Enhancing the stability of the polymeric Lewis-base-assisted dual-phase 3D CsPbBr₃–Cs₄PbBr₆ perovskite by molecular engineering and self-passivation

Abstract

Inorganic metal halide perovskites have attracted attention for use in next-generation perovskite light-emitting diodes (PeLEDs) due to their excellent optical performance. However, the performance of most PeLEDs is influenced by surface defects and carrier diffusion properties. Herein, we present a facile and effective approach to form a self-ordered macromolecular intermediate phase by incorporating high molecular weight Lewis base polyvinylpyrrolidone (HM-PVP) within perovskite films. The synergistic effect of thermodynamically controlled perovskite grain growth and grain boundary passivation enables the formation of a highly cross-linked and bridged long-range-ordered polymer–perovskite composite. Furthermore, theoretical density functional theory calculations confirmed that C=O groups in HM-PVP induce a shift of the electronic cloud toward the Pb²⁺ ions, resulting in a decrease in the perovskite surface energy and favoring thermodynamically modulated perovskite growth. Significantly, silver nanoparticle incorporation into the hole transport layer improves carrier transmission efficiency in HM-7% PVP bulk 3D perovskites and quasi-2D perovskite composite devices, exhibiting luminances of 12 000 cd m⁻² and 9500 cd m⁻² and current efficiencies of 11.5 cd A⁻¹ and 15.4 cd A⁻¹, respectively. Our results demonstrate that employing a polymeric passivating agent as a Lewis base adduct thermodynamically modulates perovskite growth and improves the perovskite film's quality for achieving highly stable PeLEDs.

---

Introduction

Metal halide perovskites have emerged as a promising candidate for use in optoelectronic devices including photovoltaics,^1–3 light-emitting devices,^4–6 and photodetectors^7,8 owing to their low cost and excellent optoelectronic properties. Development of perovskite light-emitting diodes (PeLEDs) has been mainly based on the hybrid organic–inorganic perovskite (OIP) system because of its high luminance, high external quantum efficiency (EQE), and tunable optical band gap.^9–13 Despite great progress in the performance of OIP-based PeLEDs, the intrinsic moisture and thermal instability originating from the organic component remain major concerns that restrict their practical mass production. By contrast, all-inorganic halide perovskites (CsPbX₃, where X = Cl, Br, or I) have been tested and attracted considerable attention for their application in LEDs due to their high stability and optoelectronic properties, such as high thermal and chemical stability (∼580 °C for CsPbBr₃, compared with ∼220 °C for MAPbBr₃),^14,15 narrow emission wavelength (17 ≤ full width at half maximum [FWHM] ≤ 19 nm), high photoluminescence quantum yield (PLQY),^16,17 and low exciton binding energy (40 meV).^18–20 Given these advantages, CsPbBr₃, which has high stability and efficiency, has been extensively investigated.^21–25 All-inorganic halide perovskites are highly stable against humidity and under continuous operation and are thus promising building blocks for the development of sustainable applications.

Tailoring the structurally defective grain boundaries of perovskite films is essential for improving their optoelectronic properties and device stability.^26–30 Studies have demonstrated that pinholes and discontinuous surface coverage in perovskite films provide pathways for electrical short circuits, which hinder the PeLED performance.²⁹ Furthermore, most defects, such as dangling bonds and vacancies, provide charge accumulation pathways at grain boundaries and trigger degradation of charge carrier transport and device performance.^31–35 Theoretical simulations and experimental studies have demonstrated that disordered octahedral structures can form defect states and are associated with electronic trap states and nonradiative recombination centers; they thus limit the practical application of PeLEDs. Hence, high-quality perovskite films for highly efficient optoelectronic devices must fulfil several crucial criteria, including (1) modulation of perovskite crystallization kinetics to obtain a perovskite film without pinhole defects to suppress internal structural defects, (2) spontaneous passivation of surface defects at grain boundaries, and (3) intergrain cross-linking to synergistically control perovskite growth and provide a solid foundation for enhanced stability against degradation due to a hostile environment. The inter-grain polymer cross-linking of perovskite grains reduces the number of trap sites and thermodynamically regulates the grain growth resulting in lower electrical decoupling. Yang Yang et al.^27,28 reported a polymer possessing numerous bases interacting with a perovskite precursor via Lewis adduct formation to obtain several benefits. Lewis adduct formation induces electronic cloud displacement toward the perovskite. (4) Intrinsic perovskite defect suppression can be carried out through strict screening effect management of carrier and defect coulombic interaction. Comprehensive discussion of the electro–phonon coupling with charge carriers and defect behaviors within perovskites remains scarce and needs to be addressed urgently. So far, reorientation of organic dipoles and ionic lattice-induced polaronic relaxation have resulted in a screening effect that induces increased control of carrier coulombic interaction as well as defect management.^36–38 Therefore, our PVP with N functions as an extrinsic surface passivator to provide numerous and strong Lewis adduct formation sites, thus playing two keys roles: intergrain bridging and efficient grain passivation as well as a screening effect enhancer. The correlation between the screening effect and defect passivation results in suppressed electrical decoupling, a larger charge transport rate, and lower trap site density.

Defect passivation engineering has been exploited to produce highly stable and efficient PeLEDs.^39–41 Numerous reports have indicated that Lewis base additive passivation is widely considered an effective method of passivation for PeLEDs. The functional groups in Lewis base additives, such as sulfonic groups, carboxyls, and amines, have been employed to passivate halide vacancies to form Lewis acid–base adducts through coordination with undercoordinated Pb²⁺ defects.^{3,26–30,42–46} This approach suppresses nonradiative recombination, simultaneously improves the perovskite's phase stability, and enhances the device performance.^31–35,47 Noel et al. proposed that organic Lewis bases containing pyridine and thiophene groups modulate perovskite nucleation and crystallization. The nitrogen donor of pyridine, with a strong interaction effect between N atoms and Pb atoms, helps improve the quality of perovskite crystals, whereas the sulfur donor of thiophene, with weak electronegativity difference between S atoms and Pb atoms, helps manipulate perovskite's crystallization kinetics.³ Lee et al. reported that when the oxygen donor of urea was used to slow the perovskite formation process by establishing an intermediate phase between urea and undercoordinated Pb ions, nonradiative recombination traps at grain boundaries were suppressed.²⁶ Han et al. reported that the introduction of amino-functionalized 1,4,8,11-tetraazacyclotetradecane molecules in CsPbBr₃-based perovskites facilitated defect site passivation and improved PeLED performance.⁴⁵ Despite reports highlighting the suitability of small-molecule species for stabilizing the perovskite crystal structure, practical use of small molecules has been limited by the molecules’ high volatility and diffusion coefficients; thus, their application in a device would cause degradation upon exposure to a harsh environment. By contrast, the incorporation of polymeric additives into perovskite films endows the films with a stable macromolecular intermediate phase. This unique strategy provides strong cross-linking passivation agents to bridge the perovskite grains, minimizes intergrain electrical decoupling, and greatly improves device performance.^26–30

Hence, we exploit the synergistic effect of Lewis base additives including low molecular weight N-vinyl-2-pyrrolidone (SM-VP) and high molecular weight polyvinylpyrrolidone (HM-PVP) to engineer perovskite grains with Cs₄PbBr₆-induced charge carrier and exciton binding energy engineering. Compared with SM-VP, HM-PVP possesses both a highly cross-linked structure and numerous repeating units, bridging perovskite grains and resulting in a pinhole-free perovskite film. The formation of a macromolecular intermediate is facilitated and the perovskite crystallization kinetics is controlled via Lewis base adducts (Fig. 1(a)). The coordination mechanism is confirmed by density functional theory (DFT) calculations; the HM-PVP oxygen donor coordinates with undercoordinated Pb²⁺, thereby decreasing perovskite surface energy and efficiently passivating grain boundary defect sites. The synergistic effect of HM-PVP and Cs₄PbBr₆ leads to the formation of strong excitons with a lifetime as high as 28.84 ns through suppression of intergrain electrical decoupling and high environmental stability against degradation sources. As a proof of concept, we fabricated HM-7% PVP bulk 3D perovskite and HM-7% PVP quasi-2D perovskite composite devices. Champion device luminances of 12 000 cd m⁻² and 9500 cd m⁻² and high current efficiencies peaking at 11.5 cd A⁻¹ and 15.4 cd A⁻¹ could be achieved for HM-7% PVP bulk 3D perovskites and HM-7% PVP quasi-2D perovskites, respectively, by silver nanoparticle (AgNP) incorporation into the hole transport layer which effectively improved carrier transmission efficiency. All of these constitute substantial improvements over bulk 3D CsPbBr₃–Cs₄PbBr₆ PeLEDs.¹¹ Our investigation provides insight into the thermodynamically controlled perovskite grain growth with strengthened binding energy for fabrication of reliable PeLEDs.

