Pressure-Induced Emission in Layered Indium Double Perovskites by Synergistically Manipulating Exciton Excitation and Recombination Dynamics

Introduction

Intrinsic broadband white-light emission originated from self-trapped exciton (STE) in two-dimensional (2D) halide perovskites is in high demand for next-generation artificial white-light indoors and massive energy savings.^1–3 Single-source perovskite bulk materials that emit white-light with good stability, can overcome many inherent problems of traditional mixing phosphors or light-emitting diodes such as self-absorption effects, poor color rendition, and emission color changes over time with unequal lifetimes of the individual phosphors.^4–6 Lattice dynamic studies at both the ground- and excited states have confirmed the soft and dynamic disorder characteristics in 2D perovskite lattices, which play an important role in various photophysical properties.^7,8 The short-range Holstein-like electron-phonon interaction is strong in 2D halide perovskites with significant quantum confinement and dielectric confinement effects, which leads to stable STE formation accompanied by transient elastic distortions of the inorganic sublattice. Due to the large associated lattice deformation, the radiation recombination of the STEs can create a broadband white-light emission that spans the entire visible spectrum with a significant Stokes shift. In one-dimensional (1D) all-inorganic halide CsCu₂I₃, the excitons prefer to be trapped in large distorted [Cu₂I₆]⁴⁻ tetrahedrons, leading to a sole yellow broadband emission due to the absence of free exciton emission.⁹ Except for the intrinsic exciton self-trapping, the extrinsic exciton self-trapping associated with natural lattice defects is introduced to explain inhomogeneous emission broadening, which can create a distribution of self-trapped states and wavelength-dependent luminescence decay dynamics.^10,11 Isoelectronic Sn doping in 2D perovskite PEA₂PbI₄ causes exciton localization with a large lattice deformation around the impurities, resulting in broadband red-to-near-infrared emission.¹² However, Karunadasa’s group¹³ observed that a single ensemble nonexponential photoluminescent (PL) population decayed in a layered (110) perovskite (EDBE)PbBr₄, the only report so far. Therefore, further in-depth study of emission decay dynamics in different white-light-emitting material systems is meaningful for a basic understanding of the broadband emission mechanism.

The air-stable indium double perovskites exhibit unique STE properties and strong electron-phonon coupling that can emit a broadband warm white-light and have received the greatest research interest among various types of double perovskite materials.^14–16 Although pure In-based double perovskites possess low photoluminescence quantum yields (PLQYs) because of parity-forbidden transition and small wavefunction overlap between electrons and holes, Na alloying Cs₂AgInCl₆ shows a bright broadband warm white-light emission with a high PLQY of 86% by alloying-induced electronic landscape reconfiguration.¹⁷ By incorporating bulky organic spacer cations into three-dimensional (3D) inorganic lattices, all kinds of 2D layered double perovskites are designed and synthesized, providing a versatile platform for manipulating photophysical properties. The fluorescence of 2D double perovskites is absent or weak under ambient conditions.^18–20 In previous 2D Bi-based double perovskite studies, we observed a pressure-induced broadband orange-light emission.^21,22 This phenomenon inspired us to infer that 2D In-based double perovskites might be excellent white-light emitters similar to their 3D counterparts by enhancing crystal rigidity or applying lattice compression. Halogen substitution in the organic spacers is a suitable way to improve the stability and rigidity of 2D double perovskites. This characteristic has been confirmed to result from halogen–halogen interactions that could stabilize the lattice of thermodynamically unfavored iodide phases of 2D double perovskites.^23,24 Hence, the combination of molecular and pressure engineering offers a viable way to explore the modification of optical properties and the associated STE dynamic mechanism.

Herein, we selected 3-bromopropylammonium (Br-PA) as an organic cation to construct 2D layered double perovskite, (Br-PA)₄AgInBr₈. The dark STEs dominated the optical properties of (Br-PA)₄AgInBr₈, corresponding to the absence of fluorescence. However, strong warm white-light emission with a long carrier lifetime (μs scale) was achieved by reconfiguring the electronic landscape and transient lattice distortion at the excited state using high-pressure techniques. The wavelength-independent PL decay indicated a time-averaged single ensemble luminescence dynamics, observed for the first time in double perovskites. To comprehensively reveal the physicochemical mechanism of the STE transformation from dark to bright state and the dynamic process of charge carriers, first-principles calculation as well as detailed high-pressure optical and structural characterizations were conducted, including steady-state PL and absorption spectra, time-resolved photoluminescence (TRPL) spectra, femtosecond transient absorption (TA) spectra, Raman spectra, and X-ray diffraction (XRD) measurements.

