CsPbBr3 Perovskite Powder, a Robust and Mass-Producible Single-Source Precursor: Synthesis, Characterization, and Optoelectronic Applications.

A facile synthesis method is proposed for the mass production of high-quality CsPbBr3 perovskite powder. It is shown that the proposed synthesis protocol is capable of producing polycrystalline CsPbBr3 powder in quantities greater than 10 g. The derived thin films by thermal evaporation and spin-coating are of compact morphologies (root-mean-square roughness < 4 nm) without voids and pinholes. Moreover, the thin films show obvious photoluminescence (PL) with a narrow (bandwidth < 19 nm) peak centered at ∼520 nm, which is blue-shifted compared with the PL emission of the powder at 542 nm. The powder and the spin-coated film exhibit superior PL stability under long-term ambient conditions and in thermal cycling experiments performed at temperatures up to ∼120 °C. Accordingly, optoelectronic applications including the fabrication and characteristics of the electroluminescence device, the organic-inorganic powder doped with methylammonium and formamidinium ions, and fluorescent greenish-blue quantum dots are also demonstrated. On the basis of these demonstrations, the synthesized CsPbBr3 perovskite powder can be expected to empower the advances in perovskite-related optoelectronics in the future.

Introduction

Since the discovery of facile synthesis and potential optoelectronic applications in 2009,[1] lead halide perovskites have become the subject of intense research worldwide for their prowess in versatile synthesis and record-breaking device performance, colloidal nanocrystal or quantum dot (QD) synthesis,[2] nanocrystal pinning or vapor treatment for film deposition,[3,4] solar cells with certified power conversion efficiency of 23.3%,[5] and green CsPbBr3 light-emitting diodes (LEDs) with an ultrahigh luminance over 590 000 cd/m2,[6] to name a few. Despite these astonishing success and achievements, it is well-known that poor stability is the key issue to hinder halide perovskites from practical commercialization.[7] To overcome this drawback, cesium cations were introduced to replace the organic part, ammonium (MA) CH3NH3+ or formamidinium (FA) CH(NH2)2+, in hybrid lead halides, and indeed, all-inorganic systems show higher thermal stability and thus better device reliability.[8,9] Similarly, sodium and tin were proposed to partially substitute MA and lead cations, respectively, in halide perovskites to improve the long-term stability of the device.[10,11]

To deposit a lead halide film, a most popular form for optoelectronic applications, MAX, FAX, or CsX (X denotes the halogen) is typically used as the source of large cations to be mixed with PbX2, the source of small cations, in peculiar protic solvent to form the precursor solution based on the simple one-step method. With a specific treatment or polymer addition, the starting precursor solution is compatible for accessing a smooth film on substrates by any solution process.[3−5] However, the exact molar ratio of the monovalent large cations (MA+, FA+, or Cs+) to the Pb cation in precursors is crucial and not as straightforward as the stoichiometric ratio in ABX3 halide perovskites. For instance, the ratio of MAX or CsX to PbX2 larger than 1.0 even 2.0 is frequently adopted to prevent the formation of impurity phases (ex: excess Pb metal or residual PbX2), which probably cause the exciton quenching in devices.[3−6] Therefore, when the precursor solutions are used for solid thin-film preparation, those unreacted and non-coordinating precursors are deposited as well and irremovable via any post-deposited treatments, no matter the simple one-step or sophisticated two-step coating methods are used.[12] Being an alternative, Hoffman et al. successfully obtained the continuous and smooth CsPbBr3 films by fusing the purified perovskite nanocrystals coated on TiO2 films at 250 °C.[13] However, the thermal treatment entirely caused the photoluminescence (PL) quenching of the annealed CsPbBr3 film. As a consequence, not only the optical and electrical but also the structural properties such as morphology, crystal grain size, and crystallinity of the thin films are affected.

In the present study, we demonstrated the facile synthesis of CsPbBr3 powder by dropping hydrohalic acid into the CsBr and PbBr2 solution at room temperature. By further precipitation–redispersion procedure, those excess unreactants can thus be removed. In this way, the synthesis of CsPbBr3 powder is capable for mass production for optoelectronics industry. From the PL measurement, we found that not only the CsPbBr3 powder but also the hereby spin-coated thin film are highly stable in ambient air and after thermal stress at temperature higher than 120 °C.