Fig. 1 Schematic diagram for the inter-grain cross-linking structure formation. (a) Lewis based additives (SM-VP and HM-PVP) assisted the perovskite gain growth forming the micromolecular and macromolecular intermediate phase. (b) 1D GIWAXS profile of dual-phase CsPbBr₃/Cs₄PbBr₆ composites and the different concentrations of HM-PVP perovskite films. (c) Synchrotron 2D GIWAXS pattern of HM-7% PVP perovskite films (d) and (e) UV/vis absorption and PL spectra of CsPbBr₃/Cs₄PbBr₆ and the different concentrations of SM-VP perovskite films and HM-PVP perovskite films. (f) 2D pattern PLE spectra of HM-7% PVP perovskite films.

Experimental section

Materials

A pre-patterned indium-tin-oxide (ITO) substrate (5 Ω) was purchased from Lumtic Ltd. Polyvinylpyrrolidone (PVP, M_w = 55 000), N-vinyl-2-pyrrolidone (NVP), cesium bromide (CsBr, 99.95%), lead bromide (PbBr₂, 99.999%), tris-[1-phenyl-1H-benzimidazole] (TPBi), lithium fluoride (LiF, 99.995%), phenethylammonium bromide (PEABr, 99.9%), silver nitrate (AgNO₃, ≥99.0%), and dimethyl sulfoxide (DMSO, anhydrous, 99.8%) were purchased from Sigma-Aldrich and used as received. Hydrogen peroxide (H₂O₂, 35 wt%, Acros Organics), trisodium-citrate dihydrate (Na₃CA·H₂O, 99.0%, Showa Chemical Co., Ltd), and sodium borohydride (NaBH₄, >98%, Acros Organics) were purchased. Poly(3,4-ethylenedioxythi ophene):poly-(styrenesulfonate) (PEDOT:PSS, AI4803) was purchased from Clevious. Poly(4-butylphenyl-diphenyl-amine) (poly-TPD) was purchased from American Dye Source and was used as received.

Synthesis of quasi-two dimensional perovskites and polymer-based perovskite film preparation

The CsPbBr₃ precursor solution (145 mg mL⁻¹) was prepared by mixing CsBr and PbBr₂ with a desired molar ratio of 1.8 : 1 in anhydrous DMSO. The Lewis based VP monomers and PVP polymers with an appropriate 0–7 wt% were mixed with the CsPbBr₃ precursor solution and stirred at 50 °C until all components were dissolved. For the preparation of quasi two-dimensional (Q-2D) perovskites, the PEA₂Cs_n−1Pb_nBr_3n+1 Q-2D perovskite was prepared by mixing CsBr, PbBr₂, and PEABr with a desired molar ratio of 1.2 : 1 : 0.5 in anhydrous DMSO. Afterward, the 7% PVP polymer was mixed with the PEA₂Cs_n−1Pb_nBr_3n+1 precursor solution and stirred at 50 °C until all components were dissolved. For perovskite film fabrication, the precursor solution was prepared by spin-coating at 2000 rpm for 60 s followed by annealing at 80 °C for 10 min in a nitrogen-filled glovebox.

Synthesis of AgNPs

AgNPS were synthesized according to a previously published method.⁶ First, an aqueous solution of polyvinylpyrrolidone (PVP, M_w = 58 000) with trisodium citrate (Na₃CA·H₂O 75 mM, 2 mL), H₂O₂ (30 wt%, 120 μL), ethanol (99.8%, 400 μL) and silver nitrate (AgNO₃, 0.05 M, 200 μL) was stirred at room-temperature. In order to perform silver reduction, sodium borohydride (NaBH₄, 100 mM) was swiftly added to the solution which resulted in the appearance of a light-yellow solution. Ratio tuning between NaBH₄ and H₂O₂ resulted in the formation of Ag NPs of various dimensions that was observed from the change in solution color from dark yellow to red.

Perovskite LED device fabrication

Patterned ITO substrates were cleaned by using deionized water, acetone, and IPA for 20 min and dried under a flow of N₂ gas. The substrate was sequentially treated with UV-ozone for 20 min before use. The hole injection layer (HTL) of PEDOT:PSS solutions was filtered by 0.45 μm poly(tetrafluoroethylene) syringe filters and spin-coated onto the ITO substrates at 3000 rpm for 60 s, followed by annealing at 150 °C for 10 min. A series of perovskite precursor solutions with different desired ratios of Lewis based polymers were spin-coated onto the HTL by using the method describe above. TPBI (15 nm), LiF (1 nm), and Ag (80 nm) anodes were deposited by thermal evaporation under high vacuum (∼4 × 10⁻⁶ Pa). Before taking out of the glovebox, the glass cover was used for device encapsulation by using epoxy glue. The device active area was 0.2 × 0.2 cm² as defined using the overlapping area between the ITO and Ag electrodes.

Characterization

¹H NMR data were recorded at room temperature using a Bruker AM 300 (300 MHz) spectrometer and the residual proton resonance of deuterated chloroform. Transmission electron microscopy (TEM) images were recorded using a JEM-2100F field emission electron microscope equipped with a selected area electron diffraction (SAED) system operating at 200 kV. The TEM samples were prepared by dropping dilute toluene dispersions of high purity on Formvar carbon–copper grids. The morphologies of perovskites were characterized using scanning electron microscopy (SEM; Hitachi S-520) with EDX spectroscopy. Each of the samples was coated with platinum and further analysed and characterized at an acceleration rate of 15 kV. The ultraviolet-visible (UV-vis) absorption and photoluminescence (PL) spectra were recorded using a Shimadzu UV-vis spectrophotometer and a Fluorolog-3 spectrofluorometer (Horiba Jobin Yvon). The photoluminescence quantum yield (PLQY) was determined using a research grade spectrofluorometer system (FP-8500, JASCO). The Fourier transform infrared (FT-IR) spectra were recorded using a Bio-Rad 155 FT-IR spectrometer at ambient temperature in the range of 650–4000 cm⁻¹. The XRD patterns of the prepared films were characterized using a PANalytical diffractometer (X’ Pert3 Powder). The grazing-incidence wide angle X-ray scattering (GIWAXS) patterns were recorded on a beamline BL23A1 at the National Synchrotron Radiation Research Center (NSRRC), Taiwan.

The conversion from the scattering angle (θ) to the scattering vector (q) is based on the formula, Q = 4π/λ sinθ, where λ is the X-ray wavelength. Ultraviolet photoelectron spectroscopy (UPS) was performed using valence band ultraviolet radiation (He I photon energy, 21.2 eV) under a high vacuum condition of ∼8 × 10⁻¹⁰ Pa. X-ray photoelectron spectroscopy (XPS) spectra were measured through a Thermo VG-Scientific spectrometer using a monochromatic Al KR radiation source (1486.6 eV). The electroluminescence spectra were characterized using a Hamamatsu PMA-11 photonic multichannel analyzer. The device performance was determined using a Keithley 2400 source meter and a Minolta CS200 luminance meter was used to record the current–voltage and luminance–voltage characteristics. TR-PL spectra were detected using a spectrometer (iHR320, HORIBA) coupled with a Hamamatsu C10910 streak camera and an M10913 slow single-sweep unit. (NSRRC).

