Optical Properties and First-Principles Study of CH3NH3PbBr3 Perovskite Structures.

Solution-processed organic-inorganic hybrid perovskites have attracted attention as light-harvesting materials for solar cells and photonic applications. The present study focuses on cubic single crystals and microstructures of CH3NH3PbBr3 perovskite fabricated by a one-step solution-based self-assembly method. It is seen that, in addition to the nucleation from the precursor solution, crystallization occurs when the solution is supersaturated, followed by the formation of a small nucleus of CH3NH3PbBr3 that self-assembles into bigger hollow cubes. A three-dimensional (3D) fluorescence microscopy investigation of hollow cubes confirmed the formation of hollow plates on the bottom; then, the growth starts from the perimeter and propagates to the center of the cube. Furthermore, the growth in the (001) direction follows a layer-by-layer growth model to form a complete cube, confirmed by scanning electronic microscopy (SEM) observations. Two-dimensional (2D)-3D fluorescence microscopy and photoluminescence (PL) measurements confirm a peak emission at 535 nm. To get more insights into the structural and optical properties, density functional theory (DFT) simulations were conducted. The electronic and optical properties calculated by DFT are in agreement with the obtained experimental values. The density-of-state (DOS) calculations revealed that the valence band maximum (VBM) consists of states contributed by Br and Pb, which agrees with the X-ray photoelectron spectroscopy valence band (XPS VB) measurements.

Introduction

Organic–inorganic perovskites in the form of thin films, microcrystals, nanoparticles, and bulk single crystals exhibit outstanding optoelectronic properties.[1] They are attractive candidates in many cutting-edge applications such as solar cells, light-emitting diodes (LEDs), lasers, and photodetectors[2,3] and are a competitive material to many standard semiconductors.[4−10] The properties of perovskites highly depend on their composition, crystallinity, and morphology. They belong to a large crystallographic family that has the same crystal structure as calcium titanate (CaTiO3)[11] and has the general ABX3, three-dimensional (3D) structural framework,[12] where A and B are cations of different sizes and X is an anion.[13] Different perovskite nanostructures, such as thin films for solar cells,[14] two-dimensional (2D) nanoplates,[15] one-dimensional (1D) nanowires,[16] and quantum dots,[17] have been studied at the microscale and nanoscale levels. Also, the trap-state density and carrier diffusion length have been investigated in bulk perovskite single crystals.[18] However, low-dimensional halide perovskites show optical and electrical properties that are different from those of bulk halide perovskites.[19] Hence, controlling the scale and shape of the synthesized perovskites is necessary for fundamental and applications research. The changes in the optical and electrical properties are attributed to the quantum size effects, large surface-to-volume ratio, and anisotropic geometry.[20] Several synthesis methods were used to prepare single-crystal CH3NH3PbX3, such as top-seed solution growth,[21] inverse temperature crystallization,[22] and antisolvent vapor-assisted crystallization.[23] Recently, researchers have been interested in the nucleation and growth mechanisms of perovskite structures prepared by the inverse temperature crystallization method, using grazing incidence X-ray diffraction (XRD) or in situ Fourier transform infrared spectroscopy. These techniques can accurately explain the crystallinity of the material and its chemical composition.[24] For example, Chen et al. have used filter paper inserted between substrate and precursor solution droplet to separate CH3NH3PbBr3 from the DMF solution, followed by crystallization mechanisms.[25] However, not many studies were conducted on the detailed growth mechanism of cubic CH3NH3PbBr3, evolution of its morphology, and optical properties followed by an in-depth analysis using first-principles methods.

In this work, CH3NH3PbBr3 microstructures were synthesized using a one-step solution self-assembly method. The morphology and the structure were analyzed using scanning electron microscopy (SEM) and X-ray diffraction. Scanning electron microscopy (SEM) and 3D fluorescence microscopy were used to explain the growth mechanism. Further, the optical properties were studied by photoluminescence (PL) and correlated to 2D fluorescence microscopy measurements. We also carried out first-principles-based density functional theory (DFT) simulations to explain the electronic and optical properties of cubic CH3NH3PbBr3 microstructures.

Results and Discussion

The scanning electronic microscopy (SEM) observations of CH3NH3PbBr3 structures show a wide range of shapes, cubes, plates, wires, and hallow cubes (Figure a), formed on a silicon substrate. The length range of the wires is from a few microns to more than 100 μm, and the width range is from a few hundred nanometers to 40 μm. Most of the wires were found to have rectangular cross sections, as shown in Figure a,b. Cubes and plates with sharp edges existed with different sizes. Also, hollow cubes appeared with a sharp edge (see Figure c). These hollow cubes are in the early crystallization stages due to the formation of agglomerate crystals, and it seems that growth starts from the perimeter and propagates to the center of the cube.[26]

Figure 1

(a) SEM images of CH3NH3PbBr3 perovskite structures: (b) plates and (c) hollow cubes.

