Single-Crystal Thin Films of Cesium Lead Bromide Perovskite

relative humidity at room temperature, there was no change in the XRD pattern of CsPbBr3. SCTF (Figure S8). Photophysical properties of CsPbBr3 SCTFs...
0 downloads 13 Views 4MB Size
Download PDF
This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article pubs.acs.org/JACS

Single-Crystal Thin Films of Cesium Lead Bromide Perovskite Epitaxially Grown on Metal Oxide Perovskite (SrTiO3) Jie Chen,†,‡ Darien J. Morrow,† Yongping Fu,† Weihao Zheng,§ Yuzhou Zhao,† Lianna Dang,† Matthew J. Stolt,† Daniel D. Kohler,† Xiaoxia Wang,§ Kyle J. Czech,† Matthew P. Hautzinger,† Shaohua Shen,‡ Liejin Guo,‡ Anlian Pan,§ John C. Wright,† and Song Jin*,† †

Department of Chemistry, University of WisconsinMadison, 1101 University Avenue, Madison, Wisconsin 53706, United States International Research Center for Renewable Energy, State Key Laboratory of Multiphase Flow in Power Engineering, Xi’an Jiaotong University, Shaanxi 710049, P. R. China § Key Laboratory for Micro-Nano Physics and Technology of Hunan Province, School of Physics and Electronic Science, Hunan University, Changsha 410082, P. R. China ‡

S Supporting Information *

ABSTRACT: High-quality metal halide perovskite single crystals have low defect densities and excellent photophysical properties, yet thin films are the most sought after material geometry for optoelectronic devices. Perovskite single-crystal thin films (SCTFs) would be highly desirable for high-performance devices, but their growth remains challenging, particularly for inorganic metal halide perovskites. Herein, we report the facile vapor-phase epitaxial growth of cesium lead bromide perovskite (CsPbBr3) continuous SCTFs with controllable micrometer thickness, as well as nanoplate arrays, on traditional oxide perovskite SrTiO3(100) substrates. Heteroepitaxial single-crystal growth is enabled by the serendipitous incommensurate lattice match between these two perovskites, and overcoming the limitation of island-forming Volmer−Weber crystal growth is critical for growing large-area continuous thin films. Time-resolved photoluminescence, transient reflection spectroscopy, and electrical transport measurements show that the CsPbBr3 epitaxial thin film has a slow charge carrier recombination rate, low surface recombination velocity (104 cm s−1), and low defect density of 1012 cm−3, which are comparable to those of CsPbBr3 single crystals. This work suggests a general approach using oxide perovskites as substrates for heteroepitaxial growth of halide perovskites. The highquality halide perovskite SCTFs epitaxially integrated with multifunctional oxide perovskites could open up opportunities for a variety of high-performance optoelectronics devices.

■

INTRODUCTION Lead halide perovskites with an APbX3 stoichiometry (A is methylammonium, formamidinium, or cesium; X is I−, Br−, or Cl−) have re-emerged as an exciting class of semiconductors for optoelectronic applications owing to their excellent optical and electrical properties, such as large absorption coefficients, long carrier lifetimes, and balanced hole and electron transport.1−11 They have been utilized in high-performance solar cells,11−13 light-emitting diodes (LEDs),14,15 lasers,16−18 and ultravioletto-infrared photodetectors,19,20 as well as X-ray21 and γ-ray detectors.22 Even though impressive photovoltaic power conversion efficiency exceeding 22% has been achieved in polycrystalline thin-film devices12 despite the unavoidable higher defect densities relative to the single-crystal materials, high-quality single-crystal materials with fewer grain boundaries and lower trap densities would further boost device performance, because they could have higher carrier mobility and longer carrier lifetimes than the polycrystalline films.5−8 © 2017 American Chemical Society

Therefore, bulk single crystals of halide perovskites have been grown using several methods, such as antisolvent vapor-assisted crystallization,7,23 slow crystallization upon cooling saturated solutions,24−26 and inverse temperature crystallization,5,27,28 leading to fundamental physical insights about these materials.4−9 Although a few reports have incorporated single crystals into optoelectronic devices,19,21,29 thin films are still the most sought after materials geometry for most optoelectronic devices. Historically, the growth of high-quality semiconductor singlecrystal thin films (SCTFs) is critically important for both technological applications and fundamental research. For example, the industrial integration and deeper understanding of the conventional group IV, III−V, and II−VI semiconductors relied largely on the success of high-quality thin-film Received: July 18, 2017 Published: September 5, 2017 13525

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society

Figure 1. Illustration of the heteroepitaxy and structural characterizations of CsPbBr3 nanoplates epitaxially grown on STO substrates. (a) Illustration of the incommensurate lattice match between CsPbBr3(100) and STO(100) crystallographic planes. The region shaded light yellow is the unit cell overlap area, which consists of 2 × 2 CsPbBr3 (100) unit cells or 3 × 3 STO (100) unit cells. (b) PXRD patterns of the CsPbBr3 nanostructures grown on STO(100) and STO(110) substrates. Inset is the magnified range of 13°−32° for the CsPbBr3/STO(100) sample. The peaks associated with CsPbBr3 and SrTiO3 are marked with asterisks and diamonds, respectively. (c) Optical microscopy and (d) SEM image of the CsPbBr3 nanoplates grown on STO(100) substrate. (e and f) Representative AFM images of individual CsPbBr3 nanoplates.

