the wet film with PbI2, MAI and DMSO; (iii) conversion of MAI–PbI2–DMSO into the perovskite phase with corner-sharing octahedra by removing DMSO molecules while annealing. Adapted with permission from [46]. In order to improve the crystallization quality and reduce defect states of the perovskite films, some additives have also been introduced. Yang et al. introduced additional iodide ions into the organic cation solution, which was subsequently used for the intramolecular exchange process for perovskite films. Their distinctive work lies in the fact that their iodide solution was actually triiodide (I3−) solution, which was obtained by stirring iodine (I2) in isopropanol (IPA) at 80°C and reducing some I2 molecules to I3− ions. Further investigation revealed that I3− decreased deep-level traps in bulk perovskite layers, thus leading to a PCE as high as 22.1% with a certified PCE of 19.7% [69]. However, unlike Yang's work, Liu et al. used a small amount of methylamine solution as an additive to eliminate small I2 and polyiodide colloids in the precursor solution, also leading to reduced defect density and extended carrier diffusion length in the resultant film [70]. Obviously, a thorough understanding a variety of defect states as well as the origin is desirable in the future. PHYSICS IN PEROVSKITE SOLAR CELLS Although a certified PCE of 22.1% is presented in the NREL best efficiency chart [6], some important issues still need to be clarified. In addition to the stability, the discrepancy in reverse and forward scanned current–voltage curves, so-called ‘hysteresis’, is also an obstacle. Therefore, deep understanding the physics involved in PSCs is essential for developing stable and highly efficient devices. Doping, carrier and electric properties Point defects and doping Point defects and self-doping of a multiple-component semiconductor are permanent topics, especially in understanding the physical properties of a photovoltaic material. As a ternary semiconductor, organic halide lead hybrid perovskites can be stable in a certain thermodynamic region [71]. In this region, the composition of halide perovskite materials may deviate from stoichiometry more or less, forming point doping or defects, such as ion vacancies, interstitials and substitutions. The self-doping properties of the perovskite MAPbI3 are shown in Fig. 6. Iodide interstitial (Ii), cation substitution (MAPb), cation vacancies (VMA and VPb) and anti-substitutions (e.g. IMA and IPb) will introduce shallow acceptor doping levels, while deep defects may be caused by the IMA and IPb. On the other hand, the cation interstitial (MAi and Pbi), anion vacancies (VI), cation substitution (PbMA) and anti-substitutions (e.g. MAI and PbI) will introduce shallow donor doping levels, while Pb interstitial and anti-substitution will cause deep defects. Generally speaking, it is still dependent on their formation energies in the thermodynamic region whether these self-doping levels could appear while the perovskite film is growing. The formation energy is mainly related to growth conditions. For example, the Pb vacancy easily appears in the I-rich/Pb-poor condition instead of the I-poor/Pb-rich condition. In experiments, this self-doping may be influenced by the film deposition method, precursor compositions, thermal treatment temperature and crystallization process. Wang et al. found that the doping type of MAPbI3 is strongly dependent on the molar ratio of PbI2/MAI in the one-step deposition process, and more MAI would lead to p-type perovskite [73]. They also found that the interdiffusion method is beneficial to the growth of p-type MAPbI3 film whereas the hole concentration will decrease while prolonging the annealing time. For the two-step method, the hole concentration could be adjusted by the MAI concentration, and higher hole concentration was obtained by using a more diluted MAI solution [74]. For Sn-based perovskites, the exotic self-doping ability of the Sn-based perovskites allows them to adopt electrical properties from semiconductor to metallic, which are dependent on the synthetic procedures [58,75]. In particular, easy oxidization of Sn2+ to Sn4+ will cause materials doped with Sn4+, which behave as p-type semiconductors, however, displaying metal-like conductivity, e.g. the electron and hole mobilities of MASnI3 are estimated to exceed 2000 and 300 cm2/(Vs), respectively [76]. Figure 6. View largeDownload slide (a) Calculated transition energy levels of intrinsic point defects in CH3NH3PbI3, and possible (b) donor and (c) acceptor doping by impurity atom substitution or interstitial. Adapted with permission from [71,72]. Figure 6. View largeDownload slide (a) Calculated transition energy levels of intrinsic point defects in CH3NH3PbI3, and possible (b) donor and (c) acceptor doping by impurity atom substitution or interstitial. Adapted with permission from [71,72]. Figure 7. View largeDownload slide (a) Schematic diagram of the Kelvin probe force microscopy (KPFM) for the measurement of surface contact potential across the cell. (b) KPFM results and derived electric field properties of a planar perovskite solar cell under different steady-state bias voltages. Adapted with permission from [85]. Figure 7. View largeDownload slide (a) Schematic diagram of the Kelvin probe force microscopy (KPFM) for the measurement of surface contact potential across the cell. (b) KPFM results and derived electric field properties of a planar perovskite solar cell under different steady-state bias voltages. Adapted with permission from [85]. Besides self-doping, p- or n-type doping by impurity atoms has also been theoretically studied [77]. Yan et al. found that cation substitutions with higher valent atoms or cation interstitials could give more free electrons to the perovskite system, so-called ‘n doping’ [72]. On the other hand, ion substitutions with lower valent atoms would bring about electron-deficient energy levels, capturing the electrons in the valence band maximum (VBM) and causing p doping, as shown in Fig. 6c. Recently, the doping toward the Pb atom (such as Sb, In, Rb, Cu and Ag) has been reported to enhance the efficiency and stability of the solar cell [24,78–80]. However, little experimental work could directly confirm the doping and carrier properties. Junction and electric field According to the basic principles of semiconductor physics, a junction with a built-in electric field may form when two materials with different Fermi energy levels electrically contact with each other [81]. However, this concept was not widely accepted in the early developing stage of PSCs until more and more real-space nanoscale electric measurements were carried out [82]. Kelvin probe force microscopy (KPFM) is an effective approach to investigate the potential and electric field properties inside the cell [83]. By the aid of KPFM, a potential distribution is found in the perovskite and the TiO2-based mesoscopic PSC, and a large electric field of about 106 V m−1 can be derived. A similar result was obtained in the planar cell with a ∼300-nm-thick depletion region [84]. In these measurements, no obvious depletion was found in the n–TiO2 layer. This agrees well with the junction theory since the majority electron concentration in the TiO2 is much higher than the majority hole in the perovskite absorber. To exclude the influence of surface defects for a more reliable result, steady-state bias voltages are applied at the cell to produce potential and electric field differences [85]. The KPFM setup for this measurement is depicted in Fig. 7a, where a voltage source is connected to the electrodes of the cell and the nanoscale potential across the cell is probed by the Kelvin probe. When a positive bias voltage is applied at the cell, the electric field strength in the perovskite absorber obviously decreases, in good agreement with the junction theory, as shown in Fig. 7b. Also, almost no electric field change can be observed in the TiO2 and spiro-OMeTAD layers, which further confirms that the depletion region is mainly located in the perovskite region. However, if the perovskite absorber is thick enough, the depletion will also appear in the transparent conductive oxide layer. When a negative bias voltage is applied, more significant depletion can be observed. Less electric field change is observed in TiO2 and spiro-OMeTAD layers, which further tells us that the bias voltage is mainly applied at the perovskite absorber due to its low carrier density and conductivity. Another important result is that no junction forms at the perovskite/spiro-OMeTAD interface although doped spiro-OMeTAD is usually a heavy p-type semiconductor. One reason for this is that the perovskite absorber usually has a hole density of about 1014–1016 cm−3 [86–87]. Therefore, the Fermi energy level of the perovskite absorber is about −5.2 eV [88], much closer to the highest occupied molecular orbital (HOMO) energy level of the spiro-OMeTAD; thus, no obvious junction is found. Another possible reason lies in different charge transfer mechanisms between the organic charge transporting layer (CTL) and the inorganic semiconductor, which still needs more fundamental investigation. Defect properties and relative measurement methods Carrier property is not only determined by shallow doping; deep defects also have a significant influence on the charge dynamics and the performance of the semiconductor device. For example, deep defects can be caused by self-doping under certain film deposition conditions, as shown in Fig. 8a. Besides bulk defects, defects may also exist on the crystal surface and grain boundaries due to uncoordinated atoms and active electronic sites. In experiments, the low-temperature solution process is supposed to easily bring the impurity into the perovskite crystal, which may also cause deep defects. Some characterization methods have been applied to investigate deep defects in the perovskite material, especially in MAPbI3. By replacing the oxygen atom on spiro-OMeTAD with a sulfur atom, which can effectively lower the highest occupied molecular orbital of the molecule, a stronger Pb–S interaction with perovskites occurs, thus leading to efficient hole extraction and surface trap passivation [91]. Figure 8. View largeDownload slide (a) Capacitance as a function of frequency of the PSC with a structure of ITO/PEDOT:PSS/MAPbI3/PCBM/Al; (b) Arrhenius curves derived from the capacitance spectra, giving defect activation energies of 0.24 and 0.66 eV from the curves; (c) thermally stimulated current (TSC) spectra of PSCs with different device architectures; (d) and (e) defect activation energies derived from the TSC spectra. Adapted with permission from [89–90]. Figure 8. View largeDownload slide (a) Capacitance as a function of frequency of the PSC with a structure of ITO/PEDOT:PSS/MAPbI3/PCBM/Al; (b) Arrhenius curves derived from the capacitance spectra, giving defect activation energies of 0.24 and 0.66 eV from the curves; (c) thermally stimulated current (TSC) spectra of PSCs with different device architectures; (d) and (e) defect activation energies derived from the TSC spectra. Adapted with permission from [89–90]. Thermal admittance (capacitance) spectroscopy (TAS) is an effective approach to obtain the defect distribution (i.e. defect density and activation energy) and dynamic information (i.e. capture velocity and escape frequency) in the absorber [92]. A complete device is needed in TAS measurement to ensure that the perturbation voltage can be applied at the absorber. According to the theory for admittance (capacitance) spectra, the distribution of defect density, Nt (Eω), is obtained by the following equation:   \begin{equation*} N_{\rm t}(E_{\omega}) = - \frac{{{V_D}}}{{q{W_D}}}\frac{\omega }{{KT}}\frac{{dC}}{{d\omega }}\end{equation*} where VD and WD are the heterojunction potential barrier and depletion width derived from the Mott–Schottky measurement, respectively; ω and C are the angular frequency and capacitance from capacitance spectra, respectively; K is the Boltzmann constant and T is the absolute temperature. Eω, defined as the energy above the VBM, is obtained as follows:   \begin{equation*} {E_\omega } = KT\ln \frac{{2{\beta _{{p}}}{N_{{V}}}}}{\omega }\end{equation*} where βp is the hole coefficient and NV is the effective density of states in the valence band. To obtain βp and NV, a temperature-dependent capacitance spectrogram is needed. The minimum Eω value is the energy difference between the Fermi energy and the VBM, usually larger than zero. To accurately evaluate the defect distribution, capacitance spectra at different temperatures and Mott–Schottky curves are also needed. Figure 8a gives an example of capacitance spectra (−f × dC/df) at different temperatures [89]. For each curve, a peak corresponds to a certain frequency, and the peaks appear at a lower frequency while at relatively high temperatures, or at higher frequency while at relatively low temperatures. The peak −f × dC/df at higher temperatures is much larger than that at lower temperatures, indicating that two defect energy levels with different densities may coexist in the bandgap. Arrhenius curves are thus derived with these feature frequencies in Fig. 8b. Clearly, two distinct curves exist at different temperature regions, the slopes of which tell us that the perovskite MAPbI3 has two defect energy levels of 0.24 eV (Ea) and 0.66 eV (Eb) above the VBM. Heo et al. observed a defect energy level of about 0.28 eV, generally in agreement with Ea in Fig. 8b, and the density of states (DOS) is on the order of about 1017– 1018 cm−3 eV−1 [93]. Huang and co-workers reported a defect energy level of ∼0.35 eV, and a much lower DOS of ∼1012 cm−3 eV−1 in the interdiffusion-deposited film [92,94]. The calculated defect density distribution is shown in Fig. 8c; the DOS of the defects centered at ∼0.62 eV is on the order of 1016 cm−3 eV−1. Snaith and co-workers used a transient photoluminescence (PL) method to evaluate the defect density of the perovskite film of about 1016–1017 cm−3 eV−1 [95–96]. Current–voltage measurements were used to estimate a defect density of about 1015–1016 cm−3 (∼1016–1017 cm−3 eV−1) [97]. Importantly, the defect density can be significantly decreased by increasing the perovskite crystal quality [98]. Thermally stimulated current spectroscopy (TSC) is also an effective method to evaluate the defect distribution in a high-resistivity semiconductor, II-VI and III-V semiconductor materials [99–100]. By using this method, the current originating from the carriers released from trap defects is measured while increasing the sample temperature at a certain rate. Figure 8d shows the TSC spectra of MAPbI3 perovskite devices with different structures. T1 and T2 TSC peaks at low temperature were assigned to shallow defects in the PCBM/C60 layers; however, the T3 peak was due to the lattice phase transition of MAPbI3. The T4 peak was finally assigned to deep defects in the perovskite. Arrhenius curves for different device structures were finally derived in Fig. 8e. A deep defect energy level of about 508 ± 20 meV with a density of about 1015 cm−3 (∼1016 cm−3 eV−1) was estimated and the trap density was basically consistent with that evaluated from the TAS method; however, the energy positions of the defect and the type (electron or hole) still need further clarification. However, where these defects are from, in bulk, film surface or grain boundaries, is still a question. Ion migration and its influence on doping and defect As discussed above, the carrier and defect properties of the perovskite film are related to the intrinsic doping, while the intrinsic doping is influenced by the localized element distribution inside the film. Recently, it has been found that the element distribution in the perovskite is significantly influenced by the external measurement conditions, such as bias voltage and illumination [101–103]. Ion migration is considered as the origin for element redistribution, which has been a hot research topic on the material and physics properties of the perovskite [104]. Xiao et al. first reported the ion migration and its influence on the photoelectric behaviors of the perovskite device [105]. Theoretical calculation based on DFT suggests the mechanism of ion migration, revealing that the ion migration activation energy is as low as about 0.5 eV [106]. In addition, the ions can easily migrate under an electric field of about 106 V/m, similar to the strength of the built-in electric field. Furthermore, the distribution of MA+ cations in the perovskite film was observed by using spatially resolved photothermal induced resonance (PTIR) microscopy, and the MA+ was found to migrate and accumulate at the perovskite/cathode interface, as shown in Fig. 9a. A distribution of electric potential was also observed in the film, possibly due to a p–n homojunction derived from the ion accumulation induced doping, see Fig. 9b. This implies that the ion migration under electric field will change the localized carriers and electrical properties of the film. Figure 9. View largeDownload slide (a) Photothermal induced resonance (PTIR) microscopy of MA+ group and (b) Kelvin probe contact potential in the perovskite film after electrical poling; (c) simulated current–voltage (I–V) curves of the cells with ion migration and interfacial traps; the cell was applied by different steady-state bias voltages before I–V measurement; (d) schematic diagram of the ion migration and interfacial traps. Adapted with permission from [101,107]. Figure 9. View largeDownload slide (a) Photothermal induced resonance (PTIR) microscopy of MA+ group and (b) Kelvin probe contact potential in the perovskite film after electrical poling; (c) simulated current–voltage (I–V) curves of the cells with ion migration and interfacial traps; the cell was applied by different steady-state bias voltages before I–V measurement; (d) schematic diagram of the ion migration and interfacial traps. Adapted with permission from [101,107]. More importantly, ion migration is supposed to be an origin of the anomalous photoelectric hysteresis in PSCs. Xiao et al. suggested that the doping effect caused by the ion migration is the physics mechanism of the hysteresis. The doping and the homojunction formation may change the net charge distribution inside the film. On the other hand, based on theoretical calculation, Reenen et al. investigated the interfacial charge accumulation is due to ion migration [107]. They suggested that charge accumulation alone would not lead to the hysteresis and the traps at the interface also have to be considered, as shown in Fig. 9c. A steady-state bias voltage (Vstabilized) is applied to accumulate different densities of ions at the interface, and at different Vstabilized, the cells show different I–V characteristics in Fig. 9d. In the meantime, considering the ion migration process, time-dependent I–V curves are simulated. The influence of ion migration on the electric properties of the film was also investigated by our lab-made optoelectro-modulated TPV/TPC system (M-TPV/M-TPC), as shown in Fig. 10a [108]. This measurement system can experimentally probe the charge dynamics while modulating the electric field and light illumination, and thus can provide more valuable information on the physical properties, operating mechanisms and device performance of solar cells or other related photoelectric devices [108–109]. A negative electric field was found in the perovskite film close to the TiO2/perovskite interface when a positive bias voltage was applied. By simulation, this negative electric field can be induced only by additional interfacial doping. This additional interfacial doping derived from ion migration will cause photoelectrical hysteresis, but cannot decrease the open-circuit voltage (Voc). When a constant interfacial trap is also considered, Voc would not be influenced either. Only when both the interfacial doping and trap density are influenced by the ion migration, a negative electric field and Voc will change, thus leading to obvious photoelectric hysteresis, as shown in Fig. 10b–d. It is thus concluded that the ion migration in the perovskite can change both the intrinsic doping and defect properties of the film [109–110]. This will finally influence the electric field and charge dynamic properties of PSCs. However, quantitative relationships between the ion accumulation, the change of doping and defect density have not been established yet. Figure 10. View largeDownload slide Schematic diagrams of (a) our electrically modulated transient photovoltage/photocurrent measurement system (TPV/TPC); (b) electric