Sb
3+-doped Cs
2SnCl
6 crystals were synthesized through the conventional hydrothermal method. In detail, cesium chloride (CsCl), tin chloride (SnCl
2) and antimony oxide (Sb
2O
3) were added into HCl aqueous solution in a PTFE container of hydrothermal reactors. Then the solution was kept at 180°C for 10 h to ensure the complete dissolution of all precursors. After a slow cooling down process from 180°C to room temperature for 20 h, tiny crystals could be formed in the solution. To ensure the total removal of surface-adsorbed ions, the tiny crystals were rinsed by methanol for three times. The Sb/(Sb+ Sn) molar ratios in the solution were set at 0, 0.99%, 4.76%, 9.09%, 16.66% and 23.08%. The whole procedure was conducted in air atmosphere. Here we utilized Sb
2O
3 instead of SbCl
3 as Sb
3+ precursors due to the high hygroscopicity of SbCl
3 and hence the difficulty in accurate weighing. The reason why we chose Sn (II) as the tin source will be discussed in the later part of the calculation for chemical potential (
Dm) and formation enthalpies (
DH). X-ray diffraction (XRD) measurement was applied to determine the structure of the samples and the result is shown in Fig. 1(a). Without any impurity phase, all the diffraction peaks of products matched well with Cs
2SnCl
6 crystal structure (ISCD #9023) with a space group of Fm
m. Inductively coupled plasma optical emission spectrometry (ICP-OES) was used to precisely determine the chemical compositions of products. As shown in Table S1, the ICP-OES-determined Sb/(Sb+ Sn) concentrations are 0.20%, 0.41%, 0.59%, 0.89% and 0.98%, while the feeding concentrations are 0.99%, 4.76%, 9.09%, 16.66% and 23.08%, respectively. To simplify later discussion, we label the samples as Cs
2SnCl
6:
xSb, where
x is the real concentration provided by ICP-OES. It should be noted that the Sb concentrations are fairly low compared to the Bi concentrations reported in our previous work [
28]. X-ray photoelectron spectroscopy (XPS) measurement was utilized to further verify the composition and obtain the valence state of the elements within the products. The XPS survey spectrum (Fig. 1(b)) shows the characteristic peaks for Cs, Sn, Cl and Sb. As shown in Fig. S2(a), the peaks locating at 496.3 and 487.8 eV correspond to Sn
4+ 3d3/2 and 3d5/2, respectively, proving that the Sn
2+ was completely oxidized to Sn
4+ after the reaction process in the hydrothermal reactor. The peaks locating at 540.0 and 532.9 eV correspond to Sb
3+ 3d3/2 and 3d5/2, respectively (Fig. S2(b)). There is an additional peak at 531.7eV in Fig. S2(b), which was tentatively attributed to the presence of Sb-O or Sn-O on the crystal surface due to the hydrolysis of antimony chloride and tin chloride [
29–
32]. From the XRD, ICP-OES and XPS measurements, we speculate that the Sb
3+ was incorporated into the Cs
2SnCl
6 matrix successfully. Moreover, the chemical potential window and formation enthalpies for Sb interstitial (Sb
i) and Sb-Sn substitution (Sb
Sn) were calculated (Figs. 1(c) and 1(d)). The
DH>0 means that it is not conducive to forming such products. Hence the results reveal that Sb
i can hardly form because of the too high
DH values in the whole chemical potential regions, and Sb
Sn can also hardly form under Sn
4+-rich/Sn
2+-poor conditions (i.e., in the high
Dmsn regions C-F in Fig. 1(c)). Instead, the use of Sn
4+-poor/Sn
2+-rich condition (i.e., in the low
Dmsn regions A, B, and G) is favorable for the formation of Sb
Sn, where the
DH values are low or even negative. Thereby we chose Sn (II) as the tin source for the synthesis of antimony ion doped Cs
2SnCl
6, in analogy with our previous synthesis of Cs
2SnCl
6:
xBi. The refined lattice parameters of Cs
2SnCl
6:
xSb samples increased as the Sb concentration increased (see Fig. S1 and Table S2 in the Supporting Information), which is consistent with ion radius of Sn
4+ (0.69 Å) and Sb
3+ (0.76 Å) for Sb-Sn substitution.