Effect of Cs on the Stability and Photovoltaic Performance of 2D/3D Perovskite-Based Solar Cells

One of the attractive features of hybrid perovskite is the possibility to reduce its dimensionality, which enhances the perovskite’s resistivity to moisture. In this work, we used 2D/3D perovskites to study different organic molecular spacers (aromatic ring vs cyclic ring); Cs was introduced as an additional small cation to methylammonium. It was found that Cs improves the photovoltaic performance; however, it reduces the cells’ stability because two cations having a different ionic radius are mixed, which creates strains in the perovskite structure. The aromatic ring spacers display better stability in complete cells than does the cyclo spacer. Importantly, Cs has a greater effect on the stability than does the nature of the spacer molecule. The difference in the size of the organic cations as well as the inorganic cations plays a major role in the perovskite’s stability in a film and in a complete solar cell. Three-dimensional (3D) perovskite-based solar cells achieved PCEs of around 22%; however, the moisture and long-term use remain a concern for large-scale device manufacturing. Perovskite structures with the general chemical formula of AMX3 represent a class of materials with a cubic unit cell. Here A is a monovalent cation that could be organic, as in the case of methylammonium (MA), and inorganic, such as cesium; M is a divalent metallic compound, usually Pb, and X is a monovalent halide anion such as I−, Br−, or Cl−. In recent years, low-dimensional perovskite has been embedded in several solar cell structures including planar TiO2 and mesoporous TiO2. 2,3 Low-dimensional perovskite solar cells exhibit better resistance to degradation processes because of moisture over their three-dimensional counterparts. The two-dimensional perovskite is fabricated by utilizing a large organic cation with a large ionic radius that does not fit into the 3D perovskite structure and thus creates a layered perovskite structure where the large cations act as a spacer in the inorganic lead halide octahedral framework. The inorganic framework is composed of a corner-sharing octahedron [MX6] 4− that is derived from the parent 3D perovskite. The organic cations are attracted by ionic and hydrogen bonds to the inorganic framework, and as a result, the inorganic layers are confined and stacked together. In such a layered structure, a quantum well is formed, and the dielectric constant of the organic barrier molecules is lower than the dielectric constant of the inorganic lead halide octahedral framework. Thus, the spacer molecule acts as the barrier and the inorganic framework acts as the well. The dielectric and quantum confinement in the case of low n values will result in higher band gap energies. Tsai et al. used n-butylammonium (BUA) as a spacer in a perovskite structure of the formula (BUA)2(MA)3Pb4I13 (n = 4). The low-dimensional perovskite functioned as the light harvester in a planar architecture, achieving a high PCE of 12.51%. Improved performance, compared with earlier studies, was achieved by using the hot-casting one-step deposition technique, which forms well-oriented perovskite layers in which the inorganic framework is vertical with respect to the substrate, allowing pathways for better charge transport through the perovskite film toward the contacts. The two-dimensional (2D) perovskite solar cells (PSCs) exhibited enhanced stability over their 3D counterparts. Zhang et al. fabricated the same cell structure using the same spacer but with the addition of 5% Cs, achieving 13.68% efficiency. Grancini et al. utilized ammonium valeric acid iodide (AVAI) as a cation spacer and fabricated a combination of 2D/ 3D (AVAI)2PbI4/ MAPbI3 perovskites using fully printable industrial scale processes to fabricate a solar module with a carbon-based architecture. The solar module achieved a power Received: November 30, 2017 Accepted: January 5, 2018 Published: January 5, 2018 Leter Cite This: ACS Energy Lett. 2018, 3, 366−372 © XXXX American Chemical Society 366 DOI: 10.1021/acsenergylett.7b01196 ACS Energy Lett. 2018, 3, 366−372 conversation efficiency (PCE) of 11.2% and exhibited stable PV performance under 1 sun illumination for 1 year. Recently utilized 2D/3D perovskites, based on a mixed cation, formamidinium (FA) with Cs, and mixed halides (iodide and bromide) using the BUA cation as a spacer, were used in a solar cell. The stability of these cells was measured under constant illumination, and they were exposed to 40−50% humidity; the results show slightly enhanced stability for the 2D/3D-based cells compared with the 3D-based cells. Motivated by previous studies that showed better stability for the 2D, 2D/3D, and quasi-2D perovskites over the 3D perovskite while maintaining high PV performance, we studied their effect on the stability and PV performance when using 2D/3D perovskite with different barrier molecules (aromatic vs cyclic) and mixed cations, i.e., Cs and methylammonium (MA). MA is known to be one of the driving forces for decomposing the perovskite because of its