field profile across the cell when both the interfacial doping and defect are considered, where Ea is the external modulation and Ebi is the junction built-in electric field; (c) transient photovoltage results at different times after the bias voltage at the cell is switched from 0 to 900 mV and (d) the evolution of collected photo-induced charge, peak photovoltage and photovoltage lifetime with time in dark. Adapted with permission from [108–109]. Figure 10. View largeDownload slide Schematic diagrams of (a) our electrically modulated transient photovoltage/photocurrent measurement system (TPV/TPC); (b) electric field profile across the cell when both the interfacial doping and defect are considered, where Ea is the external modulation and Ebi is the junction built-in electric field; (c) transient photovoltage results at different times after the bias voltage at the cell is switched from 0 to 900 mV and (d) the evolution of collected photo-induced charge, peak photovoltage and photovoltage lifetime with time in dark. Adapted with permission from [108–109]. STABILITY Although the high PCE of the PSCs exhibited attractive potential for commercial application, its stability issue makes people a little worried as the PSCs in the NREL efficiency chart are labeled with ‘not stabilized” [6]. Obviously, the stability issue is the most important problem to be solved at the current point. Therefore, research work on the stability of perovskite solar cells as well as relevant influencing factors has been widely investigated. Humidity, electrical field (including applied bias), UV radiation, thermal treatment and even some additives in the hole-transportation material are generally considered as common extrinsic incentives to the degradation of organic–inorganic perovskite materials and PSCs [111–112]. The humidity degradation mechanism of organic–inorganic perovskite materials is supposed to be that perovskite materials can react with water to hydrate intermediate compounds and finally decompose into PbI2 [113]. For example, CH3NH3PbI3 readily reacts with water to monohydrate CH3NH3PbI3·H2O, dihydrate (CH3NH3)4PbI6·2H2O and eventually decomposes into CH3NH3I and PbI2. An effective method to suppress this degradation is to prevent humidity penetration, i.e. some surface passivation materials including hydrophobic materials have been employed to modify the perovskite film, such as crystal crosslinkers, a non-hydrolytic Al2O3 layer, polystyrene and so on [114–116]. PSCs with encapsulation materials (e.g. a Teflon polymer layer, bifunctional polydimethylsiloxane and so on) have also been reported [117–118]. Another method is to use an insulating hygroscopic-like polymer polyethylene glycol (PEG) as the scaffold, which is supposed to work as a moisture absorber around the perovskite crystals, thus restraining the moisture from permeating and reacting with the perovskites [119]. The device based on a hygroscopic PEG scaffold can present an extraordinarily stable behavior at a humidity of 70% for 300 h. Photo-induced degradation of perovskites is mainly supposed to depend on the hybrid perovskite compositions in the crystal structure, which is also related to the light intensity and the temperature [29,120]. In particular, TiO2 is a typical photocatalyst for oxidizing water to afford hydroxyl radicals which can oxidize organic materials as well. Therefore, light irradiation containing a small amount of UV light with the existence of O2 can degrade the meso-TiO2-based perovskites film, i.e. the decomposition of iodide-substituted perovskites into HI [121]. Obviously, interfacial modification between mesoporous TiO2 scaffold and the perovskite, or using other non-UV response electron transport materials, have usually been adopted to improve the photo-induced instability of perovskites [122–123]. Very recently, Shin et al. adopted La–BaSnO3 to replace conventional TiO2, which not only presented a PCE of 21.2% but also retained 93% of its initial performance after 1000 h of full sun illumination [15]. Thermal degradation of organic–inorganic perovskite materials is that the dark-brown perovskite materials gradually turn to yellow PbI2 with increasing thermal treatment temperature [16]. CH3NH3PbI3 was reported to start decomposing at 230°C while its decomposition temperature is 290°C [124]. Of course, the thermal degradation of perovskites is closely related to the moisture and characteristics of perovskite materials [125]. Very recently, with the aid of thermal gravimetric and differential thermal analysis (TG-DTA) coupled with quadrupole mass spectrometry (MS) and first-principles calculations, Juarez-Perez et al. found that the thermal degradation of MAPbI3 and MAI releases methyliodide (CH3I) and ammonia (NH3) as degradation products, instead of the common CH3NH2 and HI [126]. Furthermore, the key factor to the degradation mechanism is also the instability of the organic cation (i.e. MA+ or FA+). Therefore, the most effective way to improve the stability of perovskite is to engineer the compositions of perovskite materials. Mei et al. reported (5-AVA)x(MA)1-xPbI3 perovskite material by partially replacing the MA with 5-aninovaleric (5-AVA) in the cuboctahedral site of typical MAPbI3, and the hole-conductor-free, fully printable mesoporous device based on this material demonstrated excellent long-term stability for more than 1000 h in ambient air under full sunlight illumination [127]. Replacing organic cations with inorganic Cs+ or Rb+ is another effective way to stabilize perovskite solar cells [14,16,128]. Besides that, the layered perovskites or named 2D perovskites coupled with carbon electrodes show better resistance to humidity than their 3D analogues [129–130]. Grancini et al. reported one-year stable perovskite devices by introducing an ultra-stable 2D/3D (HOOC(CH2)4NH3)2PbI4/CH3NH3PbI3 perov-skite junction, which is an exceptional gradually-organized multidimensional interface, thus presenting a PCE of 12.9% with carbon electrodes and 14.6% with normal mesoporous structure [131]. Very recently, the electric stability issue of PSCs has been gradually recognized. As reported, mobile ions, in particular MA+ ions, will drift along the electric field and build up certain stress regions within PSCs, leaving ion vacancies [28]. Leijtens et al. thought that ion motion alone is reversible and would not degrade PSCs, whereas combining moisture, residual solvent and electric bias may reduce the device stability [132]. In detail, electrical bias can initiate a rapid degradation of perovskite materials under humid conditions. Some investigation further revealed that the existence of grain boundaries will provide fast channels for ion transport, where the degradation usually starts [133–134]. Continuous and alternating ion transport in the perovskite will bring about the material degradation and the decay of the cell performance. The key problem lies in the fact that the solar cell has to work under electric fields and light illumination, and the working condition has to switch between day and night. This will definitely induce irreversible ion migration, and then cause stability issues. Although the mechanism of electrical-field-induced PSC performance deterioration is still under research, seeking suitable ways to suppress the ion migration and reduce the grain boundaries is imperative for improving the stability of PSCs. SUMMARY AND PROSPECTS Perovskite materials are already a kind of star material and receive much more attention from the research community, especially in the field of materials science and photovoltaics. So far, perovskite solar cells have already exhibited over 22% efficiency, comparable to CdTe and CIGS solar cells. Perovskite materials are even expected to have the potential to change the current photovoltaics landscape. As we know, the ultimate goal of a photovoltaic technology is to realize its commercialization and wide application. To PSCs, although great progress has been achieved, there is still a long way to go until its practical application in the PV market. To further enhance the cell performance and speed up the commercialization, several issues need to be addressed in the future. First of all, the stability issue is of primary importance. As inorganic–organic hybrid perovskites and devices are susceptible to environmental factors (i.e. moisture, UV light, oxygen, heating), the instability problems of PSCs under outdoor conditions must be resolved. These degradation pathways can be circumvented by appropriate encapsulation, such as using waterproof, airtight coating and UV-protective coating. In addition, the key instabilities are interfacial and electrical-field-induced instability issues, which are strongly dependent on ion migration in the perovskite itself. This ion migration may be intrinsic and cannot be solved by device encapsulation. Further investigation has revealed that ion migration may be responsible for many unusual phenomena in the perovskite and the devices, such as I-V hysteresis, the switchable photovoltaic effect, the giant dielectric constant, photo-induced phase separation, the photo-induced self-poling effect and electrical-field-driven reversible conversion between PbI2 and perovskite [135]. Therefore, a thorough understanding of the ion migration mechanism is crucial at both experimental and theoretical levels and is expected to help us develop appropriate methodologies to control ion migration in the future. Currently, material engineering and interfacial engineering are generally considered to be effective ways to reduce grain boundaries to prevent ion migration and the decomposition of the perovskite film as well as to improve the stability. Some new methodologies and technologies are also imperative in the future. Secondly, much more effort is still needed to deeply understand the photophysics of perovskites and the devices. Currently, by various time-resolved spectroscopy techniques and theoretical calculation, some physical properties of perovskite materials have been investigated, which can further confirm their superiority in the charge transfer/separation characteristics coupled with the suppressed charge recombination, an extremely shallow trap depth (10 meV) with a low trap-state concentration, ultra-long diffusion length and so on. For the whole device, however, when the doping, defects and static charge distribution vary inside the perovskite material, the variety of basic dynamic processes is still unclear, especially when the device is operating. From this perspective, in-depth understanding the intrinsic physics of perovskites and devices is the key to further improve the cell performance and the stability issue. Thirdly, developing large-area stable PSCs and modules is inevitable to realize their commercialization and wide application. Recently, people have attempted realistic ways (especially fabrication techniques) to obtain highly efficient largescale PSCs. The one-step anti-solvent spin-coating method has been proved to be a state-of-the-art technology; however, it is not suitable for largescale devices. Some research groups have developed different perovskite film deposition methods suitable for largescale devices; a certified PCE as high as 19.7% of 1-cm2-sized PSCs has been achieved [35,69,136]. Grancini et al. fabricated 10 × 10 cm2 solar modules by a fully printable industrial-scale process, presenting a PCE of 11.2% with zero loss in cell performance for over 10 000 h measured under controlled standard conditions [131]. As we know, for a largescale PSC, the defect density, the crystal size, film uniformity and the electron collecting resistances are significant challenges. Therefore, several aspects have to be considered, such as the balance between fabrication technique and cell performance, and the relationship between interface properties and device efficiency, of course, still including the hysteresis and stability. Fourthly, highly efficient PSCs are based on lead-based perovskites, all of which contain the toxic Pb element. More importantly, halide perovskites can easily dissolve in water, thus leading to more concern about the toxicity issue when largescale application of PSCs will be on the agenda. Obviously, to facilitate the practical application, the toxicity issue of the lead element cannot be ignored; however, how to handle this problem is still under debate. In fact, developing highly efficient, stable lead-free alternatives may permanently solve this issue. Currently, pure Sn-based PSCs struggle to repeat the high performance of Pb-based devices [57,80]. Therefore, further enhancement in cell performance and stability is necessary, which is expected to come from fundamental research on understanding the electron dynamics and interfacial engineering. Overall, with these mixed halide perovskite materials, higher efficiency is still possible through structural modification along with bandgap tuning. In the meantime, higher Voc is also expected by careful control of the luminescent property of the perovskite, thus leading to higher PCE. Fabricating highly efficient, stable PSCs is a systematic and multidisciplinary project, which requires contributions from semiconductor physicists, organic chemists, material scientists and even engineers. With substantial efforts toward developing hybrid perovskite materials, it is believed that a kind of reliable photovoltaic technology based on perovskites can be realized in the future. FUNDING This work was supported by the National Natural Science Foundation of China (91433205, 51627803, 51402348, 51421002, 21173260 and 11474333) and the Knowledge Innovation Program of the Chinese Academy of Sciences. REFERENCES 1. Green MA, Emery K, Hishikawa Y et al.   Solar cell efficiency tables (version 48). Prog Photovolt Res Appl  2016; 24: 905– 13. Google Scholar CrossRef Search ADS   2. Jackson P, Wuerz R, Hariskos D et al.   Effects of heavy alkali elements in Cu(In, Ga)Se2 solar cells with efficiencies up to 22.6%. Phys Status Solidi Rapid Res Lett  2016; 10: 583– 6. Google Scholar CrossRef Search ADS   3. Grätzel M, Mitzi DB, Sargent EH et al.   Materials interface engineering for solution-processed photovoltaics. Nature  2012; 488: 304– 12. Google Scholar CrossRef Search ADS PubMed  4. Eslamian M. Inorganic and organic solution-processed thin film devices. Nano-Micro Lett  2017; 9: 3. Google Scholar CrossRef Search ADS   5. Razza S, Castro-Hermosa S, Brown TM et al.   Research update: large-area deposition, coating, printing, and processing techniques for the upscaling of perovskite solar cell technology. APL Mater  2016; 4: 091508. Google Scholar CrossRef Search ADS   6. National Renewable Energy Laboratory Service. Photovoltaic Research: Efficiency Chart. https://www.nrel.gov/pv/assets/images/efficiency-chart.png (4 September 2017, date last accessed). 7. Cai M, Wu Y, Han L et al.   Cost-performance analysis of perovskite solar modules. Adv Sci  2017; 4: 1600269. Google Scholar CrossRef Search ADS   8. Saparov B, Mitzi DB. Organic-inorganic perovskites: structural versatility for functional materials design. Chem Rev  2016; 116: 4558– 96. Google Scholar CrossRef Search ADS PubMed  9. Kojima A, Teshima K, Miyasaka T et al.   Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. J Am Chem Soc  2009; 131: 6050– 1. Google Scholar CrossRef Search ADS PubMed  10. Im JH, Lee CR, Park NG et al.   6.5% efficient perovskite quantum-dot-sensitized solar cell. Nanoscale  2011; 3: 4088– 93. Google Scholar CrossRef Search ADS PubMed  11. Kim HS, Gratzel M, Park NG et al.   Lead iodide perovskite sensitized all-solid-state submicron thin film mesoscopic solar cell with efficiency exceeding 9%. Sci Rep  2012; 2: 591. Google Scholar CrossRef Search ADS PubMed  12. Lee MM, Teuscher J, Snaith HJ et al.   Efficient hybrid solar cells based on meso-superstructured organometal halide perovskites. Science  2012; 338: 643– 7. Google Scholar CrossRef Search ADS PubMed  13. Burschka J, Pellet N, Grätzel M et al.   Sequential deposition as a route to high-performance perovskite-sensitized solar cells. Nature  2013; 499: 316– 20. Google Scholar CrossRef Search ADS PubMed  14. Saliba M, Matsui T, Grätzel M et al.   Cesium-containing triple cation perovskite solar cells: improved stability, reproducibility and high efficiency. Energy Environ Sci  2016; 9: 1989– 97. Google Scholar CrossRef Search ADS PubMed  15. Shin SS, Noh JH, Seok SI et al.   Colloidally prepared La-doped BaSnO3 electrodes for efficient, photostable perovskite solar cells. Science  2017; 356: 167– 71. Google Scholar CrossRef Search ADS PubMed  16. Saliba M, Matsui T, Grätzel M et al.   Incorporation of rubidium cations into perovskite solar cells improves photovoltaic performance. Science  2016; 354: 206– 9. Google Scholar CrossRef Search ADS PubMed  17. Tsai H, Nie W, Mohite AD et al.   High-efficiency two-dimensional Ruddlesden-Popper perovskite solar cells. Nature  2016; 536: 312– 7. Google Scholar CrossRef Search ADS PubMed  18. Green MA, Emery K, Hishikawa Y et al.   Solar cell efficiency tables (version 49). Prog Photovolt Res Appl  2017; 25: 3– 13. Google Scholar CrossRef Search ADS   19. Park NG, Grätzel M, Emery K et al.   Towards stable and commercially available perovskite solar cells. Nat Energy  2016; 1: 16152. Google Scholar CrossRef Search ADS   20. Seo J, Noh JH, Seok SI. Rational strategies for efficient perovskite solar cells. Acc Chem Res  2016; 49: 562– 72. Google Scholar CrossRef Search ADS PubMed  21. Meng L, You J, Yang Y et al.   Recent advances in the inverted planar structure of perovskite solar cells. Acc Chem Res  2016; 49: 155– 65. Google Scholar CrossRef Search ADS PubMed  22. Rong Y, Liu L, Han H et al.   Beyond efficiency: the challenge of stability in mesoscopic perovskite solar cells. Adv Energy Mater  2015; 5: 1501066. Google Scholar CrossRef Search ADS   23. Zhang F, Wang S, Grätzel M et al.   Isomer-pure bis-PCBM-assisted crystal engineering of perovskite solar cells showing excellent efficiency and stability. Adv Mater  2017; 29: 1606806. Google Scholar CrossRef Search ADS   24. Shi J, Xu X, Meng Q et al.   Interfaces in perovskite solar cells. Small  2015; 11: 2472– 86. Google Scholar CrossRef Search ADS PubMed  25. Manser JS, Kamat PV. Band filling with free charge carriers in organometal halide perovskites. Nat Photon  2014; 8: 737– 43. Google Scholar CrossRef Search ADS   26. Green MA, Ho-Baillie A, Snaith HJ. The emergence of perovskite solar cells. Nat Photon  2014; 8: 506– 14. Google Scholar CrossRef Search ADS   27. Baikie T, Fang Y, Kadro JM et al.   Synthesis and crystal chemistry of the hybrid perovskite (CH3NH3)PbI3 for solid-state sensitized solar cell applications. J Mater Chem A  2013; 1: 5628– 41. Google Scholar CrossRef Search ADS   28. Stranks SD, Eperson GE, Snaith HJ et al.   Electron-hole diffusion lengths exceeding 1 micrometer in an organometal trihalide perovskite absorber. Science  2013; 342: 341– 4. Google Scholar CrossRef Search ADS PubMed  29. Stoumpos CC, Malliakas CD, Kanatzidis MG. Semiconducting tin and lead iodide perovskites with organic cations: phase transitions, high mobilities, and near-infrared photoluminescent properties. Inorg Chem  2013; 52: 9019– 38. Google Scholar CrossRef Search ADS PubMed  30. Xing G, Mathews N, Sum TC et al.   Long-range balanced electron-and hole-transport lengths in organic-inorganic CH3NH3PbI3. Science  2013; 342: 344– 7. Google Scholar CrossRef Search ADS PubMed  31. Chen Q, Zhou H, Yang Y et al.   Under the spotlight: The organic-inorganic hybrid halide perovskite for optoelectronic applications. Nano Today  2015; 10: 355– 96. Google Scholar CrossRef Search ADS   32. Gao P, Grätzel M, Nazeruddin MK. Organohalide lead perovskites for photovoltaic applications. Energy Environ Sci  2014; 7: 2448– 63. Google Scholar CrossRef Search ADS   33. Xiao M, Cheng YB, Spiccia L et al.   A fast deposition-crystallization procedure for highly efficient lead iodide perovskite thin-film solar cells. Angew Chem Int Ed  2014; 53: 9898– 903. Google Scholar CrossRef Search ADS   34. Yin M, Xie F, Han L et al.   Annealing-free perovskite films by instant crystallization for efficiency solar cells. J Mater Chem A  2016; 4: 8548– 53. Google Scholar CrossRef Search ADS   35. Li X, Bi DQ, Grätzel M et al.   A vacuum flash-assisted solution process for high-efficiency large-area perovskite solar cells. Science  2016; 353: 58– 62. Google Scholar CrossRef Search ADS PubMed  36. Xiao J, Yang Y, Meng Q et al.   Pressure-assisted CH3NH3PbI3 morphology reconstruction to improve the high performance of perovskite solar cells. J Mater Chem A  2015; 3: 5289– 93. Google Scholar CrossRef Search ADS   37. Zhu L, Shi J, Meng Q et al.   