Experimental Methods

Sample preparation

Preparation of (Br-PA)₄AgInBr₈. First, 0.5 mmol of InBr₃, 0.5 mmol of AgBr, and 2 mmol of Br-PABr were dissolved in a 3 mL HBr solution. Next, the precursor solution was heated and stirred at 100 °C for 15 min until a clear solution was formed. Colorless plate-like crystals were obtained as the solution slowly cooled to room temperature. Finally, the formed crystals were collected by filtration and washed with 10 mL diethyl ether. The preparation process of (Br-PA)₄AgBiBr₈ was the same as (Br-PA)₄AgInBr₈, except for replacing InBr₃ with BiBr₃.

High-pressure generation

A symmetric diamond anvil cell (DAC) with 400 μm flats was used to create a high-pressure environment. A preindented T301 steel gasket with a thickness of 50 μm was laser-drilled to form a hole with a diameter of 150 μm to serve as the high-pressure sample chamber. The ruby fluorescence technique was used to determine pressure in the high-pressure sample chamber. We used silicone oil as a pressure-transmitting medium (PTM) for the high-pressure PL, ultraviolet–visible (UV–vis) absorption, time-resolved PL, femtosecond TA, and powder X-ray diffraction (PXRD) experiments. Argon was used as PTM for the Raman experiments.

In situ high-pressure optical measurements

High-pressure PL measurements were carried out using an optical spectrometer and a 355 nm excitation laser with a power of 10 mW. High-pressure UV–vis absorption measurements were detected using a deuterium-halogen light source. Data acquisition was achieved using an Ocean Optics QE65000 spectrometer (Ocean Optics; Florida, United States) in high-pressure PL and absorption measurements. Time-resolved PL experiments were measured by a homemade setup with an excitation source of 375 nm pulsed diode laser (LDH-P-C-375B, 40 ps). A homemade pump–probe system coupled with a regeneratively amplified Ti:sapphire laser system (Coherent Legend Elite HE+USP-1K–III, 35 fs, 1 kHz) was used to measure the femtosecond TA spectra. High-pressure Raman measurements were performed using a 785 nm and 10 mW excitation laser and recorded by a Raman spectrometer (iHR 550, Symphony II; Horiba Jobin Yvon, Kyoto, Japan).

In situ high-pressure X-ray measurements

High-pressure PXRD experiments were performed at the 4W2 High-Pressure Station in Beijing Synchrotron Radiation Facility (BSRF), Beijing, China. The diffraction data was recorded using a Mar-345 CCD detector (MarXpert GmbH; Munich, Germany). The obtained Bragg diffraction rings were integrated into a 1D profile using the FIT2D program ( https://www.esrf.fr/computing/scientific/FIT2D/). Material Studio program ( https://www.3ds.com/products/biovia/materials-studio) was performed to refine the crystal structure.

Density functional theory (DFT) calculations

Geometry optimization and electronic structure simulation are both performed in the CP2K software package ( https://www.cp2k.org/) for 2×2×1 supercell with 584 atoms, where the lattice parameters were fixed to the experimental values at 1 atm and 8.3 GPa pressure, respectively. The molecularly optimized (MOLOPT) Godecker–Teter–Hutter (GTH) double-ζ valence single polarization short-ranged (DZVP-MOLOPT-SR-GTH) basis set, GTH pseudo-potential and Perdew-Burke-Ernzerhof (PBE) functional are adopted in the calculation, along with the grid cutoff of 550 Ry. The relaxation of atomic positions and electronic structure properties of excited states were performed by utilizing a Restricted Open-shell Kohn-Sham (ROKS) framework for singlet excited state, in which self-interaction correction (SIC) with the empirical parameter α = 0.3 is adopted to mitigate the self-interaction error of PBE functional. The organic skeleton was fixed during the relaxation for simplification.