sublimation and its reaction with oxygen, which can create a free oxygen radical, which in turn can react with the perovskite and/or with the hole transporting material (HTM), thus accelerating the decomposition process. The 2D/3D perovskite synthesized in this study is related to the formula (R-NH3)2(A)n−1PbnI3n+1, where A represents CH3NH3 + (MA) or, in the mixed cation configuration, the ratio between MA and Cs (i.e., MA1−xCsx) and R-NH3 represent the barrier molecules, which were used in this study, as shown in Figure 1: phenethylammonium iodide, C8H12NI (PEA); cyclohexylammonium iodide, C6H14NI (CHMA); and benzylammonium iodide, C7H10NI (BA). The 2D and 2D/3D perovskites were synthesized by the addition of MAI, PbI2, a barrier molecule, and Cs in appropriate stoichiometric quantities in order to achieve the desired n value. Detailed description of the sample and solar cell preparation is presented in the Supporting Information (SI). Figure 2A presents X-ray diffraction (XRD) measurements of the synthesized 2D perovskite films (n = 1) using different barriers. The XRD results confirm the formation of pure 2D perovskite by the absence of peaks that are ascribed to 3D tetragonal perovskite at angles of 14.2°, which correspond to a crystallographic plane (110), and at angles of 28.5°, which correspond to a crystallographic plane (220). The dominate reflections are related to the (00l) planes, suggesting the preferential c-axis growth of the crystals. The presence of higher-order peaks suggests that the films are highly oriented and have excellent crystallinity. The d spacings of the different perovskites are 16.3, 16.8, and 14.4 Å for PEA, CHMA, and BA, respectively. The difference in the d spacing between the benzene derivative molecules, namely, PEA and BA, is related to the additional methylene group that may influence the orientation and length between adjacent inorganic lead halide octahedral frameworks. However, the difference in d spacing between BA and CHMA is due to the aromatic ring (in BA) and the nonaromatic ring (in CHMA), which influence the C− C bond length. To elucidate the optical properties of 2D perovskite (n = 1), absorbance was measured, as shown in Figure 2B. The absorbance spectra are slightly shifted between the barriers, where BA is red-shifted compared with CHMA and PEA; these results are in good agreement with previous reports. As discussed previously, the band gap energy of low n values increased owing to quantum and dielectric confinement. By introducing different organic cations, the Pb−I−Pb angle deviated, owing to various orientations and hydrogen bonding, which cause inand out-of-plane distortions of the inorganic framework. As a result, the band gap energies are slightly shifted between the barrier molecules. In addition to 2D perovskite, where n = 1, 2D/3D perovskite (where n = 40) was synthesized in order to incorporate it into photovoltaic cells. The absorbance of the 2D/3D perovskite films displays band gap energies similar to those of the 3D perovskite (Figure 3A,B). At low n values, the band gap energy increases; however, at higher n values, there is no confinement due to the increasing thickness of the inorganic framework. Furthermore, the absorbance spectra do not exhibit excitonic Figure 1. Schematic structure of the different barriers (R-NH3): phenethylammonium iodide (PEA), benzylammonium iodide (BA), and cyclohexylammonium iodide (CHMA). Figure 2. (A) XRD patterns of the 2D layered perovskite structure n = 1 of the different barrier molecules. (B) Absorbance spectra of the corresponding barrier molecules. ACS Energy Letters Letter DOI: 10.1021/acsenergylett.7b01196 ACS Energy Lett. 2018, 3, 366−372 367 behavior, suggesting the presence of free carriers, which is opposite to the case of low n values, where the exciton binding energy is higher than kT, which results in transport properties similar to those of the 2D/3D perovskite and the 3D perovskite. Photovoltaic Performance. Following the physical and optical characterization, we incorporated the 2D/3D (n = 40) perovskite with the different barrier molecules into the solar cell structure presented in Figure 4A. A high-resolution scanning electron microscopy (HR-SEM) cross section of the studied solar cells is shown in Figure 4B. The solar cell consists of FTO glass, compact and mesoporous TiO2 layers with thicknesses of 50 and 350 nm, respectively, followed by perovskite as the light harvester, with a thickness of ca. 400 nm, and HTM (spiro-OMeTAD) on top of the perovskite. Finally, gold contacts were evaporated with a thickness of 70 nm. Previous work carried out in our lab showed that at low n values the perovskite transport properties are not good enough to deliver high PV performance, whereas in the case of high n values, i.e., 2D/3D perovskite, the PV performance can be enhanced compared with the 3D perovskite. Therefore, in this work, a 2D/3D perovskite structure, corresponding to high n values, was used in the solar cell. All solar cells were fabricated with stoichiometry corresponding to n = 40, i.e., 2D/3D perovskite. The PV performance of the PSCs is summarized in