Computational details

All the density functional theory (DFT) calculations were performed using the projector augmented wave (PAW) method,⁴⁸ as implemented in the Vienna Ab initio Simulation Package (VASP).^49–51 The Perdew–Burke–Ernzerhof functional (PBE) within the generalized gradient approximation (GGA) was used to describe the exchange–correlation effects.⁵² The optB86b-vdW functional was used to include the van der Waals interaction.^52,53 The structural relaxation was performed until the forces were less than 0.02 eV Å⁻¹ on each atom. The convergence criterion was set to a 10⁻⁴ eV change in the energy during geometry optimization. The energy cutoff of the plane-wave basis sets was 550 eV, and the Brillion zone sampled using 6 × 4 × 6 and 2 × 4 × 1 Monkhorst–Pack K-points was employed for bulk and surface calculations, respectively.⁵⁴ One excess electron was added when electron polarons were studied, and one electron was removed from the system when hole polarons were investigated. The calculated lattice parameters for the bulk CsPbBr₃ perovskite structure are a = 8.401 Å, b = 11.784 Å and c = 7.992 Å, which are in agreement with the experimental lattice parameters (a = 8.244 Å, b = 11.735 Å and c = 8.198 Å).⁵⁵ In this study we selected the CsPbBr₃(121) surface to study the interaction between the polymer and the perovskite surface based on the corresponding XRD patterns. All the calculations were carried out on the (121) surface using a (2 × 1) slab model, in which the slab consists of four molecular layers. We calculated the surface energy for each surface termination of the CsPbBr₃(121) surface as shown in Fig. S1 (ESI†) and Table S1 (ESI†). Surface number 17 had a lower surface energy (0.056 eV Å⁻²) than that of other possible surface terminations, thereby this surface was chosen for a follow-up study. The most stable termination of the (121) surface has been confirmed to have the smallest surface energy. To avoid interaction between neighboring slabs, the vacuum region was set to be 20 Å. During geometry optimization, the bottom two molecular layers were fixed, and the top two layers were relaxed. Here, we compared the adsorption energy (E_ads) of the PVP polymer at different sites and found the most favorable interaction site.

Results and discussion

The newly designed macromolecular intermediate phases are constructed with intermolecular forces between the polymeric Lewis base and perovskite forming an intergrain cross-linking structure. The key advantages of this approach are the provision of important factors to sustain suitable interaction with the perovskite precursors and modulate the perovskite crystallization; this modulation is achieved because the polymers have greater molecular length and lower molecular chain mobility than the small-molecule additives (Fig. 1(a)). The Lewis base of PVP plays a dual role of a cross-linker and a passivation agent at the perovskite grain boundaries to enhance the moisture resistance of the perovskite film. First, PVP forms a strong Lewis acid–base adduct with the perovskite precursors, and this retards the perovskite formation process through the formation of cross-links between perovskite grains. Second, the oxygen donor of PVP contributes a lone pair electron to passivate halide vacancies and form a Lewis acid–base adduct through coordination with the undercoordinated Pb²⁺ defects.⁴⁶ Hence, Lewis bases with different repeating units, SM-VP and HM-PVP, were employed to investigate the effect of Lewis acid–base adducts on the perovskite precursor behavior, as illustrated in Fig. 1(a). Various concentrations of SM-VP and HM-PVP in perovskite films were systematically employed, and the crystal growth and crystallographic orientation were determined from GIWAXS spectra. In the 1D GIWAXS profiles presented in Fig. 1(b), the diffraction peaks at the (101), (121), (202), (141), and (321) planes correspond to the orthorhombic (Pnma space group) CsPbBr₃ phase (ICSD, No. 97851), whereas the new diffraction peaks at (012), (030), (124), and (132) correspond to the hexagonal (R3cH space group) Cs₄PbBr₆ phase (ICSD no. 25124); these findings were consistent with the XRD spectra (Fig. S2 (ESI†)) and confirmed that all planes were at preferable positions for dual-phase CsPbBr₃–Cs₄PbBr₆ composites.⁵⁶ Moreover, in comparison with the pristine perovskite and that containing other concentrations of SM-VP and HM-PVP (Fig. 1(c) and Fig. S3 (ESI†)), the incorporation of 7% PVP into the perovskite strengthened the corresponding ring intensity of CsPbBr₃(121) and Cs₄PbBr₆(132) and resulted in a highly homogeneous crystallographic orientation and formation of well-ordered crystals. This observation verified that HM-PVP improved perovskite crystallization and orientation. Fig. 1(d) and (e) presents the optical absorption (UV) and PL of the dual-phase CsPbBr₃–Cs₄PbBr₆ composites and perovskite films with various SM-VP and HM-PVP concentrations. The spectrum of the dual-phase CsPbBr₃–Cs₄PbBr₆ composite has absorption peaks at 313 and 508 nm, respectively. The strong peak at 313 nm corresponds to Cs₄PbBr₆ and is associated with the Pb²⁺ centers of the isolated PbBr₆⁴⁻ octahedra that result in ¹S₀ → ³P₁ transition; by contrast, the weak peak at 508 nm is attributed to CsPbBr₃, indicating that the two phases, namely, CsPbBr₃ and Cs₄PbBr₆, coexist, as confirmed from the XRD and UV results (Fig. 1(b) and (e)). When VP and PVP were present within the perovskite films, a new characteristic peak was obtained at 362 nm, which was assigned to the n → π* transition of the carbonyl group of VP and PVP (Fig. 1(d)).⁵⁷ The spectrum of the pristine perovskite and VP-perovskite with an FWHM of 22 nm has a wide emission peak centered at 523 nm. By contrast, the HM-PVP PL intensity is clearly increased and the FWHM is narrowed to 20 nm when the concentration is increased to 7% (Fig. 1(e)). Furthermore, Fig. 1(f) presents the corresponding 2D PL excitation pattern of HM-7% PVP and shows the narrow shape of the emission band over the excitation wavelength range of 300–450 nm; when compared with the concentrations of SM-VP and HM-PVP (Fig. S4 (ESI†)), it is evident that HM-PVP not only acts as an efficient Lewis acid–base adduct but also effectively passivates the perovskite defects. Moreover, the spectrum for HM-PVP has a band gap that is not clearly shifted compared with that for the pristine perovskite and SM-VP (Fig. S5 (ESI†)), revealing that the introduction of the polymeric additive of PVP had a strong stabilizing effect on the perovskite structure.

Fig. 2 shows the morphological characterization of HM-7% PVP perovskite composites. High-resolution transmission electron microscopy (HRTEM) images showed two lattice fringes of 0.42 and 0.32 nm corresponding to the (121) and (132) crystal planes of the dual-phase CsPbBr₃–Cs₄PbBr₆ composite, respectively (Fig. 2(a)); such findings are consistent with the XRD results presented in Fig. S2 (ESI†). Furthermore, the fast Fourier transform (FFT) pattern and selected-area electron diffraction (SAED) pattern both showed that the orthorhombic–hexagonal CsPbBr₃–Cs₄PbBr₆ phases coexisted, as shown in the inset of Fig. 2(a) and (b). We conducted energy dispersive spectroscopy (EDS) analysis of the HM-7% PVP perovskite composite for elemental mapping, as shown in Fig. 2(c) and (g). The elemental mapping of Cs, Pb, and Br atoms was performed to confirm that the CsPbBr₃–Cs₄PbBr₆ composites had a well-defined crystal structure (Fig. 2(c) and (f)). Notably, the O atomic distribution overlapped with the perovskite atomic element distribution and further expanded beyond the crystal boundaries (Fig. 2(g)), indicating that HM-7% PVP not only created a bridge and formed a robust intermediate and protective phase between perovskite grains but also efficiently preserved the material's structural integrity, as revealed in Fig. S6 (ESI†).

Fig. 2 Morphological characterization of HM-7% PVP perovskite composites. (a) HRTEM image of HM-7% PVP with the inset showing photographs of FFT patterns. (b) SAED patterns of HM-7% PVP perovskite composites. (c) HRTEM image of HM-7% PVP perovskite composites. (d)–(g) EDS elemental mapping images of HM-7% PVP perovskite composites. (h) and (i) SEM and AFM surface morphology of HM-7% PVP perovskite composites.

Fig. 2(h) and (i) and Fig. S7a–h (ESI†) show the morphology of the SM-VP and HM-PVP composites with various concentrations of VP and PVP in scanning electron microscopy (SEM) and atomic force microscopy (AFM) images. The pristine CsPbBr₃–Cs₄PbBr₆, with large pinholes and agglomerations at the perovskite grain boundaries, resulted in the formation of a discontinuous perovskite film, as shown in Fig. S7a (ESI†), whereas the HM-7% PVP (Fig. 2(h)), with a pinhole-free morphology and compact structure, effectively inhibited aggregation and fissure formation, indicating that the high molecular Lewis bases dominated the smooth surface effects and resulted in uniform perovskite dispersal within the polymeric films. These results were verified from the AFM images presented in Fig. 2(i) and Fig. S7e–h (ESI†). The pristine CsPbBr₃–Cs₄PbBr₆ had a surface with relatively high surface roughness, specifically 21 nm, which resulted in high pinhole defect density at the perovskite grain boundaries (Fig. S7e (ESI†)). By contrast, the HM-7% PVP (Fig. 2(i)), with relatively low surface roughness of 14 nm, further passivated the defects in the perovskite grain boundaries.

