Stability of the Halide Double Perovskite Cs2AgInBr6

Cs2AgInBr6 is among the lead-free halide perovskites of interest, predicted by first-principles calculations to be stable with a direct band gap, but there has been only one report of its synthesis. Herein we report the formation of Cs2AgInBr6 thin films through thermal evaporation of CsBr, AgBr, and InBr3 and subsequent annealing between 130 °C and 250 °C. Cs2AgInBr6 appears stable in this temperature range. However, Cs2AgInBr6 thin films are thermodynamically unstable at room temperature, remaining cubic only long enough to be characterized but not long enough to be useful for practical devices. Cs2AgInBr6 decomposed into Cs2AgBr3, Cs3In2Br9, AgBr, and InBr3 upon cooling from 130 °C to 250 °C to room temperature. This conclusion did not depend on illumination, film thickness, annealing environment, or details of the film formation, pointing to an intrinsic thermodynamic instability of the material. Optical absorption measurements may be interpreted as Cs2AgInBr6 having a direct band gap of 1.57 ± 0.1 eV.


C. X-ray Diffraction Patterns of Films with Different Thicknesses
Films thinner and thicker than 500 nm (~100 nm and ~1000 nm) were deposited by decreasing and increasing the deposition time under identical deposition conditions to study whether film stress and thickness affect the material stability. The deposition times were 5 minutes, 30 minutes, and 55 minutes for the ~100 nm, ~ 500 nm, and ~ 1000 nm thick films, respectively. Figure S5a shows the XRD patterns from a ~1000 nm thick film as deposited at room temperature, after heating to 90 o C, 110 o C, and 150 o C, immediately after cooling to room temperature after this heating, after a few hours the same day and after one day. Figure S5b shows the XRD patterns during in situ annealing as a function of temperature. As with the 500 nm film, the as-deposited ~1000 nm thick film comprises Cs2AgBr3, Cs3In2Br9, AgBr, and InBr3, indicating only partial reaction between CsBr and AgBr and InBr3 during the deposition to form the ternary phases with no evidence of Cs2AgInBr6 formation. Cs2AgInBr6 forms upon heating to 150 o C, which is stable upon cooling to room temperature but only for a few hours (typically less than 4 hours). Eventually, the film decomposes to Cs2AgBr3, Cs3In2Br9, and InBr3. Figure S6a and S6b show the same for the thinner (~100 nm) film. The detector and the incident X-ray were positioned at lower incident angles (θ1 =5°, θ2 =27°) to increase the sampling volume and, thus, the peak intensities, which is responsible for the peak broadening in Figure S6.
The objective was to ascertain the film's stability. As with the thicker films, Cs2AgInBr6 double perovskite started to form at 110 °C, and the reaction was completed at 150 °C. When the films are measured immediately after cooling to room temperature, Cs2AgInBr6 is still observed.
However, the structure is only stable for less than 4 hours and decomposes completely after a day. In conclusion, film thickness and stress do not seem to impact material stability in the film thickness range of 100 to 1000 nm. Figure S5. (a) In situ XRD measurements of the ~1000 nm thick film (from bottom to top) as deposited at room temperature, at 90 °C, at 110 °C, at 150 °C, measured immediately when cooled to 30 °C (denoted as 0 min), after storing at room temperature for a few hours, and after storing at room temperature for a day. Simulated XRD patterns of Cs2AgInBr6 (using a = 11.2 Å and a = 11.0 Å) and precursor are also shown as references. (b) In situ XRD measurements of the ~1000 nm thick film as it is heated stepwise from room temperature to 150 °C and then cooled stepwise from 150 °C to room temperature, and after storing at room temperature for a day, along with simulated XRD patterns of Cs2AgInBr6 (using a = 11.2 Å and a = 11.0 Å). The film thickness determined by fitting the thin film interference fringes in the optical transmission was 1055 nm. Figure S6. (a) In situ XRD measurements of the ~100 nm thick film (from bottom to top) as deposited at room temperature, at 90 °C, at 110 °C, at 150 °C, measured immediately when cooled to 30 °C (denoted as 0 min), after storing at room temperature for a few hours, and after storing at room temperature for a day. Simulated XRD patterns of Cs2AgInBr6 (using a = 11.2 Å and a = 11.0 Å) and precursor are also shown as references. (b) In situ XRD measurements of the ~100 nm thick film as it is heated stepwise from room temperature to 150 °C and then cooled stepwise from 150 °C to room temperature, and after storing at room temperature for a day, along with simulated XRD patterns of Cs2AgInBr6 (using a = 11.2 Å and a = 11.0 Å).