Temperature-assisted controlling morphology and charge transport property for highly efficient perovskite solar cells. Nano Energy  2015; 15: 540– 8. Google Scholar CrossRef Search ADS   38. Chen Q, Zhou H, Yang Y et al.   Planar heterojunction perovskite solar cells via vapor assisted solution process. J Am Chem Soc  2014; 136: 622– 5. Google Scholar CrossRef Search ADS PubMed  39. Ono LK, Leyden MR, Qi Y et al.   Organometal halide perovskite thin films and solar cells by vapor deposition. J Mater Chem A  2016; 4: 6693– 713. Google Scholar CrossRef Search ADS   40. Yin J, Qu H, Zheng N et al.   Vapor-assisted crystallization control toward high performance perovskite photovoltaics with over 18% efficiency in the ambient atmosphere. J Mater Chem A  2016; 4: 13203– 10. Google Scholar CrossRef Search ADS   41. Owoade OK, Salau AM, Akande AR et al.   Annealing effect on the structural properties of the alloy of PbCl2 and NaCl thin film. Indian J Pure Appl Phys  2000; 38: 499– 503. 42. Mitzi DB, Prikas MT, Chondroudis K. Thin film deposition of organic-inorganic hybrid materials using a single source thermal ablation technique. Chem Mater  1999; 11: 542– 4. Google Scholar CrossRef Search ADS   43. Liu M, Johnston MB, Snaith HJ. Efficient planar heterojunction perovskite solar cells by vapor deposition. Nature  2013; 501: 395– 8. Google Scholar CrossRef Search ADS PubMed  44. Malinkiewicz O, Yella A, Bolink HJ et al.   Perovskite solar cells employing organic charge-transport layers. Nat Photon  2014; 8: 128– 32. Google Scholar CrossRef Search ADS   45. Leyden MR, Jiang Y, Qi Y. Chemical vapor deposition grown formamidinium perovskite solar modules with high steady state power and thermal stability. J Mater Chem A  2016; 4: 13125– 32. Google Scholar CrossRef Search ADS   46. Jeon NJ, Noh JH, Seok SI et al.   Solvent engineering for high-performance inorganic-organic hybrid perovskite solar cells. Nat Mater  2014; 13: 897– 903. Google Scholar CrossRef Search ADS PubMed  47. Fan R, Huang Y, Zhou H et al.   The progress of interface design in perovskite-based solar cells. Adv Energy Mater  2016; 6: 1600460. Google Scholar CrossRef Search ADS   48. Noh JH, Im SH, Seok SI et al.   Chemical management for colorful, efficient, and stable inorganic-organic hybrid nanostructured solar cells. Nano Lett  2013; 13: 1764– 9. Google Scholar CrossRef Search ADS PubMed  49. Yang B, Keum J, Xiao K et al.   Deciphering halogen competition in organometallic halide perovskite growth. J Am Chem Soc  2016; 138: 5028– 35. Google Scholar CrossRef Search ADS PubMed  50. Chen Y, Gao D, Liang Z et al.   Efficient and reproducible CH3NH3PbI3-x(SCN)x perovskite based planar solar cells. Chem Commun  2015; 51: 11997– 9. Google Scholar CrossRef Search ADS   51. Sun Y, Peng J, Liang Z et al.   Triple-cation mixed-halide perovskites: towards efficient, annealing-free and air-stable solar cells enabled by Pb(SCN)2 additive. Sci Rep  2017; 7: 46193. Google Scholar CrossRef Search ADS PubMed  52. Daub M, Hillebrecht H. Synthesis, single-crystal structure and characterization of (CH3NH3)2Pb(SCN)2I2. Angew Chem Int Ed  2015; 54: 11016– 7. Google Scholar CrossRef Search ADS   53. Xiao Z, Mitzi DB, Yan Y et al.   Photovoltaic properties of two-dimensional (CH3NH3)2Pb(SCN)2I2 perovskite: A combined experimental and density functional theory study. J Phys Chem Lett  2016; 7: 1213– 8. Google Scholar CrossRef Search ADS PubMed  54. Jiang Q, Rebollar D, Xu T et al.   Pseudohalide-induced moisture tolerance in perovskite CH3NH3Pb(SCN)2I thin films. Angew Chem Int Ed  2015; 54: 7617– 20. Google Scholar CrossRef Search ADS   55. Liu J, Shirai Y, Han L et al.   High-quality mixed-organic-cation perovskites from a phase-pure non-stoichiometric intermediate (FAI)1-x-PbI2 for solar cells. Adv Mater  2015; 27: 4918– 23. Google Scholar CrossRef Search ADS PubMed  56. Jain A, Voznyy O, Sargent EH. High-throughput screening of lead-free perovskite-like materials for optoelectronic applications. J Phys Chem C  2017; 121: 7183– 7. Google Scholar CrossRef Search ADS   57. Noel NK, Stranks SD, Snaith HJ et al.   Lead-free organic-inorganic tin halide perovskites for photovoltaic applications. Energy Environ Sci  2014; 7: 3061– 8. Google Scholar CrossRef Search ADS   58. Toshniwal A, Kheraj V. Development of organic-inorganic tin halide perovskites: a review. Sol Energ  2017; 149: 54– 9. Google Scholar CrossRef Search ADS   59. Mossconi E, Umari P, Angelis FD. Electronic and optical properties of mixed Sn-Pb organohalide perovskites: a first principles investigation. J Mater Chem A  2015; 3: 9208– 15. Google Scholar CrossRef Search ADS   60. Hao F, Stoumpos CC, Chang RPH et al.   Anomalous band gap behavior in mixed Sn and Pb perovskites enables broadening of absorption spectrum in solar cells. J Am Chem Soc  2014; 136: 8094– 9. Google Scholar CrossRef Search ADS PubMed  61. Vigneshwaran M, Ohta T, Hayase S et al.   Facile synthesis and characterization of sulfur doped low bandgap bismuth based perovskites by soluble precursor route. Chem Mater  2016; 28: 6436– 40. Google Scholar CrossRef Search ADS   62. Mitzi DB. Organic-inorganic perovskites containing trivalent metal halide layers: the templating influence of the organic cation layer. Inorg Chem  2000; 39: 6107– 13. Google Scholar CrossRef Search ADS PubMed  63. Slavney AH, Hu T, Karunadasa HI et al.   A bismuth-halide double perovskite with long carrier recombination lifetime for photovoltaic applications. J Am Chem Soc  2016; 138: 2138– 41. Google Scholar CrossRef Search ADS PubMed  64. Shi D, Adinolfi V, Bakr OM et al.   Low trap-state density and long carrier diffusion in organolead trihalide perovskite single crystals. Science  2015; 347: 519– 22. Google Scholar CrossRef Search ADS PubMed  65. Fu W, Li H, Chen H et al.   Controlled crystallization of CH3NH3PbI3 films for perovskite solar cells by various PbI2(X) complexes. Sol Energ Mater Sol Cells  2016; 155: 331– 40. Google Scholar CrossRef Search ADS   66. Jo Y, Oh KS, Kim DS et al.   High performance of planar perovskite solar cells produced from PbI2(DMSO) and PbI2(NMP) complexes by intramolecular exchange. Adv Mater  2016; 3: 1500768. 67. Li L, Wang X, Zhou H et al.   The additive coordination effect on hybrid perovskite crystallization and high-performance solar cell. Adv Mater  2016; 28: 9862– 8. Google Scholar CrossRef Search ADS PubMed  68. Lee JW, Kim HS, Park NG. Lewis acid-base adduct approach for high efficiency perovskite solar cells. Acc Chem Res  2016; 49: 311– 9. Google Scholar CrossRef Search ADS PubMed  69. Yang WS, Noh JH, Seok SI et al.   Iodide management in formamidinium-lead-halide-based perovskite layers for efficient solar cells. Science  2017; 356: 1376– 9. Google Scholar CrossRef Search ADS PubMed  70. Liu Z, Chen Q, Zhou H et al.   Chemical reduction of intrinsic defects in thicker heterojunction planar perovskite solar cells. Adv Mater  2016; 29: 1606774. Google Scholar CrossRef Search ADS   71. Yin WJ, Shi T, Yan Y. Unusual defect physics in CH3NH3PbI3 perovskite solar cell absorber. Appl Phys Lett  2014; 104: 063903. Google Scholar CrossRef Search ADS   72. Shi T, Yin WJ, Yan Y. Predictions for p-type CH3NH3PbI3 perovskites. J Phys Chem C  2014; 118: 25350– 4. Google Scholar CrossRef Search ADS   73. Wang Q, Shao Y, Huang J et al.   Qualifying composition dependent p and n self-doping in CH3NH3PbI3. Appl Phys Lett  2014; 105: 163508. Google Scholar CrossRef Search ADS   74. Shi J, Wei H, Meng Q et al.   Control of charge transport in the perovskite CH3NH3PbI3 thin film. ChemPhysChem  2015; 16: 842– 7. Google Scholar CrossRef Search ADS PubMed  75. Hao P, Stoumpos CC, Kanatzidis MG et al.   Lead-free solid-state organic-inorganic halide perovskite solar cells. Nat Photon  2014; 4: 489– 94. Google Scholar CrossRef Search ADS   76. Stoumpos CC, Kanatzidis MG. The renaissance of halide perovskites and their evolution as emerging semiconductors. Acc Chem Res  2015; 48: 2791– 802. Google Scholar CrossRef Search ADS PubMed  77. Oku T, Ohishi Y, Suzuki A. Effects of antimony addition to perovskite-type CH3NH3PbI3 photovoltaic devices. Chem Lett  2016; 45: 134– 6. Google Scholar CrossRef Search ADS   78. Wang ZK, Li M, Yang YG et al.   High efficiency Pb-In binary perovskite solar cells. Adv Mater  2016; 28: 6695– 703. Google Scholar CrossRef Search ADS PubMed  79. Chen Q, Cai J, Chen L et al.   Ag-incorporated organic-inorganic perovskite films and planar heterojunction solar cells. Nano Lett  2017; 17: 3231– 7. Google Scholar CrossRef Search ADS PubMed  80. Abdi-Jalebi M, Dar MI, Sadhanala A et al.   Impact of monovalent cation halide additives on the structural and optoelectronic properties of CH3NH3PbI3 perovskite. Adv Energy Mater  2016; 6: 1502472. Google Scholar CrossRef Search ADS   81. Sze SM, Kwok KN. Physics of Semiconductor Devices , 3 rd edn. New York: Wiley, 2006. Google Scholar CrossRef Search ADS   82. Shi J, Dong J, Meng Q et al.   Hole-conductor-free perovskite organic lead iodide heterojunction thin-film solar cells: High efficiency and junction property. Appl Phys Lett  2014; 104: 063901. Google Scholar CrossRef Search ADS   83. Bergmann VW, Weber SA, Ramos FJ et al.   Real-space observation of unbalanced charge distribution inside a perovskite-sensitized solar cell. Nat Commun  2014; 5: 5001. Google Scholar CrossRef Search ADS PubMed  84. Guerrero A, Juarez-Perez EJ, Bisquert J et al.   Electrical field profile and doping in planar lead halide perovskite solar cells. Appl Phys Lett  2014; 105: 133902. Google Scholar CrossRef Search ADS   85. Jiang CS, Yang M, Zhou Y et al.   Carrier separation and transport in perovskite solar cells studied by nanometer-scale profiling of electrical potential. Nat Commun  2015; 6: 8397. Google Scholar CrossRef Search ADS PubMed  86. Laban WA, Etgar L. Depleted hole conductor-free lead halide iodide heterojunction solar cells. Energy Environ Sci  2013; 6: 3249– 53. Google Scholar CrossRef Search ADS   87. Bi C, Shao Y, Huang J et al.   Understanding the formation and evolution of interdiffusion grown organolead halide perovskite thin films by thermal annealing. J Mater Chem A  2014; 2: 18508– 14. Google Scholar CrossRef Search ADS   88. Shi J, Luo Y, Meng Q et al.   Modified two-step deposition method for high-efficiency TiO2/CH3NH3PbI3 heterojunction solar cells. ACS Appl Mater Interfaces  2014; 6: 9711– 8. Google Scholar CrossRef Search ADS PubMed  89. Samiee M, Konduri S, Ganapathy B et al.   Defect density and dielectric constant in perovskite solar cells. Appl Phys Lett  2014; 105: 153502. Google Scholar CrossRef Search ADS   90. Baumann A, Väth S, Dyakonov V et al.   Identification of trap states in perovskite solar cells. J Phys Chem Lett  2015; 6: 2350– 4. Google Scholar CrossRef Search ADS PubMed  91. Chen H, Jen AKY, Chen H et al.   Molecular engineering hole-extraction materials to enable dopant-free, efficient p-i-n perovskite solar cells. Adv Energy Mater  2017; 7: 1700012. Google Scholar CrossRef Search ADS   92. Xiao Z, Dong Q, Huang J et al.   Solvent annealing of perovskite-induced crystal growth for photovoltaic-device efficiency enhancement. Adv Mater  2014; 26: 6503– 9. Google Scholar CrossRef Search ADS PubMed  93. Heo JH, Song DH, Im SH et al.   Planar CH3NH3PbI3 perovskite solar cells with constant 17.2% average power conversion efficiency irrespective of the scan rate. Adv Mater  2015; 27: 3424– 30. Google Scholar CrossRef Search ADS PubMed  94. Shao Y, Xiao Z, Huang J et al.   Origin and elimination of photocurrent hysteresis by fullerene passivation in CH3NH3PbI3 planar heterojunction solar cells. Nat Commun  2014; 5: 5784. Google Scholar CrossRef Search ADS PubMed  95. Stranks SD, Burlakov VM, Snaith HJ et al.   Recombination kinetics in organic-inorganic perovskites: excitons, free charge, and subgap states. Phys Rev Appl  2014; 2: 034007. Google Scholar CrossRef Search ADS   96. Noel NK, Abate A, Snaith HJ et al.   Enhanced photoluminescence and solar cell performance via Lewis based passivation of organic-inorganic lead halide perovskites. ACS Nano 2014; 8: 9815– 21. 97. Yang D, Yang R, Liu S et al.   Hysteresis-suppressed high-efficiency flexible perovskite solar cells using solid-state ionic-liquid for effective electron transport. Adv Mater  2016; 28: 5206– 13. Google Scholar CrossRef Search ADS PubMed  98. Liu Y, Zhang Y, Liu S et al.   Thinness- and shape-controlled growth for ultrathin single-crystalline perovskite wafers for mass production of superior photoelectronic devices. Adv Mater  2016; 28: 9204– 9. Google Scholar CrossRef Search ADS PubMed  99. Baldini A, Bruzzi M. Thermally stimulated current spectroscopy: Experimental techniques for the investigation of silicon detectors. Rev Sci Instrum  1993; 64: 932– 6. Google Scholar CrossRef Search ADS   100. Micocci G, Rizzo A, Tepore A. Trapping center parameters in indium selenide single crystals by thermally stimulated current measurements. J Appl Phys  1983; 564: 1924– 9. Google Scholar CrossRef Search ADS   101. Yuan Y, Chae J, Huang J et al.   Photovoltaic switching mechanism in lateral structure-hybrid perovskite solar cells. Adv Energy Mater  2015; 5: 1500615. Google Scholar CrossRef Search ADS   102. deQuilettes DW, Zhang W, Strank SD et al.   Photo-induced halide redistribution in organic-inorganic perovskite films. Nat Commun  2016; 7: 11683. Google Scholar CrossRef Search ADS PubMed  103. Deng Y, Xiao Z, Huang J. Light-induced self-poling effect on organometal trihalide perovskite solar cells for increased device efficiency and stability. Adv Energy Mater  2015; 5: 1500721. Google Scholar CrossRef Search ADS   104. Yuan Y, Huang J. Ion migration in organometal trihalide perovskite and its impact on photovoltaic efficiency and stability. Acc Chem Res  2016; 49: 286– 93. Google Scholar CrossRef Search ADS PubMed  105. Xiao Z, Yuan Y, Huang J et al.   Giant switchable photovoltaic effect in organometal trihalide perovskite devices. Nat Mater  2015; 14: 193– 8. Google Scholar CrossRef Search ADS PubMed  106. Eames C, Frost JM, Islam MS et al.   Ionic transport in hybrid lead iodide perovskite solar cells. Nat Commun  2015; 6: 7497. Google Scholar CrossRef Search ADS PubMed  107. Reenen SV, Kemerink M, Snaith HJ, Modeling anomalous hysteresis in perovskite solar cells. J Phys Chem Lett  2015; 6: 3808– 14. Google Scholar CrossRef Search ADS PubMed  108. Shi J, Li D, Meng Q et al.   Opto-electro-modulated transient photovoltage and photocurrent system for investigation of charge transport and recombination in solar cells. Rev Sci Instrum  2016; 87: 123107. Google Scholar CrossRef Search ADS PubMed  109. Shi J, Zhang H, Meng Q et al.   Microscopic charge transport and recombination processes behind the photoelectric hysteresis in perovskite solar cells. Small  2016; 32: 5288– 94. Google Scholar CrossRef Search ADS   110. Shi J, Xu X, Meng Q et al.   Intrinsic slow charge response in the perovskite solar cells: Electron and ion transport. Appl Phys Lett  2015; 107: 163901. Google Scholar CrossRef Search ADS   111. Wang D, Wright M, Elumalai NK et al.   Stability of perovskite solar cells. Sol Energ Mater Sol Cells  2016; 147: 255– 75. Google Scholar CrossRef Search ADS   112. Yue Y, Salim NT, Han L et al.   Enhanced stability of perovskite solar cells through corrosion-free pyridine derivatives in hole-transporting materials. Adv Mater  2016; 28: 10738– 43. Google Scholar CrossRef Search ADS PubMed  113. Frost JM, Butler KT, Walsh A et al.   Atomistic origins of high-performance in hybrid halide perovskite solar cells. Nano Lett  2014; 14: 2584– 90. Google Scholar CrossRef Search ADS PubMed  114. Wang Q, Dong Q, Huang J et al.   Thin insulating tunneling contacts for efficient and water-resistant perovskite solar cells. Adv Mater  2016; 28: 6734– 9. Google Scholar CrossRef Search ADS PubMed  115. Zhao Y, Yu D, Zhao Q et al.   A polymer scaffold for self-healing perovskite solar cells. Nat Commun  2015; 7: 1038. 116. Yue Y, Yang X, Han L et al.   Selective deposition of insulating metal oxide in perovskite solar cells with enhanced device performance. ChemSusChem  2015; 8: 2625– 9. Google Scholar CrossRef Search ADS PubMed  117. Kim IS, Martinson ABF. Stabilizing hybrid perovskites against moisture and temperature via non-hydrolytic atomic layer deposited overlayers. J Mater Chem A  2015; 3: 20092– 6. Google Scholar CrossRef Search ADS   118. Li X, Dar MI, Grätzel M et al.   Improved performance and stability of perovskite solar cells by crystal crosslinking with alkylphosphonic acid ω;-ammonium chlorides. Nat Chem  2015; 7: 703– 11. Google Scholar CrossRef Search ADS PubMed  119. Zhao Y, Wei J, Yu D et al. A polymer scaffold for self-healing perovskite solar cells. Nat Commun  2016; 7: 10228. Google Scholar CrossRef Search ADS PubMed  120. Leijtens T, Hoke ET, Snaith HJ et al.   Stability of metal halide perovskite solar cells. Adv Energy Mater  2015; 5: 1500963. Google Scholar CrossRef Search ADS   121. Ito S, Tanaka S, Manabe K et al.   Effects of surface blocking layer of Sb2S3 on nanocrystalline TiO2 for CH3NH3PbI3 perovskite solar cells. J Phys Chem C  2014; 118: 16995– 7000. Google Scholar CrossRef Search ADS   122. Li W, Li J, Wang L et al.   Effect of cesium chloride modification on the film morphology and UV-induced stability of planar perovskite solar cells. J Mater Chem A  2016; 4: 11688– 95. Google Scholar CrossRef Search ADS   123. Li W, Zhang W, Snaith HJ et al.   Enhanced UV-light stability of planar heterojunction perovskite solar cells with caesium bromide interface modification. Energy Environ Sci  2016; 9: 490– 8. Google Scholar CrossRef Search ADS   124. Dualeh A, Gao P, Grätzel M et al.   Thermal behavior of methylammonium lead-trihalide perovskite photovoltaic light harvesters. Chem Mater  2014; 26: 6160– 4. Google Scholar CrossRef Search ADS   125. Misra RK, Aharon S, Katz EA et al.   Temperature- and component-dependent degradation of perovskite photovoltaic materials under concentrated sunlight. J Phys Chem Lett  2015; 6: 326– 30. Google Scholar CrossRef Search ADS PubMed  126. Juarez-Perez EJ, Hawash Z, Qi Y et al.   Thermal degradation of CH3NH3PbI3 perovskite into NH3 and CH3I gases observed by coupled thermogravimetry-mass spectrometry analyses. Energy Environ Sci  2016; 9: 3406– 10. Google Scholar CrossRef Search ADS   127. Mei A, Li X, Han H et al.   A hole-conductor-free, fully printable mesoscopic perovskite solar cell with high stability. Science  2014; 345: 295– 8. Google Scholar CrossRef Search ADS PubMed  128. Li Z, Yang M, Zhu K et al.   Stabilizing perovskite structures by tuning tolerance factor: formation of formamidinium and cesium lead iodide solid-state alloys. Chem Mater  2016; 28: 284– 92. Google Scholar CrossRef Search ADS   129. Dou L, Wong AB, Yang P et al.   Atomically thin two-dimensional organic-inorganic hybrid perovskites. Science  2015; 349: 1518– 21. Google Scholar CrossRef Search ADS PubMed  130. Smith IC, Hoke ET, Karunadasa HI et al.   A layered hybrid perovskite solar-cell absorber with enhanced moisture stability. Angew Chem Int Ed  2014; 53: 11232– 5. Google Scholar CrossRef Search ADS   131. Grancini G, Roldán-Carmona C, Nazeeruddin MK et al.   One-year stable perovskite solar cells by 2D/3D interface engineering. Nat Commun  2017; 8: 15684. Google Scholar CrossRef Search ADS PubMed  132. Leijtens T, Hoke ET, Petrozza A et al.   Mapping electric field-induced switchable poling and structural degradation in hybrid lead halide perovskite thin films. Adv Energy Mater  2015; 5: 1500962. Google Scholar CrossRef Search ADS   133. Yun JS, Seidel J, Seok SI et al.   Critical role of grain boundaries for ion migration in formamidium and methylammonium lead halide perovksite solar cells. Adv Energy Mater  2016; 6: 1600330. Google Scholar CrossRef Search ADS   134. Shao Y, Fang Y, Huang J et al.   Grain boundary dominated ion migration in polycrystalline organic-inorganic halide perovskite films. Energy Environ Sci  2016; 9: 1752– 9. Google Scholar CrossRef Search ADS   135. Xing J, Wang Q, Huang J et al.  . Ultrafast ion migration in hybrid perovskite polycrystalline thin films under light and suppression in single crystals. Phys Chem Chem Phys  2016; 18: 30184– 90. 136. Wu Y, Yang X, Han L et al.   Perovskite solar cells with 18.21% efficiency and area over 1 cm2 fabricated by heterojunction engineering. Nat Energy  2016; 1: 16148. Google Scholar CrossRef Search ADS   © The Author(s) 2017. Published by Oxford University Press on behalf of China Science Publishing & Media Ltd. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png National Science Review Oxford University Press

