Our research addresses knowledge gaps in the optical properties of Cs2AgBiCl6, as previous studies have reported inconsistent results regarding its bandgap energy (ranging from 2.08 to 2.89 eV) and the origin of its characteristic orange photoluminescence. These inconsistencies have hindered the material’s application in optoelectronic devices. In this work, we investigate Cs2AgBiCl6 thin films deposited via reactive co-evaporation, focusing on three key objectives: (1) developing a reliable synthesis route to phase-pure materials, (2) resolving discrepancies in reported optical properties through rigorous characterization techniques, and (3) exploring the potential of Yb-doped Cs2AgBiCl6 for optoelectronic applications, specifically, high-efficiency downconversion of UV and blue photons to near-infrared (NIR) photons.
We deposited Cs2AgBiCl6 films using reactive co-evaporation in a high vacuum deposition chamber, where BiCl3, CsCl, and AgCl precursor powders were evaporated onto glass substrates maintained at 30 °C. The precursor molecules condense and react on the glass substrates during the deposition and post-deposition annealing to form the target material. Evaporation rates are measured using quartz crystal microbalances and actively controlled using closed-loop feedback by manipulating the evaporation crucible temperatures. Evaporation rates are typically set to achieve stoichiometric films, but excess reactants are also used in some depositions. We find that stoichiometric evaporation of precursors (2:1:1 ratio of CsCl: AgCl: BiCl3) resulted in films with impurity phases, primarily Cs3BiCl6, as confirmed by X-ray diffraction (XRD) and Raman spectroscopy. The presence of this impurity phase persisted even after annealing under various conditions. We discovered that using 25% excess BiCl3 followed by annealing yielded phase-pure Cs2AgBiCl6 films, as evidenced by XRD patterns matching the expected cubic (Fm3̅m) structure without impurity peaks. We attribute this requirement for excess BiCl3 to its volatility during deposition or annealing, suggesting some BiCl3 evaporates before fully reacting. This insight is crucial for the reproducible synthesis of phase-pure materials.
Raman spectroscopy results validated the phase purity, showing characteristic peaks at 114, 215, and 283 cm-1 corresponding to T2g, Eg, and A1g modes of Cs2AgBiCl6, respectively. Scanning electron microscopy (SEM) revealed that films deposited with stoichiometric BiCl3 developed porosity upon annealing, which we attribute to the evaporation of unreacted BiCl3, while films made with excess BiCl3 showed improved morphology with well-defined grains approximately 100 nm in size.
Previous studies have reported bandgap values ranging from 2.08 to 2.89 eV. Through careful optical characterization, including correction for thin-film interference effects and rigorous Tauc analysis, we determined that Cs2AgBiCl6 has an indirect bandgap of 2.68 eV and a direct transition at 2.9 eV. Optical absorption revealed a characteristic exciton peak at 370 nm (3.35 eV), associated with the Bi 6s→6p transition in the BiCl6³⁻ octahedra. We also addressed the longstanding debate regarding the origin of the orange photoluminescence (PL) in Cs2AgBiCl6, which peaks at approximately 655 nm (1.9 eV) with a broad emission lineshape (full width at half-maximum of ~250 nm). This emission is significantly red-shifted from the bandgap energy, ruling out direct band-edge recombination as its origin. Previous studies have attributed this emission to trap states, self-trapped excitons (STEs), or indirect bandgap recombination. Time-resolved photoluminescence (TRPL) measurements interpreted using a physically-based kinetic model of exciton decay provided compelling evidence that the orange emission likely comprises contributions from both self-trapped excitons and radiative defect states. Our model accounts for five possible decay processes: (1) nonradiative recombination, (2) nonradiative recombination via trap states, (3) second-order radiative band-edge recombination, (4) first-order radiative recombination via trap states, and (5) second-order recombination via self-trapped excitons. By analyzing PL decay curves using this model, we found that the relative contributions of self-trapped excitons and defect emissions depend strongly on synthesis conditions. For instance, as-deposited films showed emission dominated by first-order radiative processes (attributed to defects), while annealed films exhibited increased contributions from second-order processes associated with self-trapped excitons. This finding suggests a resolution to the apparent contradictions in previous literature regarding the origin of orange emission, suggesting that both mechanisms can contribute to the observed PL, with their relative importance determined by material processing. Our fluence-dependent PL measurements further support this conclusion, showing a nonlinear (second-order) dependence of PL intensity on excitation power. This behavior is inconsistent with purely trap-mediated emission, which would be expected to show first-order kinetics, and suggests significant contributions from bimolecular processes such as self-trapped exciton recombination.
Our interest in Cs2AgBiCl6 is motivated by its potential application as a host for Yb, a well-known luminophore that enables downconversion and quantum cutting—a process where one ultraviolet photon generates two near-infrared photons. In this energy transfer mechanism, the perovskite host absorbs blue photons and transfers the energy to Yb ions, which then relax (2F5/2→2F7/2) and emit near-infrared photons (1.25 eV). We doped Cs2AgBiCl6 by coevaporating YbCl3 with CsCl, AgCl and BiCl3. Yb-doped films achieved a photoluminescence quantum yield (PLQY) of 50%, attributed to downconversion. However, PLQY for these films decayed to 30% within one week of exposure to ambient conditions. Hypothesizing that this degradation is a surface-related phenomenon, we implemented a surface passivation strategy on both the top and bottom surfaces of the films. We investigated various passivation strategies using halide-based compounds related to the constituent elements of our perovskite materials. These included simple halide compounds (CsCl, BiCl3, and AgCl) as well as more complex cesium-bismuth-silver halide compounds, including Cs3BiCl6, Cs3Bi2Cl9, Cs2AgCl3, and CsAgCl2. Cs3BiCl6 passivation applied to both the top and bottom surfaces of Yb-doped Cs2AgBiCl6 films increased PLQY to 75%, with stability maintained for over 6 months.
To investigate PLQY decay mechanisms in non-passivated films, we performed time-dependent attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy on films deposited on trapezoidal-shaped GaAs ATR crystals. Our analysis revealed that water molecules from ambient air adsorb onto the film surface and react with BiCl3 in the perovskite to form BiOCl and HCl. Simultaneously, CO2 adsorption leads to carbonate formation (evidenced by an emerging infrared peak at ~1440 cm⁻¹), facilitated by the presence of HCl. We verified the role of excess BiCl3 by comparing films with stoichiometric BiCl3 content, which showed significantly lower carbonate formation that remained stable over time. Passivated films exhibited similar behavior, confirming that BiCl3 reaction with water vapor is the primary mechanism behind carbonate formation and subsequent PLQY degradation.
This talk will summarize our investigation of vapor-deposited Cs2AgBiCl6 thin films and address several key knowledge gaps in the literature. We have established that phase-pure films require excess BiCl3 during synthesis, determined precise values for the direct and indirect bandgaps, and provided a physical model that explains the origin of the characteristic orange emission as a combination of self-trapped exciton and defect emissions. Furthermore, we demonstrated the potential of Yb-doped Cs2AgBiCl6 for optoelectronic applications, achieving enhanced near-infrared PLQY, as high as 75%, and stability through surface passivation strategies. Our FTIR analysis revealed that the degradation of PLQY in Yb-doped Cs2AgBiCl6 films is primarily due to the reaction of water with excess BiCl3, leading to the formation of BiOCl and HCl, as well as carbonate formation from atmospheric CO2 adsorption.