Experimental Section

Synthesis of CsPbBr3 Powder

To synthesize the CsPbBr3 powder, 1.9 mmol PbBr2 and CsBr were added in 3 mL of dimethyl sulfoxide (DMSO) and the solution was vigorously stirred for 30 min. Subsequently, 3 mL of hydrobromic acid (HBr) was added dropwise into the transparent solution with continuous stirring and the orange turbid solution was immediately observed indicating the formation of CsPbBr3. In fact, HBr was also reported to be added in the precursor solution to reduce the voids and pinholes of the spin-coated MAPbBr3 films.[14,15] After centrifuging the solution and discarding the supernatant, the precipitate was washed twice by adding certain amount of ethanol to remove unreacted agents. Finally, orange CsPbBr3 powder with a net weight of ∼900 mg was obtained by vacuum-drying the precipitate overnight. As shown in Figure S1, the product can be easily scaled up over 12 g by increasing the amounts of precursors for mass production. On the basis of our preliminary characterization, the massive amount of the powder precursor sealed in the glass bottle can be stored at room temperature in the ambient environment without degradation for over 6 months.

Film Deposition

The dry CsPbBr3 powder was then subjected to film deposition. By thermal evaporation at 5 × 10–6 Torr, a high-quality film with the thickness exceeding 100 nm can be readily deposited on various substrates such as silicon or indium–tin–oxide (ITO)/glass. Also, 290 mg (0.5 mmol) of the CsPbBr3 powder precursor was dissolved in 1.5 mL of DMSO for film deposition by solution process. It should be mentioned that unlike those strongly emissive CsPbBr3 nanocrystals dispersed in hexane or toluene, our CsPbBr3 DMSO solution is colorless and transparent under normal light irradiation (see Figure S2). Poly(ethylene oxide) (PEO) was then added into the solution in the weight ratio of CsPbBr3/PEO = 1:0.07, from which the yellowish CsPbBr3 films can be obtained by spin-coating at 3000 rpm for 80 s and baking at 60 °C for 30 min on substrates pretreated with O2-plasma.

Synthesis of CsPbBr3 QDs

To synthesize the CsPbBr3 QDs, 0.2 mmol CsPbBr3 powder was mixed with 0.25 mL of oleylamine and 0.5 mL of oleic acid in 5 mL of dimethylformamide under stirring at 60 °C for 3 h to completely dissolve the powder. After cooling down to room temperature, 1 mL of the precursor solution was injected into 10 mL of toluene to obtain the well-dispersed QDs.

Results and Discussion

Characterization of CsPbBr3 Powder and Films

The thermogravimetric analysis (TGA) and differential thermal analysis (DTA) curves of the perovskite powder are shown in Figure . The abrupt weight loss at 214 °C should be attributed to the ignition of DMSO or its related complex because of the exothermic peak shown in the DTA curve. Very low weight loss of ∼0.17% at this stage indicates that twice washing by ethanol is sufficient to remove DMSO residues in the produced powder. The onset of decomposition is at ∼510 °C because of the evaporation of PbBr2, and the obvious weight loss at ∼568 °C is due to the melting of powder.[16] The high-temperature sustainability of the powder is also a unique merit to remove those unremovable impurities by precipitation purification during synthesis. As shown, the weight loss is only 5.8% of its original weight as the temperature increases to 600 °C, indicating the high thermal stability of CsPbBr3.

Figure 1

TGA and DTA curves of perovskite powder.