Results and Discussion

The diversity of crystal structure and chemical components of 2D layered double perovskites offer favorable opportunities to manipulate their optoelectronic properties and electronic structure by controlling the assembly of organic and inorganic building blocks. The 2D Ruddlesden-Popper-type perovskite (Br-PA)₄AgInBr₈ crystallizes in a monoclinic space group P2/m, forming colorless thick plates.²⁵ The alternating [InBr₆]³⁻ and [AgBr₆]⁵⁻ octahedra are fully ordered and corner-shared by the bridged Br atoms in the inorganic monolayer. The inorganic monolayer is separated by the bilayered Br-PA organic cations in the out-of-plane direction, forming a distorted sandwich-like structure due to the tilting of the inorganic units, where the [InBr₆]³⁻ units maintain a nearly regular octahedral geometry compared to the axial compressed [AgBr₆]⁵⁻ octahedra with elongated equatorial Ag-Br bonds and short axial Ag-Br bonds ( Supporting Information Figures S1 and S2). At ambient conditions, the absorption spectrum of Br-PA₄AgInBr₈ displays a steep absorption edge located at 370 nm, corresponding to a large band gap and the colorless feature of crystals ( Supporting Information Figure S3). However, its fluorescence could not be detected, indicating the existence of dark STE states with a nonradiative transition in Br-PA₄AgInBr₈, consistent with earlier reported 2D In-based double perovskite analogs.^19,26 This phenomenon indicates that merely increasing a little rigidity of the crystal structure makes it difficult to achieve highly efficient exciton emission in 2D In-based double perovskites through organic molecular engineering. Reconfiguring electronic states and the dynamic process of bound excitons are particularly important for exploring the efficient emission of 2D In-based double perovskites.

Reconfiguring the electronic landscape is meaningful to investigate the possibility of efficient luminescence in 2D In-based double perovskites as potential optoelectronic materials. Therefore, we first carried out in situ high-pressure steady-state PL spectrum measurements for (Br-PA)₄AgInBr₈ and (Br-PA)₄AgBiBr₈ single crystals under the same test conditions to trace the effect of pressure and metal composition on fluorescence properties. Upon compression, (Br-PA)₄AgInBr₈ exhibited a broadband emission that centered at around 725 nm and spanned the visible and near-infrared regions from 450 to 1000 nm with a large Stokes shift of 380 nm and a full width at half-maximum (FWHM) of 255 nm at 0.5 GPa, which are the typical features of the STE emission in various 2D perovskites with distorted inorganic frameworks (Figure 1a).^1,17 With a further increase in the applied pressure, the PL intensity displayed a significant increase process accompanied by a continuous blueshift of the emission center and decreased FWHM before 8.0 GPa ( Supporting Information Figure S4). Compared with a slight increase of the PL intensity in the low-pressure region (>3 GPa), the sharp PL enhancement process was mainly concentrated in a small pressure range from 3 to 8.3 GPa, indicating that lattice precompression was crucial to boosting the radiative recombination of the STEs. Upon further compression, the PL intensity entered a progressive weakening process in a large pressure range, while the FWHM of the emission spectra increased, and the emission center slightly redshifted with a small fluctuation. Strong emission still existed at 34.1 GPa, the applied highest pressure for the measured PL spectra. As a comparison, the overall evolution trend of the emission of (Br-PA)₄AgBiBr₈ was consistent with that of (Br-PA)₄AgInBr₈ under high pressure ( Supporting Information Figures S5 and S6). However, the pressure effect on emission intensity and center shift in (Br-PA)₄AgInBr₈ was significantly stronger than that of (Br-PA)₄AgBiBr₈. A series of PL micrographs under different pressures confirmed the considerable changes in emission brightness and color, which intuitively reflected their evolution differences between (Br-PA)₄AgInBr₈ and (Br-PA)₄AgBiBr₈ double perovskites (Figure 1b,c). Compared with the orange emission of (Br-PA)₄AgBiBr₈, (Br-PA)₄AgInBr₈ could emit white-light, with an intensity of one order of magnitude higher (Figure 1d). The chromaticity coordinates of the emission of (Br-PA)₄AgInBr₈ were continuously tuned and recorded in Supporting Information Figure S7, corresponding to the fluorescence color transition from light orange to white. The strongest emission obtained a larger contribution from the red region of the visible spectrum at 8.3 GPa, bringing about a “warm” white-light with the International Commission on Illumination (CIE) coordinate of (0.44, 0.50) and a correlated color temperature (CCT) of 3590 K, suitable for indoor lighting applications. This result indicated that layered In-based double perovskites exerted much better luminescence properties and application potential than their Bi-based analogs ( Supporting Information Figure S8). Therefore, exploring the physicochemical mechanisms of efficient emission of (Br-PA)₄AgInBr₈ was crucial for understanding the root causes of the differences and further material design.

