One-Dimensional Iodoantimonate(III) and Iodobismuthate(III) Supramolecular Hybrids with Diiodine: Structural Features, Stability and Optical Properties

Two isostructural pairs of supramolecular iodoantimonate(III) and iodobismuthate(III) complexes with I2 units “trapped” in solid state via halogen bonding—Cat3[[M2I9](I2)} (Cat = tetramethylammonium and 1-methylpyridinium, M = Sb(III) and Bi(III)) were prepared. For all compounds, values of optical band gaps were determined, together with thermal stability; the complexes were additionally characterized by Raman spectroscopy.


Introduction
Within the past decades, there has been an ongoing growth of interest on halide complexes of p-block elements. The motivation of research in this field is mostly explained by the fact that iodometalates (in particular, iodoplumbates(II) [1][2][3][4]) can be used as components of solar cells and/or photodetectors-so-called perovskite photovoltaics constituted an area that has developed extremely rapidly [5][6][7][8][9][10][11][12].
One of the problems in this area is the fact that "true" isotropic 3D halometalates (with all covalent bonds), ideally suitable for the use as light absorbers in photovoltaic devices, are extremely rare: those are formed by either Sn(II) or Pb(II) with a very limited number of small organic (such as methylammonium [13,14]) or inorganic (such as cesium [3,15]) cations. This fact encourages the search for alternative candidates in this role. Halometalates of Sb, Bi and other p-elements have been extensively tested [16], mostly revealing low-tomoderate efficiencies. This is generally explained by the less favorable molecular structure of these compounds (those are 0D, 1D or, rarely, 2D).
There are different approaches towards the overcoming of this obstacle. Recently, Shevelkov et al., presented the first supramolecular iodobismuthate(III) complexes containing diiodine units [17][18][19]. In such hybrids, I 2 is "captured" in solid state, forming halogen bonding with iodometalate anions and acting therefore as "linker". Such structures feature therefore higher dimensionality of halometalate framework which is, as mentioned above, desirable for the enhancement of photovoltaic properties. It is also notable that these hybrids usually feature lower optical band gaps than "pure iodobismuthates" (<1.7 eV) and, at the same time, can be thermally stable enough to be used in solar cells.
Performing research in the field of polyhalogen-halometalates since the last years, we confirmed that representatives of this class indeed can be suitable for photovoltaic applications (mostly as photodetectors) [20,21]. Our work with polyiodo-iosobismuthates(III) resulted in the discovery of new structural types, including 1D and 2D supramolecular frameworks [22]. However, the number of these compounds is yet small. Moreover, those are virtually unknown for the closest neighbor of Bi(III)-antimony, which also readily forms halide complexes. On this reason, we decided to try preparing the pairs of polyiodo-iodometalates of Sb(III) and Bi(III) in order to compare their structural features (suggesting that those can be isostructural), as well as physical properties.

Experimental Part
The chemicals were purchased from commercial sources and used without additional purification. Solvents were purified according to the standard procedures. The 1-methylpyridinium iodide was prepared by reaction of pyridine and methyl iodide (1:1.05) following the standard method of alkylation of substituted pyridines. In all cases, concentrate HI was used.

Synthesis of 2
The procedure was the same as for 1, using Bi 2

Raman Spectroscopy, Thermogravimetric Analysis, Difuse Reflectance Spectroscopy
Details are given in Supplementary Materials.