The mechanism of forming a stable dual-phase CsPbBr₃–Cs₄PbBr₆ composite was based on introducing a long alkyl chain PVP with a steric hindrance effect and thus preventing aggregation of the perovskite crystal. The HM-PVP passivation mechanism was further investigated by determining the solubility parameter under saturated conditions of CsBr−HM-PVP and PbBr₂−HM-PVP in DMSO solution in order to avoid the solvation effects (Fig. S8 (ESI†)). The CsBr−HM-PVP mixed solution was observed to be a turbid liquid, whereas HM-PVP−PbBr₂ was a precipitate-free and homogeneous solution, thereby demonstrating strong interaction between HM-PVP and Pb atoms.^45,46 The acylamino group of PVP donated oxygen and nitrogen lone-pair electrons and thus provided many coordination centers that could act as a Lewis base center. The resonance effect strengthened the basicity of the oxygen atom and increased its electronic cloud density, facilitating the Lewis acid–base adduct with a Pb²⁺ coordinating center.^47,58 Moreover, the long-backbone PVP, with a substantially electronegative acylamino group structure, acted as an intermediate linker to attract more perovskite moieties to the coordination center, triggering the formation and growth of a dual-phase CsPbBr₃–Cs₄PbBr₆ composite, as depicted in Fig. 3(a). To verify the potential coordination bonds between the Lewis bases (SM-VP and HM-PVP) and the perovskite, Fourier transform infrared spectroscopy (FTIR), nuclear magnetic resonance (NMR), and X-ray photoelectron spectroscopy (XPS) were performed (Fig. 3(c)–(f)). Fig. 3(e) shows absorption peaks at 1710 cm⁻¹ (C=O stretching), 1340–1508 cm⁻¹ (C–H stretching), and 1240 cm⁻¹ (C–N stretching) that corresponded to the acylamino groups of PVP. In comparison with the spectra of SM-3.5% VP and HM-3.5% PVP (Fig. S9 (ESI†) and Fig. 3(e)), the spectrum of HM-7% PVP has a large shift from 1710 to 1720 cm⁻¹ corresponding to the C=O vibration, which is ascribed to the HM-PVP having long-range-ordering polymer chains and thus providing a large degree of coordination for sharing the lone electron pair on the C=O group and further delocalizing to the empty orbital of Pb²⁺, resulting in proton-accepting hydrogen bond formation. Furthermore, the pristine PVP and HM-PVP perovskite were dissolved in deuterated DMSO-d₆ solution, after which ¹H and ¹³C NMR were performed. The characteristic resonances of C=O groups (δ = 2.5) and N atoms (δ = 3.42) corresponded to the acylamino group of PVP. The chemical shift in the spectrum of HM-7% PVP was considerable (Δδ ≈ 0.3 ppm). Additionally, a resonance peak shifted up-field (Fig. 3(f)). Moreover, the corresponding ¹³C NMR spectra (Fig. S10 (ESI†)) revealed that HM-PVP caused an up-field shift of Δδ ≈ 1 ppm in the main chemical shift of 174 ppm (C=O group) compared with the pristine PVP, indicating that the long-chain molecules had more main functional groups (C=O) than the short-chain molecules, assisting the resonance-strengthening effects between the Pb cations of the perovskite and oxygen anions of PVP.⁴⁷ The XPS spectra indicated the binding strength of Cs 3d, Pb 4f, and Br 3d orbital species and confirmed the incorporation of VP and PVP into the CsPbBr₃–Cs₄PbBr₆ perovskite (Fig. 3(c), (d) and Fig. S11 (ESI†)). The spectrum of the pristine perovskite has two peaks located at 138.64 and 143.67 eV, corresponding to the Pb 4f_7/2 and Pb 4f_5/2 orbitals, respectively. The Br 3d_5/2 peak at 68.98 eV corresponds to the Br⁺ cation region. Notably, when VP was present in the perovskite (Fig. S11 (ESI†)), the SM-7% VP spectrum showed a slight red-shift in the Pb 4f and Br 3d peak region. By contrast, for HM-7% PVP with a long-range order and well-ordered chain segments, the spectrum was shifted considerably to higher binding energy in the Pb 4f and Br 3d peak region, which resulted from an increase in the cationic charge of the Pb ions, indicating that the shifts in binding energy stabilized perovskite formation (Fig. 3(c) and (d)). We further used DFT calculation to investigate the interaction between the Lewis base additives (SM-VP and HM-PVP) and perovskites (Fig. 3(b)). The adsorption structures of VP monomer, dimer, and trimer forms were optimized. The adsorption energy (E_ads) of VP adsorption on the CsPbBr₃(121) perovskite surface was calculated:

E_ads = E_total − (E_molecule + E_surface) (1)

where E_total, E_surface, and E_molecule are the total energies of the optimized system, clean surface, and isolated VP (monomer, dimer, and trimer units) molecules, respectively. As illustrated in Fig. 3(b), the VP monomer bound to the surface of Pb atoms through the O atom of the acylamino group with a bond length of 2.55 Å, showing a coordination bonding between the VP monomer and CsPbBr₃(121) perovskite surface, whereas the bond length between Pb and O atoms shortened from 2.57 to 2.55 Å. In addition to the dative bonding, CH⋯Br hydrogen bonding occurs between the VP monomer and surface, with the CH⋯Br distance being elongated from 2.30 to 2.83 Å (Table 1 and Table S2 (ESI†)). We calculated the energies of VP dimer and trimer adsorption on the CsPbBr₃(121) surface to mimic VP polymer adsorption. The dimer and trimer adsorption energies were −1.69 and −3.02 eV, respectively, as shown in Fig. 3(b) and Table S3 (ESI†), demonstrating that the adsorption energy increased with increasing VP length. The VP trimer has more acylamino groups that can interact with the surface through Pb/Cs⋯O dative bonding and CH⋯Br interaction, resulting in higher adsorption energy compared with that of the VP monomer. Table 1 shows the lengths of oxygen bonds between the protons in the VP trimer and lead and cesium in the perovskite, which are more crucial; these lengths were calculated to be 2.47 and 3.21 Å, respectively, indicating that the shortened distance favors a shift of the electron cloud from oxygen to lead due to electronic affinity and results in a shortened interatomic distance and stronger Pb–O interaction. Because of its high steric hindrance, the binding probability between the perovskite and VP does not increase even on adding larger amount of VP units. Therefore, the stable adsorption onto perovskite structures was optimized when the VP form was a trimer system. Furthermore, Fig. S12 (ESI†) shows the electron density difference plot of the VP monomer before and after adsorption on the CsPbBr₃(121) surface, illustrating the electron accumulation between O and Pb atoms and H and Br atoms. It indicates that the VP monomer interacts with the CsPbBr₃(121) surface through O⋯Pb dative bonds and H⋯Br hydrogen bonds. The electron distribution of the coordination bond is closer to that of the oxygen atom due to the higher negativity of the oxygen atom. Moreover, we calculated the density of states before and after adsorption on the CsPbBr₃(121) perovskite surface to understand the influence of VP monomer adsorption on the electronic properties of perovskites. As illustrated in Fig. S13 (ESI†), after VP monomer or multimer adsorption, the Fermi level was slightly shifted toward the conduction band. The calculated band gap of the pristine CsPbBr₃(121) perovskite surface was 1.595 eV, and the band gap was slightly shifted from 1.595 to 1.594 eV after VP monomer or multimer adsorption. This revealed that no great change occurred in the band gap compared with that of the pristine surface; thus, the length of the VP on the perovskite surface does not affect the optical properties of the perovskite. The detailed computational information and absorption energy calculations are summarized in Table 1 and Table S2 (ESI†).