growth.30−33 The growth of large-area single-crystal thin films of halide perovskites would be very desirable, but it remains challenging, particularly for the all-inorganic metal halide perovskites, which have competitive optoelectronic properties and better stability as compared to the organic−inorganic hybrid counterparts.4,34−36 The strategies of introducing an ultrasonic pulse during antisolvent diffusion growth29 or inverse temperature crystallization from solutions coupled with a space confinement37,38 were recently used to fabricate perovskite SCTFs, but these approaches are limited to hybrid perovskites, such as MAPbX3. Due to the lower solubility of inorganic perovskites and the more complex Cs−Pb−X phase diagram, the solution growth of inorganic perovskites is generally more challenging. For example, there are three phases of CsPbBr3, CsPb2Br5, and Cs4PbBr6 and the crystallization of a specific phase requires careful control of the stoichiometry in the precursor solution, which changes during the crystal growth.23,39,40 Therefore, developing an effective way to grow large-area inorganic halide perovskite SCTFs is highly desirable but remains difficult. Vapor-phase epitaxial (VPE) growth is a powerful technique that has demonstrated great success in high-quality singlecrystal thin-film growth of conventional semiconductors.31−33 VPE growth is not only effective in controlling morphology and reducing defects but also enables the monolithic integration of single-crystalline semiconductors with versatile epitaxial substrates to create advanced devices. Important examples include the epitaxial growth of III−V semiconductors on silicon wafers for solar cells and LEDs,30,41 the epitaxial growth of multifunctional perovskite oxides (LaAlO3,42 SrRuO3,43 etc.) on SrTiO3(100) for ferroelectrics and superconductors, and the epitaxial integration of SrTiO3 on Si(001) as photocathodes for efficient and stable water reduction.44 The VPE growth technique has catalyzed the proliferation of semiconductor heterostructure devices in the past decades; however, very little

progress has been made in halide perovskite materials.34,45 The perovskite structure of SrTiO3 (STO) has been a universal substrate for the epitaxial growth of many perovskite oxides with diverse properties,42,43,46 yet it has not been used as the substrates for metal halide perovskite growth, despite the shared structure type and name. Moreover, STO is a multifunctional wide band gap semiconductor that has been widely used as an electronic ceramic,40 photoelectrochemical water splitting material,47 and electron transport layer in perovskite photovoltaics.48 However, STO has a very large lattice constant mismatch with CsPbBr3 (a = 3.905 vs 5.830 Å) despite the fact that both have the cubic perovskite structural type, and STO has large surface energy and poor wetting by dissimilar materials49 that is not conducive to continuous thinfilm overgrowth. In this paper, we report a facile method to grow large-area CsPbBr3 SCTFs with controllable micrometer thickness on STO(100) substrates by vapor-phase epitaxial growth. Epitaxial CsPbBr3 nanoplates arrays could also be grown on STO(100), which can be explained by a Volmer−Weber growth mode. We found that the key to growing continuous CsPbBr3 SCTFs is to elevate the reaction temperature to enhance the diffusion of the adatoms and accelerate the nucleation of preferred epitaxial crystallites. Electrical transport measurement shows that such epitaxial CsPbBr3 SCTF has a low bulk defect density of 1012 cm−3, and time-resolved photoluminescence and transient reflection spectroscopy reveal comparable photophysical properties, such as carrier mobility, charge carrier recombination rate, and surface recombination velocity, relative to CsPbBr3 single crystals. These high-quality CsPbBr3 SCTFs integrated with semiconducting SrTiO3 as the epitaxial substrate could enable the fabrication of diverse high-performance optoelectronics devices. 13526

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society

Figure 2. Structural characterizations of CsPbBr3 SCTFs epitaxially grown on STO(100) substrates. (a) Cross-sectional SEM images and photographs of the CsPbBr3 SCTFs with varying thickness of 1, 2, 5, and 7 μm. (b) Top down SEM image of the CsPbBr3 SCTF with 7 μm thickness. The inset shows a magnified top down SEM image of the SCTF. (c) PXRD pattern of the corresponding sample. The left inset magnifies the 2θ range of 12°−27°, and the right inset shows the θ-rocking curve of the CsPbBr3(200) peak. (d) The (110) pole figure of the 7 μm CsPbBr3 SCTF sample.

■

RESULTS AND DISCUSSION Epitaxial Growth of CsPbBr3 Nanoplate Arrays on STO. As a representative metal oxide perovskite, SrTiO3 (space group Pm3m ̅ , a = 3.905 Å) possesses the same crystal structure as the cubic phase of CsPbBr3 (space group Pm3̅m, a = 5.830 Å) but with a much smaller lattice constant. Despite such a large difference in lattice constants, we recognized that the lattice constant of CsPbBr3 happens to be ∼150% that of STO. Therefore, as illustrated in Figure 1a, two unit cells of CsPbBr3 can match with three unit cells of STO in an incommensurate fashion with a very low lattice mismatch factor f of 0.47% [f = (1 − doverlayer/dsubstrate), where d is the lattice spacing] in both CsPbBr3[100]∥STO[100] and CsPbBr3[010]∥STO[010] directions on their a crystallographic planes. We further calculated the overlap area based on the 2 × 2 (100) unit cell area of the overlayer material (CsPbBr3). The basic unit of the overlap area (the region shaded light yellow in Figure 1a) is 135.96 Å2, which is much smaller than that of our previous epitaxial growth of CsPbBr3 nanowires on mica(001) (753 Å2).34 Therefore, the heteroepitaxial growth of CsPbBr3 on STO would be plausible. Our initial chemical vapor deposition (CVD) experiments on single-crystal STO(100) substrates led to epitaxial growth of CsPbBr3 nanostructures, but not continuous SCTFs. A critical step for the successful epitaxial growth is that, before every growth, the STO substrates should be etched by immersing in hot water and then in buffered HF solution to make an atomically flat Ti−O-terminated surface.50 We first present the optimized results of epitaxial growth of CsPbBr3 nanoplates on STO(100) by CVD using solid CsPbBr3 precursor at 320 °C for 60 min [see the Supporting Information (SI) for experimental details and Figure S1, a schematic of the CVD