Results and Discussion
There is a general method for preparation of polyhalogen-halometalates. There are three components: (1) metal source (oxide, oxohalide or halide), which is converted into halometalate by addition of either hydrohalic acid or other source of halide ligand, (2) halide salt of certain cation and (3) dihalogen. Depending on the precursors, synthesis can be conducted in acidic aqueous solutions or in organic solvents, such as acetonitrile. Preparation of 1-4 follows this paradigm as well, with no specific details (see Experimental Section). The yields can be regarded as good.
The main feature of the pairs 1-2 and 3-4 is that each of them is isostructural. In all cases, the halometalate building blocks are binuclear [Sb 2 I 9 ] 3− or [Bi 2 I 9 ] 3− anions consisting of two {MI 6 } octahedra connected via shared face. This type of anion is very common in structural chemistry of Group 15 halometalates [26][27][28][29][30]. The M-I term and M-µ 2 -I bond lengths in 1-4 are given in Table 1. In the crystal structure of 1, the [Sb 2 I 9 ] 3− units have orientational disorder over two closed positions with 0.97/0.03 occupancies. All TMA + cations are disordered over two different orientations with different occupancies. In the case of complex 2, the quality of crystalline material was too low to extract enough data for well-solved crystal structure. Complex 2 has practically the same unit cell parameters as complex 1. Therefore, the use of structural atomic coordinates from 1 in the refinement of 2 gave the structural model with R c.a. 12%.
The main structural feature of 1-4 is the presence of diiodine units which, as follows from analysis of interatomic distances (<3.98 Å, which is the sum of Bondi's van der Waals radii [31]), likely interact with terminal iodide ligands of [M 2 I 9 ] 3− anions. The geometric parameters of M-I···I and I-I···I fragments are given in Table 1.
Complexes of general formula {[M 2 X 9 ](X 2 )} were known earlier for M = Sb [32] and Bi [33], demonstrating isomerism due to diverse systems of halogen···halogen interactions. This feature appears also in the pairs 1-2 and 3-4: mutual orientation of [M 2 I 9 ] 3− and I 2 building blocks vary, resulting in assembly of 1D supramolecular chains of different shape (Figures 1 and 2). The crystal packings for the pairs 1-2 and 3-4 are given in Supplementary Materials. building blocks vary, resulting in assembly of 1D supramolecular chains of different shape (Figures 1 and 2). The crystal packings for the pairs 1-2 and 3-4 are given in Supplementary Materials.  Powder X-ray diffractometry data indicate that 1-4 are precipitated as pure single phases (see Supplementary Materials, Figures S1-S4). This fact makes possible in-depth investigation of physical properties of these compounds.
Raman spectra of 1-4 are presented on Figures 3 and 4. The main and most characteristic spectral feature appears at 175-178 cm −1 ; it corresponds to the diiodine fragment. Interestingly, this band is not metal-sensitive (the differences are minor), and it is not strongly affected by the variations in Iterm···II2 distances (Table 1). This fact agrees well with the Raman data obtained by us earlier for other diiodine-iodobismuthates(III): the range of shifts for incorporated I2 is rather narrow (170-175 cm −1 [22,34]). The bands at <150 cm −1 are very weak (also similar to the earlier results [22]); those correspond to different modes of {MI6} octahedra.
Overall, these observations confirm that the use of Raman spectroscopy can shed light on the nature of polyhalide units incorporated into the polyhalogen-halometalate framework or just indicate their presence (in express analysis of the outcomes of such syntheses). building blocks vary, resulting in assembly of 1D supramolecular chains of different shape (Figures 1 and 2). The crystal packings for the pairs 1-2 and 3-4 are given in Supplementary Materials.  Powder X-ray diffractometry data indicate that 1-4 are precipitated as pure single phases (see Supplementary Materials, Figures S1-S4). This fact makes possible in-depth investigation of physical properties of these compounds.
Raman spectra of 1-4 are presented on Figures 3 and 4. The main and most characteristic spectral feature appears at 175-178 cm −1 ; it corresponds to the diiodine fragment. Interestingly, this band is not metal-sensitive (the differences are minor), and it is not strongly affected by the variations in Iterm···II2 distances (Table 1). This fact agrees well with the Raman data obtained by us earlier for other diiodine-iodobismuthates(III): the range of shifts for incorporated I2 is rather narrow (170-175 cm −1 [22,34]). The bands at <150 cm −1 are very weak (also similar to the earlier results [22]); those correspond to different modes of {MI6} octahedra.