Fig. 3 PVP-passivated perovskite composites. (a) Mechanism of PVP-induced perovskite stability. (b) Binding energy of a different number of VP units adsorbed on the CsPbBr₃(121) perovskite surface via the DFT simulation. Color coding of the atoms: red, O; white, H; blue, N; light gray, C; purple, Cs; brown, Br; gray, Pb. XPS spectra of (c) Pb 4f and (d) Br 3d regions result from perovskite and the different concentrations of HM-PVP perovskite films. (e) FTIR spectra of PVP and the different concentrations of HM-PVP perovskite films. (f) ¹H NMR spectra of PVP and HM-PVP perovskite films.

Table 1 The calculated geometrical parameters (interatomic distances in Å) of the different number of VP units adsorbed on the CsPbBr₃(121) surface, number in parentheses represents the total number of distances less than 4 Å

Distance*
	d O–Pb	d H–Br	d O–Cs
Monomer	2.55	2.83–3.97 (2)	—
Dimer	2.54–2.57 (2)	3.04–3.96 (6)	—
Trimer	2.47–2.55 (2)	2.67–3.83 (8)	3.21–3.29 (2)

---

To investigate the carrier's recombination pathways at the polymer–perovskite interface upon light irradiation, we first added or removed one electron from the neutral CsPbBr₃(121) perovskite surface system to create electron and hole polarons.^59–61 Bader charge analysis demonstrated that Pb atoms gained 0.52 electrons upon electron addition, whereas Br atoms lost 0.81 electrons upon removal of an electron; hence, electron and hole polarons were mainly localized on the Pb and Br atoms, respectively, indicating that electrons were transferred from Br atoms to Pb atoms when the perovskite was illuminated. We also optimized the adsorption structures of the VP monomer, dimer, and trimer on perovskite surfaces containing hole and electron polarons (Fig. S14 (ESI†)). Table S4 (ESI†) indicates that the adsorption energies of the VP monomer, dimer, and trimer on the surface with hole and electron polarons were −0.94, −2.23, and −3.09 eV, respectively. Similar to VP adsorption on the neutral surface, VP trimer adsorption on the surface with hole polarons was more stable than VP monomer adsorption. Furthermore, we observed that the strength of VP adsorption on the surface with hole polarons was greater than that on the neutral CsPbBr₃(121) perovskite surface and CsPbBr₃ surface with electron polarons. The hole polarons were close to the VP adsorption site, whereas the electron polarons were far away, resulting in enhanced stability of the charge-separated state. The detailed calculation of charge differences for VP adsorption on both neutral and charged systems is presented in Table S4 (ESI†). In addition, we investigated the impact on the exciton recombination dynamics of CsPbBr₃–Cs₄PbBr₆, SM-VP, and HM-PVP by using time-resolved photoluminescence (TRPL). The PL decay was conducted through biexponential fitting by using the following equations:

where A₁ and A₂ are the fractional intensities, and τ₁ and τ₂ are the fast and slow decay components, respectively.^6,13 Fig. S15a (ESI†) and Fig. 4(a)–(d) show the 2D contour plots of the TRPL of CsPbBr₃–Cs₄PbBr₆ with different concentrations of SM-VP and HM-PVP. Fig. 4(d) and (e) shows that the HM-PVP perovskite film with relatively high PVP content (7%) had a longer carrier lifetime (28.84 ns) than that of the pristine CsPbBr₃–Cs₄PbBr₆ (16.47 ns; Fig. S15a (ESI†) and Fig. 4(e)) and the films with other concentrations of SM-VP and HM-PVP. This was ascribed to the uniformly compact structure of HM-PVP, which effectively passivated the perovskite surface defects (Fig. 2(h) and (i)) and increased the exciton recombination kinetics, resulting in long-term stability of the PL lifetime and possessing a high PLQY value (Table S5 (ESI†)). The improved carrier transport led to significantly improved long-term stability in comparison to that of 0D of MAPbBr₃ (∼27 ns) and CsPbI₃ (∼100 ns) or CsPbBr₃ NCs (∼1 ns) originating from a longer polymeric chain to passivate more efficiently the CsPbBr₃–Cs₄PbBr₆ perovskite structure and isolate more efficiently the Cs₄PbBr₆ phase.^62–64 Overall, perovskite electrical decoupling may occur and originate from the secondary pure organic phase formed at the interface between the perovskite and conventional passivating molecule. To minimize this drawback, our novel Lewis base HM-PVP not only mitigates the formation of an organic insulating phase at the perovskite grain interface but also helps prevent intergrain electrical decoupling.^29,33 Moreover, we also measured the exciton binding energy of CsPbBr₃–Cs₄PbBr₆, SM-VP, and HM-PVP from 80 K to 300 K based on the temperature-dependent PL spectra (Fig. 4(f)–(i)). The exciton binding energy was determined by using the following equations:^62–64where I₀ is the PL intensity at 0 K, E_b is the exciton binding energy, and K_b is the Boltzmann constant. As shown in Fig. S15b (ESI†), the pristine CsPbBr₃–Cs₄PbBr₆ shows a low E_b value of 46 meV. By contrast, the E_b value tends to increase from 46 meV to 60 meV upon the addition of SM-VP and HM-PVP into CsPbBr₃–Cs₄PbBr₆(Fig. 4(f)–(i)). Among them, HM-7% PVP possesses a high E_b value of 60 meV which is much larger than those of other concentrations of SM-VP and HM-PVP and previously reported CsPbBr₃ (40 meV). This confirms that the stronger EB of CsPbBr₃–Cs₄PbBr₆ not only originates from the polymeric passivation of the CsPbBr₃–Cs₄PbBr₆ octahedra but also from strongly confined excitons in the Cs₄PbBr₆ isolated octahedra.^64–66 In addition, we consider that the exciton binding energy is associated with the dielectric-confinement effect. The dielectric confinement effect is induced by the dielectric constant mismatch between the inorganic well and surrounding organic ligands, which in turn tunes the perovskite vicinity polarizability, thus reducing the dielectric inhomogeneity for the bulk 3D perovskite.^66–68 It means that an increase in the dielectric constant of organic cations can weaken the dielectric-confinement and then result in decreased exciton binding energy. Therefore, we speculate that dielectric constant engineering plays two key roles in altering the exciton binding energy value along with the dielectric-confinement effect. First, the major phase of CsPbBr₃ (ε = 20.67) exhibits a higher dielectric constant than that of Cs₄PbBr₆ (ε = 5.79),⁶⁶ which increases the dielectric screening of electron–hole Coulomb interaction and thus decreases the overall exciton binding energy significantly. Second, PVP possessing a low dielectric constant of 7.7 compared to that of CsPbBr₃ (ε = 20.67) and present in large amount, not only effectively passivated the bulk 3D perovskite defect density but also simultaneously regularized the dielectric confinement, thereby strengthening exciton coulombic interaction and obtaining higher exciton binding energy.

Fig. 4 2D pseudo-color plot of SM-VP, and HM-PVP. (a) SM-3.5% VP. (b) SM-7% VP. (c) HM-3.5% PVP. (d) HM-7% PVP. (e) TRPL decay curves of CsPbBr₃–Cs₄PbBr₆, SM-VP, and HM-PVP. (f)–(i) Temperature-dependent PL spectra of SM-3.5% VP, SM-7% VP, HM-3.5% PVP, and HM-7% PVP.