setup]. The optical microscopy image (Figure 1c) and scanning electron microscopy (SEM) image (Figure 1d) clearly show that the majority of the products are well-orientated rectangular plates with lateral dimensions of about a few to tens of micrometers on the STO substrate. The powder X-ray diffraction (PXRD) pattern (Figure 1b) reveals that the predominant diffraction peaks of the as-grown CsPbBr3/ STO(100) sample are well-indexed to the (100) and (200) planes of the cubic phase CsPbBr3, indicating that the [001] direction of as-grown CsPbBr3 nanoplates is vertical to the STO(100) substrate. The absence of other diffraction peaks from CsPbBr3 confirms the hypothesized epitaxial alignments of the two perovskite structures; even though there are many nanoplates, their crystallographic orientations are all aligned as if they are part of a single crystal. We need to note that, upon closer examination of the PXRD (Figure 1b, inset), a very small (110) peak is still visible, which could be attributed to additional minor morphologies other than the plates (highlighted in Figure S2a, SI). A few CsPbBr3 nanoplates with a 45° angle to the majority ones were also occasionally observed (Figure S2a, SI). These plates likely have a less thermodynamically favorable epitaxial relationship, in which CsPbBr3[100]∥STO[110] and CsPbBr3[010]∥STO[110] have an f of −5.57%, as illustrated in Figure S2b (SI). We further performed the CVD growth of CsPbBr3 on STO(110) substrates to support the epitaxy mechanism. It can be expected that the epitaxial alignments of CsPbBr3[110]∥STO[110] and CsPbBr3[001]∥STO[100] and the epitaxy of CsPbBr 3 [100]∥STO[110] and CsPbBr 3[010]∥STO[110] would be the preferred orientations (Figure S3a, SI). Indeed, Figure S3b−d (SI) shows a large majority of CsPbBr3 prisms and nanoplates grown on STO(110), corresponding to the two 13527

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society

section SEM images (Figure 2a) show that 12, 14, 17, and 20 min growth times yielded CsPbBr3 SCTFs with thicknesses of approximately 1, 2, 5, and 7 μm, respectively. Note that the film thicknesses were measured from the thicker ends of the corresponding samples and the films on the other ends on the 10 mm long substrate are slightly thinner, which causes the slight color variations seen in the photographs in Figure 2a. This is because the 10 mm substrate is too long relative to the small size of our CVD system. But the film thickness variation is not very large, and the SCTF is uniform in a large area of up to at least hundreds of micrometers (see Figure S5 of the SI for a large-scale SEM cross section image of the 7 μm thick SCTF), which is comparable in size to the reported MAPbBr3 SCTFs.37 Figure 2b shows that the surface of the 7 μm thick continuous film sample is mostly smooth with small concave textures, suggesting that the thickening of the SCTFs follows the same layer-by-layer homoepitaxial growth mode as discussed previously for the thick nanoplates. The surface roughness (Sq) of the 7 μm thick SCTF was estimated to be around 26 nm (Figure S6a,b, SI), which is comparable to the reported ultrasmooth halide perovskite thin film.53 Higher-resolution SEM and AFM images of a 1 μm thick sample (Figure S6c,d, SI) also show a layer-by-layer growth feature and further corroborate the growth mode. The PXRD pattern for the 7 μm CsPbBr3 SCTF sample on STO shows only (100) and (200) peaks (Figure 2c). The lack of a detectable (110) peak (Figure 2c, left inset) implies that this sample has a better epitaxial quality than that grown at a lower temperature (Figure 1c). This result means that the large, irregular nanocrystals can be avoided by high temperature growth. The rocking curve for the (200) diffraction peak (the right inset in Figure 2c) has a narrow full width at halfmaximum (fwhm) of 0.18°, indicating good single crystallinity of the thin film. Also note that, in contrast to the nanoplates on the STO sample (Figure 1b), the diffraction peaks from STO are barely detectable, which further confirms that the thick and continuous CsPbBr3 SCTF covers and masks the STO substrate very well. We also obtained a (110) pole figure of the sample to further evaluate the single crystallinity and domain orientation (Figure 2d). Four poles with a 90° interval were observed with all of them symmetrically equivalent to the (110) plane. This result conclusively confirms the single crystalline quality of our CsPbBr3 thin film, and the two cubic perovskite structures are aligned along their crystallographic axes, as illustrated in Figure 1a. The pole figure confirms that the high-temperature epitaxy at 450 °C led to the growth of only one single-crystal domain of the cubic CsPbBr3 thin film on the STO substrate. This result is consistent with our hypothesis that the thermodynamically favorable orientations could also have faster growth kinetics. At higher growth temperature, the preferred epitaxial orientation will be dominant, because it is both thermodynamically and kinetically favorable. To monitor the growth before the continuous film formation and understand the growth process better, we performed the reactions with a shorter growth time (7−10 min). After a 7 min growth, the initial nucleated CsPbBr3 nanostructures have a much higher density (Figure S7b, SI) than the nanoplate arrays after a longer growth time of 60 min at lower temperature (Figure S7a, SI). The nanostructures then quickly merged together to form an interconnected network after 10 min growth time (Figure S7c, SI), due to the higher adatoms diffusion that increased the lateral growth rate at higher temperature. Interestingly, the