• Download figure
• Download PowerPoint

Large tunability of the optical properties derived from a considerable response of the crystal structure to pressure. To understand the source of the pressure-induced white-light emission and structure-property relationships in 2D (Br-PA)₄AgInBr₈, in situ synchrotron high-pressure powder XRD experiments were conducted with the applied pressure up to 28 GPa ( Supporting Information Figure S12). As the pressure increased, all diffraction peaks continuously shifted to a larger 2θ range. The sudden appearance and disappearance of diffraction peaks were not observed during the compression processes, implying that the crystal structure always maintained initial symmetry without structural phase transition. However, we noted that the relative intensity of the (110) diffraction peak increased as a function of pressure, accompanied by a decrease in the (111) reflection until they reached an equivalent intensity at ∼3 GPa ( Supporting Information Figure S13), reflecting that the changes in the relative position between atoms and the internal distortion of the octahedra. In addition, the broadening and weakening of anisotropic diffraction peaks were obvious with increasing pressure, implying pressure-induced lattice distortions, associated with the inorganic octahedral framework and decreased crystallinity.^27–31 As the pressure was increased to 10 GPa, (Br-PA)₄AgInBr₈ was partially amorphous, evidenced by the remaining two diffraction broadbands. Under higher pressure, however, the noted disappearance of all diffraction peaks confirmed the formation of a completely amorphous phase originating from the disordered arrangement of inorganic octahedra with highly distorted organic cations. However, the strong emission of (Br-PA)₄AgInBr₈ still existed in the amorphous phase, indicating that its luminescence properties had a high defect tolerance. Upon decompression, the sample maintained a partially amorphous phase with the absence of long-range order, as reflected by weak and broadened diffraction patterns ( Supporting Information Figure S14). Raman spectroscopy was employed as a more sensitive tool to detect the bond vibrations associated with the local structural distortions and the deformation of the organic moieties in organic-inorganic hybrids than that of XRD. In the low-frequency range of the Raman spectra, a dominant vibrational mode at 166.4 cm⁻¹ was observed and assigned to the stretching vibrations of the [AgBr₆]⁵⁻ and [InBr₆]³⁻ octahedra with the A_1g symmetry, while other two weak vibrational modes were attributed to the stretching (E_g, 123 cm⁻¹) and breathing (T_2g, 107.6 cm⁻¹) vibrations of the inorganic octahedra at ambient conditions.^32,33 The redshift of the vibrational mode confirmed the softening of the crystal lattices by inserting elastic organic layers compared with its 3D analogs ( Supporting Information Figure S15).³⁴ Upon compression, the A_1g mode splitting that creates a new peak at the low-frequency shoulder illustrated a discontinuous evolution process of the inorganic octahedra at 2.5 GPa. At the same time, the synchronous transformations of the organic moieties in the high-frequency range were monitored, demonstrating that the distortion of organic molecules began to appear with the progressive lattice hardening upon compression, consistent with the changes observed with the XRD pattern. Upon decompression, the Raman spectra returned to their original state, confirming the reversibility of the local structure ( Supporting Information Figure S16).

The sandwiched hard-soft superlattices with discrepant compressibility between the organic and inorganic layers led us to infer that there occurred a unique structural evolution in the hybrid system by applying hydrostatic pressure. The unit cell volume exhibited a continuous compression with a bulk modulus (B₀) of 20.43 GPa, confirming the high compressibility and soft features of 2D double perovskites ( Supporting Information Figure S17). Considerable anisotropic compressibility in different directions was observed. In the initial compression stage, lattice contraction was dominated by the quasi-uniaxial shrinkage of the flexible organic layers along the out-of-plane direction. The elastic organic moiety as a buffer layer could absorb most of the compressive stress by deformation and shrinkage.³⁵ With increasing pressure, the hardness of the organic moiety increased and reached a comparable magnitude to that of the inorganic moiety, in turn, an approximate isotropic compression occurred accompanied by severe lattice distortions.^36,37 The refinement results suggested that lattice contraction resulted in significant effects on the deformation of the inorganic framework. The Ag–Br–In bond angle within the inorganic layer increased sharply before 3 GPa, followed by a slow increase under higher pressures ( Supporting Information Figure S18), implying that the distortion of the inter-octahedra in the inorganic monolayer was decreased. On the contrary, the distortion of the intra-octahedra increased significantly before 3 GPa, confirmed by the reduced apical Br–In–Br (Br–Ag–Br) band angle in the out-of-plane direction that mainly contributed to the changes in the relative intensity between (110) and (111) reflections. The organic cations interacted with the inorganic layers through hydrogen-bonding networks between the NH₃⁺ groups and the Br⁻ ions, where the hydrogen bonds between N and apical Br, were the strongest. Upon compression, the shrinkage of the organic layer in the out-of-plane direction would increase the penetration of the NH₃⁺ groups into the cavities between the inorganic octahedra,^38–41 while the contraction in the inorganic layer and enhanced hydrogen bonds hindered the penetration of the organic spacer, which, in turn, resulted in a large intraoctahedral distortion. As the compression of organic molecules reached a saturated state, the denser octahedral packing and the synergistic compression of the inorganic and organic parts slowed down the intraoctahedral distortion speed accompanied by the intensification of organic molecular twisting.