From the scanning electron microscopy (SEM) image shown in the inset of Figure , the CsPbBr3 powder comprised tetragonal-shaped microcrystals with sizes ranging from a few to tens of micrometers. The X-ray diffraction (XRD) patterns of the synthesized powder and evaporated and spin-coated films are presented in Figure . The peaks at 2θ = 21.52°, 26.36°, 30.40°, 30.72°, and 34.37° are in good agreement with diffractions from (020), (022), (004), (220), and (222) planes of orthorhombic polycrystalline CsPbBr3 perovskite which features the splitting peaks at 30°–31°, without detectable secondary phases such as PbBr2 and PEO.[17] It indicates that the synthesis powder and hereby derived thin films are high-quality CsPbBr3 polycrystals and inherit the superior properties widely reported in general. More specifically, from the total X-ray photoelectron spectroscopy (XPS) spectrum of the as-synthesized powder shown in Figure S3a, the molar ratio of incorporated Cs, Pb, and Br elements is 1:0.98:3.05, highly close to the ideal stoichiometry of CsPbBr3. Low intensity of peaks for carbon and oxygen elements corresponds to minimal residues of DMSO and ethanol. In the Pb 4f core-level spectrum (Figure S3b), the peak corresponding to 4f7/2 is centered at ∼138 eV and no obvious signal was detected in the energy range for metal Pb,[18,19] which means the peak is mainly associated with the stoichiometric CsPbBr3.

Figure 2

XRD patterns of perovskite CsPbBr3 powder and the derived films. The XRD spectra indicate that the perovskites are with an orthorhombic crystal structure. Inset is the SEM image of the powder.

As shown in Figure , the optical absorption spectra of the spin-coated and evaporated films exhibit a distinct absorption peak at 515 and 516 nm, respectively, while the strong excitonic emission peaks at 519 and 522 nm with narrow full width at half-maximum (fwhm) of 17.6 and 18.9 nm are depicted in the PL spectra. The nominal shift of the absorption peak for the evaporated film compared to that of the spin-coated film maybe attributed to the small discrepancy of crystal grain sizes and uniformity,[20] which also results in the shift of the PL peak. The band gap energy of 2.356 and 2.360 eV is thus determined for the evaporated and spin-coated films from the transformed Kubelka–Munk function (see Figure S4).[21,22] The PL spectrum of the powder is also shown for comparison of the peak shift. Considering the weaker PL intensity of the powder compared to that of the thin film, the large blue shift of PL peak position is probably a result of defect-related bound exciton emission in the powder.[23]

Figure 3

Optical absorption and PL spectra of the CsPbBr3 powder and the derived evaporated and spin-coated films. Inset is an image of the CsPbBr3 powder under UV irradiation.

For lead halide perovskites, the deposition of a smooth and continuous thin film for optoelectronic applications is always challenging and never being straightforward, as the homogeneous and even morphology is the essential factor toward efficient device applications. The topographical atomic force microscopy (AFM) images of the evaporated and spin-coated thin films are measured and shown in Figure . The root-mean-square (rms) roughnesses of the evaporated and spin-coated films on Si substrates and the spin-coated ones on the ITO/glass substrate are 3.5, 2.9, and 3.1 nm, respectively, which is homogeneous and smooth enough for most optoelectronic devices. As shown, small and uniform crystalline grains can be obtained by thermal evaporation with the deposition rate lower than 0.2 Å/s. Despite the similar grain size shown in Figure b,c, the slight rougher surface of CsPbBr3 films spin-coated on ITO/glass should be attributed to the roughness (∼5 nm) of underneath ITO films.[24] Except the minor voids of spin-coated films on ITO/glass, no pin holes responsible for leading electrical short are found in all deposited films. The thickness of both evaporated and spin-coated films was determined as ∼114 nm by scratching a part of the film to create the difference of height (see Figure S5).

Figure 4

AFM images of (a) evaporated and (b) spin-coated CsPbBr3 films on silicon substrates and (c) spin-coated one on the ITO/glass substrate. The scale bar represents 1 μm.