# Inorganic–organic halide perovskites for new photovoltaic technology

, Volume Advance Article – Aug 29, 2017
18 pages

/lp/ou_press/inorganic-organic-halide-perovskites-for-new-photovoltaic-technology-zVy9su5Sst
Publisher
Oxford University Press
ISSN
2095-5138
eISSN
2053-714X
D.O.I.
10.1093/nsr/nwx100
Publisher site
See Article on Publisher Site

### Abstract

Abstract Perovskite solar cells (PSCs) have received worldwide attention due to excellent power-to-electricity conversion efficiency (PCE). Currently, a PCE record of as high as 22.1% has been achieved, comparative to those of copper indium gallium selenide and CdTe solar cells. However, there are still some critical issues to solve, especially the stability issue at the current state, in order to promote PSC commercialization. In this respect, various device architectures and the working principle are briefly introduced. Then, key features of inorganic–organic hybrid perovskites and related film deposition methods are discussed. Typically, experimental and theoretical efforts to reveal the physical issues of the device are emphasized as well as the cell stability. Finally, a short summary, challenge and prospect are given. inorganic–organic hybrid perovskite materials, perovskite solar cells, heterojunction, photophysics, stability INTRODUCTION Thin-film solar cells are a promising type of cost-competitive solar power via cost-effective materials and fabrication technology, comparable with commercial silicon-based photovoltaics [1]. The bandgaps of CdTe (1.45 eV) and copper indium gallium selenide (CIGS, 1.04–1.68 eV) semiconductors can well meet the spectrum of sunlight, guaranteeing a power-to-electricity conversion efficiency (PCE) of over 22% for CdTe and CIGS thin-film solar cells; however, vacuum deposition and/or high-temperature processes are still needed [2–3]. Solution-processed technology represents the lowest-cost production method for thin-film solar cells, including spin-coating, blade-coating, spraying, inkjet printing, gravure printing, slot-dye coating and so on [4–5]. Recently, solution-processed perovskite solar cells (PSCs) have attracted worldwide attention due to their extremely high efficiency exceeding 22%, comparative to those of CIGS and CdTe solar cells and far exceeding dye-sensitized solar cells (DSCs) and organic photovoltaics (OPVs) [6]. It is very likely that a PCE of 25% may be achieved in the future [7]. Surprisingly, this progress has been achieved in only a few years along with rapid development of different device architectures, processing techniques and novel materials. The excellent cell performance of PSCs mainly benefits from inorganic–organic hybrid perovskite materials with advantages of ease of fabrication, panchromatic sunlight absorption, high carrier mobility, superior carrier diffusion length, long carrier lifetimes and so on [8]. The first PSC based on CH3NH3PbX3 (X = Cl and Br) as sensitizers was reported by Kojima et al. in 2009, exhibiting 3.8% PCE, which was further improved to 6.54% by Im et al. in 2011; however, a rapid degradation of the cell performance perplexed scientists [9–10]. The breakthrough came in 2012 when Kim et al. fabricated all solid-state mesoscopic heterojunction PSCs by using spiro-OMeTAD, presenting 9.7% PCE and relatively good stability [11]. Since then, some milestones in PCEs have emerged successively, which further stimulated the development of PSCs. In 2012, Lee et al. fabricated a ‘meso-superstructured’ PSC with 10.9% PCE by replacing the mesoporous TiO2 with Al2O3 [12]. The successful application of this meso-structured insulating scaffold demonstrated that the CH3NH3PbI3-xClx perovskite material is extraordinarily effective, and can transport both electrons and holes between cell terminals and function well in a simplified device architecture even without scaffolds. In 2013, Burschka et al. reported a sequential two-step dipping method to fabricate PSCs with 15% and a certified 14.14% PCEs [13]. Subsequently, by the aid of compositional engineering and the ‘anti-solvent one step deposition method’, Saliba and Shin's groups independently achieved over 21% PCE [14–15]. In the meantime, the stability of the PSCs was improved by introducing Cs+ and Rb+ cations, adopting 2D perovskite films or using La–BaSnO3 to replace conventional TiO2 [15–17]. In addition, a certified 19.7% PCE for PSCs with an aperture area exceeding 1 cm2 was also obtained [18]. Although the high efficiencies of PSCs have been frequently refreshed, the perovskite layers are subject to compositional degradation under high humidity and thermal treatment. Recent investigation reveals that the PSC may suffer from a change in the semiconductor electric properties while the cell operates. Therefore, long-term stability becomes a primary concern of the community [19]. On one hand, thermal stability is particularly important for solar cell operation; on the other hand, solar modules have to operate continuously between temperatures of −40 and 85°C as well as high solar irradiance over a long period of time [14]. In this review, we will briefly introduce the device architectures and the working principle. The key features of perovskite materials and film deposition techniques will also be reviewed. Typically, experimental and theoretical efforts to reveal the physical issues of the device are emphasized as well as cell stability. Finally, a short conclusion and prospect will be given. DEVICE ARCHITECTURES The huge success of PSCs not only lies in their high efficiencies but also their diversity in device architectures, all of which are derived from the distinctive optoelectric properties of the perovskite materials. It is also found that the evolution of device structures is also accompanied by the development history of the devices. In this section, we will briefly introduce different PSC structures and the development history of the devices. The first PSC employed a mesoporous TiO2 layer as a scaffold, which supported hybrid perovskite materials as sensitizers and transported photo-generated electrons [9]. Due to the fine carrier transportation ability of inorganic–organic hybrid perovskites, the device architecture has evolved and different PSC structures have emerged [20]. Mesoscopic structure employs a semiconductor oxide scaffold, which is fully infiltrated with the perovskite and can extend over the nanostructure to form a capping layer, so-called ‘mesoporous structure’, see Fig. 1a. A specific mesoporous PSC structure worth mentioning is based on a mesoporous Al2O3 insulating scaffold [12]. PSCs can also be fabricated without a mesoporous scaffold, so-called ‘planar PSCs’ [21]. Considering the position of the electron- or hole-transporting layer in front of the incident light, planar PSCs can be classified into n-i-p and p-i-n (inverted) structures; see Fig. 1b. Regardless of the light irradiation direction, the PSC generally consists of a conductive substrate (fluorine-doped tin oxide (FTO), indium-tin-oxide (ITO) or flexible ITO-polyethylene terephthalate (PET)), an electron-transporting layer (ETL), a perovskite layer, a hole-transporting layer (HTL) and a back electrode (Au or Ag). When PSCs are defined by whether they use an HTL or not, the HTL-free PSC is also an important branch; the HTL-free full-printable PSCs in particular exhibit good stability [22]. At present, mesoscopic systems exhibit a certified PCE of 22.1% on the National Renewable Energy Laboratory's best efficiency chart, slightly superior to planar embodiments with a PCE of 20.8% [6,23]. Figure 1. View largeDownload slide Schemes of basic PSC structures: (a) mesoporous PSCs; (b) planar PSCs with n-i-p and p-i-n cell architectures; (c) and (d) energy-level diagrams (eV) of routine PSC materials. Figure 1. View largeDownload slide Schemes of basic PSC structures: (a) mesoporous PSCs; (b) planar PSCs with n-i-p and p-i-n cell architectures; (c) and (d) energy-level diagrams (eV) of routine PSC materials. PSCs have been proved to be a free carrier-based device, in which the charge separation and transportation processes are more like a heterojunction solar cell [24–25]. To achieve efficient PSCs, the energy levels of these different components need to be well matched (see Fig. 1c, d). In addition, the key materials, fabrication techniques and interface engineering have to be optimized. In the meantime, a better understanding the dynamic processes of the device is also needed. Detailed discussion will be presented in the following sections. PEROVSKITE MATERIALS AND FILM FABRICATION TECHNIQUES Generally, optoelectrical materials with different chemical compositions exhibit different trends in their application. The hybrid organic–inorganic perovskite structures closely bridge organic and inorganic materials through a well-defined molecular assembly. In this section, we will review the basic properties of perovskite materials and the recent efforts in understanding the relationship between the composition, properties and fabrication techniques. Engineering perovskite materials as light absorbers Brief introduction to perovskites The perovskite structure was first recognized as mineral calcium titanate (CaTiO3). Halide perovskite, as a new branch of perovskite materials, exhibits attractive photovoltaic characteristics [26]. So far, all inorganic CsBX3 (B = Sn, Pb; X = Cl, Br, I) and inorganic–organic hybrid perovskites have been successfully used in photovoltaic devices. Stoichiometric inorganic–organic hybrid perovskites, ABX3 (A = organic cation, widely used methylammonium (MA) or formamidinium (FA); B = Pb, Sn; X = Cl, Br and I), are generally used to describe this kind of perovskite, which has a 3D structure with corner-sharing anionic BX6 octahedra with the X atoms located in the corners and the B atom in the middle of the octahedra, and A cations located at interstices surrounded by eight octahedra in the cuboctahedral gap; see Fig. 2a. These hybrid perovskite materials present fascinating characteristics for high-performance photovoltaic devices, i.e. adjustable bandgap (optimal 1.4– 1.5 eV), high absorption coefficient (104–105 cm−1), low exciton binding energy (<50 meV), ambipolar charge transport capability and long carrier diffusion length (∼175 μm) [8,27,28]. In addition, relatively good crystallization and easy film formation derived from a solution-processed method at low temperature make them highly competitive, promising low-cost photovoltaic materials. Figure 2. View largeDownload slide (a) The basic perovskite structure ABX3 and their extended network structure connected by the corner-shared octahedral; (b) basic structures of 2D organic–inorganic perovskite with bilayer intercalated organic molecules; (c) calculated t and μ factors for 12 halide perovskites; EA represents CH3CH2NH3 (Note: redrawn figure). Adapted with permission from [26]. Figure 2. View largeDownload slide (a) The basic perovskite structure ABX3 and their extended network structure connected by the corner-shared octahedral; (b) basic structures of 2D organic–inorganic perovskite with bilayer intercalated organic molecules; (c) calculated t and μ factors for 12 halide perovskites; EA represents CH3CH2NH3 (Note: redrawn figure). Adapted with permission from [26]. Generally speaking, the crystallographic stability and possible structure of perovskite ABX3, even selecting A, B and X ions, have to consider an octahedral factor μ and a tolerance factor t:   $$\mu \ = \frac{{{R_{\rm{B}}}}}{{{R_{\rm{X}}}}}\$$ (1)  $$t\ = \frac{{{R_{\rm{A}}} + {R_{\rm{X}}}}}{{\sqrt 2 \left( {{R_{\rm{B}}} + {R_{\rm{X}}}} \right)}}\$$ (2)where RA, RB and RX represent the ionic radii of A, B and X ions. If t is in the range of 0.89–1.0, the perovskite material usually adopts a cubic structure, as shown in Fig. 2c; however, a lower t value will result in less symmetrical tetragonal or orthorhombic structures. When t is close to unity, the ideal cubic perovskite structure can be obtained. Calculated t and μ factors for a range of halide perovskites are shown in Fig. 2c [26]. Despite these constraints, transitions between such structures while heating are common for any given halide perovskites, such as MAPbI3 (Fig. 3a) and FAPbI3 (Fig. 3b), and the high-temperature phase is generally cubic [29]. In fact, a cation is generally considered to be the charge compensation in the lattice, and does not contribute to the band structure. However, variation of the A cation's size will strongly influence the optical property due to its deformation of the MX64− octahedral network. Namely, smaller A cations (i.e. MA or FA) can maintain a 3D structure, whereas larger A cations (long chain alkyl or aryl ammonium) will result in lower-dimensional layered or confined perovskites, as seen in Fig. 2b. Figure 3. View largeDownload slide Graphical schemes of the structural transformation of (a) the MAPbI3 perovskite and (b) the FAPbI3 perovskite. Adapted with permission from [29]. Figure 3. View largeDownload slide Graphical schemes of the structural transformation of (a) the MAPbI3 perovskite and (b) the FAPbI3 perovskite. Adapted with permission from [29]. The 3D halide perovskites have a high extinction coefficient (α) up to 105 cm−1 [30–31]. For example, the light absorption capability of MAPbI3 is even higher than that of GaAs, which contributes to the best-performing single-junction solar cell, and two orders higher than that of crystal silicon [6]. According to the first-principles calculation, the electronic structure of MAPbI3 is dominated by characteristics of the Pb–I bond, the valence-band maximum of which is characterized as an anti-bonding hybrid state of Pb 6s and I 5p orbitals, while the conduction band minimum is characterized as a non-bonding hybrid state of Pb 6p and I 5p orbitals [32]. The high α is attributed to the high density of states in the conduction band minimum (CBM) of the perovskite, and direct p–p transition may be the origin. Due to the high α and strongly reflective back electrode, a 500-nm-thick perovskite film is enough to capture most incident photons with energies lower than the energy gap (Eg). Perovskite film fabrication techniques The morphology and crystallization of perovskite films will strongly influence the cell performance. In the meantime, high-quality perovskite film is susceptible to film deposition conditions and techniques. For the moment, three main film deposition methods are usually adopted: one-step precursor solution deposition, two-step sequential deposition (including the vapor-assisted solution process) and vapor deposition, as shown in Fig. 4. Figure 4. View largeDownload slide Three general perovskite film fabrication methods: (a) one-step precursor deposition method (CBZ: chlorobenzene); (b) sequential deposition method; (c) dual-source vapor deposition method. Adapted with permission from [33,35,43]. Figure 4. View largeDownload slide Three general perovskite film fabrication methods: (a) one-step precursor deposition method (CBZ: chlorobenzene); (b) sequential deposition method; (c) dual-source vapor deposition method. Adapted with permission from [33,35,43]. One-step precursor solution deposition At the initial development stage of PSCs, the one-step precursor solution deposition method was used most often; it was carried out by directly spin-coating the mixed solution on the substrate, and the precursor solution was obtained by dissolving all the precursor salts in high-boiling-point polar solvent (i.e. dimethylformamide (DMF), dimethylsulfoxide (DMSO), γ-butyrolactone (GBL)), as shown in Fig. 4a [11]. Aiming at uncovered pinhole areas derived from large MAPbI3 grains, Zhao et al. developed a one-step, solvent-induced, fast crystallization method (the so-called ‘anti-solvent method’) involving spin-coating a DMF solution of MAPbI3 followed immediately by exposure to chlorobenzene to induce perovskite crystallization for planar PSCs [33]. So far, by using the anti-solvent method combined with compositional engineering, over 21% PCE has been achieved [14–17,34]. In addition, a simple alternative vacuum flash-assisted deposition method was reported by Li et al., which enabled a certified record PCE of 19.6% for a 1-cm2-sized device and a maximum efficiency of 20.5% without hysteresis [35]. Overall, the anti-solvent one-step method can clearly induce rapid nucleation and slow crystal growth processes during the perovskite film formation, thus providing an effective way to achieve high-quality perovskite crystals. In the meantime, this method is suitable for all PSC structures (planar vs. mesoporous), and it may be one possible way to realize the large-area manufacturing. Two-step deposition method A sequential two-step dipping technique was first reported by Mitzi in 1998, then Burschka et al. in 2013 [13]. In a typical two-step method, PbI2 is firstly spin-coated onto the substrate, followed by exposure to the organic ions either in solution or in vapor to afford the perovskite film (Fig. 4b). This method presented the first 15% PCE with a certified 14.14% PCE [13]. Subsequently, aiming at solving incomplete conversion of PbI2 and uncontrollable crystal size and surface morphology, different post-treatment methods have also been developed; e.g., Zhu et al. and Xiao et al. reported a pressure-assisted morphology reconstruction method and temperature-assisted sequential deposition method to control film morphology and interfacial charge transport property for high-efficiency PSCs, respectively [36–37]. The vapor-assisted solution process (VASP) can be considered as another kind of two-step method, as shown in Fig. 4b [38]. This method is different from the conventional two-step solution process and vacuum deposition, and takes advantage of the kinetic reactivity of methylamine iodide (MAI) and the thermodynamic stability of perovskites during the film growth process and affords well-defined grain structures with sizes up to micron-scale, full surface coverage and smooth surface without pinholes. Currently, various VASP film growth techniques have been developed, such as hybrid chemical vapor deposition (HCVD), hybrid physical chemical vapor deposition (HPCVD), low-pressure chemical vapor deposition (LPCVD), in situ tubular chemical vapor deposition (ITCVD) and modified chemical vapor transport (mCVT) [39]. 