types of lattice match mentioned above. As expected, the PXRD of the CsPbBr3/STO(110) sample (Figure 1b, green trace) shows two sets of diffraction peaks of (hk0) and (00l) families, which agree with the orientations of prisms and nanoplates, respectively. Figure 1d−f reveals more details about the morphologies and growth mode for CsPbBr3 nanoplates. The thickness and lateral dimensions of CsPbBr3 nanoplates vary between 40 and 200 nm and between 1 and 10 μm, respectively. This variation suggests that these nanoplates nucleate at different times and are at different growth stages. We suggest that the lateral growth of CsPbBr3 nanoplates follows a Volmer−Weber (V− M) island formation heteroepitaxial growth mode,51 as confirmed by the appearance of a rectangular shape with zigzag edges that are formed by the merging of neighboring nanoplates (islands). The V−M growth mode usually results from the stronger adatom cohesive force of the overlayer material compared to the surface adhesive force between overlayer and substrate. In our case, the large surface energy (∼1 eV)49 and lack of wetting layer on the STO(100) surface may be the main reasons for the hypothesized V−M growth mode. An atomic force microscopy image (Figure 1e) shows that the majority of nanoplates are smooth and flat, but a few nanoplates have concave surfaces (Figure 1f). These concave nanoplates are usually thicker than the smooth nanoplates, and they are more likely to be observed after a longer growth time. These observations strongly suggest that the vertical growth of these CsPbBr3 nanoplates occurs via a layer-by-layer inverted wedding cake homoepitaxial growth mechanism. Similar concave features have also been observed in the homoepitaxial growth of ZnO nanostructures via the inverted wedding cake growth mode, where the new atomic layers first nucleated at the edge, propagated inward, and then became limited by the interlayer mass transport.52 Epitaxial Growth of Continuous CsPbBr3 SCTFs on STO. We then attempted to produce CsPbBr3 SCTFs on STO(100) substrates by increasing the CVD reaction time at 320 °C. However, simply prolonging the reaction time failed to yield continuous thin films. Instead, as shown in Figure S4 (SI), two types of orientated nanoplates and many other morphologies formed on the STO substrate after a 90 min growth. Interestingly, the nanoplates of the most preferred epitaxial orientations are much larger than those with the less preferred orientation (denoted with red circles). We expect the crystallites with more thermodynamically favorable heteroepitaxial orientations to have faster growth kinetics, such as high nucleation and growth rate. Moreover, comparing the products of 60 and 90 min growth [Figure 1c and Figure S4 (SI)] shows that the lateral size of the nanoplates does not significantly increase, indicating a slow lateral growth of the nanoplates. This failure probably results from the slow diffusion of the adatoms on the STO surface in the V−M growth mode, which makes the thickening more dominant than the lateral growth. We found the key to overcoming the limitation of the V−M crystal growth and growing continuous CsPbBr3 SCTFs is to raise the reaction temperature to 450 °C to enhance the diffusion of the adatoms, to increase the nucleation density, and to avoid randomly grown irregular nanostructures. As show in Figure 2a,b, large-area continuous CsPbBr3 SCTFs with various thicknesses can be readily grown on 5 × 10 mm STO(100) substrates by controlling the growth time, which is much shorter than that needed for nanoplate growth. The cross13528

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society

Figure 3. (a) PL spectra of a 7 μm VPE-grown CsPbBr3 SCTF excited by a 442 nm laser together with that of a solution-grown single crystal (SC), showing similar PL profile and intensity. The PL intensities are not normalized. (b) TRPL of a 7 μm VPE-grown SCTF and a solution-grown single crystal excited by 400 nm laser at a low pump intensity of 25 nJ/cm2.

Figure 4. Pseudocolor representations of the transient reflectance spectra for (a) a 7 μm VPE-grown CsPbBr3 SCTF and (b) a solution-grown single crystal. (c) Normalized transient reflection spectroscopy of the SCTF and SC samples, under an excitation fluence of 14 μJ/cm2, and their fits to the diffusion model (solid lines).

nanostructures formed at 450 °C are not well-faceted, in contrast to the those formed at lower temperature (Figure S7a, SI), yet the (110) pole figure of this 10 min growth sample (Figure S7d, SI) confirms that such a structure is still a singlecrystalline epitaxial network with the same orientation, just as the continuous SCTF (Figure 2d). This result strongly suggests that the overwhelmingly higher density nucleation and subsequent faster crystal growth at higher temperature only follow the more preferred epitaxial orientation. Thus, the epitaxy growth can overcome the island-forming V−M growth mode and a continuous single-crystal film can be grown in a short period of time at higher temperature. The VPE-grown CsPbBr3 products (both nanoplates and SCTFs) have excellent phase purity, which is an important advantage of the vaporphase growth of CsPbBr3 over the solution growth methods. In addition, the CsPbBr3 SCTFs have excellent air and moisture stability. After 3 months of exposure to ambient air with ∼20% relative humidity at room temperature, there was no change in the XRD pattern of CsPbBr3 SCTF (Figure S8, SI). Photophysical Properties of CsPbBr3 SCTFs. The optical spectroscopy and electrical studies on these CsPbBr3 SCTFs (see the Supporting Information for experimental details) show that their semiconducting and photophysical properties are comparable to the solution-grown CsPbBr3 single crystals. As a comparison, the CsPbBr3 single crystals were grown following a reported antisolvent method23 (see “Materials synthesis” in the SI for details). The CsPbBr3 SCTFs exhibit a bright photoluminescence (PL) emission centered at 525 nm with a fwhm of ∼18 nm (Figure 3a), which is consistent with the previous reports34,35 and the spectrum of

solution-grown single crystals. To study the charge carrier recombination rate of the CsPbBr3 SCTF and a solution-grown single crystal, time-resolved photoluminescence (TRPL) decays were measured under a low pump intensity (25 nJ/cm2) and fit individually to a single exponential decay function (Figure 3b). The PL lifetimes of the SCTF and solution-grown single crystal are both ∼10 ns (inset table in Figure 3b), which are comparable to the previous reports on CsPbBr3 crystals54−56 [see Table S1 (SI) for a summary of the reported lifetime values on CsPbBr3 single crystals], implying that the carrier recombination rate in the SCTFs is as low as that in the solution-grown single crystals. It has been suggested that surface trap states and grain boundaries are the limiting factors for the carrier lifetime and have detrimental effects on photovoltaic performance of the halide perovskite materials.1,11,57 Moreover, recent experiments showed that the recombination on the surface has a more profound influence than recombination at grain boundaries.57 We therefore performed surface-sensitive ultrafast transient reflection (TR) spectroscopy on both a CsPbBr3 SCTF and a solution-grown single crystal to measure the surface recombination velocity and the carrier diffusion constant in the region near the sample surface. The representative pseudocolor TR spectra of the SCTF and solution-grown single crystal are shown in parts a and b of Figure 4, respectively. The corresponding reflectance change (ΔR/R) kinetics extracted from the entire spectra are provided in Figure 4c. We fit the experimental data using the diffusion model developed by Hoffman et al.,58 which has been successfully applied to perovskite materials.4,8,57 In this model, the carriers diffuse 13529

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society

Figure 5. (a) I−V characteristics of the Au/CsPbBr3/Au device under dark and illuminated conditions (1 sun of illumination). The inset shows the Au/CsPbBr3/Au device structure for electrical transport and photocurrent measurements. The gap between the Au electrodes is 50 μm. (b) Photoresponse of the Au/CsPbBr3/Au device under an applied bias of 2.5 V and 1 sun of light illumination (100 mW/cm2). (c) High electric field I−V curve of the device under the dark condition displayed on a log scale to reveal the three transport regimes and the trap-filled limit voltage (VTFL).

