Compound characterization data — Glazer et al., 2026

Cs₂AgInCl₆ and alkali-ion-alloyed derivatives
PXRD and Rietveld refinement of Cs₂AgInCl₆	The toluene-assisted product contains only trace quantities of non-perovskite phases and has an approximately cubic lattice parameter a = 10.464 Å and unit-cell volume 1146.7 Å³
PXRD of (Cs₂₋ₓRbₓ)AgInCl₆ (0.25 ≤ x ≤ 1)	Reflections remain dominated by the Fm3̅m double-perovskite structure. All double-perovskite reflections shift systematically to higher 2θ with increasing Rb⁺ substitution, consistent with contraction of the cubic cell and A-site substitution. AgCl reflections are also present. Weak RbCl reflections in CsRbAgInCl₆ suggest finite Rb⁺ solubility on the A-site, at least when the material is prepared by mechanosynthesis. The (400)/(220) peak-area ratio increases at higher Rb⁺ loading. The (311) reflection is almost entirely absent in Cs₂AgInCl₆ but becomes non-zero as Rb⁺ loading increases; the (311)/(220) ratio increases through 25 mol% Rb⁺ and then plateaus
PXRD of Cs₂(Ag₁₋ᵧNaᵧ)InCl₆ (0 ≤ y ≤ 1)	At Na⁺ loadings ≥75 mol%, non-perovskite reflections are observed. Double-perovskite reflections shift linearly to lower angle as Na⁺ loading increases below 75 mol%; at ≥75 mol%, the (422) reflection continues to shift to lower angle to a lesser extent, suggesting that further Na⁺ incorporation into the double-perovskite phase is incomplete and some Na⁺ is instead segregated into impurity phases. The resulting unit-cell expansion is assigned to Na⁺ replacement of Ag⁺ on the B′ site
¹³³Cs MAS NMR (11.7 T, 15 kHz)	Cs₂AgInCl₆ shows one narrow resonance at 107.7 ppm, consistent with a single crystallographic Cs⁺ site in a well-ordered material. Na⁺ substitution shifts this resonance to lower frequency, indicating shielding of ¹³³Cs nuclei. At 12.5 mol% Rb⁺, a slight shift to lower frequency is also observed; at higher Rb⁺ loading the signal shifts to higher frequency than in Cs₂AgInCl₆, indicating reduced ¹³³Cs shielding. The authors conclude that at low loading Rb⁺ substitutes both Ag⁺ on the B′ site and Cs⁺ on the A-site, but for loadings ≥12.5 mol% Rb⁺ preferentially substitutes for Cs⁺. Substitution by either Rb⁺ or Na⁺ broadens the signal relative to Cs₂AgInCl₆, which the authors assign to static disorder and locally inequivalent Cs⁺ environments
⁸⁷Rb MAS NMR (11.7 T, 15 kHz)	(Cs₁.₇₅Rb₀.₂₅)AgInCl₆ shows two relatively narrow resonances. The more intense peak at −17.5 ppm is assigned to A-site Rb⁺ with six nearest-neighbour Cs⁺ ions; the peak at −16.4 ppm is assigned to A-site Rb⁺ with five Cs⁺ and one Rb⁺ nearest neighbour. (Cs₁.₅Rb₀.₅)AgInCl₆ shows three distinct peaks assigned to Rb⁺ having zero, one, or two nearest Rb⁺ neighbours. Relative to (Cs₁.₇₅Rb₀.₂₅)AgInCl₆, its zero- and one-Rb⁺-neighbour signals shift and broaden slightly, reflecting local structural distortion and cubic-symmetry breaking. At x = 1, the resolved resonances are replaced by one broad asymmetric signal, assigned to overlapping signals from all possible first-neighbour Cs⁺/Rb⁺ arrangements with second-neighbour shift dispersion. No significant resonance assignable to B′-site Rb⁺ was observed in (Cs₁.₇₅Rb₀.₂₅)AgInCl₆; the authors note that the spectra are not fully quantitative because site-dependent quadrupolar coupling gives different signal build-up rates
¹¹⁵In MAS NMR (11.7 T, 20 kHz)	Cs₂(Ag₁₋ᵧNaᵧ)InCl₆ resolves In³⁺ environments having different nearest-neighbour Na⁺/Ag⁺ configurations. In the (Cs₂₋ₓRbₓ)AgInCl₆ series, no corresponding resolved environments are observed; instead, Rb⁺ incorporation broadens the single Cs₂AgInCl₆ resonance. The authors identify this contrast as evidence for different Rb⁺ and Na⁺ site selectivity and for local cubic-symmetry breaking induced by A-site Rb⁺ alloying