18.9% PCE was obtained under ambient atmosphere by a modified VASP method [40]. VASP presents a simple, controllable and versatile method for high-quality perovskite films and high-efficiency PV devices. Vapor deposition method Vapor deposition techniques are widely used in the semiconductor industry for largescale production in optoelectronic applications. Perovskite films derived from the vapor deposition process under vacuum were first reported by Salau et al. [41], then by Mitzi et al. [42]. In 2013, Liu et al. modified this technique as a dual-source vapor deposition method for pinhole-free MAPbI1-xClx perovskite films with hundreds of nanometer-scale crystals for planar PSCs; see Fig. 4c [43]. Malinkiewicz et al. fabricated an inverted p-i-n planar PSC with sublimated MAPbI3 perovskite layer [44]. Recently, the chemical vapor deposition (CVD) method was also reported by precisely controlling the crystallization process [45]. Vapor deposition techniques are suitable for multilayered thin-film structures and a variety of substrates; however, this technique demands high vacuum. Compositional engineering As we know, tailoring the bandgap of the light absorber is important for minimizing the energy loss and improving the cell performance. A material engineering in terms of AMX3 compositional control is necessary; namely, the absorption range and the Eg value of these perovskites can be adjusted by ion substitutions [28,47]. In addition, the Eg value of a semiconductor is related to the lattice size and the atomic interaction according to the tight-binding approximation theory. MAPbI3 can absorb the widest range of visible light, showing an Eg value of about 1.5 eV; however, the Eg value of MAPbBr3 is about 2.3 eV. Noh et al. reported (MAPbI3)1-x(MAPbBr3)x mixed perovskite by material engineering, the bandgap (Eg) of which can be changed from 1.57 to 2.29 eV by the following equation Eg(x) = 1.57 + 0.039 x + 0.33 x2. This mixed perovskite with x ≥ 0.2 exhibits higher stability toward the humidity than those with x < 0.2 [48]. In the chlorine substitution of MAPbI3, the main role of the chloride is not to regulate the bandgap. Instead, the introduction of chloride into MAPbI3 mainly improves the film morphology, preferential orientation, crystallinity, grain size and so on, which can further improve carrier transport across the heterojunction interface, passivate the grain boundary and result in higher cell performance [49]. Partial replacement of halides with the pseudohalide SCN−, MAPbI3-x(SCN)x, has also been reported; this exhibited good humidity-resistance characteristics, thus leading to enhanced moisture stability of the PSCs [50–51]. In fact, based on the investigation of single crystals grown under the same conditions, 2D structured (MA)2Pb(SCN)2I2 is suggested instead [52]. Xiao et al. further investigated this 2D (MA)2Pb(SCN)2I2 by the aid of experimental measurement and DFT calculation, and found that its bandgap is about 2.0 eV, larger than that of MAPbI3; in the meantime, the carriers are confined within the 2D layers [53]. PCEs of 8.3% were reported for mixed (MA)2Pb(SCN)2I2/MAPbI3 devices [54]. When larger-sized FA (HC(NH2)2+) occupies the A site of the perovskite lattice, the Eg value can be reduced to about 1.47 eV, closer to the desired Eg of 1.4 eV. In addition, FAPbI3 has other advantages over MAPbI3, such as a smaller polarization dipole, stronger p-type characteristics and higher thermal stability. For FAPbI3-based PSCs, a higher photocurrent can be obtained but with a lower photovoltage, and a non-perovskite phase is found to restrict the cell performance [55]. Therefore, the Eg needs to be optimized for a balance. (FAPbI3)0.85(MAPbBr3)0.15 mixed perovskite derived from FAPbI3, FAPbBr3, MAPbI3 and MAPbBr3 can afford a complete perovskite phase, in which the PCE was increased to 20.8% with a negligible hysteresis [15]. Very recently, PSCs based on triple and quadruple cation perovskites (CsRbMAFAPb(I/Br)3) have been fabricated; which exhibited a PCE of 21.6% with an average PCE of 20.2% and stabilized 19.0% on a 0.5 cm2 cell. In addition, polymer-coated cells can maintain 95% of their initial performance at 85°C after 500 aging hours under full illumination and maximum power point tracking [14,16]. Obviously, compositional engineering is feasible to fabricate highly efficient PSCs with better stability and reproducibility. Replacing the metal cation is another method of narrowing the bandgap; in the meantime, developing lead-free perovskite photovoltaics is also in line with the environmental regulations. Jain et al. used high-throughput density functional theory calculations to screen lead-free perovskite-like materials and found that about 10 compounds (A = Na, K, Rb, Cs, Cu and Ag; B = Ga, In and Sb; X = Cl, Br and I) have bandgaps in the range of 1.5–2.5 eV and exhibit potential application in lead-free photovoltaic devices [56]. Pure Sn halide perovskite itself only exhibited a PCE of over 6% and poor stability [57–58]. However, according to the first-principles calculation, when Pb atoms are partially substituted by Sn atoms, the Eg decreases instead, probably due to weaker atomic interactions between Sn and halides [59]. Eg can be further reduced to 1.16 eV when mixed Sn/Pb atoms occupy the B site, and a steady decrease in Eg is also verified by theoretical calculation and experimental measurement [60]. Although a higher photocurrent could be obtained by decreasing the Eg with Sn substitution, a much lower device performance was presented instead due to high hole doping or defects, and the existence of Sn atoms also resulted in low stability [57]. Bismuth (Bi3+) can be used as a nontoxic replacement for Pb-based perovskite due to its similar electronic structure to Pb2+. Recently, some efforts have been made to incorporate Bi3+ as the B-site cation in the ABX3 perovskite framework. (MA)3Bi2I9 perovskite possesses a bandgap of 2.1 eV and cannot realize the high cell performance; however, the bandgap was significantly reduced to 1.45 eV after sulfur doping, which was supposed to be more suitable for application in photovoltaics [61]. Some work reported that, in the formation of 2D Bi3+-halide perovskites, Bi3+ has been incorporated into the lattice with vacancy formation on the B site and an organic cation as the template [62]. Slavney et al. reported a double-perovskite structure Cs2AgBiBr6 with an indirect bandgap of 1.95 eV, suitable for a tandem solar cell [63]. So far, compositional engineering with A and X site substitutions has brought about tremendous success. For a single-crystal MAPbI3 perovskite, the absorption edge is red-shifted with Eg lower than 1.5 eV. This may be due to the difference in lattice size and defect properties between the surface and the bulk of perovskite crystals [64]. Currently, the cell performance based on single-crystal perovskites is still unsatisfactory. Crystal growth engineering A high-quality perovskite film is the key to high-efficiency PSCs, and this kind of high-quality perovskite film refers to films with fewer defects and trap states derived from pinholes, crystal defects and grain boundaries. As we know, inorganic–organic halide perovskite films can be easily obtained in mild conditions due to a low lattice energy. Therefore, whatever fabrication method is adopted, it is necessary to effectively control the perovskite growth process. In fact, when the perovskite formation rate is high, it is hard to control crystal growth. The Lewis acid–base adduct method has been recognized as a viable way to regulate perovskite crystal growth. A Lewis acid is defined as an electron-pair acceptor and a base as an electron-pair donor. The Lewis acid–base reaction will lead to either a redox reaction or adduct formation, and adducts are generally a kind of coordination compounds. For inorganic–organic halide perovskites, Pb(II) halides are typical Lewis acids, which can afford adducts with monodentate or bidentate ligands in 1: 1, 1: 2 or 2: 1. Monodentate or bidentate ligands (Lewis bases) usually contain O-donors (e.g. DMSO, N-methyl-2-pyrrolidone (NMP)), S-donors (e.g. thiourea, thioacetamide) and N-donors (e.g. pyridine). Initially, this adduct-induced perovskite film growth method is supposed to strongly retard the rapid crystallization of PbI2, which is based on the strong interaction between Pb2+ and S = O from DMSO. Jeon et al. first prepared the PbI2(DMSO) precursor, which was further applied to fabricate the FAPbI3/MAPbBr3 by direct intramolecular exchange with FAI/MABr, presenting a PCE of 20.2%, as shown in Fig. 5 [46]. Then, the mixture of DMSO and DMF was directly adopted to give an adduct of MAI⋅PbI2⋅DMSO, since PbI2 prefers to coordinating with DMSO relative to DMF due to its stronger Lewis basicity [65]. This exchange does not induce volume expansion or change the film thickness due to the similar molecular volumes of DMSO and FAI, and the perovskite film is extremely uniform and flat as well. Recently, Jo et al. reported that a PbI2(NMP) complex (NMP = N-methyl-2-pyrrolidone) as an intermediate can also convert into high-quality perovskites. Larger perovskite grain size was obtained from the PbI2(NMP) film than from the PbI2(DMSO), which can reduce bulk defects and improve charge carrier mobility, thus leading to a PCE of 19.5% [66–67]. When the FAPbI3 film was fabricated by this adduct method, however, thiourea as an S-donor was preferred rather than DMSO as an O-donor due to its stronger coordination ability, and a 1:1:1 adduct of FAI⋅PbI2⋅thiourea is suggested. The resulting FAPbI3 films showed better cell performance because of faster charge transport and the reduced trap states [68]. In addition, this kind of Lewis acid–base intermediate adducts plays a crucial role in controlling the morphology of perovskites. Crystal growth engineering via Lewis acid–base adducts is expected to afford homogeneous films with large grain sizes since the removal of the Lewis base in the adduct can also reduce perovskite crystallization rate. Figure 5. View largeDownload slide Scheme for perovskite formation via the MAI–PbI2–DMSO intermediate: (i) PbI2 with edge-sharing [PbI6]4− octahedral layers; (ii) the formation of flat MAI–PbI2–DMSO intermediate phase by MAI and DMSO molecules intercalated between the layers while toluene is introduced to the wet film with PbI2, MAI and DMSO; (iii) conversion of MAI–PbI2–DMSO into the perovskite phase with corner-sharing octahedra by removing DMSO molecules while annealing. Adapted with permission from [46]. Figure 5. View largeDownload slide Scheme for perovskite formation via the MAI–PbI2–DMSO intermediate: (i) PbI2 with edge-sharing [PbI6]4− octahedral layers; (ii) the formation of flat MAI–PbI2–DMSO intermediate phase by MAI and DMSO molecules intercalated between the layers while toluene is introduced to the wet film with PbI2, MAI and DMSO; (iii) conversion of MAI–PbI2–DMSO into the perovskite phase with corner-sharing octahedra by removing DMSO molecules while annealing. Adapted with permission from [46]. In order to improve the crystallization quality and reduce defect states of the perovskite films, some additives have also been introduced. Yang et al. introduced additional iodide ions into the organic cation solution, which was subsequently used for the intramolecular exchange process for perovskite films. Their distinctive work lies in the fact that their iodide solution was actually triiodide (I3−) solution, which was obtained by stirring iodine (I2) in isopropanol (IPA) at 80°C and reducing some I2 molecules to I3− ions. Further investigation revealed that I3− decreased deep-level traps in bulk perovskite layers, thus leading to a PCE as high as 22.1% with a certified PCE of 19.7% [69]. However, unlike Yang's work, Liu et al. used a small amount of methylamine solution as an additive to eliminate small I2 and polyiodide colloids in the precursor solution, also leading to reduced defect density and extended carrier diffusion length in the resultant film [70]. Obviously, a thorough understanding a variety of defect states as well as the origin is desirable in the future. PHYSICS IN PEROVSKITE SOLAR CELLS Although a certified PCE of 22.1% is presented in the NREL best efficiency chart [6], some important issues still need to be clarified. In addition to the stability, the discrepancy in reverse and forward scanned current–voltage curves, so-called ‘hysteresis’, is also an obstacle. Therefore, deep understanding the physics involved in PSCs is essential for developing stable and highly efficient devices. Doping, carrier and electric properties Point defects and doping Point defects and self-doping of a multiple-component semiconductor are permanent topics, especially in understanding the physical properties of a photovoltaic material. As a ternary semiconductor, organic halide lead hybrid perovskites can be stable in a certain thermodynamic region [71]. In this region, the composition of halide perovskite materials may deviate from stoichiometry more or less, forming point doping or defects, such as ion vacancies, interstitials and substitutions. The self-doping properties of the perovskite MAPbI3 are shown in Fig. 6. Iodide interstitial (Ii), cation substitution (MAPb), cation vacancies (VMA and VPb) and anti-substitutions (e.g. IMA and IPb) will introduce shallow acceptor doping levels, while deep defects may be caused by the IMA and IPb. On the other hand, the cation interstitial (MAi and Pbi), anion vacancies (VI), cation substitution (PbMA) and anti-substitutions (e.g. MAI and PbI) will introduce shallow donor doping levels, while Pb interstitial and anti-substitution will cause deep defects. Generally speaking, it is still dependent on their formation energies in the thermodynamic region whether these self-doping levels could appear while the perovskite film is growing. The formation energy is mainly related to growth conditions. For example, the Pb vacancy easily appears in the I-rich/Pb-poor condition instead of the I-poor/Pb-rich condition. In experiments, this self-doping may be influenced by the film deposition method, precursor compositions, thermal treatment temperature and crystallization process. Wang et al. found that the doping type of MAPbI3 is strongly dependent on the molar ratio of PbI2/MAI in the one-step deposition process, and more MAI would lead to p-type perovskite [73]. They also found that the interdiffusion method is beneficial to the growth of p-type MAPbI3 film whereas the hole concentration will decrease while prolonging the annealing time. For the two-step method, the hole concentration could be adjusted by the MAI concentration, and higher hole concentration was obtained by using a more diluted MAI solution [74]. For Sn-based perovskites, the exotic self-doping ability of the Sn-based perovskites allows them to adopt electrical properties from semiconductor to metallic, which are dependent on the synthetic procedures [58,75]. In particular, easy oxidization of Sn2+ to Sn4+ will cause materials doped with Sn4+, which behave as p-type semiconductors, however, displaying metal-like conductivity, e.g. the electron and hole mobilities of MASnI3 are estimated to exceed 2000 and 300 cm2/(Vs), respectively [76]. Figure 6. View largeDownload slide (a) Calculated transition energy levels of intrinsic point defects in CH3NH3PbI3, and possible (b) donor and (c) acceptor doping by impurity atom substitution or interstitial. Adapted with permission from [71,72]. Figure 6. View largeDownload slide (a) Calculated transition energy levels of intrinsic point defects in CH3NH3PbI3, and possible (b) donor and (c) acceptor doping by impurity atom substitution or interstitial. Adapted with permission from [71,72]. Figure 7. View largeDownload slide (a) Schematic diagram of the Kelvin probe force microscopy (KPFM) for the measurement of surface contact potential across the cell. (b) KPFM results and derived electric field properties of a planar perovskite solar cell under different steady-state bias voltages. Adapted with permission from [85]. Figure 7. View largeDownload slide (a) Schematic diagram of the Kelvin probe force microscopy (KPFM) for the measurement of surface contact potential across the cell. (b) KPFM results and derived electric field properties of a planar perovskite solar cell under different steady-state bias voltages. Adapted with permission from [85]. Besides self-doping, p- or n-type doping by impurity atoms has also been theoretically studied [77]. Yan et al. found that cation substitutions with higher valent atoms or cation interstitials could give more free electrons to the perovskite system, so-called ‘n doping’ [72]. On the other hand, ion substitutions with lower valent atoms would bring about electron-deficient energy levels, capturing the electrons in the valence band maximum (VBM) and causing p doping, as shown in Fig. 6c. Recently, the doping toward the Pb atom (such as Sb, In, Rb, Cu and Ag) has been reported to enhance the efficiency and stability of the solar cell [24,78–80]. However, little experimental work could directly confirm the doping and carrier properties. Junction and electric field According to the basic principles of semiconductor physics, a junction with a built-in electric field may form when two materials with different Fermi energy levels electrically contact with each other [81]. However, this concept was not widely accepted in the early developing stage of PSCs until more and more real-space nanoscale electric measurements were carried out [82]. Kelvin probe force microscopy (KPFM) is an effective approach to investigate the potential and electric field properties inside the cell [83]. By the aid of KPFM, a potential distribution is found in the perovskite and the TiO2-based mesoscopic PSC, and a large electric field of about 106 V m−1 can be derived. A similar result was obtained in the planar cell with a ∼300-nm-thick depletion region [84]. In these measurements, no obvious depletion was found in the n–TiO2 layer. This agrees well with the junction theory since the majority electron concentration in the TiO2 is much higher than the majority hole in the perovskite absorber. To exclude the influence of surface defects for a more reliable result, steady-state bias voltages are applied at the cell to produce potential and electric field differences [85]. The KPFM setup for this measurement is depicted in Fig. 7a, where a voltage source is connected to the electrodes