To identify the phases of the crystalline CH3NH3PbBr3 perovskite, the powder X-ray diffraction (XRD) spectrum is presented in Figure . The XRD patterns of these microstructures exhibit a cubic phase structure of CH3NH3PbBr3 with the Pm3̅m space group symmetry. The diffraction peaks correspond to the planes (001), (002), (210), (211), (022), and (003). The cubic phase of CH3NH3PbBr3 appears in the absence of other impurities and the lattice parameter a is equal to 5.9403 Å, which is consistent with the literature.[26] We note that all of the shapes present the same X-ray diffraction because the temperature was kept constant during the growth; in contrast, Peng et al. showed that rising temperature can affect the structure, shape, and size of perovskite nanostructures.[26] Based on the diffraction peaks in the XRD spectrum, the (001) series of crystal planes contribute to the growth in the thickness direction.[27]

Figure 2

XRD patterns of CH3NH3PbBr3 microstructures at room temperature.

To explain the growth mechanism through surface evolution, several plates, cubes, and hollow cubes were observed using 2D and 3D fluorescence microscopy coupled with SEM observations. The schematic representation in Figure a presents the growth and crystallization mechanisms of CH3NH3PbBr3 structures.

Figure 3

(a) Schematic scenario of the growth process of CH3NH3PbBr3 microstructures. (b, c) SEM and 3D fluorescence microscopy images of perovskite showing hollowed interior, respectively. (d) 2D fluorescence microscopy images of CH3NH3PbBr3 plates. (e) Schematic representation of the layer-by-layer (Frank–van der Merwe) growth model of CH3NH3PbBr3 single crystal in the (001) direction.

The growth starts when crystallization occurs in the supersaturated CH3NH3Br·PbBr2·DMF precursor solution, and the CH3NH3PbBr3 molecules condense into small seeds. These CH3NH3PbBr3 seeds coalesce into bigger particles in a short time. Then, CH3NH3PbBr3 particles gradually self-assemble into a hollow structure like a hollow cage and the growth starts from the perimeter and propagates to the center (see Figure b), giving forms of hollow cubes when the growth is not finished due to the lack of crystal. These crystals are twisted, and their faces are peculiarly inclined toward each other. 3D fluorescence microscopy and SEM observations, as shown in Figure b,c, confirm the presence of hollow cubes with the formation of a hollow plate at the bottom, and then CH3NH3PbBr3 crystals accumulate in a layered stacked structure and continue to grow in the (001) direction until the final cubic single crystal is formed. Indeed, the CH3NH3PbBr3 crystals in the (001) direction were grown by a layer-by-layer model, also known as the Frank–van der Merwe growth model, until the formation of the complete cube. To explain the growth mechanism of these structures in the (100) direction, a schematic scenario is presented in Figure e. Small plates of the CH3NH3PbBr3 crystal appeared on the bottom of the substrate on the (100) facet to play a role of independent seed crystal to form the frame of the cube by self-assembly, as described in Figure e. During this step, new layers arise on the top to form cubic plates. Indeed, this growth model of the CH3NH3PbBr3 (100) facet is proceeded by a layer-by-layer assembly. In general, growing a macroscopic film needs a balance of surface energies of the substrate γB and the deposit γA and the energy of the interface γ* formed between the two (Figure e), which are controlled by the change in Gibbs free energy needed for the creation of the surface or interface.[28,29] The layer-by-layer growth will be characterized by the balance of energies that will support the increase of the area of the deposit (and the interface) over leaving an exposed substrate surface (γA + γ* < γB). This growth will result in the completion of one layer before the nucleation of subsequent layers. This proposed model was confirmed by 3D fluorescence microscopy and SEM observations, which clearly indicate the formation of layers, as shown in Figure c, as well as support the approach of Chen et al.[25] Also, it seems that these cubic plates will connect with the nearest ones to form bigger cubes and cover a larger surface. Furthermore, 2D and 3D fluorescence microscopy shows uniform green emission with the same geometrical form as shown in the SEM images. To confirm this green emission, the CH3NH3PbBr3 structures were characterized using photoluminescence (PL). The PL and the optical absorption of CH3NH3PbBr3 were studied and are presented in Figure . The optical absorption shows a peak at 523 nm. The PL emission has a sharp emission peak at 535 nm with full width at half-maximum (FWHM) equal to 21 nm. The emission wavelength at 535 nm is in agreement with the green emission from a fluorescence microscope. Furthermore, the emission peak of the CH3NH3PbBr3 structures is comparable to the emission peak measured on thin films and quantum dots by Zhang et al. and Gonzalez-Carrero et al.[30,31]

Figure 4

PL of CH3NH3PbBr3 microstructures at room temperature (green line). Optical absorption of perovskite deposed on a quartz substrate (red line).

