High Quality Factor in Solution-Processed Inorganic Microcavities Embedding CsPbBr3 Perovskite Nanocrystals

Optical microcavities grant manipulation over light–matter interactions and light propagation, enabling the fabrication of foundational optical and optoelectronic components. However, the materials used for high-performing systems, mostly bulk inorganics, are typically costly, and their processing is hardly scalable. In this work, we present an alternative way to fabricate planar optical resonators via solely solution processing while approaching the performances of conventional systems. Here, we couple fully solution-processed high dielectric contrast inorganic Bragg mirrors obtained by sol–gel deposition with the remarkable photoluminescence properties of CsPbBr3 perovskite nanocrystals. The approach yields microcavities with a quality factor of ∼220, which is a record value for solution-processed inorganic structures, and a strong emission redistribution resulting in a 3-fold directional intensity enhancement.


CALCULATION OF ABBE NUMBERS: Equation S1.
Where nc, nd and nF are the refractive indices at the Fraunhofer wavelengths, 656.3 nm, 587.56 nm and 486.1 nm respectively. Figure S1. Schematic of the fabrication process from sol preparation to deposition, drying and annealing for a hybrid thin film. Figure S2. a) Schematic of NC multilayer deposition via LbL spin-casting; b) Normalized optical absorption and photoluminescence spectra for a 5-layers NC stack as compared to NC solution.

OPTICAL ABSORPTION AND PHOTOLUMINESCENCE OF CsPbBr3 NCs
The perovskite nanocrystal (NC) film (solid black line) shows a structured and broad optical absorption with onset at about 510 nm similar to the solution absorption spectrum (dashed black line). The PL spectrum of the film (solid red line) consists in a sharp peak centered at 525 nm having a full width half maximum (FWHM) of 18 nm, while in solution (dashed red line) we observe a sharp peak centered at 503 with a FWHM of 19 nm. The PL red-shift is due to a change of dielectric medium in agreement with previous report in literature 1 and, we hypothesize the presence of a coalescence phenomenon passing from solution to solid state which can cause the mean average size of nanocrystals to increase. Such coalescence is suggested by the appearance of a secondary weak absorption peak at 519 nm in film. This phenomenon is well known for perovskite nanoplatelets 2 under irradiation. However, considering the two almost identical S4 FHWM, we observe no significant increase in polydispersity, thus signifying preservation of optical quality. Furthermore, to confirm the optical quality of the system, we acquired AFM images on both the DBR and the DBR after the deposition of a nanocrystal multilayer, and calculated average surface roughness. To validate our assumption regarding photoluminescence intensity, as already mentioned in the paper, we must demonstrate: 1) Linear dependence between excitation power and intensity of PL emission for NCs.
2) Absence of non-negligible TiO2 absorbance and light scattering phenomena at the pump wavelength S7 Figure S6. a) PL of the MC at increasing excitation power; b) maxima of emission intensity as a function of laser pump power, the data points were fit with a linear function.
In regards of (1), Figure S6a shows acquired PL spectra for a generic spot on the MC as a function of excitation laser power, while panel b shows the PL intensity maxima as a function of laser power. The intensity of emission is linearly dependent on the laser pump power. To highlight the linear behavior, we report the linear fit bearing R-Square value of 0.998, which validates our hypothesis. The data shown in Figure S6 were collected on a different area of the very same sample presented in the main text, hence the variation in spectral features.
We also report, in figure S7, the variation of PL peak linewidth with increasing power for the same spot analyzed in figure S6.
S8 Figure S7. PL linewidth as a function of laser power.
We observe a slight decrease of about 1 nm in the PL peak linewidth which is comparable however with the resolution of our spectrometer.
In regards of (2), we measured the intensity of a laser beam at a fixed power. We then proceeded by shining the same beam through a DBR and collect again the signal from the laser. Finally, we compared the intensity loss with the reflectance of photonic structures superimposing the wavelength of the laser emission. Results are reported in Fig. S8. we magnify on the spectral range of interest where we can observe how interposing a DBR between laser and detection causes a decrease of intensity of 25% which is consistent with a 25% reflectance at 405 nm due to the interference fringes. This confirms the linear dependence between laser power attenuation and reflectance of photonic structure, and excludes major parasitic scattering and optical absorption phenomena, for instance from the TiO2 films. Figure S9. Angle resolved photoluminescence for references and MC samples. However, in the transmittance angular dispersion of the microcavity we do not observe optical features ascribable to cavity resonances due to lack of transmitted light for the specific system.

PHOTOLUMINESCENCE OF REFERENCES AND MICROCAVITY
To this end, we acquired new transmission measurements using a high-sensitivity (i.e. high dynamic range) detector in the region of interest (Fig. S11). S11 Figure S11. Angle resolved transmittance of the MC.
Although still being hardly detectable, we are able to observe additional spectral features within the MC. We can observe three different cavity resonances centered at 544, 563 and 586 nm with similar spacing compared to the ones visible in reflectance measurements. However, for our samples, reflectance is more suitable for a comprehensive description of the system.