Tables 1 and 2, and the current−voltage (J−V) curves of the bestperforming cells are presented in Figure 5A,B. The 2D/3D PSCs without Cs and with Cs (i.e., mixed cations, Cs, and MA) have a PV performance similar to their 3D counterparts. The Cs concentration used in these cells is 10%, which gives the best PV performance; other PV results for different Cs concentrations can be found in Table 1S in the Supporting Information. The addition of Cs to the perovskite structure enhances the PV performance significantly, both with opencircuit voltage (Voc) and short-circuit current (Jsc), resulting in higher PCEs of these cells compared with the case without Cs. The best PV performance for the 2D/3D system was observed with 10% Cs and PEA as the barrier molecule. The cell achieved 14.33% efficiency with a Voc value of 0.92 V, a Jsc value of 25.5 mA/cm, and a fill factor of 61%. The external Figure 3. (A) Absorbance spectra of the 2D/3D (n = 40) perovskite structure of the different barrier molecules and of the 3D perovskite. (B) Absorbance spectra of 2D/3D (n = 40) perovskite with the addition of 10% Cs, using the different barrier molecules, and of the 3D mixed with the addition of 10% Cs. Figure 4. (A) Solar cell used in this study. (B) HR-SEM cross section of the solar cell with the 2D/3D (n = 40) perovskite, where the CHMA is the barrier molecule. Table 1. Average Photovoltaic Results and the Best Photovoltaic Results (in Brackets) of the Different Barrier Molecules and the 3D Perovskite barrier molecule Voc (V) Jsc (mA/cm ) fill factor (%) efficiency (%) hysteresis loss (%) PEA 0.86 ± 0.04 20.8 ± 1.8 56 ± 3 9.9 ± 0.6 12.15 (0.85) (23.1) (56.3) (11.06) BA 0.85 ± 0.04 20.6 ± 1.8 57 ± 5 10.0 ± 1.0 11.01 (0.90) (22.7) (57.8) (11.89) CHMA 0.86 ± 0.06 20.4 ± 3.5 59 ± 4 10.2 ± 0.9 5.24 (0.86) (21.6) (61.8) (11.54) 3D perovskite 0.90 ± 0.06 20.0 ± 2.0 60 ± 7 10.7 ± 1.2 14.3 (0.94) (22.1) (56.8) (11.86) ACS Energy Letters Letter DOI: 10.1021/acsenergylett.7b01196 ACS Energy Lett. 2018, 3, 366−372 368 quantum efficiency (EQE) spectra are presented in Figure 5C,D; the EQE shape is in good agreement with the absorbance spectra, and the trend of the integrated Jsc matches the trends measured under AM1.5G standard irradiation. The differences between the integrated Jsc (results shown in Tables 2S and 3S) and the Jsc value, measured by the solar simulator, are due to the reason that during the solar simulator measurements the cells operate at a maximum power point, whereas in the case of the EQE measurements, the cells operate at a point lower than the maximum power point. In order to quantify the hysteresis of these cells, we developed eq 1, which takes the area below the J−V curves and subtracts the forward curve scan from the reverse curve scan. The hysteresis loss calculated using eq 1 is shown in Table 1. In general, the hysteresis loss from those cells based on a mixed cation (with Cs) is higher compared with the MA-based cells (without Cs). The proposed mechanism for the hysteresis in PSCs is mainly due to trap-assisted recombination at the grain boundaries, which is amplified by ion migration toward the grain boundaries. As discussed, the perovskite structure of mixed cation (MA and Cs) could be more distorted than that Table 2. Average Photovoltaic Results and the Best Photovoltaic Results (in Brackets) of the Different Barrier Molecules and 3D Perovskite Using 10% Cs barrier molecule Voc (V) Jsc (mA/cm ) fill factor (%) efficiency (%) hysteresis loss (%) PEA + 10% Cs 0.90 ± 0.02 24.0 ± 1.0 59 ± 2 12.6 ± 0.9 24.49 (0.92) (25.5) (61.0) (14.33) BA + 10% Cs 0.86 ± 0.04 21.4 ± 1.8 55 ± 3 10.2 ± 0.6 22.72 (0.93) (22.4) (52.5) (11.00) CHMA + 10% Cs 0.86 ± 0.01 21.9 ± 1.5 56 ± 2 10.6 ± 1.3 11.9 (0.90) (23.6) (58.9) (12.58) 3D + 10% Cs 0.87 ± 0.08 23.9 ± 0.7 54 ± 4 11.3 ± 1.9 21.33 (0.99) (24.9) (59.0) (14.61) Figure 5. (A) Current density−voltage (J−V) curves of the best-preforming solar cells based on the 2D/3D perovskite (n = 40) using the different barriers and the 3D perovskite. (B) Current density−voltage (J−V) curves of the best-preforming solar cells based on the 2D/3D mixed cation perovskite (n = 40) using the different barriers and of 3D mixed cation perovskite. (C) EQE of the best-preforming solar cells based on the 2D/3D perovskite (n = 40) using the different barriers and the 3D perovskite. (D) EQE of the best-preforming solar cells based on the 2D/3D mixed cation perovskite (n = 40) using the different barriers and the 3D mixed cation perovskite. ACS Energy Letters Letter DOI: 10.1021/acsenergylett.7b01196 ACS Energy Lett. 2018, 3, 366−372 369 without incorporating Cs; therefore, films based on these perovskite structures could potentially have higher defect densities and, therefore, more pronounced hysteresis. Furthermore, because the barrier molecule is shorter, the hysteresis loss is smaller, e.g., BA and CHMA barriers. The reason for that could be due to charge accumulation, which is influenced by the barrier molecule. As the barrier molecule becomes longer, the charge accumulation in the perovskite layer becomes higher, which results in more pronounced hysteresis.