of the cell and the nanoscale potential across the cell is probed by the Kelvin probe. When a positive bias voltage is applied at the cell, the electric field strength in the perovskite absorber obviously decreases, in good agreement with the junction theory, as shown in Fig. 7b. Also, almost no electric field change can be observed in the TiO2 and spiro-OMeTAD layers, which further confirms that the depletion region is mainly located in the perovskite region. However, if the perovskite absorber is thick enough, the depletion will also appear in the transparent conductive oxide layer. When a negative bias voltage is applied, more significant depletion can be observed. Less electric field change is observed in TiO2 and spiro-OMeTAD layers, which further tells us that the bias voltage is mainly applied at the perovskite absorber due to its low carrier density and conductivity. Another important result is that no junction forms at the perovskite/spiro-OMeTAD interface although doped spiro-OMeTAD is usually a heavy p-type semiconductor. One reason for this is that the perovskite absorber usually has a hole density of about 1014–1016 cm−3 [86–87]. Therefore, the Fermi energy level of the perovskite absorber is about −5.2 eV [88], much closer to the highest occupied molecular orbital (HOMO) energy level of the spiro-OMeTAD; thus, no obvious junction is found. Another possible reason lies in different charge transfer mechanisms between the organic charge transporting layer (CTL) and the inorganic semiconductor, which still needs more fundamental investigation. Defect properties and relative measurement methods Carrier property is not only determined by shallow doping; deep defects also have a significant influence on the charge dynamics and the performance of the semiconductor device. For example, deep defects can be caused by self-doping under certain film deposition conditions, as shown in Fig. 8a. Besides bulk defects, defects may also exist on the crystal surface and grain boundaries due to uncoordinated atoms and active electronic sites. In experiments, the low-temperature solution process is supposed to easily bring the impurity into the perovskite crystal, which may also cause deep defects. Some characterization methods have been applied to investigate deep defects in the perovskite material, especially in MAPbI3. By replacing the oxygen atom on spiro-OMeTAD with a sulfur atom, which can effectively lower the highest occupied molecular orbital of the molecule, a stronger Pb–S interaction with perovskites occurs, thus leading to efficient hole extraction and surface trap passivation [91]. Figure 8. View largeDownload slide (a) Capacitance as a function of frequency of the PSC with a structure of ITO/PEDOT:PSS/MAPbI3/PCBM/Al; (b) Arrhenius curves derived from the capacitance spectra, giving defect activation energies of 0.24 and 0.66 eV from the curves; (c) thermally stimulated current (TSC) spectra of PSCs with different device architectures; (d) and (e) defect activation energies derived from the TSC spectra. Adapted with permission from [89–90]. Figure 8. View largeDownload slide (a) Capacitance as a function of frequency of the PSC with a structure of ITO/PEDOT:PSS/MAPbI3/PCBM/Al; (b) Arrhenius curves derived from the capacitance spectra, giving defect activation energies of 0.24 and 0.66 eV from the curves; (c) thermally stimulated current (TSC) spectra of PSCs with different device architectures; (d) and (e) defect activation energies derived from the TSC spectra. Adapted with permission from [89–90]. Thermal admittance (capacitance) spectroscopy (TAS) is an effective approach to obtain the defect distribution (i.e. defect density and activation energy) and dynamic information (i.e. capture velocity and escape frequency) in the absorber [92]. A complete device is needed in TAS measurement to ensure that the perturbation voltage can be applied at the absorber. According to the theory for admittance (capacitance) spectra, the distribution of defect density, Nt (Eω), is obtained by the following equation:   \begin{equation*} N_{\rm t}(E_{\omega}) = - \frac{{{V_D}}}{{q{W_D}}}\frac{\omega }{{KT}}\frac{{dC}}{{d\omega }}\end{equation*} where VD and WD are the heterojunction potential barrier and depletion width derived from the Mott–Schottky measurement, respectively; ω and C are the angular frequency and capacitance from capacitance spectra, respectively; K is the Boltzmann constant and T is the absolute temperature. Eω, defined as the energy above the VBM, is obtained as follows:   \begin{equation*} {E_\omega } = KT\ln \frac{{2{\beta _{{p}}}{N_{{V}}}}}{\omega }\end{equation*} where βp is the hole coefficient and NV is the effective density of states in the valence band. To obtain βp and NV, a temperature-dependent capacitance spectrogram is needed. The minimum Eω value is the energy difference between the Fermi energy and the VBM, usually larger than zero. To accurately evaluate the defect distribution, capacitance spectra at different temperatures and Mott–Schottky curves are also needed. Figure 8a gives an example of capacitance spectra (−f × dC/df) at different temperatures [89]. For each curve, a peak corresponds to a certain frequency, and the peaks appear at a lower frequency while at relatively high temperatures, or at higher frequency while at relatively low temperatures. The peak −f × dC/df at higher temperatures is much larger than that at lower temperatures, indicating that two defect energy levels with different densities may coexist in the bandgap. Arrhenius curves are thus derived with these feature frequencies in Fig. 8b. Clearly, two distinct curves exist at different temperature regions, the slopes of which tell us that the perovskite MAPbI3 has two defect energy levels of 0.24 eV (Ea) and 0.66 eV (Eb) above the VBM. Heo et al. observed a defect energy level of about 0.28 eV, generally in agreement with Ea in Fig. 8b, and the density of states (DOS) is on the order of about 1017– 1018 cm−3 eV−1 [93]. Huang and co-workers reported a defect energy level of ∼0.35 eV, and a much lower DOS of ∼1012 cm−3 eV−1 in the interdiffusion-deposited film [92,94]. The calculated defect density distribution is shown in Fig. 8c; the DOS of the defects centered at ∼0.62 eV is on the order of 1016 cm−3 eV−1. Snaith and co-workers used a transient photoluminescence (PL) method to evaluate the defect density of the perovskite film of about 1016–1017 cm−3 eV−1 [95–96]. Current–voltage measurements were used to estimate a defect density of about 1015–1016 cm−3 (∼1016–1017 cm−3 eV−1) [97]. Importantly, the defect density can be significantly decreased by increasing the perovskite crystal quality [98]. Thermally stimulated current spectroscopy (TSC) is also an effective method to evaluate the defect distribution in a high-resistivity semiconductor, II-VI and III-V semiconductor materials [99–100]. By using this method, the current originating from the carriers released from trap defects is measured while increasing the sample temperature at a certain rate. Figure 8d shows the TSC spectra of MAPbI3 perovskite devices with different structures. T1 and T2 TSC peaks at low temperature were assigned to shallow defects in the PCBM/C60 layers; however, the T3 peak was due to the lattice phase transition of MAPbI3. The T4 peak was finally assigned to deep defects in the perovskite. Arrhenius curves for different device structures were finally derived in Fig. 8e. A deep defect energy level of about 508 ± 20 meV with a density of about 1015 cm−3 (∼1016 cm−3 eV−1) was estimated and the trap density was basically consistent with that evaluated from the TAS method; however, the energy positions of the defect and the type (electron or hole) still need further clarification. However, where these defects are from, in bulk, film surface or grain boundaries, is still a question. Ion migration and its influence on doping and defect As discussed above, the carrier and defect properties of the perovskite film are related to the intrinsic doping, while the intrinsic doping is influenced by the localized element distribution inside the film. Recently, it has been found that the element distribution in the perovskite is significantly influenced by the external measurement conditions, such as bias voltage and illumination [101–103]. Ion migration is considered as the origin for element redistribution, which has been a hot research topic on the material and physics properties of the perovskite [104]. Xiao et al. first reported the ion migration and its influence on the photoelectric behaviors of the perovskite device [105]. Theoretical calculation based on DFT suggests the mechanism of ion migration, revealing that the ion migration activation energy is as low as about 0.5 eV [106]. In addition, the ions can easily migrate under an electric field of about 106 V/m, similar to the strength of the built-in electric field. Furthermore, the distribution of MA+ cations in the perovskite film was observed by using spatially resolved photothermal induced resonance (PTIR) microscopy, and the MA+ was found to migrate and accumulate at the perovskite/cathode interface, as shown in Fig. 9a. A distribution of electric potential was also observed in the film, possibly due to a p–n homojunction derived from the ion accumulation induced doping, see Fig. 9b. This implies that the ion migration under electric field will change the localized carriers and electrical properties of the film. Figure 9. View largeDownload slide (a) Photothermal induced resonance (PTIR) microscopy of MA+ group and (b) Kelvin probe contact potential in the perovskite film after electrical poling; (c) simulated current–voltage (I–V) curves of the cells with ion migration and interfacial traps; the cell was applied by different steady-state bias voltages before I–V measurement; (d) schematic diagram of the ion migration and interfacial traps. Adapted with permission from [101,107]. Figure 9. View largeDownload slide (a) Photothermal induced resonance (PTIR) microscopy of MA+ group and (b) Kelvin probe contact potential in the perovskite film after electrical poling; (c) simulated current–voltage (I–V) curves of the cells with ion migration and interfacial traps; the cell was applied by different steady-state bias voltages before I–V measurement; (d) schematic diagram of the ion migration and interfacial traps. Adapted with permission from [101,107]. More importantly, ion migration is supposed to be an origin of the anomalous photoelectric hysteresis in PSCs. Xiao et al. suggested that the doping effect caused by the ion migration is the physics mechanism of the hysteresis. The doping and the homojunction formation may change the net charge distribution inside the film. On the other hand, based on theoretical calculation, Reenen et al. investigated the interfacial charge accumulation is due to ion migration [107]. They suggested that charge accumulation alone would not lead to the hysteresis and the traps at the interface also have to be considered, as shown in Fig. 9c. A steady-state bias voltage (Vstabilized) is applied to accumulate different densities of ions at the interface, and at different Vstabilized, the cells show different I–V characteristics in Fig. 9d. In the meantime, considering the ion migration process, time-dependent I–V curves are simulated. The influence of ion migration on the electric properties of the film was also investigated by our lab-made optoelectro-modulated TPV/TPC system (M-TPV/M-TPC), as shown in Fig. 10a [108]. This measurement system can experimentally probe the charge dynamics while modulating the electric field and light illumination, and thus can provide more valuable information on the physical properties, operating mechanisms and device performance of solar cells or other related photoelectric devices [108–109]. A negative electric field was found in the perovskite film close to the TiO2/perovskite interface when a positive bias voltage was applied. By simulation, this negative electric field can be induced only by additional interfacial doping. This additional interfacial doping derived from ion migration will cause photoelectrical hysteresis, but cannot decrease the open-circuit voltage (Voc). When a constant interfacial trap is also considered, Voc would not be influenced either. Only when both the interfacial doping and trap density are influenced by the ion migration, a negative electric field and Voc will change, thus leading to obvious photoelectric hysteresis, as shown in Fig. 10b–d. It is thus concluded that the ion migration in the perovskite can change both the intrinsic doping and defect properties of the film [109–110]. This will finally influence the electric field and charge dynamic properties of PSCs. However, quantitative relationships between the ion accumulation, the change of doping and defect density have not been established yet. Figure 10. View largeDownload slide Schematic diagrams of (a) our electrically modulated transient photovoltage/photocurrent measurement system (TPV/TPC); (b) electric field profile across the cell when both the interfacial doping and defect are considered, where Ea is the external modulation and Ebi is the junction built-in electric field; (c) transient photovoltage results at different times after the bias voltage at the cell is switched from 0 to 900 mV and (d) the evolution of collected photo-induced charge, peak photovoltage and photovoltage lifetime with time in dark. Adapted with permission from [108–109]. Figure 10. View largeDownload slide Schematic diagrams of (a) our electrically modulated transient photovoltage/photocurrent measurement system (TPV/TPC); (b) electric field profile across the cell when both the interfacial doping and defect are considered, where Ea is the external modulation and Ebi is the junction built-in electric field; (c) transient photovoltage results at different times after the bias voltage at the cell is switched from 0 to 900 mV and (d) the evolution of collected photo-induced charge, peak photovoltage and photovoltage lifetime with time in dark. Adapted with permission from [108–109]. STABILITY Although the high PCE of the PSCs exhibited attractive potential for commercial application, its stability issue makes people a little worried as the PSCs in the NREL efficiency chart are labeled with ‘not stabilized” [6]. Obviously, the stability issue is the most important problem to be solved at the current point. Therefore, research work on the stability of perovskite solar cells as well as relevant influencing factors has been widely investigated. Humidity, electrical field (including applied bias), UV radiation, thermal treatment and even some additives in the hole-transportation material are generally considered as common extrinsic incentives to the degradation of organic–inorganic perovskite materials and PSCs [111–112]. The humidity degradation mechanism of organic–inorganic perovskite materials is supposed to be that perovskite materials can react with water to hydrate intermediate compounds and finally decompose into PbI2 [113]. For example, CH3NH3PbI3 readily reacts with water to monohydrate CH3NH3PbI3·H2O, dihydrate (CH3NH3)4PbI6·2H2O and eventually decomposes into CH3NH3I and PbI2. An effective method to suppress this degradation is to prevent humidity penetration, i.e. some surface passivation materials including hydrophobic materials have been employed to modify the perovskite film, such as crystal crosslinkers, a non-hydrolytic Al2O3 layer, polystyrene and so on [114–116]. PSCs with encapsulation materials (e.g. a Teflon polymer layer, bifunctional polydimethylsiloxane and so on) have also been reported [117–118]. Another method is to use an insulating hygroscopic-like polymer polyethylene glycol (PEG) as the scaffold, which is supposed to work as a moisture absorber around the perovskite crystals, thus restraining the moisture from permeating and reacting with the perovskites [119]. The device based on a hygroscopic PEG scaffold can present an extraordinarily stable behavior at a humidity of 70% for 300 h. Photo-induced degradation of perovskites is mainly supposed to depend on the hybrid perovskite compositions in the crystal structure, which is also related to the light intensity and the temperature [29,120]. In particular, TiO2 is a typical photocatalyst for oxidizing water to afford hydroxyl radicals which can oxidize organic materials as well. Therefore, light irradiation containing a small amount of UV light with the existence of O2 can degrade the meso-TiO2-based perovskites film, i.e. the decomposition of iodide-substituted perovskites into HI [121]. Obviously, interfacial modification between mesoporous TiO2 scaffold and the perovskite, or using other non-UV response electron transport materials, have usually been adopted to improve the photo-induced instability of perovskites [122–123]. Very recently, Shin et al. adopted La–BaSnO3 to replace conventional TiO2, which not only presented a PCE of 21.2% but also retained 93% of its initial performance after 1000 h of full sun illumination [15]. Thermal degradation of organic–inorganic perovskite materials is that the dark-brown perovskite materials gradually turn to yellow PbI2 with increasing thermal treatment temperature [16]. CH3NH3PbI3 was reported to start decomposing at 230°C while its decomposition temperature is 290°C [124]. Of course, the thermal degradation of perovskites is closely related to the moisture and characteristics of perovskite materials [125]. Very recently, with the aid of thermal gravimetric and differential thermal analysis (TG-DTA) coupled with quadrupole mass spectrometry (MS) and first-principles calculations, Juarez-Perez et al. found that the thermal degradation of MAPbI3 and MAI releases methyliodide (CH3I) and ammonia (NH3) as degradation products, instead of the common CH3NH2 and HI [126]. Furthermore, the key factor to the degradation mechanism is also the instability of the organic cation (i.e. MA+ or FA+). Therefore, the most effective way to improve the stability of perovskite is to engineer the compositions of perovskite materials. Mei et al. reported (5-AVA)x(MA)1-xPbI3 perovskite material by partially replacing the MA with 5-aninovaleric (5-AVA) in the cuboctahedral site of typical MAPbI3, and the hole-conductor-free, fully printable mesoporous device based on this material demonstrated excellent long-term