Air Stability

As shown in Figure , the air stability of the CsPbBr3 powder on an Al foil and the spin-coated film on the Si substrate were evaluated by recording the long-term PL intensity of samples exposing in an ambient environment at 25 °C and with 50–70% relative humidity. It should be mentioned that the air exposure time-dependent PL intensity of the evaporated film was not measured because of its less PL efficiency under the irradiation of a UV (λ ≈ 400 nm) LED, the excitation source. Therefore, the characteristics of the measured film in the following discussion solely referred to those of the spin-coated film. In Figure , PL intensities exceeding 83% of the original value from the powder can be retained for up to ∼100 h. Though the intensity seems to slowly increase after ∼70 h, the fluctuation is not as significant as that from the film which rapidly increases to >140% within 1.5 h and becomes ∼198% after 100 h. The discrepancy of these results can be explained by the prepared sample condition for analysis. The powder was pressed to be a thick film with a thickness over 300 μm, whereas the thickness of a spin-coated film is only about 100 nm. That is, the effect of air exposure is quickly responded in the thin spin-coated film, whereas most of the beneath powder does not directly expose to the air and the measured PL only shows the ensemble result. On the basis of this demonstration, the CsPbBr3 powder and spin-coated film are immune to environmental degradation from air and moisture and can be directly exposed in ambient air for several days at least. On the other hand, the increase of PL intensity with the increase of air exposure time has been reported in the printed 3–4 μm-thick MAPbBr3/PEO composite film and the authors directly related this phenomenon to the efficiency improvement of derived perovskite LEDs.[25] On the basis of the increase of PL lifetime, the authors attributed the PL enhancement by air exposure to the decrease of defect density due to self-healing of the perovskite lattice upon moisture.

Figure 5

PL intensity ratio of CsPbBr3 powder and spin-coated film as a function of air exposure time.

Thermal Stability

The thermal stability of the powder and spin-coated film was also evaluated by recording the PL intensity in thermal cycling experiments. To avoid the influence of air and moisture exposure, samples were heated on a hotplate in a nitrogen-filled glovebox with both oxygen and moisture concentration at ≤1 ppm. As shown in Figure , the measurement begins at 30 °C and the temperature is increased to 120 or 140 °C and then decreased to 30 °C. As the temperature elevates, the PL intensity of the powder gradually reduced to 90% at 140 °C because of the thermal quenching effect. Noticeably, the subsequent decrease of temperature does not significantly recover the PL intensity, consistent with the observation in alkyl-phosphate-coated CsPbBr3 QDs by Xuan et al.[26] Unlike organic and inorganic lead halide nanocrystals undergoing severe PL quenching due to thermally assisted defect trapping, phase transition, or structural decomposition at high temperatures,[27,28] the CsPbBr3 powder exhibits the much stable thermal stability, even though the irreversible PL intensity. In the present research, the color and lattice structure of the powder do not have apparent difference after thermal stressing. Therefore, the superior thermal stability of the pristine CsPbBr3 powder without sophisticated decoration or treatment should be attributed to the reduction of nonradiative defects and impurities via purification in the synthesis procedure.

Figure 6

Variation of PL intensity ratio of CsPbBr3 powder and spin-coated film with measurement temperature. Arrows indicate the testing sequence.

For the CsPbBr3 film, a very different variation of PL intensity was observed. With increasing temperature, the intensity initially increases at temperatures below 90 °C and rapidly decreases to 58% at 120 °C. The initial intensity improvement of PL intensity at temperatures below 90 °C can be simply explained as the annealing effect, and the thermal quenching effect influences the PL when the temperature continues to increase. The observed variation well explains that the typical annealing temperature for the CsPbBr3 films is below 100 °C. Later, in the process of cooling down to room temperature, the remarkable enhancement of PL intensities beyond the initial values was recorded. Being a possible interpretation, it was reported that the existence of second-phase CsPb2Br5 during the synthesis of CsPbBr3 QDs could be beneficial for the device fabrication.[29] With a higher amount of halogen atoms, the CsPb2Br5 near the crystal grain boundary could prevent the excitons being trapped by the defects (the so-called self-passivation effect), and optoelectronic devices with higher efficiency could thus be achieved. Therefore, the thermally induced phase transition from CsPbBr3 to CsPb2Br5 might be responsible for the observed PL enhancement of spin-coated CsPbBr3 film during cooling. However, from the XRD spectra of the samples before and after thermal stress (see Figure S6), no apparent difference is detected probably because of a small amount of CsPb2Br5 crystal.