To elucidate the photophysical mechanism governing carrier dynamics and pressure-driven STE transformations from dark to bright states in (Br-PA)₄AgInBr₈, we conducted in situ high-pressure TRPL experiments. Initially, before reaching 3 GPa, the PL decay signal was undetectable, correlating with weak PL results and a carrier recombination process dominated by ultrafast nonradiative transitions ( Supporting Information Figure S19). Upon further compression, a longer-lived tail of the PL decay emerged and prolonged gradually with increasing pressure, indicating the dark STEs turned into bright STEs, consistent with the observed pressure-induced emission phenomenon (Figure 2a).^42,43 The behavior of a fast initial drop, followed by a slow decay in PL intensity was also observed in 3D double perovskites.^42,44,45 Notably, an ultrafast component, with a rapid intensity drop within 3 ns, persisted throughout the compression process (Figure 2b). The lifetime of the ultrafast component was affected by pressure with an effective value of 0.4 ns at 7.1 GPa ( Supporting Information Figure S20), but its evolution process under high pressure could not be traced due to the limitation by instrument response, where more details can be given by femtosecond TA technique discussed below. The ultrafast component of the PL decay is attributable to the nonradiative recombination of the photoexcited carriers trapped in volume defects, deemed a subpopulation f_D of the dark STE states.^45,46 The proportion of the ultrafast decay f_D is significantly decreasing during the whole compression process, indicating that the ultrafast nonradiative transitions were effectively suppressed by applying appropriate pressure, which, in turn, promoted a transformation from the dark STE states to the bright STE states. The slow decay process of the slow component was fitted by a single exponential function with a pressure-induced prolonged lifetime of 0.67 μs at 3.7 GPa to 3.02 μs at 6.2 GPa (Figure 2c).²² When the applied pressure exceeded 7 GPa, a new transition pathway for the STEs emerged, characterized by an intermediate lifetime ( Supporting Information Figure S21). Therefore, the PL decay tail was fitted with a biexponential function. The lifetime of the long-lived component decreased progressively with increasing pressure, while the lifetime of the intermediate component increased dramatically before 15 GPa, followed by a gradually decreasing process upon further compression ( Supporting Information Figure S22). In addition, compared with the long-lived component, the proportion of the intermediate-lived components increased rapidly and reached a plateau of ∼15 GPa (Figure 2d). This additional channel was attributed to trapping processes associated with surface defects resulting from pressure-induced amorphization of crystals, leading to significant nonradiative losses that compete with the efficient radiative decay of bright STE states.²² For these two types of layered double perovskites, the long lifetime of the slow component originating from the bright STEs increased under high pressure, but (Br-PA)₄AgInBr₈ was one order of magnitude longer than that of (Br-PA)₄AgBiBr₈ ( Supporting Information Figures S23 and S24). Noteworthy, the two fast processes derived from lattice defects in (Br-PA)₄AgBiBr₈ always existed during the entire compression process. The suppression of the ultrafast component and the absence of the intermediate fast component in (Br-PA)₄AgInBr₈ created favorable conditions for efficient STE emission before 7 GPa, providing a distinct advantage.

