Apr 24, 2018 - Rubidium as an Alternative Cation for Efficient Perovskite Light-. Emitting Diodes. Anil Kanwat, Eric Moyen, Sinyoung Cho, and Jin Jang*.
Research Article Cite This: ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
www.acsami.org
Rubidium as an Alternative Cation for Efficient Perovskite LightEmitting Diodes Anil Kanwat, Eric Moyen, Sinyoung Cho, and Jin Jang* Advanced Display Research Center (ADRC), Department of Information Display, Kyung Hee University, Dongdaemoon-ku, Seoul 130-701, Korea S Supporting Information *
ABSTRACT: Incorporation of rubidium (Rb) into mixed lead halide perovskites has recently achieved record power conversion efficiency and excellent stability in perovskite solar cells. Inspired by these tremendous advances in photovoltaics, this study demonstrates the impact of Rb incorporation into MAPbBr3-based light emitters. Rb partially substitutes MA (methyl ammonium), resulting in a mixed cation perovskite with the formula MA(1−x)RbxPbBr3. Pure MAPbBr3 crystallizes into a polycrystalline layer with highly defective sub-micrometer grains. However, the addition of a small amount of Rb forms MA(1−x)RbxPbBr3 nanocrystals (10 nm) embedded in an amorphous matrix of MA/Rb Br. These nanocrystals grow into defect-free sub-micrometer-sized crystallites with further addition of Rb, resulting in a 3-fold increase in exciton lifetime when the molar ratio of MABr/RbBr is 1:1. A thin film fabricated with a 1:1 molar ratio of MABr/ RbBr showed the best electroluminescent properties with a current efficiency (CE) of 9.45 cd/A and a luminance of 7694 cd/m2. These values of CE and luminance are, respectively, 19 and 10 times larger than those achieved by pure MAPbBr3 devices (0.5 cd/A and 790 cd/m2). We believe this work provides important information on the future compositional optimization of Rb+based mixed cation perovskites for obtaining high-performance light-emitting diodes. KEYWORDS: light-emitting diode, mixed halide cations, nanocrystals, perovskite, polycrystalline
■
of Pb+ ions at grain boundaries, thus reducing the density of nonradiative recombination centers. Perovskite thin films obtained by the NCP method exhibit a poor surface coverage, resulting in a high leakage current when integrated into PeLED devices. An excess of MABr (MABr/PbBr2, 3:1) can improve the surface coverage. The excess of MABr forms a matrix which passivates 12 nm-sized MAPbBr3 NCs and reduces the leakage current.23 Such thin films enabled to obtain a current efficiency (CE) as high as 34.46 Cd A−1 when integrated into a LED. These thin films are as thick as 800 nm, which could be the reason PeLEDs have lower brightness. Alternatively, methyl ammonium (MA) can be substituted by organo-ammonium halides such as butyl ammonium bromide (BABr) and phenethylammonium bromide to constrain the growth of NCs during the film formation.24−26 Such films are 70 nm (BABr) and 50 nm (PEABr) thick with a high surface coverage and a very low surface roughness. NCs can be as small as 10 nm (BABr) or 6.4 nm (PEABr), and EQEs are as high as 9.3 (BaBr) or 7.0% (PEABr). However, nanocrystalline thin films tend to exhibit a high density of surface defects by increasing the surface/volume ratio, which limits the PLQY and therefore the performance of the PeLEDs. Owing to the sensitivity of perovskite, defects
INTRODUCTION Organic−inorganic metal halide perovskites (OMHPs) or allinorganic metal halide perovskites with the structure ABX3 [where A and B are cations and X is an anion (a halide or oxide)] represent a new and disruptive technology in the field of optoelectronics.1−6 Perovskites exhibit excellent optoelectronic properties such as high optical absorption coefficients, long diffusion lengths for charge carriers, high photoluminescence quantum yields (PLQYs), and easy band gap tunability.7−15 These properties make them ideal candidates for solar cells and light-emitting diode (LED) applications. OMHPs have revolutionized the technology hole-only devices (HODs) dye-sensitized solar cells in 2009 with a power conversion efficiency (PCE) of only 3.8% to solid-state solar cells in 2016 with a certified PCE of 22.1%.16 Perovskites for electroluminance (EL) have recently attracted the attention of many researchers worldwide. EL from perovskite LEDs (PeLEDs) was first reported in 2015 by Tan et al.1 The authors demonstrated green and red LEDs with external quantum efficiencies (EQEs) of 0.1 and 0.76% by utilizing MAPbBr3 and MAPbI3−xClx OMHP, respectively. This opened the doors to the extensive study of green emission PeLEDs.17−27 More recently, an EQE of 8.53% has been achieved by a nanocrystal pinning (NCP) method.6 Pinning of nanocrystallites (NCs) results in the formation of 100 nm-sized crystallites. A nonstoichiometric ratio of precursors (MABr/PbBr2 = 1.05:1) reduces the concentration © XXXX American Chemical Society
Received: January 28, 2018 Accepted: April 24, 2018 Published: April 24, 2018 A
DOI: 10.1021/acsami.8b01292 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 1. Evolution of the morphology of MAPbBr3 thin films upon Rb incorporation. (a) XRD spectra of MAPbBr3 and (b) shift of the (100) XRD peak with increasing RbBr concentration. TEM images of (c) a pure MAPbBr3 film (S1) and MA1−xRbxPbBr3 films; MABr/RbBr of (d) 1:0.33 (S2), (e) 1:0.66 (S3), and (f) 1:1 (S4). The insets are selected area electron diffraction patterns.