To further understand the dependence of the intensity of the PL peak on the excitation energy densities, we investigated the emission characteristics of CH3NH3PbBr3 structures at room temperature using different laser densities, as shown in Figure . At room temperature, no notable shift in the position of the single PL peak at 535 nm was observed over a range of excitation densities. The PL intensity of the perovskite structures shows an over-linear dependence on the laser density with a slope of 0.6 in the power dependence line until the laser power of 2.01 mW (inset in Figure ). This increase in the carrier density leads to the saturation of nonradiative recombination centers, which improves the effective internal quantum efficiency.[32] Zhang et al. showed a PL dependence on laser energy with a slope of 0.7 for CsPbBr6 microdisks using a 405 nm excitation laser.[33] Furthermore, by increasing the laser power, the linewidth remains constant around 20 nm (FWHM).

Figure 5

PL emission vs laser power of CH3NH3PbBr3 microstructures. The inset graphs show PL intensity and FWHM vs laser density.

Since we have obtained the cubic phase for CH3NH3PbBr3 perovskites in our experiments, the unit cell of a simple cubic structure is considered for the calculations, and the optimized crystal structure is shown in Figure . The room-temperature crystal structure of CH3NH3PbBr3 is cubic with the Pm3̅m space group, and we have obtained a bulk lattice parameter of 5.92 Å after optimization, which agrees with the experimentally reported value of 5.94 Å.[34]

Figure 6

Optimized structure of the CH3NH3PbBr3 unit cell. The atoms are carbon (silver), nitrogen (blue), hydrogen (pink), Pb (green), and Br (brown).

The total density of states (DOS) and the projected density of states calculated for the individual atoms plotted using GGA + U are presented in Figure a at a temperature of 0 K. We can see that the main contribution close to the valence band maximum (VBM) comes from the halogen (Br) 4p states. The experimental DOS measured by the XPS VB is presented in Figure b. It can be seen that the features of the XPS VB and theoretical DOS show a good agreement over a wide energy range. The VBM also has a smaller contribution from the Pb 6s and 6p orbitals. The CH3NH3PbBr3 microstructures show a band gap of about 2.3 eV.

Figure 7

Comparison of (a) calculated total and projected density of states (DOS) of CH3NH3PbBr3 using GGA + U. (b) Experimental DOS measured by the XPS VB.

The insulating nature of CH3NH3PbBr3 is evident from the DOS, and the calculated electronic band structure is presented in Figure a. The band structure is plotted along the cubic high-symmetry points and presents a direct band gap of about 2 eV at R, which is comparable to that obtained by previous experiments, where the reported value is 2.32 eV,[35] Here, the 10% error bar is acceptable because DFT is known to underestimate the band gaps in these systems. Moreover, our DFT results provide a qualitative understanding of the conducting properties of the system.

Figure 8

(a) Band structure of CH3NH3PbBr3 plotted along high-symmetry lines for the simple cubic structure. (b) Theoretical and experimental absorption.

The theoretical and experimental absorption coefficients are presented in Figure b. The absorption coefficient is an important optical parameter that shows the extent to which light will penetrate the material before it is absorbed. The absorption coefficient depends on the material thickness and on the wavelength range of the incident light. We found that both the theoretical absorption and experimental absorption have the same band gap of ∼2.3 eV of CH3NH3PbBr3 structures.

We also calculated optical constants such as dielectric constants and refractive index (n) of CH3NH3PbBr3 to understand the optical response in detail. The dielectric constant is calculated using the expression ε = ε1 + iε2, where ε1 is the real part that represents the capacitance and ε2 is the imaginary part that represents the loss of electromagnetic energy in the system. The dielectric constant represents the polarization of the system in the presence of a static electric field. The calculated values of dielectric function are plotted in Figure (solid red line). For the real and imaginary parts, the peaks are found between 2–3 and 3–4 eV, respectively. The maximum value of static dielectric constant obtained is about 7 eV for both real and imaginary parts. For comparison, the real and imaginary parts of the modeled ε of CH3NH3PbI3 are included in Figure (dashed black line) from ref (36). The overall ε spectral shapes of these two materials look similar. However, CH3NH3PbBr3 shows a larger band gap as expected[37,38] than that for CH3NH3PbI3 with a band gap of ∼1.5 eV. The calculated refractive index is shown in Figure (blue line), which shows the largest peak similar to that of the real part of dielectric constant. The refractive index reduces for higher energy values; it is around 2.48 at 535 nm for CH3NH3PbBr3 structures, while it is 2.29 for the CH3NH3PbBr3 thin film.[39] All single crystals exhibit 2–8% higher refractive index. A higher refractive index of around 25% is also observed for the CH3NH3PbBr3 perovskite single crystals compared to the equivalent thin films.[40,41]

Figure 9

Comparison of the calculated ε1 and ε2 values of CH3NH3PbBr3 (solid red line) and CH3NH3PbI3 (dashed black line) reported in ref (36), and the calculated refractive index of CH3NH3PbBr3 (blue line).