• Download figure
• Download PowerPoint

To gain a deeper understanding of the dynamics of dark and bright STE states and the ultrafast photophysical processes in (Br-PA)₄AgInBr₈, the femtosecond TA technique was employed. At ambient conditions, (Br-PA)₄AgInBr₈ exhibited a broadband positive photoinduced absorption (PIA) across the probe region from 450 to 1000 nm, which provided direct evidence for STE formation, corresponding to the dark STE states (Figure 3a).^42,47,48 The STE formation time was determined to be 320 fs according to the ultrafast rise of the PIA signal at different wavelengths with identical patterns, indicating there was no potential barrier between the free excitons and STEs ( Supporting Information Figure S25a). The decay of the PIA band was traced by performing the global fitting with three different processes: an ultrafast lifetime of τ₁: 0.27 ps, an intermediate lifetime of τ₂: 4.47 ps, and a long lifetime of τ₃: 6.95 ns. The ultrafast component was attributed to hot-carrier relaxation, a phenomenon previously observed in 3D double perovskite nanocrystals characterized by ultrafast hot-carrier cooling (>1 ps).^49,50 The hot carriers make partial contributions to the PIA signal, and its ultrafast relaxation processes compete with the carrier trapping processes arising from the lattice defects, favorable to the light-emitting devices. The intermediate component was assigned to the intrinsic carrier trapping derived from the volume defects, corresponding to the ultrafast component in TRPL.^42,51 The long-lived component was ascribed to the nonradiative recombination of the dark STEs.⁴² Upon compression, the strong broadband PIA reemerged and maintained the same region as the ambient phase, corresponding to the bright STEs, but its PIA decay was significantly slower than that of the initial dark STE states (Figure 3b). The PIA decay process could be fitted by four components. The lifetime of the initial two fast components was τ₁: 0.31 ps and τ₂: 5.02 ps, respectively. The improved lifetime of the hot-carrier relaxation suggested that pressure can be an effective way to slow down the hot-carrier cooling and assist in exploring the hot-carrier cooling mechanisms, such as pressure-induced phonon bottleneck effect in Sr₂IrO₄.⁵² The additional fast component possessed a middle lifetime of τ₃: 121.3 ps, attributed to surface defects trapping.^22,42,49 The ultralong lifetime τ₄ (11.03 ns) was attributed to the radiative recombination of the bright STEs, as reflected by the long-lived component of TRPL spectra at high pressure and evidenced by the pressure-induced bright STEs. The common feature was a long lifetime for the dark and bright STEs, demonstrating that the STE lifetime was not dependent on its bright or dark state. The bright STE formation time (360 fs) was longer than that of the dark STEs, ascribed to the reduced electron-phonon coupling strength due to lattice contraction. The formation process of the STEs in (Br-PA)₄AgInBr₈ is significantly faster compared to its 2D Ag-Bi double perovskite analogs (400 fs) ( Supporting Information Figure S25b),²² proving that the 2D Ag-In double perovskite was more favorable to form stable STEs, responsible for strong warm white-light emission. The same decay curves at different detected wavelengths confirmed that the PIA signal stems from the same excited state ( Supporting Information Figure S26).^53,54

• Download figure
• Download PowerPoint

To offer a fundamental picture of the photophysical mechanism in this complex system and understand pressure effects on the electronic landscape, we performed DFT calculations of the electronic band structure. Previous works demonstrated an inversion symmetrically-induced parity-forbidden transitions in 3D Ag-In double perovskites with a direct bandgap primarily responsible for the inferior absorption and PLQY behaviors.⁵⁵ Compared with its 3D counterparts, the octahedral distortions and monolayered inorganic framework originated from a dimensional reduction in (Br-PA)₄AgInBr₈ engendered significant changes in the electronic band structure. The first change was that the crystal inversion symmetry was broken due to the octahedral deformation, which eliminated the possibility that the parity-forbidden transition dominated the absent emission and nonradiative recombination process of the STEs. The second important change was the direct-to-indirect bandgap transition associated with the inorganic monolayer lattice (Figure 4a,b). The conduction band minimum (CBM) of (Br-PA)₄AgInBr₈ was mainly derived from In s and Br p orbitals, while the valence band maximum (VBM) was composed of Ag d and Br p orbitals. The organic layers made a negligible contribution to the CB and VB edges, which served as barriers that weakened the interactions between adjacent inorganic layers. An evident feature of the band edges was the lack of electronic band dispersion with a small bandwidth, indicating heavy effective masses for electrons and holes. The photoexcited electrons in (Br-PA)₄AgInBr₈ were strongly confined in isolated [InBr₆]³⁻ octahedra, whereas the generated holes were mainly confined in the In-Br plane of a single [AgBr₆]⁵⁻ octahedron perpendicular to the a-axis direction, corresponding to an Ag-to-In (metal-to-metal) charge transfer (MMCT) character (Figure 4c). The 0D electronic dimensionality of (Br-PA)₄AgInBr₈ satisfied the prerequisite for forming stable STEs.⁵⁶ However, highly localized electrons and holes that were distributed in two different octahedra resulted in an inferior spatial overlap of the electron-hole wavefunctions and weak charge interactions.^17,56 The transition charge density between ground and excited states and the radiative recombination rate was proportional to the transition dipole moment (TDM). The calculated nonzeroTDM along different directions confirmed parity-allowed transitions, but the weak intensity stemmed from the poor orbital overlap between electron and hole, suggesting weak edge-to-edge transitions and oscillator strength associated with the optical absorption coefficients, thus accounting for the nonfluorescent behavior and the absence of absorption tail with an indirect bandgap characteristic. Upon compression, the calculated TDM was significantly improved, indicating the tuned intra- and inter-octahedral configuration, and the lattice contraction enhanced wavefunction overlapped and optical activity of the STE state, in turn, resulted in the promoted transition charge density and radiative recombination rate of carriers. The absorption tail appeared and gradually increased as a function of pressure, which further confirmed the increase of TDM ( Supporting Information Figure S10). To investigate the fundamental sources of stronger emissions in 2D Ag-In double perovskite and compare with the 2D Ag-Bi double perovskite analogs, we replaced In atoms in (Br-PA)₄AgInBr₈ with Bi atoms at 8.3 GPa, then optimized the crystal structure and calculated the electronic band structure and TDM ( Supporting Information Figure S27). Despite the indirect bandgap being reproducible in (Br-PA)₄AgBiBr₈ without spin–orbit coupling, the TDM of (Br-PA)₄AgBiBr₈ was much smaller than that of (Br-PA)₄AgInBr₈, confirming the significant advantage of In element for efficient STE emission.