± 1.5 and 26.9 ± 1.2 cm2 V−1 s−1 for VPE-grown SCTF and solution-grown SC, respectively. It is encouraging that the average S and D values for the two types of CsPbBr3 samples are within a factor of 2 of each other, indicating the comparable physical properties of the CsPbBr3 SCTF grown by vaporphase epitaxy to the solution-grown CsPbBr3 single crystals. Moreover, we find the SCTF has slower surface recombination kinetics than the solution-grown SC. Transport Properties of CsPbBr3 SCTFs. To quantitatively characterize the defect density in the CsPbBr3 SCTFs, we built an Au/CsPbBr3/Au device for photocurrent and electrical transport measurements and calculated the trap density using the space-charge-limited-current (SCLC) method. The device was fabricated by thermal evaporation of Au through a shadow mask onto a 7 μm thick CsPbBr3 SCTF with a channel width of 50 μm between the two electrodes (inset of Figure 5a; see the Supporting Information for device fabrication and measurement details). Typical linear and symmetrical current versus voltage (I−V) curves were observed in the dark and under illumination (Figure 5a). The device exhibited a fast photoresponse with a response time of only several milliseconds [see Figures 5b and S13 (SI) for details], suggesting a low density of deep trap states.59 We further estimated the trap density (nt) using the I−V curve measured under the dark condition following the SCLC method previously used for hybrid perovskite single crystals.5 As shown in the log I − log V plot in Figure 5c, with increasing applied bias, the I−V displays three linear regions: the Ohm’s law region (red) at low voltage, the trap-filling region (blue) after reaching the trap-filled limit voltage (VTFL), and the Child’s regime (green), where all the defect states are filled.54,59 The presence of these three sections in the dark I−V curve qualitatively indicates a low defect density and poor intrinsic conductivity in this CsPbBr3 SCTF.7,59 The trap density can be calculated by the equation nt = 2εε0VTFL/(ed2), where ε is the dielectric constant, e is the elemental charge, and d is the device channel width. Here we take ε = 16.46 for CsPbBr354 and d = 50 μm and determine the VTFL to be 1.64 V from Figure 5c. The calculated nt is 1.5 × 1012 cm−3, which is comparable to that of the CsPbBr3 single crystals and several orders of magnitude lower than that of polycrystalline halide perovskite films.5,7,28,59,60 Furthermore, we calculated the electron mobility (μ) from the Child region according to Mott−Gurney’s equation μ = (8JDd3)/(9εε0V2), where JD is the current density at an applied voltage (V) in the Child region, μ is the electron mobility, and d is the gap distance of the two Au electrodes. The values we used are JD =

according to Fick’s second law with a diffusion coefficient, D. Carriers that diffuse to the surface then recombine with a timedependent flux of J(t) = SNsurf(t), where Nsurf(t) is the surface carrier density and S is the surface recombination velocity. The normalized carrier density (N) as a function of time (t) and distance away from surface (z) is given by N (z ,t ) =

⎧ N0 −z 2 /4Dt ⎪ ⎛ z ⎞ ⎨W ⎜α Dt − ⎟ e ⎪ 2 2 Dt ⎠ ⎩ ⎝ ⎛ z ⎞ ⎟− + W ⎜α Dt + ⎝ 2 Dt ⎠

S

2D S D

−α

⎡ ⎛ ⎛S z ⎞ ⎟ − W⎜ Dt ⎢W ⎜α Dt + ⎠ ⎝D ⎣ ⎝ 2 Dt +

⎫ z ⎞⎤⎪ ⎟⎥⎬ ⎪ 2 Dt ⎠⎦⎭

where W(X) ≡ exp(X2)[1 − erf(X)], D is the diffusion coefficient, N is the carrier density, S is the surface recombination velocity, and α is the absorption coefficient at the pump excitation wavelength. We only consider the surface reflectance change as a function of N(0,t). In this model, the majority of carriers recombine at the surface and/or diffuse into bulk, where recombination is much slower. The surface carriers recombine with a time-dependent flux of J(t) = SNsurf(t), where Nsurf(t) is the surface carrier density and S is the surface recombination velocity. The time-resolved TR spectra were fitted following these models (the solid lines in Figure 4c), and the fitted results of S and D for both types of samples are shown in the inset table of Figure 4c. More details on the TR spectroscopy measurements and modeling are provided in the Supporting Information (Figures S9−S12, Table S2, and the associated text), including a demonstration of signal vs pump fluence linearity (Figure S12, SI). It is worth noting that our decay transients constrain the exact values of S and D only modestly, as shown in Figure S11 (SI). Despite this uncertainty, it is still fair to compare these two types of CsPbBr3 samples using this self-consistent measurement. The average S and D values are (1.5 ± 0.16) × 104 cm s−1 and 0.35 ± 0.04 cm2 s−1 for the VPE-grown SCTF and (3.0 ± 0.16) × 104 cm s−1 and 0.69 ± 0.03 cm2 s−1 for the solutiongrown SC, respectively. On the basis of the Einstein relationship, the corresponding carrier mobilities (μ) are 13.6 13530

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society 0.68 A cm−2, V = 10 V, ε = 16.46, ε0 = 8.854 × 10−14 F cm−1, and d = 50 μm. Note here we calculated JD by using an estimated area of 1 mm × 7 μm, where 1 mm is the length of the deposited Au electrode (while the width is 50 μm) and 7 μm is the thickness of the CsPbBr3 SCTF. Therefore, the ascalculated μ value is 518 cm2 V−1 s−1. This value is a little higher than that reported for a solution-grown CsPbBr3 single crystal,39 possibly due to the smoothness and single crystallographic orientation of the present CsPbBr3 SCTF. These results show that the epitaxially grown CsPbBr3 SCTFs have an optoelectronic quality comparable to that of the CsPbBr3 single crystals54,59 [see Table S3 (SI) for a summary of reported nt values of CsPbBr3 single crystals], which make single-crystal CsPbBr3 thin films promising for high-performance optoelectronics applications.