Conclusions

In summary, perovskite CH3NH3PbBr3 microstructures were synthesized using a one-step solution self-assembly method. The morphology of these microstructures consists of a mixture of plates and cubes. The growth process of the CH3NH3PbBr3 single crystal was presented and was supported by 3D fluorescence microscopy coupled to SEM. We found that after crystallization of CH3NH3PbBr3, hollow plates are formed on the substrate, and then a layer-by-layer growth model was used grow CH3NH3PbBr3 cubes in the (001) direction. Fluorescence microscopy shows a uniform green emission, which is confirmed by the PL emission at 535 nm. The electronic band structure, density of states, and optical properties calculated using the DFT method are in good agreement with the experimental results. The shapes and the optical properties of these perovskite microstructures are promising for next-generation light-emitting devices and nanolaser structures.

Experimental Section

Synthesis of Hybrid Organic–Inorganic Perovskite CH3NH3PbBr3

The hybrid organic–inorganic perovskite under this study is CH3NH3PbBr3. The CH3NH3PbBr3 microstructures (hollow cubes, plates, cubes, and wires) were synthesized using a one-step solution self-assembly method, which has been reported in refs[42, 43]. CH3NH3Br and PbBr2 were independently dissolved in N,N-dimethylformamide (DMF) with the same concentration of 0.2 M. These two solutions were mixed at room temperature with a 1:1 volume ratio to form a CH3NH3Br–PbBr2 solution with concentration equal to 0.1 M. The diluted solution was dip-cast onto a glass or silicon substrate, which was placed on a Teflon stage in a beaker. Dichloromethane (DCM, CH2Cl2) was placed in a beaker and sealed with a porous parafilm to control the evaporation speed. After 24 h, CH3NH3PbBr3 perovskites microstructures were successfully synthesized on the silicon substrate.

Physical Characterization

The fabricated structures were then characterized using SEM and X-ray to study their morphology and crystallinity. A Jeol scanning electron microscope (SEM) operating at a 20 keV beam energy was used to analyze the structures. The X-ray powder diffraction measurements were performed on a Bruker system (D8 Avance) at room temperature and using a Cu Kα source with λ = 1.54056 Å. Room-temperature absorption measurements were carried out using a UV–vis Varian Cary 5000 spectrophotometer. PL measurements were performed at room temperatures using a Jobin Yvon LabRAM HR 800 UV system. A laser emitting at 407 nm was used as an excitation source for PL measurements. The incident laser power for the measurements was set to 5 mW. A fluorescence microscope, Olympus BX51, was used to study the structures in 2D and 3D with an excitation laser wavelength of 407 nm. X-ray photoelectron spectroscopy (XPS) studies were carried out in a Kratos AXIS Supra DLD spectrometer equipped with a monochromatic Al Kα X-ray source (hν = 1486.6 eV) operating at 45 W, a multichannel plate, and a delay line detector under a vacuum pressure of ∼10–9 mbar. All spectra were recorded using an aperture slot of 300 μm x 700 μm. Survey spectra were collected using a pass energy of 160 eV and a step size of 1 eV. A pass energy of 20 eV and a step size of 0.1 eV were used for the high-resolution spectra. For XPS analysis, samples were mounted in floating mode to avoid differential charging. Charge neutralization was required for all samples. Binding energies were referenced to the C 1s binding energy of adventitious carbon contamination, which was taken to be 284.8 eV.

Computational Methodology

We have carried out density functional theory calculations on bulk CH3NH3PbBr3 to get further insights into the experimentally observed properties employing the plane-wave pseudopotential code, Vienna Ab initio Simulation Package (VASP).[44,45] The exchange and correlation are described in the generalized gradient approximation (GGA).[46] The pseudopotentials are described using the projected augmented wave (PAW) method with Perdew–Burke–Ernzerhof (PBE) formalism.[47] A kinetic energy cutoff of 650 eV is used to expand the plane waves included in the basis set. Since it is well known that GGA underestimates the band gap of halide perovskite structures,[48] we have employed the Hubbard approximation with U parameter = 8 eV as implemented in the Dudarev approach in VASP.[49] The Brillouin zone is sampled using a Monkhorst Pack grid of 8 × 8 × 8. The energy and force relaxations were performed within tolerances of 1 × 10–6 eV and 1 × 10–3 eV/Å, respectively.