stability for more than 1000 h in ambient air under full sunlight illumination [127]. Replacing organic cations with inorganic Cs+ or Rb+ is another effective way to stabilize perovskite solar cells [14,16,128]. Besides that, the layered perovskites or named 2D perovskites coupled with carbon electrodes show better resistance to humidity than their 3D analogues [129–130]. Grancini et al. reported one-year stable perovskite devices by introducing an ultra-stable 2D/3D (HOOC(CH2)4NH3)2PbI4/CH3NH3PbI3 perov-skite junction, which is an exceptional gradually-organized multidimensional interface, thus presenting a PCE of 12.9% with carbon electrodes and 14.6% with normal mesoporous structure [131]. Very recently, the electric stability issue of PSCs has been gradually recognized. As reported, mobile ions, in particular MA+ ions, will drift along the electric field and build up certain stress regions within PSCs, leaving ion vacancies [28]. Leijtens et al. thought that ion motion alone is reversible and would not degrade PSCs, whereas combining moisture, residual solvent and electric bias may reduce the device stability [132]. In detail, electrical bias can initiate a rapid degradation of perovskite materials under humid conditions. Some investigation further revealed that the existence of grain boundaries will provide fast channels for ion transport, where the degradation usually starts [133–134]. Continuous and alternating ion transport in the perovskite will bring about the material degradation and the decay of the cell performance. The key problem lies in the fact that the solar cell has to work under electric fields and light illumination, and the working condition has to switch between day and night. This will definitely induce irreversible ion migration, and then cause stability issues. Although the mechanism of electrical-field-induced PSC performance deterioration is still under research, seeking suitable ways to suppress the ion migration and reduce the grain boundaries is imperative for improving the stability of PSCs. SUMMARY AND PROSPECTS Perovskite materials are already a kind of star material and receive much more attention from the research community, especially in the field of materials science and photovoltaics. So far, perovskite solar cells have already exhibited over 22% efficiency, comparable to CdTe and CIGS solar cells. Perovskite materials are even expected to have the potential to change the current photovoltaics landscape. As we know, the ultimate goal of a photovoltaic technology is to realize its commercialization and wide application. To PSCs, although great progress has been achieved, there is still a long way to go until its practical application in the PV market. To further enhance the cell performance and speed up the commercialization, several issues need to be addressed in the future. First of all, the stability issue is of primary importance. As inorganic–organic hybrid perovskites and devices are susceptible to environmental factors (i.e. moisture, UV light, oxygen, heating), the instability problems of PSCs under outdoor conditions must be resolved. These degradation pathways can be circumvented by appropriate encapsulation, such as using waterproof, airtight coating and UV-protective coating. In addition, the key instabilities are interfacial and electrical-field-induced instability issues, which are strongly dependent on ion migration in the perovskite itself. This ion migration may be intrinsic and cannot be solved by device encapsulation. Further investigation has revealed that ion migration may be responsible for many unusual phenomena in the perovskite and the devices, such as I-V hysteresis, the switchable photovoltaic effect, the giant dielectric constant, photo-induced phase separation, the photo-induced self-poling effect and electrical-field-driven reversible conversion between PbI2 and perovskite [135]. Therefore, a thorough understanding of the ion migration mechanism is crucial at both experimental and theoretical levels and is expected to help us develop appropriate methodologies to control ion migration in the future. Currently, material engineering and interfacial engineering are generally considered to be effective ways to reduce grain boundaries to prevent ion migration and the decomposition of the perovskite film as well as to improve the stability. Some new methodologies and technologies are also imperative in the future. Secondly, much more effort is still needed to deeply understand the photophysics of perovskites and the devices. Currently, by various time-resolved spectroscopy techniques and theoretical calculation, some physical properties of perovskite materials have been investigated, which can further confirm their superiority in the charge transfer/separation characteristics coupled with the suppressed charge recombination, an extremely shallow trap depth (10 meV) with a low trap-state concentration, ultra-long diffusion length and so on. For the whole device, however, when the doping, defects and static charge distribution vary inside the perovskite material, the variety of basic dynamic processes is still unclear, especially when the device is operating. From this perspective, in-depth understanding the intrinsic physics of perovskites and devices is the key to further improve the cell performance and the stability issue. Thirdly, developing large-area stable PSCs and modules is inevitable to realize their commercialization and wide application. Recently, people have attempted realistic ways (especially fabrication techniques) to obtain highly efficient largescale PSCs. The one-step anti-solvent spin-coating method has been proved to be a state-of-the-art technology; however, it is not suitable for largescale devices. Some research groups have developed different perovskite film deposition methods suitable for largescale devices; a certified PCE as high as 19.7% of 1-cm2-sized PSCs has been achieved [35,69,136]. Grancini et al. fabricated 10 × 10 cm2 solar modules by a fully printable industrial-scale process, presenting a PCE of 11.2% with zero loss in cell performance for over 10 000 h measured under controlled standard conditions [131]. As we know, for a largescale PSC, the defect density, the crystal size, film uniformity and the electron collecting resistances are significant challenges. Therefore, several aspects have to be considered, such as the balance between fabrication technique and cell performance, and the relationship between interface properties and device efficiency, of course, still including the hysteresis and stability. Fourthly, highly efficient PSCs are based on lead-based perovskites, all of which contain the toxic Pb element. More importantly, halide perovskites can easily dissolve in water, thus leading to more concern about the toxicity issue when largescale application of PSCs will be on the agenda. Obviously, to facilitate the practical application, the toxicity issue of the lead element cannot be ignored; however, how to handle this problem is still under debate. In fact, developing highly efficient, stable lead-free alternatives may permanently solve this issue. Currently, pure Sn-based PSCs struggle to repeat the high performance of Pb-based devices [57,80]. Therefore, further enhancement in cell performance and stability is necessary, which is expected to come from fundamental research on understanding the electron dynamics and interfacial engineering. Overall, with these mixed halide perovskite materials, higher efficiency is still possible through structural modification along with bandgap tuning. In the meantime, higher Voc is also expected by careful control of the luminescent property of the perovskite, thus leading to higher PCE. Fabricating highly efficient, stable PSCs is a systematic and multidisciplinary project, which requires contributions from semiconductor physicists, organic chemists, material scientists and even engineers. With substantial efforts toward developing hybrid perovskite materials, it is believed that a kind of reliable photovoltaic technology based on perovskites can be realized in the future. FUNDING This work was supported by the National Natural Science Foundation of China (91433205, 51627803, 51402348, 51421002, 21173260 and 11474333) and the Knowledge Innovation Program of the Chinese Academy of Sciences. REFERENCES 1. Green MA, Emery K, Hishikawa Y et al.   Solar cell efficiency tables (version 48). Prog Photovolt Res Appl  2016; 24: 905– 13. Google Scholar CrossRef Search ADS   2. Jackson P, Wuerz R, Hariskos D et al.   Effects of heavy alkali elements in Cu(In, Ga)Se2 solar cells with efficiencies up to 22.6%. Phys Status Solidi Rapid Res Lett  2016; 10: 583– 6. Google Scholar CrossRef Search ADS   3. Grätzel M, Mitzi DB, Sargent EH et al.   Materials interface engineering for solution-processed photovoltaics. Nature  2012; 488: 304– 12. Google Scholar CrossRef Search ADS PubMed  4. Eslamian M. Inorganic and organic solution-processed thin film devices. Nano-Micro Lett  2017; 9: 3. Google Scholar CrossRef Search ADS   5. Razza S, Castro-Hermosa S, Brown TM et al.   Research update: large-area deposition, coating, printing, and processing techniques for the upscaling of perovskite solar cell technology. APL Mater  2016; 4: 091508. Google Scholar CrossRef Search ADS   6. National Renewable Energy Laboratory Service. Photovoltaic Research: Efficiency Chart. https://www.nrel.gov/pv/assets/images/efficiency-chart.png (4 September 2017, date last accessed). 7. Cai M, Wu Y, Han L et al.   Cost-performance analysis of perovskite solar modules. Adv Sci  2017; 4: 1600269. Google Scholar CrossRef Search ADS   8. Saparov B, Mitzi DB. Organic-inorganic perovskites: structural versatility for functional materials design. Chem Rev  2016; 116: 4558– 96. Google Scholar CrossRef Search ADS PubMed  9. Kojima A, Teshima K, Miyasaka T et al.   Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. J Am Chem Soc  2009; 131: 6050– 1. Google Scholar CrossRef Search ADS PubMed  10. Im JH, Lee CR, Park NG et al.   6.5% efficient perovskite quantum-dot-sensitized solar cell. Nanoscale  2011; 3: 4088– 93. Google Scholar CrossRef Search ADS PubMed  11. Kim HS, Gratzel M, Park NG et al.   Lead iodide perovskite sensitized all-solid-state submicron thin film mesoscopic solar cell with efficiency exceeding 9%. Sci Rep  2012; 2: 591. Google Scholar CrossRef Search ADS PubMed  12. Lee MM, Teuscher J, Snaith HJ et al.   Efficient hybrid solar cells based on meso-superstructured organometal halide perovskites. Science  2012; 338: 643– 7. Google Scholar CrossRef Search ADS PubMed  13. Burschka J, Pellet N, Grätzel M et al.   Sequential deposition as a route to high-performance perovskite-sensitized solar cells. Nature  2013; 499: 316– 20. Google Scholar CrossRef Search ADS PubMed  14. Saliba M, Matsui T, Grätzel M et al.   Cesium-containing triple cation perovskite solar cells: improved stability, reproducibility and high efficiency. Energy Environ Sci  2016; 9: 1989– 97. Google Scholar CrossRef Search ADS PubMed  15. Shin SS, Noh JH, Seok SI et al.   Colloidally prepared La-doped BaSnO3 electrodes for efficient, photostable perovskite solar cells. Science  2017; 356: 167– 71. Google Scholar CrossRef Search ADS PubMed  16. Saliba M, Matsui T, Grätzel M et al.   Incorporation of rubidium cations into perovskite solar cells improves photovoltaic performance. Science  2016; 354: 206– 9. Google Scholar CrossRef Search ADS PubMed  17. Tsai H, Nie W, Mohite AD et al.   High-efficiency two-dimensional Ruddlesden-Popper perovskite solar cells. Nature  2016; 536: 312– 7. Google Scholar CrossRef Search ADS PubMed  18. Green MA, Emery K, Hishikawa Y et al.   Solar cell efficiency tables (version 49). Prog Photovolt Res Appl  2017; 25: 3– 13. Google Scholar CrossRef Search ADS   19. Park NG, Grätzel M, Emery K et al.   Towards stable and commercially available perovskite solar cells. Nat Energy  2016; 1: 16152. Google Scholar CrossRef Search ADS   20. Seo J, Noh JH, Seok SI. Rational strategies for efficient perovskite solar cells. Acc Chem Res  2016; 49: 562– 72. Google Scholar CrossRef Search ADS PubMed  21. Meng L, You J, Yang Y et al.   Recent advances in the inverted planar structure of perovskite solar cells. Acc Chem Res  2016; 49: 155– 65. Google Scholar CrossRef Search ADS PubMed  22. Rong Y, Liu L, Han H et al.   Beyond efficiency: the challenge of stability in mesoscopic perovskite solar cells. Adv Energy Mater  2015; 5: 1501066. Google Scholar CrossRef Search ADS   23. Zhang F, Wang S, Grätzel M et al.   Isomer-pure bis-PCBM-assisted crystal engineering of perovskite solar cells showing excellent efficiency and stability. Adv Mater  2017; 29: 1606806. Google Scholar CrossRef Search ADS   24. Shi J, Xu X, Meng Q et al.   Interfaces in perovskite solar cells. Small  2015; 11: 2472– 86. Google Scholar CrossRef Search ADS PubMed  25. Manser JS, Kamat PV. Band filling with free charge carriers in organometal halide perovskites. Nat Photon  2014; 8: 737– 43. Google Scholar CrossRef Search ADS   26. Green MA, Ho-Baillie A, Snaith HJ. The emergence of perovskite solar cells. Nat Photon  2014; 8: 506– 14. Google Scholar CrossRef Search ADS   27. Baikie T, Fang Y, Kadro JM et al.   Synthesis and crystal chemistry of the hybrid perovskite (CH3NH3)PbI3 for solid-state sensitized solar cell applications. J Mater Chem A  2013; 1: 5628– 41. Google Scholar CrossRef Search ADS   28. Stranks SD, Eperson GE, Snaith HJ et al.   Electron-hole diffusion lengths exceeding 1 micrometer in an organometal trihalide perovskite absorber. Science  2013; 342: 341– 4. Google Scholar CrossRef Search ADS PubMed  29. Stoumpos CC, Malliakas CD, Kanatzidis MG. Semiconducting tin and lead iodide perovskites with organic cations: phase transitions, high mobilities, and near-infrared photoluminescent properties. Inorg Chem  2013; 52: 9019– 38. Google Scholar CrossRef Search ADS PubMed  30. Xing G, Mathews N, Sum TC et al.   Long-range balanced electron-and hole-transport lengths in organic-inorganic CH3NH3PbI3. Science  2013; 342: 344– 7. Google Scholar CrossRef Search ADS PubMed  31. Chen Q, Zhou H, Yang Y et al.   Under the spotlight: The organic-inorganic hybrid halide perovskite for optoelectronic applications. Nano Today  2015; 10: 355– 96. Google Scholar CrossRef Search ADS   32. Gao P, Grätzel M, Nazeruddin MK. Organohalide lead perovskites for photovoltaic applications. Energy Environ Sci  2014; 7: 2448– 63. Google Scholar CrossRef Search ADS   33. Xiao M, Cheng YB, Spiccia L et al.   A fast deposition-crystallization procedure for highly efficient lead iodide perovskite thin-film solar cells. Angew Chem Int Ed  2014; 53: 9898– 903. Google Scholar CrossRef Search ADS   34. Yin M, Xie F, Han L et al.   Annealing-free perovskite films by instant crystallization for efficiency solar cells. J Mater Chem A  2016; 4: 8548– 53. Google Scholar CrossRef Search ADS   35. Li X, Bi DQ, Grätzel M et al.   A vacuum flash-assisted solution process for high-efficiency large-area perovskite solar cells. Science  2016; 353: 58– 62. Google Scholar CrossRef Search ADS PubMed  36. Xiao J, Yang Y, Meng Q et al.   Pressure-assisted CH3NH3PbI3 morphology reconstruction to improve the high performance of perovskite solar cells. J Mater Chem A  2015; 3: 5289– 93. Google Scholar CrossRef Search ADS   37. Zhu L, Shi J, Meng Q et al.   Temperature-assisted controlling morphology and charge transport property for highly efficient perovskite solar cells. Nano Energy  2015; 15: 540– 8. Google Scholar CrossRef Search ADS   38. Chen Q, Zhou H, Yang Y et al.   Planar heterojunction perovskite solar cells via vapor assisted solution process. J Am Chem Soc  2014; 136: 622– 5. Google Scholar CrossRef Search ADS PubMed  39. Ono LK, Leyden MR, Qi Y et al.   Organometal halide perovskite thin films and solar cells by vapor deposition. J Mater Chem A  2016; 4: 6693– 713. Google Scholar CrossRef Search ADS   40. Yin J, Qu H, Zheng N et al.   Vapor-assisted crystallization control toward high performance perovskite photovoltaics with over 18% efficiency in the ambient atmosphere. J Mater Chem A  2016; 4: 13203– 10. Google Scholar CrossRef Search ADS   41. Owoade OK, Salau AM, Akande AR et al.   Annealing effect on the structural properties of the alloy of PbCl2 and NaCl thin film. Indian J Pure Appl Phys  2000; 38: 499– 503. 42. Mitzi DB, Prikas MT, Chondroudis K. Thin film deposition of organic-inorganic hybrid materials using a single source thermal ablation technique. Chem Mater  1999; 11: 542– 4. Google Scholar CrossRef Search ADS   43. Liu M, Johnston MB, Snaith HJ. Efficient planar heterojunction perovskite solar cells by vapor deposition. Nature  2013; 501: 395– 8. Google Scholar CrossRef Search ADS PubMed  44. Malinkiewicz O, Yella A, Bolink HJ et al.   Perovskite solar cells employing organic charge-transport layers. Nat Photon  2014; 8: 128– 32. Google Scholar CrossRef Search ADS   45. Leyden MR, Jiang Y, Qi Y. Chemical vapor deposition grown formamidinium perovskite solar modules with high steady state power and thermal stability. J Mater Chem A  2016; 4: 13125– 32. Google Scholar CrossRef Search ADS   46. Jeon NJ, Noh JH, Seok SI et al.   Solvent engineering for high-performance inorganic-organic hybrid perovskite solar cells. Nat Mater  2014; 13: 897– 903. Google Scholar CrossRef Search ADS PubMed  47. Fan R, Huang Y, Zhou H et al.   The progress of interface design in perovskite-based solar cells. Adv Energy Mater  2016; 6: 1600460. Google Scholar CrossRef Search ADS   48. Noh JH, Im SH, Seok SI et al.   Chemical management for colorful, efficient, and stable inorganic-organic hybrid nanostructured solar cells. Nano Lett  2013; 13: 1764– 9. Google Scholar CrossRef Search ADS PubMed  49. Yang B, Keum J, Xiao K et al.   Deciphering halogen competition in organometallic halide perovskite growth. J Am Chem Soc  2016; 138: 5028– 35. Google Scholar CrossRef Search ADS PubMed  50. Chen Y, Gao D, Liang Z et al.   Efficient and reproducible CH3NH3PbI3-x(SCN)x perovskite based planar solar cells. Chem Commun  2015; 51: 11997– 9. Google Scholar CrossRef Search ADS   51. Sun Y, Peng J, Liang Z et al.   Triple-cation mixed-halide perovskites: towards efficient, annealing-free and air-stable solar cells enabled by Pb(SCN)2 additive. Sci Rep  2017; 7: 46193. Google Scholar CrossRef Search ADS PubMed  52. Daub M, Hillebrecht H. Synthesis, single-crystal structure and characterization of (CH3NH3)2Pb(SCN)2I2. Angew Chem Int Ed  2015; 54: 11016– 7. Google Scholar CrossRef Search ADS   53. Xiao Z, Mitzi DB, Yan Y et al.   