• Download figure
• Download PowerPoint

Large elastic lattice distortions in the excited state originating from strong electron-phonon coupling play an important role, arising from Stokes-shifted broadband STE emission. To investigate pressure effects on the excited state, the configurational coordinate diagram (CCD) was constructed to describe the formation of the STEs based on the energetics. At ambient conditions, we observed an intersection near the STE configuration between the excited- and ground-state curves, which meant that the photoexcited electrons and holes would nonradiatively, recombine via this interaction point, emitting some phonons (Figure 4d).^17,56,57 The calculated self-trapping energy E_st (the energy difference between STE and free exciton), and lattice deformation energy E_d (the increased ground state energy due to lattice deformation) are 1.83 and 0.90 eV, respectively, indicating the self-trapping energy E_st made a major contribution to the large Stokes shift of the STE broadband emission. Under high pressure, the interaction between excited- and ground-state curves in (Br-PA)₄AgInBr₈ was eliminated with the decreased self-trapping energy E_st (1.03 eV) and lattice deformation energy E_d (0.61 eV), suggesting an effective radiative transition pathway and transformation from dark to bright state for the STEs (Figure 4e). In 2D Ag-Bi double perovskites, an energy barrier between the intercept and the excited-state minimum could form without the elimination of intersection under high pressure, which was a prominent disadvantage compared with 2D Ag-In double perovskites in suppressing nonradiative transitions of the STEs.²² The decreasing coordinate difference ΔQ (the lattice changes between the ground- and excited-state equilibrium position) proved the transient elastic structural deformation in excited states was suppressed remarkably due to the lattice contraction and improved structure rigidity, corresponding to a weakened electron-phonon coupling strength. Stable STE formation was attributed to the strong electron-phonon coupling and soft lattice in metal halide double perovskite. Excessive electron-phonon coupling brought about quenching or extremely weak strength for the STE emission, whereas intermediate carrier-phonon coupling could result in an efficient STE emission.¹⁷ In principle, the Huang-Rhys factor S could be used to describe the strength of electron-phonon coupling, which followed the standard form of the CCD and could be obtained by the approximate relationship between the Stokes shift energy and longitudinal optical phonon energy (ΔE_stokes = 2Sћω_LO).⁵⁸ The estimated value of the Huang-Rhys factor was 48.3 at 0.5 GPa based on our optical measurement data, confirming the giant electron-phonon coupling strength in 2D Ag-In double perovskites ( Supporting Information Figure S28). Compared with the weak PL in PEA₄AgBiBr₈ with a relatively small Huang-Rhys factor of 27.8,²² the absence of STE emission was reasonable due to excessive interaction between electrons and phonons. With the increase of pressure, the Huang-Rhys factor decreased gradually, where the fluctuation at 2.5 GPa ascribed to the two-step compression of the lattice and did not affect the overall evolution trend, suggesting that the electron-phonon coupling strength was weakened upon compression. According to the Franck–Condon factor, the Huang–Rhys factor S affects PL emission, which has a negative correlation with PL intensity.⁵⁹ The optimal value of the Huang–Rhys factor was 28.7 at 8.3 GPa. This result proved that the appropriate magnitudes of the Huang–Rhys factor S and carrier-phonon coupling strength were crucial for achieving efficient STE emission, while an overly large S led to the curves cross between the ground- and excited states and large energy loss derived from the self-trapping energy E_st of excitons in the excited state and lattice deformation energy E_d of the crystal distortion in the ground state, promoting energy dissipation by phonons. The Huang–Rhys factor could be used as the figure of merit to design efficient broadband emission from STEs. The optimal Huang–Rhys factor of 2D Ag-In double perovskites was much greater than its Ag-Bi analogs (optimal S = 9, orange emission), critical to increasing the degree of broadening associated with the white-light emission from the STEs.