ORCID

Jie Chen: 0000-0002-2007-0896 Darien J. Morrow: 0000-0002-8922-8049 Yongping Fu: 0000-0003-3362-2474 Matthew J. Stolt: 0000-0002-6704-0862 Liejin Guo: 0000-0002-3671-5628 Anlian Pan: 0000-0003-3335-3067 Song Jin: 0000-0001-8693-7010 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research is supported by the US Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering, under award DE-FG02-09ER46664. J.C. acknowledges support from the China Scholarship Council. L.D. and M.J.S. are thankful for NSF graduate fellowships for support. W.Z., X.W., and A.P. appreciate the support from the National Natural Science Foundation of China (Nos. 51525202, 61574054).

■

CONCLUSION In conclusion, we show that large-area continuous single-crystal thin films of CsPbBr3 with controllable micrometer thickness, as well as aligned CsPbBr3 nanoplate arrays, can be grown on SrTiO3(100) substrates by a facile vapor-phase epitaxial growth for the first time. The heteroepitaxial growth is enabled by the serendipitous but excellent incommensurate lattice match between the CsPbBr3 perovskite and the perovskite oxide SrTiO3 and by the understanding of the crystal growth mechanisms that allows us to overcome the limitation of the island-forming Volmer−Weber crystal growth. The resulting CsPbBr3 SCTFs exhibit excellent crystal quality and photophysical and electrical properties, such as a slow charge carrier recombination rate, low surface recombination velocity, a low bulk defect density, comparable to those of CsPbBr3 single crystals. This approach of heteroepitaxially growing halide perovskites on perovskite oxides could be general, because there are many different perovskite oxides with slightly different lattice constants61 that could be lattice-matched in a similar incommensurate fashion with the diverse 3D AMX3 (or even 2D) perovskite materials with different A cations, M ions (Pb, Sn and others), and halide anions.1−3 This demonstration of the controllable growth of CsPbBr3 SCTF/SrTiO3 heterostructures opens up avenues for developing halide perovskites integrated on a myriad of oxide perovskites with diverse semiconducting, magnetic, and ferroelectric properties,46 which could potentially further boost the device performance of perovskite solar cells and enable diverse optoelectronics applications with high performance.

■

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b07506. Experimental methods and setup, additional AFM and SEM images, PXRD patterns, fitting methods for the TR spectra, response time from the I−V curve under dark/ illumination, tables summarizing the PL lifetimes and trap density values of the reported CsPbBr3 singlecrystalline materials, and detailed illustrations of the lattice match models (PDF)

■

REFERENCES

(1) Huang, J.; Yuan, Y.; Shao, Y.; Yan, Y. Nat. Rev. Mater. 2017, 2, 17042. (2) Stranks, S. D.; Snaith, H. J. Nat. Nanotechnol. 2015, 10, 391. (3) Zhang, W.; Eperon, G. E.; Snaith, H. J. Nat. Energy 2016, 1, 16048. (4) Zhu, H.; Trinh, M. T.; Wang, J.; Fu, Y.; Joshi, P. P.; Miyata, K.; Jin, S.; Zhu, X. Y. Adv. Mater. 2017, 29, 1603072. (5) Dong, Q.; Fang, Y.; Shao, Y.; Mulligan, P.; Qiu, J.; Cao, L.; Huang, J. Science 2015, 347, 967. (6) Guo, Z.; Wan, Y.; Yang, M.; Snaider, J.; Zhu, K.; Huang, L. Science 2017, 356, 59. (7) Shi, D.; Adinolfi, V.; Comin, R.; Yuan, M.; Alarousu, E.; Buin, A.; Chen, Y.; Hoogland, S.; Rothenberger, A.; Katsiev, K.; Losovyj, Y.; Zhang, X.; Dowben, P. A.; Mohammed, O. F.; Sargent, E. H.; Bakr, O. M. Science 2015, 347, 519. (8) Yang, Y.; Yan, Y.; Yang, M.; Choi, S.; Zhu, K.; Luther, J. M.; Beard, M. C. Nat. Commun. 2015, 6, 7961. (9) Xing, G.; Mathews, N.; Sun, S.; Lim, S. S.; Lam, Y. M.; Grätzel, M.; Mhaisalkar, S.; Sum, T. C. Science 2013, 342, 344. (10) Zhu, H.; Miyata, K.; Fu, Y.; Wang, J.; Joshi, P. P.; Niesner, D.; Williams, K. W.; Jin, S.; Zhu, X.-Y. Science 2016, 353, 1409. (11) Nie, W.; Tsai, H.; Asadpour, R.; Blancon, J.-C.; Neukirch, A. J.; Gupta, G.; Crochet, J. J.; Chhowalla, M.; Tretiak, S.; Alam, M. A.; Wang, H.-L.; Mohite, A. D. Science 2015, 347, 522. (12) Yang, W. S.; Park, B.-W.; Jung, E. H.; Jeon, N. J.; Kim, Y. C.; Lee, D. U.; Shin, S. S.; Seo, J.; Kim, E. K.; Noh, J. H.; Seok, S. I. Science 2017, 356, 1376. (13) Research Cell Record Efficiency Chart. http://www.nrel.gov/ pv/ (accessed June 2017). (14) Xiao, Z.; Kerner, R. A.; Zhao, L.; Tran, N. L.; Lee, K. M.; Koh, T.-W.; Scholes, G. D.; Rand, B. P. Nat. Photonics 2017, 11, 108. (15) Zhang, L.; Yang, X.; Jiang, Q.; Wang, P.; Yin, Z.; Zhang, X.; Tan, H.; Yang, Y. M.; Wei, M.; Sutherland, B. R.; Sargent, E. H.; You, J. Nat. Commun. 2017, 8, 15640. (16) Zhu, H.; Fu, Y.; Meng, F.; Wu, X.; Gong, Z.; Ding, Q.; Gustafsson, M. V.; Trinh, M. T.; Jin, S.; Zhu, X.-Y. Nat. Mater. 2015, 14, 636. (17) Eaton, S. W.; Lai, M.; Gibson, N. A.; Wong, A. B.; Dou, L.; Ma, J.; Wang, L. W.; Leone, S. R.; Yang, P. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 1993. (18) Fu, Y.; Zhu, H.; Schrader, A. W.; Liang, D.; Ding, Q.; Joshi, P.; Hwang, L.; Zhu, X. Y.; Jin, S. Nano Lett. 2016, 16, 1000. (19) Fang, Y.; Dong, Q.; Shao, Y.; Yuan, Y.; Huang, J. Nat. Photonics 2015, 9, 679.