CsPbBr₃–Cs₄PbBr₆-based LEDs were prepared using various concentrations of SM-VP and HM-PVP, as shown in Fig. 5(a)–(c) and Table S6 (ESI†). The band structure of CsPbBr₃–Cs₄PbBr₆, SM-VP, and HM-PVP was measured using ultraviolet photoelectron spectroscopy and optical absorption spectroscopy (Fig. S5 and S16, ESI†). The valence band and conduction band energy levels of HM-7% PVP were estimated to be −6 and −3.67 eV, respectively. The electroluminescence (EL) spectra of the CsPbBr₃–Cs₄PbBr₆, SM-VP, and HM-PVP devices are shown in Fig. S17a (ESI†). The spectrum of HM-7% PVP has a narrow emission peak at 525 nm with a FWHM of 20 nm, thereby presenting stronger EL intensity than SM-VP and CsPbBr₃–Cs₄PbBr₆; the trends are consistent with the PL spectra presented in Fig. 1(e) and (f). The dual synergistic effect of hole injection enhancement and luminescence quenching suppression results in formation of high performance devices. Notably, the underlying mechanism for insufficient hole injection and luminescence quenching originates from the mismatching of HIL stacking at the PEDOT:PSS anode. To overcome this drawback, we propose to embed PEDOT-PSS with AgNPs to improve band gap matching and carrier injection efficiency. We fabricate a LED device with the following architecture: ITO/PEDOT:PSS-AgNPs/HM-7%PVP-CsPbBr₃–Cs₄PbBr₆/TPBi/LiF/Ag as the cathode (Fig. 5(a)). The current–voltage (I–V) curve of CsPbBr₃–Cs₄PbBr₆ based on SM-VP and HM-PVP devices is displayed in Fig. 5(b). The HM-PVP device had a lower current density than that of the CsPbBr₃–Cs₄PbBr₆ and SM-VP devices. The HM-7% PVP device effectively suppressed current leakage and charge injection imbalance because the perovskite structure had a uniform and compact morphology (Fig. 2(h)). The HM-7% PVP device (Fig. 5(b) and Fig. S17b, c (ESI†), and Table S6 (ESI†)) exhibited a maximum luminance of 12 000 cd m⁻², a current efficiency of 11.5 cd A⁻¹ and an EQE value of 3.17%. This device performance reflected a 4.68-, 31.08-fold, and 25.36-fold improvement compared with that of the SM-7% VP device (2560 cd m⁻², 0.37 cd A⁻¹, and 0.125%, respectively) as well as 5.78-, 41.07-fold, and 29.35-fold improvement compared with that of the CsPbBr₃–Cs₄PbBr₆ device (2075 cd m⁻², 0.28 cd A⁻¹, and 0.108%, respectively). These values are a substantial improvement from those of previously reported bulk 3D CsPbBr₃–Cs₄PbBr₆ PeLEDs (EQE value of 2.4 × 10⁻⁵).¹¹ Significantly, HM-7% PVP devices also show a longer operational lifetime (T₅₀) of 5 min than that of the CsPbBr₃–Cs₄PbBr₆ device (Fig. S17d (ESI†)). In addition, we further investigated the high concentrations of HM-PVP up to 14% through SEM morphology (Fig. S18a, ESI†) and device results (Fig. S18b–d (ESI†)). The HM-14% PVP device had a larger current density and provided substantial current leakage resulting in limiting the charge injection into PeLEDs (Fig. S18b, ESI†). These poor device results can be attributed to the non-uniform perovskite films. At higher concentrations and PVP/perovskite ratios, the pinholes and discontinuous surface coverage become much larger, providing pathways for electrical short circuits and increasing the charge injection imbalance compared to that in our optimized HM-7% PVP device (Fig. 5(b), (c), and Fig. S17b, c, S18b–d, ESI†). To analyze the molecular engineering effect between CsPbBr₃–Cs₄PbBr₆, SM-VP and HM-PVP on the trap density of states, the current density–voltage curve of CsPbBr₃–Cs₄PbBr₆ devices without and with SM-VP and HM-PVP composites are displayed in Fig. 5(d)–(f) and Fig. S18a, d (ESI†). According to the formula N_t = 2V_TFLεε₀/eL², the trap density (N_t) was calculated from the trap-filled limit voltage (V_TFL).⁶⁹ The dark current characteristic of hole-only devices was conducted under applied bias. The CsPbBr₃–Cs₄PbBr₆ devices with a high V_TFL characteristic (0.91 V) exhibited large trap density (3.46 × 10¹⁷ cm⁻³). Upon addition of 3.5% HM-PVP and 7% HM-PVP into CsPbBr₃–Cs₄PbBr₆ films, the V_TFL of the optimized HM-7% PVP devices dramatically decreased from 0.91 to 0.13 and a lower trap density (4.95 × 10¹⁶ cm⁻³) was provided compared to that of pristine CsPbBr₃–Cs₄PbBr₆ devices and SM-VP devices (Fig. 5(d)), Fig. S18a and d (ESI†), thereby demonstrating that the V_TFL characteristic is associated with the VP monomer length results in effectively inhibited trap density (Fig. 3(b), Fig. S1 (ESI†) and Table S2, (ESI†)).

Fig. 5 Schematic of SM/HM bulk 3D perovskite and HM-quasi 2D perovskite composites. (a) Schematic energy level diagram of the HM-7% PVP device structure. (b) and (c) Current density–voltage and luminance–voltage of CsPbBr₃–Cs₄PbBr₆, SM-VP, and HM-PVP, the inset displaying a photograph of HM-7% PVP operated at 3 V. (d)–(f) J–V characteristics of the hole-only device (ITO/PEDOT:PSS-AgNPs/CsPbBr₃–Cs₄PbBr₆, 3.5% HM-PVP and 7% HM-PVP/Au) (g) Schematic energy level diagram of the HM-7% PVP quasi 2D perovskite structure (h) Current density–voltage–luminance of the HM-7% PVP quasi 2D perovskite, the inset displaying a photograph of the HM-7% PVP quasi 2D perovskite operating at 3.25 V. (i) Current efficiency of HM-7% PVP quasi 2D perovskites.

In comparison to bulk 3D perovskites, layered perovskites such as the quasi two-dimensional (2D) perovskite possesses highly confined excitons that prevent charge diffusion and trapping within the active layer trap site. The dimensionality reduction of the perovskite not only leads to improved stability but also enhanced charge confinement and reduced charge trapping. According to our previous studies,⁵ the polymeric Lewis base donating strength strongly impact the Lewis acid–base adduct formation with perovskites. The energy involved influences the perovskite crystallization kinetics through an increase in the crystal formation energy barrier. The quasi 2D perovskite is encapsulated with the polymer through Lewis acid–base interaction between Pb and O from polyethylene glycol.⁷⁰ We hypothesize that the Lewis acid–base interaction occurs between Lewis acid Pb and Lewis base nitrogen atoms of PVP through adduct formation, thereby leading to perovskite grain passivation and improved quasi 2D PeLED performances.⁵ Based on the aforementioned hypothesis, the inclusion of Cs₄PbBr₆ into the HM-7%PVP-CsPbBr₃ composite helps to regulate the quasi 2D perovskite layer lattice as confirmed from the UV spectra with a characteristic peak at 316 nm as shown in Fig. S19a (ESI†). We fabricated a LED device with the following architecture: ITO/PEDOT:PSS-AgNPs/poly-TPD/PVK/HM-7%PVP-CsPbBr₃–Cs₄PbBr₆ Quasi-2D/TPBi/LiF/Ag as the cathode (Fig. 5(g)). Significantly, the luminance performance of HM-7%PVP-CsPbBr₃–Cs₄PbBr₆ quasi-2D not only increased to 9500 cd m⁻² (Fig. 5(h)) but also considerably improved the charge transport with a current efficiency and an EQE value as high as 15.4 cd A⁻¹ and 5.13%, respectively (Fig. 5(i) and Fig. S20b (ESI†)). These current efficiency and EQE values display a 1.34-fold and 1.62-fold improvement in comparison to those of the pristine HM-7% bulk 3D CsPbBr₃–Cs₄PbBr₆ device (11.5 cd A⁻¹ and 3.17%) (Fig. 5(c), Fig. S17b, c, and S20b (ESI†)). The current work shows that the synergistic effect of Cs₄PbBr₆ with PVP not only induced an improved dielectric screening with higher exciton binding energy but also highly passivated perovskite grains through PVP-CsPbBr₃ Lewis acid–base adduct chemistry. Finally, the dual effect induces a strengthened charge dielectric screening with strong dielectric confinement for enhanced perovskite device performances and lifetime.