Photovoltaic properties of two-dimensional (CH3NH3)2Pb(SCN)2I2 perovskite: A combined experimental and density functional theory study. J Phys Chem Lett  2016; 7: 1213– 8. Google Scholar CrossRef Search ADS PubMed  54. Jiang Q, Rebollar D, Xu T et al.   Pseudohalide-induced moisture tolerance in perovskite CH3NH3Pb(SCN)2I thin films. Angew Chem Int Ed  2015; 54: 7617– 20. Google Scholar CrossRef Search ADS   55. Liu J, Shirai Y, Han L et al.   High-quality mixed-organic-cation perovskites from a phase-pure non-stoichiometric intermediate (FAI)1-x-PbI2 for solar cells. Adv Mater  2015; 27: 4918– 23. Google Scholar CrossRef Search ADS PubMed  56. Jain A, Voznyy O, Sargent EH. High-throughput screening of lead-free perovskite-like materials for optoelectronic applications. J Phys Chem C  2017; 121: 7183– 7. Google Scholar CrossRef Search ADS   57. Noel NK, Stranks SD, Snaith HJ et al.   Lead-free organic-inorganic tin halide perovskites for photovoltaic applications. Energy Environ Sci  2014; 7: 3061– 8. Google Scholar CrossRef Search ADS   58. Toshniwal A, Kheraj V. Development of organic-inorganic tin halide perovskites: a review. Sol Energ  2017; 149: 54– 9. Google Scholar CrossRef Search ADS   59. Mossconi E, Umari P, Angelis FD. Electronic and optical properties of mixed Sn-Pb organohalide perovskites: a first principles investigation. J Mater Chem A  2015; 3: 9208– 15. Google Scholar CrossRef Search ADS   60. Hao F, Stoumpos CC, Chang RPH et al.   Anomalous band gap behavior in mixed Sn and Pb perovskites enables broadening of absorption spectrum in solar cells. J Am Chem Soc  2014; 136: 8094– 9. Google Scholar CrossRef Search ADS PubMed  61. Vigneshwaran M, Ohta T, Hayase S et al.   Facile synthesis and characterization of sulfur doped low bandgap bismuth based perovskites by soluble precursor route. Chem Mater  2016; 28: 6436– 40. Google Scholar CrossRef Search ADS   62. Mitzi DB. Organic-inorganic perovskites containing trivalent metal halide layers: the templating influence of the organic cation layer. Inorg Chem  2000; 39: 6107– 13. Google Scholar CrossRef Search ADS PubMed  63. Slavney AH, Hu T, Karunadasa HI et al.   A bismuth-halide double perovskite with long carrier recombination lifetime for photovoltaic applications. J Am Chem Soc  2016; 138: 2138– 41. Google Scholar CrossRef Search ADS PubMed  64. Shi D, Adinolfi V, Bakr OM et al.   Low trap-state density and long carrier diffusion in organolead trihalide perovskite single crystals. Science  2015; 347: 519– 22. Google Scholar CrossRef Search ADS PubMed  65. Fu W, Li H, Chen H et al.   Controlled crystallization of CH3NH3PbI3 films for perovskite solar cells by various PbI2(X) complexes. Sol Energ Mater Sol Cells  2016; 155: 331– 40. Google Scholar CrossRef Search ADS   66. Jo Y, Oh KS, Kim DS et al.   High performance of planar perovskite solar cells produced from PbI2(DMSO) and PbI2(NMP) complexes by intramolecular exchange. Adv Mater  2016; 3: 1500768. 67. Li L, Wang X, Zhou H et al.   The additive coordination effect on hybrid perovskite crystallization and high-performance solar cell. Adv Mater  2016; 28: 9862– 8. Google Scholar CrossRef Search ADS PubMed  68. Lee JW, Kim HS, Park NG. Lewis acid-base adduct approach for high efficiency perovskite solar cells. Acc Chem Res  2016; 49: 311– 9. Google Scholar CrossRef Search ADS PubMed  69. Yang WS, Noh JH, Seok SI et al.   Iodide management in formamidinium-lead-halide-based perovskite layers for efficient solar cells. Science  2017; 356: 1376– 9. Google Scholar CrossRef Search ADS PubMed  70. Liu Z, Chen Q, Zhou H et al.   Chemical reduction of intrinsic defects in thicker heterojunction planar perovskite solar cells. Adv Mater  2016; 29: 1606774. Google Scholar CrossRef Search ADS   71. Yin WJ, Shi T, Yan Y. Unusual defect physics in CH3NH3PbI3 perovskite solar cell absorber. Appl Phys Lett  2014; 104: 063903. Google Scholar CrossRef Search ADS   72. Shi T, Yin WJ, Yan Y. Predictions for p-type CH3NH3PbI3 perovskites. J Phys Chem C  2014; 118: 25350– 4. Google Scholar CrossRef Search ADS   73. Wang Q, Shao Y, Huang J et al.   Qualifying composition dependent p and n self-doping in CH3NH3PbI3. Appl Phys Lett  2014; 105: 163508. Google Scholar CrossRef Search ADS   74. Shi J, Wei H, Meng Q et al.   Control of charge transport in the perovskite CH3NH3PbI3 thin film. ChemPhysChem  2015; 16: 842– 7. Google Scholar CrossRef Search ADS PubMed  75. Hao P, Stoumpos CC, Kanatzidis MG et al.   Lead-free solid-state organic-inorganic halide perovskite solar cells. Nat Photon  2014; 4: 489– 94. Google Scholar CrossRef Search ADS   76. Stoumpos CC, Kanatzidis MG. The renaissance of halide perovskites and their evolution as emerging semiconductors. Acc Chem Res  2015; 48: 2791– 802. Google Scholar CrossRef Search ADS PubMed  77. Oku T, Ohishi Y, Suzuki A. Effects of antimony addition to perovskite-type CH3NH3PbI3 photovoltaic devices. Chem Lett  2016; 45: 134– 6. Google Scholar CrossRef Search ADS   78. Wang ZK, Li M, Yang YG et al.   High efficiency Pb-In binary perovskite solar cells. Adv Mater  2016; 28: 6695– 703. Google Scholar CrossRef Search ADS PubMed  79. Chen Q, Cai J, Chen L et al.   Ag-incorporated organic-inorganic perovskite films and planar heterojunction solar cells. Nano Lett  2017; 17: 3231– 7. Google Scholar CrossRef Search ADS PubMed  80. Abdi-Jalebi M, Dar MI, Sadhanala A et al.   Impact of monovalent cation halide additives on the structural and optoelectronic properties of CH3NH3PbI3 perovskite. Adv Energy Mater  2016; 6: 1502472. Google Scholar CrossRef Search ADS   81. Sze SM, Kwok KN. Physics of Semiconductor Devices , 3 rd edn. New York: Wiley, 2006. Google Scholar CrossRef Search ADS   82. Shi J, Dong J, Meng Q et al.   Hole-conductor-free perovskite organic lead iodide heterojunction thin-film solar cells: High efficiency and junction property. Appl Phys Lett  2014; 104: 063901. Google Scholar CrossRef Search ADS   83. Bergmann VW, Weber SA, Ramos FJ et al.   Real-space observation of unbalanced charge distribution inside a perovskite-sensitized solar cell. Nat Commun  2014; 5: 5001. Google Scholar CrossRef Search ADS PubMed  84. Guerrero A, Juarez-Perez EJ, Bisquert J et al.   Electrical field profile and doping in planar lead halide perovskite solar cells. Appl Phys Lett  2014; 105: 133902. Google Scholar CrossRef Search ADS   85. Jiang CS, Yang M, Zhou Y et al.   Carrier separation and transport in perovskite solar cells studied by nanometer-scale profiling of electrical potential. Nat Commun  2015; 6: 8397. Google Scholar CrossRef Search ADS PubMed  86. Laban WA, Etgar L. Depleted hole conductor-free lead halide iodide heterojunction solar cells. Energy Environ Sci  2013; 6: 3249– 53. Google Scholar CrossRef Search ADS   87. Bi C, Shao Y, Huang J et al.   Understanding the formation and evolution of interdiffusion grown organolead halide perovskite thin films by thermal annealing. J Mater Chem A  2014; 2: 18508– 14. Google Scholar CrossRef Search ADS   88. Shi J, Luo Y, Meng Q et al.   Modified two-step deposition method for high-efficiency TiO2/CH3NH3PbI3 heterojunction solar cells. ACS Appl Mater Interfaces  2014; 6: 9711– 8. Google Scholar CrossRef Search ADS PubMed  89. Samiee M, Konduri S, Ganapathy B et al.   Defect density and dielectric constant in perovskite solar cells. Appl Phys Lett  2014; 105: 153502. Google Scholar CrossRef Search ADS   90. Baumann A, Väth S, Dyakonov V et al.   Identification of trap states in perovskite solar cells. J Phys Chem Lett  2015; 6: 2350– 4. Google Scholar CrossRef Search ADS PubMed  91. Chen H, Jen AKY, Chen H et al.   Molecular engineering hole-extraction materials to enable dopant-free, efficient p-i-n perovskite solar cells. Adv Energy Mater  2017; 7: 1700012. Google Scholar CrossRef Search ADS   92. Xiao Z, Dong Q, Huang J et al.   Solvent annealing of perovskite-induced crystal growth for photovoltaic-device efficiency enhancement. Adv Mater  2014; 26: 6503– 9. Google Scholar CrossRef Search ADS PubMed  93. Heo JH, Song DH, Im SH et al.   Planar CH3NH3PbI3 perovskite solar cells with constant 17.2% average power conversion efficiency irrespective of the scan rate. Adv Mater  2015; 27: 3424– 30. Google Scholar CrossRef Search ADS PubMed  94. Shao Y, Xiao Z, Huang J et al.   Origin and elimination of photocurrent hysteresis by fullerene passivation in CH3NH3PbI3 planar heterojunction solar cells. Nat Commun  2014; 5: 5784. Google Scholar CrossRef Search ADS PubMed  95. Stranks SD, Burlakov VM, Snaith HJ et al.   Recombination kinetics in organic-inorganic perovskites: excitons, free charge, and subgap states. Phys Rev Appl  2014; 2: 034007. Google Scholar CrossRef Search ADS   96. Noel NK, Abate A, Snaith HJ et al.   Enhanced photoluminescence and solar cell performance via Lewis based passivation of organic-inorganic lead halide perovskites. ACS Nano 2014; 8: 9815– 21. 97. Yang D, Yang R, Liu S et al.   Hysteresis-suppressed high-efficiency flexible perovskite solar cells using solid-state ionic-liquid for effective electron transport. Adv Mater  2016; 28: 5206– 13. Google Scholar CrossRef Search ADS PubMed  98. Liu Y, Zhang Y, Liu S et al.   Thinness- and shape-controlled growth for ultrathin single-crystalline perovskite wafers for mass production of superior photoelectronic devices. Adv Mater  2016; 28: 9204– 9. Google Scholar CrossRef Search ADS PubMed  99. Baldini A, Bruzzi M. Thermally stimulated current spectroscopy: Experimental techniques for the investigation of silicon detectors. Rev Sci Instrum  1993; 64: 932– 6. Google Scholar CrossRef Search ADS   100. Micocci G, Rizzo A, Tepore A. Trapping center parameters in indium selenide single crystals by thermally stimulated current measurements. J Appl Phys  1983; 564: 1924– 9. Google Scholar CrossRef Search ADS   101. Yuan Y, Chae J, Huang J et al.   Photovoltaic switching mechanism in lateral structure-hybrid perovskite solar cells. Adv Energy Mater  2015; 5: 1500615. Google Scholar CrossRef Search ADS   102. deQuilettes DW, Zhang W, Strank SD et al.   Photo-induced halide redistribution in organic-inorganic perovskite films. Nat Commun  2016; 7: 11683. Google Scholar CrossRef Search ADS PubMed  103. Deng Y, Xiao Z, Huang J. Light-induced self-poling effect on organometal trihalide perovskite solar cells for increased device efficiency and stability. Adv Energy Mater  2015; 5: 1500721. Google Scholar CrossRef Search ADS   104. Yuan Y, Huang J. Ion migration in organometal trihalide perovskite and its impact on photovoltaic efficiency and stability. Acc Chem Res  2016; 49: 286– 93. Google Scholar CrossRef Search ADS PubMed  105. Xiao Z, Yuan Y, Huang J et al.   Giant switchable photovoltaic effect in organometal trihalide perovskite devices. Nat Mater  2015; 14: 193– 8. Google Scholar CrossRef Search ADS PubMed  106. Eames C, Frost JM, Islam MS et al.   Ionic transport in hybrid lead iodide perovskite solar cells. Nat Commun  2015; 6: 7497. Google Scholar CrossRef Search ADS PubMed  107. Reenen SV, Kemerink M, Snaith HJ, Modeling anomalous hysteresis in perovskite solar cells. J Phys Chem Lett  2015; 6: 3808– 14. Google Scholar CrossRef Search ADS PubMed  108. Shi J, Li D, Meng Q et al.   Opto-electro-modulated transient photovoltage and photocurrent system for investigation of charge transport and recombination in solar cells. Rev Sci Instrum  2016; 87: 123107. Google Scholar CrossRef Search ADS PubMed  109. Shi J, Zhang H, Meng Q et al.   Microscopic charge transport and recombination processes behind the photoelectric hysteresis in perovskite solar cells. Small  2016; 32: 5288– 94. Google Scholar CrossRef Search ADS   110. Shi J, Xu X, Meng Q et al.   Intrinsic slow charge response in the perovskite solar cells: Electron and ion transport. Appl Phys Lett  2015; 107: 163901. Google Scholar CrossRef Search ADS   111. Wang D, Wright M, Elumalai NK et al.   Stability of perovskite solar cells. Sol Energ Mater Sol Cells  2016; 147: 255– 75. Google Scholar CrossRef Search ADS   112. Yue Y, Salim NT, Han L et al.   Enhanced stability of perovskite solar cells through corrosion-free pyridine derivatives in hole-transporting materials. Adv Mater  2016; 28: 10738– 43. Google Scholar CrossRef Search ADS PubMed  113. Frost JM, Butler KT, Walsh A et al.   Atomistic origins of high-performance in hybrid halide perovskite solar cells. Nano Lett  2014; 14: 2584– 90. Google Scholar CrossRef Search ADS PubMed  114. Wang Q, Dong Q, Huang J et al.   Thin insulating tunneling contacts for efficient and water-resistant perovskite solar cells. Adv Mater  2016; 28: 6734– 9. Google Scholar CrossRef Search ADS PubMed  115. Zhao Y, Yu D, Zhao Q et al.   A polymer scaffold for self-healing perovskite solar cells. Nat Commun  2015; 7: 1038. 116. Yue Y, Yang X, Han L et al.   Selective deposition of insulating metal oxide in perovskite solar cells with enhanced device performance. ChemSusChem  2015; 8: 2625– 9. Google Scholar CrossRef Search ADS PubMed  117. Kim IS, Martinson ABF. Stabilizing hybrid perovskites against moisture and temperature via non-hydrolytic atomic layer deposited overlayers. J Mater Chem A  2015; 3: 20092– 6. Google Scholar CrossRef Search ADS   118. Li X, Dar MI, Grätzel M et al.   Improved performance and stability of perovskite solar cells by crystal crosslinking with alkylphosphonic acid ω;-ammonium chlorides. Nat Chem  2015; 7: 703– 11. Google Scholar CrossRef Search ADS PubMed  119. Zhao Y, Wei J, Yu D et al. A polymer scaffold for self-healing perovskite solar cells. Nat Commun  2016; 7: 10228. Google Scholar CrossRef Search ADS PubMed  120. Leijtens T, Hoke ET, Snaith HJ et al.   Stability of metal halide perovskite solar cells. Adv Energy Mater  2015; 5: 1500963. Google Scholar CrossRef Search ADS   121. Ito S, Tanaka S, Manabe K et al.   Effects of surface blocking layer of Sb2S3 on nanocrystalline TiO2 for CH3NH3PbI3 perovskite solar cells. J Phys Chem C  2014; 118: 16995– 7000. Google Scholar CrossRef Search ADS   122. Li W, Li J, Wang L et al.   Effect of cesium chloride modification on the film morphology and UV-induced stability of planar perovskite solar cells. J Mater Chem A  2016; 4: 11688– 95. Google Scholar CrossRef Search ADS   123. Li W, Zhang W, Snaith HJ et al.   Enhanced UV-light stability of planar heterojunction perovskite solar cells with caesium bromide interface modification. Energy Environ Sci  2016; 9: 490– 8. Google Scholar CrossRef Search ADS   124. Dualeh A, Gao P, Grätzel M et al.   Thermal behavior of methylammonium lead-trihalide perovskite photovoltaic light harvesters. Chem Mater  2014; 26: 6160– 4. Google Scholar CrossRef Search ADS   125. Misra RK, Aharon S, Katz EA et al.   Temperature- and component-dependent degradation of perovskite photovoltaic materials under concentrated sunlight. J Phys Chem Lett  2015; 6: 326– 30. Google Scholar CrossRef Search ADS PubMed  126. Juarez-Perez EJ, Hawash Z, Qi Y et al.   Thermal degradation of CH3NH3PbI3 perovskite into NH3 and CH3I gases observed by coupled thermogravimetry-mass spectrometry analyses. Energy Environ Sci  2016; 9: 3406– 10. Google Scholar CrossRef Search ADS   127. Mei A, Li X, Han H et al.   A hole-conductor-free, fully printable mesoscopic perovskite solar cell with high stability. Science  2014; 345: 295– 8. Google Scholar CrossRef Search ADS PubMed  128. Li Z, Yang M, Zhu K et al.   Stabilizing perovskite structures by tuning tolerance factor: formation of formamidinium and cesium lead iodide solid-state alloys. Chem Mater  2016; 28: 284– 92. Google Scholar CrossRef Search ADS   129. Dou L, Wong AB, Yang P et al.   Atomically thin two-dimensional organic-inorganic hybrid perovskites. Science  2015; 349: 1518– 21. Google Scholar CrossRef Search ADS PubMed  130. Smith IC, Hoke ET, Karunadasa HI et al.   A layered hybrid perovskite solar-cell absorber with enhanced moisture stability. Angew Chem Int Ed  2014; 53: 11232– 5. Google Scholar CrossRef Search ADS   131. Grancini G, Roldán-Carmona C, Nazeeruddin MK et al.   One-year stable perovskite solar cells by 2D/3D interface engineering. Nat Commun  2017; 8: 15684. Google Scholar CrossRef Search ADS PubMed  132. Leijtens T, Hoke ET, Petrozza A et al.   Mapping electric field-induced switchable poling and structural degradation in hybrid lead halide perovskite thin films. Adv Energy Mater  2015; 5: 1500962. Google Scholar CrossRef Search ADS   133. Yun JS, Seidel J, Seok SI et al.   Critical role of grain boundaries for ion migration in formamidium and methylammonium lead halide perovksite solar cells. Adv Energy Mater  2016; 6: 1600330. Google Scholar CrossRef Search ADS   134. Shao Y, Fang Y, Huang J et al.   Grain boundary dominated ion migration in polycrystalline organic-inorganic halide perovskite films. Energy Environ Sci  2016; 9: 1752– 9. Google Scholar CrossRef Search ADS   135. Xing J, Wang Q, Huang J et al.  . Ultrafast ion migration in hybrid perovskite polycrystalline thin films under light and suppression in single crystals. Phys Chem Chem Phys  2016; 18: 30184– 90. 136. Wu Y, Yang X, Han L et al.   Perovskite solar cells with 18.21% efficiency and area over 1 cm2 fabricated by heterojunction engineering. Nat Energy  2016; 1: 16148. Google Scholar CrossRef Search ADS   © The Author(s) 2017. Published by Oxford University Press on behalf of China Science Publishing & Media Ltd. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

### Journal

National Science ReviewOxford University Press

Published: Aug 29, 2017

## You’re reading a free preview. Subscribe to read the entire article.

### DeepDyve is your personal research library

It’s your single place to instantly
that matters to you.

over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month ### Explore the DeepDyve Library ### Search Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly ### Organize Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place. ### Access Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals. ### Your journals are on DeepDyve Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more. All the latest content is available, no embargo periods. DeepDyve ### Freelancer DeepDyve ### Pro Price FREE$49/month
\$360/year

Save searches from
PubMed

Create lists to

Export lists, citations