Overall, these observations confirm that the use of Raman spectroscopy can shed light on the nature of polyhalide units incorporated into the polyhalogen-halometalate framework or just indicate their presence (in express analysis of the outcomes of such syntheses). Powder X-ray diffractometry data indicate that 1-4 are precipitated as pure single phases (see Supplementary Materials, Figures S1-S4). This fact makes possible in-depth investigation of physical properties of these compounds.
Raman spectra of 1-4 are presented on Figures 3 and 4. The main and most characteristic spectral feature appears at 175-178 cm −1 ; it corresponds to the diiodine fragment. Interestingly, this band is not metal-sensitive (the differences are minor), and it is not strongly affected by the variations in I term ···I I2 distances (Table 1). This fact agrees well with the Raman data obtained by us earlier for other diiodine-iodobismuthates(III): the range of shifts for incorporated I 2 is rather narrow (170-175 cm −1 [22,34]). The bands at <150 cm −1 are very weak (also similar to the earlier results [22]); those correspond to different modes of {MI 6 } octahedra.
Overall, these observations confirm that the use of Raman spectroscopy can shed light on the nature of polyhalide units incorporated into the polyhalogen-halometalate framework or just indicate their presence (in express analysis of the outcomes of such syntheses).  Diffuse reflectance spectroscopy is a powerful method for evaluation of optical band gap values of both bulk phases and thin films, and this parameter is crucial for estimation of whether obtained materials are promising for further use in photovoltaics. On this reason, we characterized 1-4 by this technique. Results are demonstrated on Figure 5 (for 1) and in Supplementary Materials (for 2-4, respectively; Figures S5-S10). As follows from these data, the Eg values for 1-4 are 1.40, 1.45, 1.42 and 1.49 eV, respectively, being lower than one for BiI3 (≈ 1.7 eV [16]). On one hand, these data agree well with those obtained for other representatives of this class of compounds [22], on another, those clearly  Diffuse reflectance spectroscopy is a powerful method for evaluation of optical band gap values of both bulk phases and thin films, and this parameter is crucial for estimation of whether obtained materials are promising for further use in photovoltaics. On this reason, we characterized 1-4 by this technique. Results are demonstrated on Figure 5 (for 1) and in Supplementary Materials (for 2-4, respectively; Figures S5-S10). As follows from these data, the Eg values for 1-4 are 1.40, 1.45, 1.42 and 1.49 eV, respectively, being lower than one for BiI3 (≈ 1.7 eV [16]). On one hand, these data agree well with those obtained for other representatives of this class of compounds [22], on another, those clearly for BiI 3 (≈ 1.7 eV [16]). On one hand, these data agree well with those obtained for other representatives of this class of compounds [22], on another, those clearly demonstrate that optical properties of 1-4 make these compounds suitable for further photovoltaic tests. demonstrate that optical properties of 1-4 make these compounds suitable for further photovoltaic tests. Another very important parameter is thermal stability of halometalate hybrids. For its estimation, we used thermogravimetric analysis (TGA). As follows from these data (see Supplementary Materials, Figures S11-S14), antimony-and bismuth-containing complexes demonstrate different behavior. For Sb(III) derivatives, the loss of incorporated I2 is gradual, and it begins almost simultaneously with the start of heating. For io- Another very important parameter is thermal stability of halometalate hybrids. For its estimation, we used thermogravimetric analysis (TGA). As follows from these data (see Supplementary Materials, Figures S11-S14), antimony-and bismuth-containing complexes demonstrate different behavior. For Sb(III) derivatives, the loss of incorporated I 2 is gradual, and it begins almost simultaneously with the start of heating. For iodobismuthates(III), this stage of thermal decomposition has a narrower temperature range ( Figures S12 and S14), and it can be stated that 2 and 4 are stable at least up to 80 and 110 • C, respectively. These facts allow us suggesting that 4 is the most suitable candidate for possible photovoltaic studies.

Conclusions
We expanded the series of diiodine-halometalates of Group 15 elements (antimony and bismuth) by two isostructural pairs belonging to two different isomers of {[M 2 I 9 ](I 2 )} 3n− type. Although all prepared complexes feature optical band gaps favorable for photovoltaic applications, only one compound (4) demonstrated higher thermal stability making such tests practically reasonable. We believe, however, that the studies of polyiodo-iodometalates in terms of photovoltaics will continue within the upcoming years.