Conclusions

In summary, we present an effective strategy to obtain an intergrain cross-linked network to improve the stability of perovskites and device efficiency by incorporating Lewis-base SM-VP and HM-PVP systems. The formation of a self-ordered macromolecular intermediate phase originates from perovskite precursor–Lewis base adduct bonds, thereby modulating the growth kinetics of polycrystalline perovskites. Through a nucleation energy barrier increase, perovskite precursor diffusion is strengthened and leads to fabrication of highly crystalline perovskite films. Furthermore, theoretical DFT calculations confirmed that the C=O group in PVP not only contributes to densification of the electronic cloud in the vicinity of the sp orbital of Pb²⁺ ions but also considerably reduces the perovskite surface energy. Grain passivation suppresses nonradiative recombination centers at perovskite grain boundaries and endows our HP-7% PVP with a long carrier lifetime (28.84 ns). Significantly, silver nanoparticle incorporation into the hole transport layer improves carrier transmission efficiency in bulk 3D CsPbBr₃–Cs₄PbBr₆ and CsPbBr₃–Cs₄PbBr₆ quasi-2D perovskites. As a result, our champion devices constructed using HM-7% PVP bulk 3D and HM-7% PVP quasi-2D perovskite composites exhibit luminances of 12 000 cd m⁻² and 9500 cd m⁻², and current efficiencies of 11.5 cd A⁻¹ and 15.4 cd A⁻¹, respectively. Our study shows that in future, highly stable and reliable PeLEDs can be constructed using polymeric passivating agents as a Lewis base adduct to modulate perovskite growth and improve the perovskite film crystallinity.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

F.-C. Liang and Z.-L. Yan contributed equally to this work. F.-C. Liang acknowledges the financial support provided by the Ministry of Science and Technology of Taiwan (contracts: MOST 111-2222-E-027-001). C.-C. Kuo acknowledges the financial support provided by the Ministry of Science and Technology of Taiwan (contracts: MOST 109-2221-E-027-114-MY3) and the National Taipei University of Technology – Shenzhen University Joint Research Program (NTUT-SZU-111-01), NTUT-SZU Joint Research Program. The authors thank Ms C.-Y. Chien and S.-J. Ji of the Ministry of Science and Technology (National Taiwan University) for their assistance in FE-TEM and EDS.