D. Cs2AgInBr6 Films' Optical Properties
The interference fringes are modeled in the transmission spectrum using a method reported by Swanepoel. 1 We calculated the thin film interference fringes using an optical model of an absorbing thin film on a thick finite substrate ( Figure S7). We then subtracted the calculated fringes from the measured extinction to determine the corrected film absorbance.
First, the baseline of the measured transmission is set to (S10) by subtracting, from the measured extinction spectrum, the difference between the measured value of the transmission at the peak of the fringe that appears at the largest and . In this equation, s is the glass refractive index (1.5). The transmission is calculated from and where and are the real and imaginary components of the film's complex refractive index (̃= − ), and d is the film thickness. is calculated as follows. First, fringe maxima, , and minima, , are identified, and their dependence is fit to a polynomial. Features suspected to be absorption peaks in the absorbing region are avoided, and only the minima and maxima above the nonabsorbing region (in this case > 800 nm) are used. in the nonabsorbing region is calculated from and in the absorbing region is calculated from The corrected absorbance is calculated using where is the experimentally measured baseline corrected transmission.

Fig. S7
Schematic of the optical model used to calculate the thin film interference fringes and correct the measured transmission (extinction). Figure S8 shows the measured and corrected absorbance (transmission) and fitted refractive index for an example film. This film was deposited as described in the main text Experimental Methods Section with a target thickness of 500 nm. It was annealed in air postdeposition by heating it slowly (~10 min) to 150 °C and cooling it slowly (~40 min) to room temperature. The absorbance (transmission) was measured immediately (within 2 minutes) after cooling to room temperature. Figure S8a shows the measured and modeled transmission (using Swanepoel's method). The modeled fringes were subtracted from the experimental transmission S11 spectrum, and the remaining was converted to absorbance. Figure S8b shows the absorbance spectrum before (as-measured) and after correction. The corrected film absorption starts to increase at 800 nm (1.55 eV). The film thickness and refractive index that best fit the fringes were 510 nm and 1.87 ±0.03, respectively ( Figure S8c). The values of the refractive index below 800 nm are extrapolations. The refractive index is expected to rise slightly as absorption begins.
Such small rises have only a small effect on the thickness but may affect the locations of the interference fringes, leading to imperfect subtraction. For this reason, the features in the absorption spectra should be interpreted with caution as they may result from small residual fringes remaining after subtraction. Figure S9 shows the optical absorbance of a 100 nm thick Cs2AgInBr6 film immediately (within 2 minutes) after cooling to room temperature, 70 minutes later, and one day later. Figure   S10 shows the same for a 1055 nm thick film. The corrected absorbance of the 1055 nm thick Cs2AgInBr6 film is also shown in Figure S10. The 100 nm thick film data was not corrected because there were no complete fringes. Like the 510 nm thick film, the transmission and extinction of the 1055 nm thick film exhibit thin film interference fringes superimposed on absorption that appears to rise at around 800 nm.
The insets in S9 and S10 show Tauc plots. Only the data near where absorption goes to   S16 Fig. S12. Optical absorbance (more correctly extinction) of a 1055 nm thick Cs2AgInBr6 film deposited and annealed at 150 o C immediately (within 2 minutes) after cooling to room temperature, 70 minutes later, and one day later. The corrected absorbance is also shown. Inset is the direct transition Tauc Plot of the 1055 nm thick Cs2AgInBr6 film deposited and annealed at 150 o C measured immediately (within 2 minutes) after cooling to room temperature. The line in the offset is a plausible extrapolation and intercepts the energy axis at 1.60 eV. The small peaks below 1.5 eV in the inset are residual due to incomplete subtraction of the interference fringes.

E. Cs2AgInBr6 optical extinction as a function of time
Optical absorbances of Cs2AgInBr6 films were measured as a function of time as indicators of Cs2AgInBr6 stability at room temperature. Figures S11-S13 show the optical extinction of the 1050 nm, 510 nm, and 100 nm thick Cs2AgInBr6 films as a function of time at room temperature after they were annealed in the air at 150 °C and then cooled to room temperature in the air. Figure S13. The optical absorbance of the thicker Cs2AgInBr6 film (1050 nm) stored at room temperature as a function of time after it has been annealed in air at 150 o C and cooled to room temperature in the wavelength ranges of (a) 300 to 2000 nm and (b) 300 to 800 nm expanded for clarity. Figure S14. The optical absorbance of the 51010 nm thick Cs2AgInBr6 film stored at room temperature as a function of time after it has been annealed in air at 150 o C and cooled to room temperature in the wavelength ranges of (a) 300 to 2000 nm and (b) 300 to 800 nm expanded for clarity. Figure S15. The optical absorbance of the thinner Cs2AgInBr6 film (~100 nm) stored at room temperature as a function of time after it has been annealed in air at 150 o C and cooled to room temperature in the wavelength ranges of (a) 300 to 2000 nm and (b) 300 to 800 nm expanded for clarity. Figure S16. SEM of a Cs2AgInBr6 film determined to be 510 nm by fitting the thin film interference fringes in the optical transmission.