performances. Photophysical properties of the thin films are analyzed by PL, optical absorbance, and time-resolved photoluminescence (TRPL) with different Rb concentrations into MAPbBr3. Furthermore, structural studies are carried out by transmission electron microscopy (TEM), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS).
impede fast in humidity, which limits the device stability. Thus, larger crystallites have recently been studied, as they pose ultralow trap densities that are on the order of 1010 cm−3.28,29 For example, an antisolvent vapor post-treatment of a CsPbBr3 thin film increases its grain size from 250 nm to 5 μm.28 The average exciton lifetime increases from 46.4 to 180.6 μs because of the reduction of surface trap density; therefore, performances and stability of the PeLEDs are improved.30 However, this method adds additional steps to the fabrication process. Alternatively, it has recently been shown that Rb incorporation into MAPbI3 thin films prepared by the conventional one-step approach can induce a significant reduction of trap density, resulting in high photocurrents when integrated in perovskite solar cells.31 As such reduction of defects could also significantly improve the performances of PeLEDs, here we study the incorporation of Rb into MAPbBr3 thin films and its consequences on PeLED
■
RESULTS AND DISCUSSION Two perovskite layers, MAPbBr3 and MAxRb1−xPbBr3, were deposited from solution by first mixing their respective precursors in two solvents: dimethylformamide (DMF) and dimethyl sulfoxide (DMSO). The precursors used for MAPbBr3 are MABr and PbBr2. For MAxRb1−xPbBr3, a third precursor, RbBr, is added. The MAPbBr3 solution was prepared with a molar ratio of MABr/PbBr2 of 2:1 to achieve a nonstoichiometric layer with excess MA. Three MAxRb1−xPbBr3 B
DOI: 10.1021/acsami.8b01292 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 2. Chemical binding energy states of Rb3d of Rb-incorporated MAPbBr3 thin films. XPS spectra of MA1−xRbxPbBr3 thin films with increasing content of Rb; MABr/RbBr of (a) 1:0 (S1), (b) 1:0.33 (S2), (c) 1:0.66 (S3), and (d) 1:1 (S4).
(S3) resulted in the formation of larger crystals of about 200 nm (Figure 1e). The crystals are facetted and appear darker than the background, indicating that they contain heavier elements (most likely Pb). Diffraction patterns obtained for the dark regions (Figure 1e) indicate good crystallinity with a negligible amount of defects or dislocations. The diffraction patterns of the surrounding light regions reveal that the region is amorphous (Figure 1e, right inset). As the Rb concentration is increased further (S4), slightly larger crystals, assembled in clusters, were formed (Figure 1f). The diffraction pattern recorded on the cluster (inset of Figure 1f) is typical of a polycrystalline structure. The chemical composition of the crystallites and matrix of S4 was determined by TEM elemental dispersive spectroscopy (EDS) (Figure S2 and Table S2a−c). The crystallites contain Pb while the only residual traces of Pb were found in the surrounding matrix, and both contain Br (S1 and S4). However, the strong overlap of the Rb and Br peaks (S4) in EDS makes it difficult to determine the actual composition of Rb in MA1−xRbxPbBr3. However, on the basis of the XRD results, we propose that the substitution of MA by Rb forms MA1−xRbxPbBr3 crystallites. To confirm the substitution of MA by Rb, we carried out XPS on all four samples. Expectedly, no Rb was found in S1 (Figure 2a), whereas the intensity of the Rb3d 3/2 (109.8 eV) and 5/2 (111.3 eV) peaks gradually increased with increasing Rb concentration (Figure 2b−d).34 With increasing Rb content, the peaks of Pb, Br, and N (see Figures S3 and S4) gradually shifted to higher energies (∼1 eV). This indicates a reorganization of their electronic clouds, which coincides with the lattice distortion observed in XRD (Figure 1a). This confirms that the NCs and crystallites observed in the TEM images of S2, S3, and S4 contain Rb and thus have a MAxRb1−xPbBr3 composition. Because the Br XPS peaks shift without broadening upon Rb insertion, Rb could be present in the amorphous matrix as well.21 Figure 3a,b shows the UV−vis absorbance and the photoluminescence (PL) spectra of S1, S2, S3, and S4 thin films. The absorbance spectra and the PL spectra show a similar blue shift with increasing Rb+ content. The optical band gaps, derived from the absorbance spectra (according to the Tauc method), increased from 2.32 eV (S1) to 2.36 eV (S4). The widening of the optical band gap could be the result of orientational or electrostatic disorder in the Rb-doped MAPbBr3 lattice as well as the formation of smaller NCs and increased exciton confinement upon Rb incorporation (in the case of S2).23,28,35 Similarly to previous reports, the PL peak of