Fabrication and Characterization of EL Devices

On the basis of the above results, the deposition conditions of the spin-coated CsPbBr3 films were then applied to fabricate the electroluminescence (EL) devices. The device structure can be simply described as ITO/poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS)/CsPbBr3/2,2′,2″-(1,3,5-benzinetriyl)-tris(1-phenyl-1-H-benzimidazole) (TPBi)/Ca/Al where the spin-coated PEDOT:PSS (Clevios P VP 4083) and thermally evaporated TPBi (E-Ray Optoelectronics Technology Co.) served for hole injection and electron transporting, respectively. The luminance–current density–voltage (L–J–V) curves of the EL device are shown in Figure . As shown, a very low light turn-on voltage of about 3.3 V is observed accompanying with the maximum luminance of ∼3700 cd/m2 at 8.0 V. From the difference of current turn-on (2.3 V) to light turn-on (3.3 V) voltages, it is anticipated that the device performance can be further improved by inserting a hole-transport layer such as poly(N,N′-bis(4-butylphenyl)-N,N′-bis(phenyl)benzidine) or poly(9,9′-dioctylfluorene-co-N-(4-butylphenyl)diphenyl amine). In the inset of Figure , the EL spectrum is peaked at 519.5 nm, which is very close to the PL peak at 519 nm. Furthermore, the spectral width of EL spectrum is slightly broadened to 18.4 nm, which is also similar to that of PL spectrum.

Figure 7

L–J–V characteristics of EL device derived from perovskite powder. Inset is the comparison of normalized PL and EL spectra.

Demonstration of MA-, FA-Doped CsPbBr3 Powder, and CsPbBr3 QDs

Finally, to demonstrate the flexibility of the facile synthesis in this study, partial CsBr precursor was replaced with MABr and FABr in the synthesis procedure to produce MA0.3Cs0.7PbBr3 and FA0.3Cs0.7PbBr3 powder. The corresponding XRD patterns are compared with that of pristine CsPbBr3 powder and are shown in Figure , and obvious peak shifts to higher angles are observed, indicating the successful incorporation of MA and FA atoms into the lattice. The appearance of MACsPbBr3 and FACsPbBr3 powder is slightly darker than that of CsPbBr3. With the same total molar amount of precursors of large cations (CsBr, MABr, and FABr), the produced MACsPbBr3 and FACsPbBr3 powder is fewer, that is, the molar ratio of large cations to the small cation (PbBr2) has to be optimized in the synthesis of MACsPbBr3 and FACsPbBr3 powder. On the other hand, the CsPbBr3 powder was adopted to replace the CsBr and PbBr2 in the room temperature synthesis of CsPbBr3 QDs based on the protocol previously reported,[30] as mentioned in the Experimental Section. However, different to the reported green-emitting QDs, our QDs exhibit strong greenish-blue luminance under UV irradiation as shown in Figure . The emission is peaked at 492 nm with a fwhm of 31 nm.

Figure 8

XRD patterns of CsPbBr3, MACsPbBr3, and FACsPbBr3 powder. Inset is the photograph of the powder under normal illumination.

Figure 9

Absorption and PL spectra of the powder-derived CsPbBr3 QDs. Insets are the photographs of the QD solution under normal light and UV irradiation, respectively.

Conclusions

In conclusion, we have demonstrated the facile synthesis and characteristics of CsPbBr3 powder and the evaporated and spin-coated thin films. The produced amount of high-quality CsPbBr3 powder, namely the single-source precursor, can easily exceed 10 g, after twice purification to remove unreactants and unwanted side products. With proper deposition conditions, smooth and continuous films with rms roughnesses below 4 nm can be obtained by thermal evaporation or spin-coating. According to the results of long-term measurement of PL intensity in ambient air, the powder and spin-coated films are very stable without significant PL degradation for ∼100 h. Their high thermal stabilities are also demonstrated in the thermal cycling experiment. Finally, the EL device with a spin-coated CsPbBr3 emissive layer exhibits low light turn-on voltage of 3.3 V and a maximum luminance of ∼3700 cd/m2. On the basis of these demonstrations, the facile synthesis of purified lead halide perovskite powder is promising and highly viable for next-generation optoelectronics.