AUTHOR INFORMATION

Corresponding Author

*[email protected] 13531

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532

Article

Journal of the American Chemical Society (20) Dou, L.; Yang, Y. M.; You, J.; Hong, Z.; Chang, W. H.; Li, G.; Yang, Y. Nat. Commun. 2014, 5, 5404. (21) Wei, W.; Zhang, Y.; Xu, Q.; Wei, H.; Fang, Y.; Wang, Q.; Deng, Y.; Li, T.; Gruverman, A.; Cao, L.; Huang, J. Nat. Photonics 2017, 11, 315. (22) Yakunin, S.; Dirin, D. N.; Shynkarenko, Y.; Morad, V.; Cherniukh, I.; Nazarenko, O.; Kreil, D.; Nauser, T.; Kovalenko, M. V. Nat. Photonics 2016, 10, 585. (23) Rakita, Y.; Kedem, N.; Gupta, S.; Sadhanala, A.; Kalchenko, V.; Böhm, M. L.; Kulbak, M.; Friend, R. H.; Cahen, D.; Hodes, G. Cryst. Growth Des. 2016, 16, 5717. (24) Baikie, T.; Fang, Y.; Kadro, J. M.; Schreyer, M.; Wei, F.; Mhaisalkar, S. G.; Graetzel, M.; White, T. J. J. Mater. Chem. A 2013, 1, 5628. (25) Dang, Y.; Liu, Y.; Sun, Y.; Yuan, D.; Liu, X.; Lu, W.; Liu, G.; Xia, H.; Tao, X. CrystEngComm 2015, 17, 665. (26) Lian, Z.; Yan, Q.; Gao, T.; Ding, J.; Lv, Q.; Ning, C.; Li, Q.; Sun, J. L. J. Am. Chem. Soc. 2016, 138, 9409. (27) Dirin, D. N.; Cherniukh, I.; Yakunin, S.; Shynkarenko, Y.; Kovalenko, M. V. Chem. Mater. 2016, 28, 8470. (28) Saidaminov, M. I.; Abdelhady, A. L.; Murali, B.; Alarousu, E.; Burlakov, V. M.; Peng, W.; Dursun, I.; Wang, L.; He, Y.; Maculan, G.; Goriely, A.; Wu, T.; Mohammed, O. F.; Bakr, O. M. Nat. Commun. 2015, 6, 7586. (29) Peng, W.; Wang, L.; Murali, B.; Ho, K. T.; Bera, A.; Cho, N.; Kang, C. F.; Burlakov, V. M.; Pan, J.; Sinatra, L.; Ma, C.; Xu, W.; Shi, D.; Alarousu, E.; Goriely, A.; He, J. H.; Mohammed, O. F.; Wu, T.; Bakr, O. M. Adv. Mater. 2016, 28, 3383. (30) Adachi, S. Properties of Semiconductor Alloys: Group-IV, III-V and II-VI Semiconductors; John Wiley & Sons, 2009; Vol. 28. (31) Meyerson, B. S. Appl. Phys. Lett. 1986, 48, 797. (32) Tsukazaki, A.; Ohtomo, A.; Onuma, T.; Ohtani, M.; Makino, T.; Sumiya, M.; Ohtani, K.; Chichibu, S. F.; Fuke, S.; Segawa, Y.; Ohno, H.; Koinuma, H.; Kawasaki, M. Nat. Mater. 2004, 4, 42. (33) Yoon, J.; Jo, S.; Chun, I. S.; Jung, I.; Kim, H. S.; Meitl, M.; Menard, E.; Li, X.; Coleman, J. J.; Paik, U.; Rogers, J. A. Nature 2010, 465, 329. (34) Chen, J.; Fu, Y.; Samad, L.; Dang, L.; Zhao, Y.; Shen, S.; Guo, L.; Jin, S. Nano Lett. 2017, 17, 460. (35) Fu, Y.; Zhu, H.; Stoumpos, C. C.; Ding, Q.; Wang, J.; Kanatzidis, M. G.; Zhu, X.; Jin, S. ACS Nano 2016, 10, 7963. (36) Kulbak, M.; Gupta, S.; Kedem, N.; Levine, I.; Bendikov, T.; Hodes, G.; Cahen, D. J. Phys. Chem. Lett. 2016, 7, 167. (37) Chen, Y. X.; Ge, Q. Q.; Shi, Y.; Liu, J.; Xue, D. J.; Ma, J. Y.; Ding, J.; Yan, H. J.; Hu, J. S.; Wan, L. J. J. Am. Chem. Soc. 2016, 138, 16196. (38) Rao, H. S.; Li, W. G.; Chen, B. X.; Kuang, D. B.; Su, C. Y. Adv. Mater. 2017, 29, 1602639. (39) Saidaminov, M. I.; Haque, M. A.; Almutlaq, J.; Sarmah, S.; Miao, X.-H.; Begum, R.; Zhumekenov, A. A.; Dursun, I.; Cho, N.; Murali, B.; Mohammed, O. F.; Wu, T.; Bakr, O. M. Adv. Opt. Mater. 2017, 5, 1600704. (40) Ling, Y.; Tan, L.; Wang, X.; Zhou, Y.; Xin, Y.; Ma, B.; Hanson, K.; Gao, H. J. Phys. Chem. Lett. 2017, 8, 3266. (41) Soref, R. A. Proc. IEEE 1993, 81, 1687. (42) Wu, S.; Luo, X.; Turner, S.; Peng, H.; Lin, W.; Ding, J.; David, A.; Wang, B.; Van Tendeloo, G.; Wang, J.; Wu, T. Phys. Rev. X 2013, 3, 041027. (43) Lee, D.; Lu, H.; Gu, Y.; Choi, S.-Y.; Li, S.-D.; Ryu, S.; Paudel, T. R.; Song, K.; Mikheev, E.; Lee, S.; Stemmer, S.; Tenne, D. A.; Oh, S. H.; Tsymbal, E. Y.; Wu, X.; Chen, L.-Q.; Gruverman, A.; Eom, C. B. Science 2015, 349, 1314. (44) Ji, L.; McDaniel, M. D.; Wang, S.; Posadas, A. B.; Li, X.; Huang, H.; Lee, J. C.; Demkov, A. A.; Bard, A. J.; Ekerdt, J. G.; Yu, E. T. Nat. Nanotechnol. 2015, 10, 84. (45) Wang, Y.; Sun, X.; Shivanna, R.; Yang, Y.; Chen, Z.; Guo, Y.; Wang, G. C.; Wertz, E.; Deschler, F.; Cai, Z.; Zhou, H.; Lu, T. M.; Shi, J. Nano Lett. 2016, 16, 7974.