solutions were prepared by mixing MABr, RbBr, and PbBr2, with molar ratios of MABr/RbBr of 1:0.33, 1:0.66, and 1:1, respectively. As RbBr has limited solubility in DMF, DMSO was added to DMF (6:4 v/v ratio) to increase its solubility (Figure S1). Samples with molar ratios of MABr/RbBr of 1:0, 1:0.33, 1:0.66, and 1:1 will be denoted as S1, S2, S3, and S4 in the paper, respectively. Thin films prepared in this paper are coated on indium tin oxide (ITO)/poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) and annealed at 80 °C for 20 min unless otherwise specified. Figure 1a shows the XRD patterns of S1, S2, S3, and S4. The diffraction pattern of S1 depicts the typical cubic phase of MAPbBr3.1,11 After Rb+ incorporation, the diffraction peaks shifted toward higher diffraction angles. Similar shifts are reported in the literature with smaller cation doping.21,32 The (100) peak of MAPbBr3 at 15.01° shifted to 15.36° as the concentration of RbBr increased, indicating lattice constant reduction upon Rb incorporation. The reduction in the lattice constant can be attributed to the smaller size of Rb with respect to MA. Shrinkage of the perovskite lattice upon smaller cation incorporation was previously reported when MA or Cs was added into formamidinium (FAPbBr3) or Cs was added to MAPbBr3.21,32,33 Thus, Rb+ seems to partially replace MA to form a mixed halide perovskite with a MA 1−xRbxPbBr3 composition. Moreover, as shown in detail in Figure 1b, XRD peaks exhibited similar full width at half-maximum (fwhm) values for S1, S3, and S4, except for S2, where the fwhm value was higher than the rest. This indicates that the minimal amount of Rb in MAPbBr3 induces the formation of crystallites that are small enough to induce a widening of the XRD peaks, that is, well below 50 nm. It is not possible to distinguish whether the formation of NCs is a direct consequence of the addition of Rb or the result of a nonstoichiometric ratio of precursors as reported by Lee et al.23 Thus, TEM images are obtained to thoroughly examine the impact of Rb doping on morphologies. Pure MAPbBr3 (S1) is composed of the crystal grains of 3 V), charge injections reduce significantly, implying that charge injections are efficient when the trap density of crystallite is reduced. Hole and electron injections from the respective electrode are differ by 1 order, which reveals that PeLEDs could be improved further upon charge balance statics.
added. The MAPbBr3 solution was prepared with a molar ratio of MABr/PbBr2 of 2:1 to achieve a nonstoichiometric layer with excess MA. Three MAxRb1−xPbBr3 solutions were prepared by mixing MABr, RbBr, and PbBr2, with molar ratios of MABr/RbBr of 1:0.33, 1:0.66, and 1:1, respectively. As RbBr has limited solubility in DMF, DMSO was added to DMF (6:4 v/v ratio) to increase its solubility (Figure S1). After 1 day of stirring, MA1−xRbxPbBr3 (x = 1) solution started to precipitate. The precipitated solution was decanted, filtered, and finally used for thin-film preparations. Device Fabrication. The ITO substrates were cleaned by methanol and IPA for 15 min each, followed by UV−ozone treatment for 15 min. PEDOT:PSS (40 nm thick) was spin-coated on ITO and annealed at 170 °C for 10 min. MAPbBr3/MA1−xRbxPbBr3 thin films (110 nm thick) were obtained by spin-coating on PEDOT:PSS. Thin films were annealed at 80 °C for 20 min in a N2-filled glovebox. To remove the good solvent (DMSO) by bad solvent (CB) for fast recrystallization, 0.5 mL of CB was dropped at 30th s during spincoating. Finally, the electron transport layer, TPBi (40 nm), and top electrode Liq (1.5 nm) and Al (100 nm) were thermally deposited (1.0 × 10−7) through a shadow mask. HODs and EODs utilizing TCTA and PVP are deposited by a shadow mask and solutions process, respectively. The first (second) layer in a double (triple)-layered perovskite coating was prepared by washing the perovskite precursors by CB at 30th s and DMF at 45th s; as a result, an ultrathin perovskite could be produced, as discussed in Figure S7. Characterization and Instrumentation. EL Device Characterization. The J−V, luminance−voltage (L−V), and EL spectra were measured from a Konica Minolta CS-100A luminance meter and a CS2000A spectrometer coupled with a Keithley 2635A voltage and current source meter. Thin-Film Characterization. The absorbance and PL of perovskite thin films were measured with a Scinco S-4100 UV−visible spectrophotometer and a JASCO FP-6500 spectrofluorometer, respectively. The TEM images of perovskite thin films using FEI Titan TM 80−300 were obtained at an accelerating voltage of 80 keV to prevent beam-induced local amorphization of the perovskite thin film. The AFM images of perovskite thin films coated on ITO/ PEDOT:PSS were obtained from XE-100. The XRD data of the perovskite thin film were obtained by ATX-G (Rigaku, Japan). The XPS experiments were performed in an UHV multipurpose surface analysis system (Sigma Probe, Thermo, UK) operated at base pressures