References

1. A. Green, A. Ho-Baillie and H. J. Snaith, Nat. Photonics, 2014, 8, 506–514.
2. J. You, L. Meng, T. B. Song, T. F. Guo, Y. M. Yang, W. H. Chang, Z. Hong, H. Chen, H. Zhou, Q. Chen, Y. Liu, N. De Marco and Y. Yang, Nat. Nanotechnol., 2016, 11, 75–81.
3. N. K. Noel, A. Abate, S. D. Stranks, S. P. Elizabeth, V. M. Burlakov, A. Goriely and H. J. Snaith, ACS Nano., 2014, 8, 9815–9821.
4. X. Yang, X. Zhang, J. Deng, Z. Chu, Q. Jiang, J. Meng, P. Wang, L. Zhang, Z. Yin and J. You, Nat. Commun., 2018, 9, 570.
5. J.-S. Benas, F.-C. Liang, W.-C. Chen, C.-W. Hung, J.-Y. Chen, Y. Zhou, S.-T. Han, R. Borsali and C.-C. Kuo, Chem. Eng. J., 2022, 431, 133701.
6. W.-C. Chen, C.-W. Hung, C.-H. Chang, F.-C. Liang, J.-S. Benas, Z.-L. Yan, B.-H. Lin, J.-H. Lin and C.-C. Kuo, Chem. Eng. J., 2022, 443, 136496.
7. J. Song, L. Xu, J. Li, J. Xue, Y. Dong, X. Li and H. Zeng, Adv. Mater., 2016, 28, 4861–4869.
8. Z. Yang, M. Wang, H. Qiu, X. Yao, X. Lao, S. Xu, Z. Lin, L. Sun and J. Shao, Adv. Funct. Mater., 2018, 28, 1705908.
9. L. N. Quan, R. Q. Bermudez, O. Voznyy, G. Walters, A. Jain, J. Z. Fan, X. Zheng, Z. Yang and E. H. Sargent, Adv. Mater., 2017, 29, 1605945.
10. M. I. Saidaminov, O. F. Mohammed and O. M. Bakr, ACS Energy Lett., 2017, 2, 889–896.
11. J. Xu, W. Huang, P. Li, D. R. Onken, C. Dun, Y. Guo, K. B. Ucer, C. Lu, H. Wang, S. M. Geyer, R. T. Williams and D. L. Carroll, Adv. Mater., 2017, 29, 1703703.
12. Q. Zhang, M. M. Tavakoli, L. Gu, D. Zhang, L. Tang, Y. Gao, J. Guo, Y. Lin, S. F. Leung, S. Poddar, Y. Fu and Z. Fan, Nat. Commun., 2019, 10, 727.
13. Z.-L. Yan, F.-C. Liang, C.-Y. Yeh, D. Kurniawan, J.-S. Benas, W.-C. Chen, C.-J. Cho, W.-H. Chiang, R.-J. Jeng and C.-C. Kuo, Chem. Eng. J., 2022, 444, 136518.
14. T. Matsushima, F. Bencheikh, T. Komino, M. R. Leyden, A. S. D. Sandanayaka, C. Qin and C. Adachi, Nature, 2019, 572, 502–506.
15. S. Lee, J. H. Par, B. R. Lee, E. D. Jung, J. C. Yu, D. Di Nuzzo, R. H. Friend and M. H. Song, J. Phys. Chem. Lett., 2017, 8, 1784–1792.
16. M. Kulbak, S. Gupta, N. Kedem, I. Levine, T. Bendikov, G. Hodes and D. Cahen, J. Phys. Chem. Lett., 2016, 7, 167–172.
17. Q. Van Le, M. Park, W. Sohn, H. W. Jang and S. Y. Kim, Adv. Electron. Mater., 2017, 3, 1600448.
18. L. Protesescu, S. Yakunin, M. I. Bodnarchuk, F. Krieg, R. Caputo, C. H. Hendon, R. X. Yang, A. Walsh and M. V. Kovalenko, Nano Lett., 2015, 15, 3692–3696.
19. A. Swarnkar, R. Chulliyil, V. K. Ravi, M. Irfanullah, A. Chowdhury and A. Nag, Angew. Chem., Int. Ed., 2015, 54, 15424–15428.
20. M. Kulbak, D. Cahen and G. Hodes, J. Phys. Chem. Lett., 2015, 6, 2452–2456.
21. F.-C. Liang, F.-C. Jhuang, Y.-H. Fang, J.-S. Benas, W.-C. Chen, Z.-L. Yan, W.-C. Lin, C.-J. Su, Y. Sato, T. Chiba, J. Kido and C.-C. Kuo, Adv. Mater., 2022 DOI:10.1002/adma.202207617.
22. J. Song, T. Fang, J. Li, L. Xu, F. Zhang, B. Han, Q. Shan and H. Zeng, Adv. Mater., 2018, 30, 1805409.
23. C.-L. He, F.-C. Liang, L. Veeramuthua, C.-J. Cho, J.-S. Benas, Y.-R. Tzeng, Y.-L. Tseng, W.-C. Chen, A. Rwei and C.-C. Kuo, Adv. Sci., 2021, 8, 2102275.
24. K. Lin, J. Xing, L. N. Quan, F. P. G. de Arquer, X. Gong, J. Lu, L. Xie, W. Zhao, D. Zhang, C. Yan, W. Li, X. Liu, Y. Lu, J. Kirman, E. H. Sargent, Q. Xiong and Z. Wei, Nature, 2018, 562, 245–248.
25. Z.-L. Yan, J.-S. Benas, C.-C. Chueh, W.-C. Chen, F.-C. Liang, Z.-X. Zhang, B.-H. Lin, C.-J. Su, T. Chiba, J. Kido and C.-C. Kuo, Chem. Eng. J., 2021, 414, 128774.
26. J.-W. Lee, S.-H. Bae, Y.-T. Hsieh, N. De Marco, M. Wang, P. Sun and Y. Yang, Chem, 2017, 3, 290–302.
27. L. Zuo, H. Guo, D. W. deQuilettes, S. Jariwala, N. De Marco, S. Dong, R. DeBlock, D. S. Ginger, B. Dunn, M. Wang and Y. Yang, Sci. Adv., 2017, 3, 1700106.
28. T. H. Han, J. W. Lee, C. Choi, S. Tan, C. Lee, Y. Zhao, Z. Dai, N. De Marco, S. J. Lee, S. H. Bae, Y. Yuan, H. M. Lee, Y. Huang and Y. Yang, Nat. Commun., 2019, 10, 520.
29. Y. Ling, Y. Tian, X. Wang, J. C. Wang, J. M. Knox, F. Perez-Orive, Y. Du, L. Tan, K. Hanson, B. Ma and H. Gao, Adv. Mater., 2016, 28, 8983–8989.
30. Y. Zong, Y. Zhou, Y. Zhang, Z. Li, L. Zhang, M.-G. Ju, M. Chen, S. Pang, X. C. Zeng and N. P. Padture, Chem, 2018, 4, 1404–1415.
31. D. W. de Quilettes, S. M. Vorpahl, S. D. Stranks, H. Nagaoka, G. E. Eperon, M. E. Ziffer, H. J. Snaith and D. S. Ginger, Science, 2015, 348, 683–686.
32. S. Y. Leblebici, L. Leppert, Y. Li, S. E. Reyes-Lillo, S. Wickenburg, E. Wong, J. Lee, M. Melli, D. Ziegler, D. K. Angell, D. F. Ogletree, P. D. Ashby, F. M. Toma, J. B. Neaton, I. D. Sharp and A. Weber-Bargioni, Nat. Energy, 2016, 1, 16093.
33. Y. Kutes, Y. Zhou, J. L. Bosse, J. Steffes, N. P. Padture and B. D. Huey, Nano Lett., 2016, 16, 3434–3441.
34. J. Wu, J. Shi, Y. Li, H. Wu, Y. Luo, D. Li and Q. Meng, Adv. Energy Mater., 2019, 9, 1901352.
35. L. Xu, S. Yuan, H. Zeng and J. Song, Mater. Today Nano, 2019, 6, 100036.
36. H. Zhu, K. Miyata, Y. Fu, J. Wang, P. P. Joshi, D. Niesner, K. W. Williams, S. Jin and X.-Y. Zhu, Science, 2016, 353, 1409–1413.
37. I. Anusca, S. Balčiūnas, P. Gemeiner, S. Svirskas, M. Sanlialp, G. Lackner, C. Fettkenhauer, J. Belovickis, V. Samulionis, M. Ivanov, B. Dkhil, J. Banys, V. V. Shvartsman and D. C. Lupascu, Adv. Energy Mater., 2017, 7, 1700600.
38. R. Su, Z. Xu, J. Wu, D. Luo, Q. Hu, W. Yang, X. Yang, R. Zhang, H. Yu, T. P. Russell, Q. Gong, W. Zhang and R. Zhu, Nat. Commun., 2021, 12, 2479.
39. W. Xu, Q. Hu, S. Bai, C. Bao, Y. Miao, Z. Yuan, T. Borzda, A. J. Barker, E. Tyukalova, Z. Hu, M. Kawecki, H. Wang, Z. Yan, X. Liu, X. Shi, K. Uvdal, M. Fahlman, W. Zhang, M. Duchamp, J.-M. Liu, A. Petrozza, J. Wang, L.-M. Liu, W. Huang and F. Gao, Nat. Photonics, 2019, 13, 418–424.
40. X. Liu, Z. Yu, T. Wang, K. L. Chiu, F. Lin, H. Gong, L. Ding and Y. Cheng, Adv. Energy. Mater., 2020, 10, 2001958.
41. L. Xu, J. Li, B. Cai, J. Song, F. Zhgang, T. Fang and H. Zeng, Nat. Commun., 2020, 11, 3902.
42. Y. Ke, N. Wang, D. Kong, Y. Cao, Y. He, L. Zhu, Y. Wang, C. Xue, Q. Peng, F. Gao, W. Huang and J. Wang, J. Phys. Chem. Lett., 2019, 10, 380–385.
43. Y. Tian, C. Zhou, M. Worku, X. Wang, Y. Ling, H. Gao, Y. Zhou, Y. Miao, J. Guan and B. Ma, Adv. Mater., 2018, 30, 1707093.
44. C. Xu, Z. Zhang, S. Zhang, H. Si, S. Ma, W. Fan, Z. Xiong, Q. Liao, A. Sattar, Z. Kang and Y. Zhang, Adv. Funct. Mater., 2021, 31, 2009425.
45. B. Han, S. Yuan, T. Fang, F. Zhang, Z. Shi and J. Song, ACS Appl. Mater. Interfaces, 2020, 12, 14224–14232.
46. S. Wang, A. Wang, X. Deng, L. Xie, A. Xiao, C. Li, Y. Xiang, T. Li, L. Ding and F. Hao, J. Mater. Chem. A, 2020, 8, 12201–12225.
47. B. Li, Y. Zhang, L. Fu, T. Yu, S. Zhou, L. Zhang and L. Yin, Nat. Commun., 2018, 9, 1076.
48. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758.
49. G. Kresse and J. Furthmüller, Efficiency of Ab-Initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set, Comput. Mater. Sci., 1996, 6, 15–50.
50. G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169.
51. G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49, 14251–14269.
52. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
53. J. Klimeš, D. R. Bowler and A. Michaelides, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 195131.
54. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188.
55. C. C. Stoumpos, C. D. Malliakas, J. A. Peters, Z. Liu, M. Sebastian, J. Im, T. C. Chasapis, A. C. Wibowo, D. Y. Chung, A. J. Freeman, B. W. Wessels and M. G. Kanatzidis, Cryst. Growth Des., 2013, 13, 2722–2727.
56. M. D. Bastiani, I. Dursun, Y. Zhang, B. A. Alshankiti, X.-H. Miao, J. Yin, E. Yengel, E. Alarousu, B. Turedi, J. M. Almutlaq, M. I. Saidaminov, S. Mitra, I. Gereige, A. Alsaggaf, Y. Zhu, Y. Han, I. S. Roqan, J.-L. Bredas, O. F. Mohammed and O. M. Bakr, Chem. Mater., 2017, 29, 7108–7113.
57. X. Li, Z. Wen, S. Ding, F. Fang, B. Xu, J. Sun, C. Liu, K. Wang and X. W. Sun, Adv. Opt. Mater., 2020, 8, 2000232.
58. C. Li, Z. Chen, H. Yi, Y. Cao, L. Du, Y. Hu, F. Kong, R. K. Campen, Y. Gao, C. Du, G. Yin, I. Y. Zhang and Y. Tong, Angew. Chem., Int. Ed., 2020, 59, 15902–15907.
59. W. P. D. Wong, J. Yin, B. Chaudhary, X. Y. Chin, D. Cortecchia, S.-Z. A. Lo, A. C. Grimsdale, O. F. Mohammed, G. Lanzani and C. Soci, ACS Mater. Lett., 2020, 2, 20–27.
60. L. Zhang and P. H.-L. Sit, et al. , J. Mater. Chem. A, 2017, 5, 9042–9049.
61. C. Peng, J. Chen, H. Wang and P. Hu, J. Phys. Chem. C, 2018, 122, 27340–27349.
62. M. Liu, H. Zhang, D. Gedamu, P. Fourmont, H. Rekola, A. Hiltunen, S. G. Cloutier, R. Nechache, A. Priimagi and P. Vivo, Small, 2019, 15, 1900801.
63. M. Lu, Y. Zhang, S. Wang, J. Guo, W. W. Yu and A. L. Rogach, Adv. Funct. Mater., 2019, 29, 1902008.
64. S. Sun, M. Lu, X. Gao, Z. Shi, X. Bai, W. W. Yu and Y. Zhang, Adv. Sci., 2021, 8, 2102689.
65. X. Wang, Y. Liu, N. Liu, R. Sun, W. Zheng, H. Liu and Y. Zhang, J. Mater. Chem. A, 2021, 9, 4658–4663.
66. R. Sun, N. Liu, W. Zheng, J. Zhang, N. Li, H. Lian, H. Liu and Y. Zhang, Chem. Mater., 2021, 33, 3721–3728.
67. B. Kang and K. S. Koushik, J. Phys. Chem. Lett., 2018, 9, 830–836.
68. Y. Jiang, M. Cui, S. Li, C. Sun, Y. Huang, J. Wei, L. Zhang, M. Lv, C. Qin, Y. Liu and M. Yuan, Nat. Commun., 2021, 12, 336.
69. S. Lee, J. H. Park, Y. S. Nam, B. R. Lee, B. D. Zhao, D. Di. Nuzzo, E. D. Jung, H. Jeon, J. Y. Kim, H. Y. Jeong, R. H. Friend and M. H. Song, ACS Nano, 2018, 12, 3417–3423.
70. Y. H. Chao, J. C. Chen, D. L. Yang, Y. J. Tseng, C. H. Hsu and J. Y. Chen, Adv. Funct. Mater., 2022, 32, 2112521.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2tc03690h

‡ F.-C. Liang and Z.-L. Yan contributed equally to this work.

---