(46) Schlom, D. G.; Chen, L.-Q.; Eom, C.-B.; Rabe, K. M.; Streiffer, S. K.; Triscone, J.-M. Annu. Rev. Mater. Res. 2007, 37, 589. (47) Wang, Q.; Hisatomi, T.; Jia, Q.; Tokudome, H.; Zhong, M.; Wang, C.; Pan, Z.; Takata, T.; Nakabayashi, M.; Shibata, N.; Li, Y.; Sharp, I. D.; Kudo, A.; Yamada, T.; Domen, K. Nat. Mater. 2016, 15, 611. (48) Bera, A.; Wu, K.; Sheikh, A.; Alarousu, E.; Mohammed, O. F.; Wu, T. J. Phys. Chem. C 2014, 118, 28494. (49) Sano, T.; Saylor, D. M.; Rohrer, G. S. J. Am. Ceram. Soc. 2003, 86, 1933. (50) Connell, J. G.; Isaac, B. J.; Ekanayake, G. B.; Strachan, D. R.; Seo, S. S. A. Appl. Phys. Lett. 2012, 101, 251607. (51) Venables, J. A.; Spiller, G. D. T.; Hanbucken, M. Rep. Prog. Phys. 1984, 47, 399. (52) Yin, X.; Geng, D.; Wang, X. Angew. Chem., Int. Ed. 2016, 55, 2217. (53) Zhang, W.; Saliba, M.; Moore, D. T.; Pathak, S. K.; Horantner, M. T.; Stergiopoulos, T.; Stranks, S. D.; Eperon, G. E.; AlexanderWebber, J. A.; Abate, A.; Sadhanala, A.; Yao, S.; Chen, Y.; Friend, R. H.; Estroff, L. A.; Wiesner, U.; Snaith, H. J. Nat. Commun. 2015, 6, 6142. (54) Song, J.; Cui, Q.; Li, J.; Xu, J.; Wang, Y.; Xu, L.; Xue, J.; Dong, Y.; Tian, T.; Sun, H.; Zeng, H. Adv. Opt. Mater. 2017, 5, 1700157. (55) Yettapu, G. R.; Talukdar, D.; Sarkar, S.; Swarnkar, A.; Nag, A.; Ghosh, P.; Mandal, P. Nano Lett. 2016, 16, 4838. (56) Cha, J. H.; Han, J. H.; Yin, W.; Park, C.; Park, Y.; Ahn, T. K.; Cho, J. H.; Jung, D. Y. J. Phys. Chem. Lett. 2017, 8, 565. (57) Yang, Y.; Yang, M.; Moore, D. T.; Yan, Y.; Miller, E. M.; Zhu, K.; Beard, M. C. Nat. Energy 2017, 2, 16207. (58) Hoffman, C. A.; Jarašiu̅nas, K.; Gerritsen, H. J.; Nurmikko, A. V. Appl. Phys. Lett. 1978, 33, 536. (59) Zhang, D.; Yang, Y.; Bekenstein, Y.; Yu, Y.; Gibson, N. A.; Wong, A. B.; Eaton, S. W.; Kornienko, N.; Kong, Q.; Lai, M.; Alivisatos, A. P.; Leone, S. R.; Yang, P. J. Am. Chem. Soc. 2016, 138, 7236. (60) Saidaminov, M. I.; Adinolfi, V.; Comin, R.; Abdelhady, A. L.; Peng, W.; Dursun, I.; Yuan, M.; Hoogland, S.; Sargent, E. H.; Bakr, O. M. Nat. Commun. 2015, 6, 8724. (61) Verma, A. S.; Jindal, V. K. J. Alloys Compd. 2009, 485, 514.

13532

DOI: 10.1021/jacs.7b07506 J. Am. Chem. Soc. 2017, 139, 13525−13532