Previous studies on the PL onset and decay times of the STEs indicated the broadband white-light emission of perovskites derived from an energy distribution of STE states with different self-trapping depths because of different electron-phonon coupling strength in various local environments in the excited state.^1,10,11 Therefore, the STE emission exhibited wavelength-dependent dynamics, where the long wavelength emission presented slower onset and decay rates compared with the shorter wavelength portion of the emission spectra.^10,29,60 Contrarily, the identical PL decay curves probed at different wavelengths indicated wavelength-independent time decay in (Br-PA)₄AgInBr₈ (Figure 5a). Apart from the amplitude decay, the shape of the TRPL spectra remained unchanged over a very long period from 3 to 2000 ns, implying a single ensemble emission.^13,61 To further confirm this behavior, the time-resolved emission spectra were constructed by integrating PL decay curves with different emission wavelengths, where all decay was recorded under identical experimental conditions. As shown in Figure 5b, the time-resolved emission spectra maintained unchanged shapes and emission centers, further confirming single ensemble decay dynamics and ruling out a distribution of subensembles with wavelength-dependent decay dynamics. The initial ultrafast decay of the TRPL within 3 ns was attributed to the trapping state of lattice defects, which acted as nonradiative recombination centers for photoexcited charge carriers.⁶¹ This fast nonradiative relaxation did not affect the emission line shape but caused PL decay. Application of pressure could create new ensemble emissions in halide perovskites such as pressure-driven reverse intersystem crossing and band-edge states reconfiguration.^62,63 In this system, the single ensemble property of the broadband emission in (Br-PA)₄AgInBr₈ was not affected by the changes in pressure ( Supporting Information Figure S29). According to early kinetic analysis and mechanism speculation in layered perovskites,¹³ the single ensemble emission from (Br-PA)₄AgInBr₈ might stem from ultrafast structural fluctuations on a much faster time scale than that of luminescence emission, resulting in STEs experiencing all transient local environments with a wide distribution of energy prior to PL decay, thereby emerging as a time-averaged single ensemble PL dynamics (Figure 5c). The occurrence of a singular emissive ensemble could be replicated in (Br-PA)₄AgBiBr₈, corroborating its status as a prevalent attribute in 2D layered double perovskites ( Supporting Information Figure S30).

• Download figure
• Download PowerPoint

Conclusion

In summary, by finely controlling the transient lattice distortion and carrier dynamics at the excited state using pressure engineering, strong single-ensemble white-light emission was achieved in 2D double perovskite (Br-PA)₄AgInBr₈. We have confirmed the existence of the dark STE states with strong electron-phonon coupling and weak TDM in 2D Ag-In double perovskites, secretly associated with the absence of fluorescence at ambient conditions. Pressure-induced conversion of the STEs from dark to bright states and large optimal Huang–Rhys factors were key to bringing about efficient white-light emission. The wavelength-independent PL decay indicated that the broadband white-light emission stems from a time-averaged single ensemble PL dynamics. The optoelectronic properties and electronic structures of 2D double perovskites were dominated by the alternating inorganic building blocks. Manipulating the composition of bimetallic combinations allowed for the engineering of specific band-edge states and interactions between exciton and phonon, which, in turn, enabled the regulation of the optoelectronic properties of these 2D double perovskites. In this study, compared to the Bi element in one of the two main trivalent cations of 2D double perovskite materials, the In element exhibited greater advantages and potential in designing solid-state luminescent materials, reflected by the pressure-induced strong white-light emission, long carrier lifetime, large TDM, and intersection elimination between the excited- and ground-state. Precompression in the initial stage was a common feature in the pressure-induced emission phenomenon of 2D perovskite materials. For materials design criteria, pressure sensitivity and electron-phonon coupling strength could be modified by combining different physical and chemical methods, including metal alloying, organic molecular engineering, crystal rigidity improvement, strain, and so on. We believe that the revealed photophysical mechanism and carrier dynamics for the bright STE states in this work could serve as a blueprint for the assembly and fine-tuning of many other 2D Ag-In double perovskites toward the realization of effective broadband emission by manipulating the electron-phonon coupling and electron-hole wavefunction overlap.

Supporting Information

Supporting Information is available and includes absorption spectra, XRD patterns, Raman spectra, TRPL spectra, carrier lifetime, and DFT calculation results.

Conflict of Interest

The authors declare no competing interests.