Syntheses and crystal structures of the quaternary thiogermanates Cu4FeGe2S7 and Cu4CoGe2S7

The isostructural crystal structures of Cu4FeGe2S7 and Cu4CoGe2S7 were solved and refined. All the metal cations form MS4 tetrahedra and share corners to create a three-dimensional, non-centrosymmetric structure.

The quaternary thiogermanates Cu 4 FeGe 2 S 7 (tetracopper iron digermanium heptasulfide) and Cu 4 CoGe 2 S 7 (tetracopper cobalt digermanium heptasulfide) were prepared in evacuated fused-silica ampoules via high-temperature, solidstate synthesis using stoichiometric amounts of the elements at 1273 K. These isostructural compounds crystallize in the Cu 4 NiSi 2 S 7 structure type, which can be considered as a superstructure of cubic diamond or sphalerite. The monovalent (Cu + ), divalent (Fe 2+ or Co 2+ ) and tetravalent (Ge 4+ ) cations adopt tetrahedral geometries, each being surrounded by four S 2À anions. The divalent cation and one of the sulfide ions lie on crystallographic twofold axes. These tetrahedra share corners to create a three-dimensional framework structure. All of the tetrahedra align along the same crystallographic direction, rendering the structure non-centrosymmetric and polar (space group C2). Analysis of X-ray powder diffraction data revealed that the structures are the major phase of the reaction products. Thermal analysis indicated relatively high melting temperatures, near 1273 K.

Chemical context
The title compounds belong to the family of quaternary thiogermanates, which can be constructed from different [Ge x S y ] zbuilding blocks, such as [GeS 4 ] 4À (Aitken et al., 2001) and [Ge 2 S 6 ] 4À (Choudhury et al., 2015). Two GeS 4 tetrahedra can share a corner to create [Ge 2 S 7 ] 6À units, which are featured in the title compounds. Cu 4 FeGe 2 S 7 and Cu 4 CoGe 2 S 7 also belong to the family of diamond-like semiconductors (DLSs), with structures that can be derived from the cubic or hexagonal (Frondel & Marvin, 1967) forms of diamond. The synthesis of new diamond-like materials is guided by valence electron principles and Pauling's rules, and the resulting DLSs can be binary, ternary, or quaternary, depending on the number of elements employed in the reaction (Parthé, 1964;Pamplin, 1981;Goryunova, 1965). Increasing the number of elements in the formula allows for greater tunability of the material's properties; thus, quaternary DLSs are a particularly appealing class of materials. As a result of their technologically relevant properties, these materials are of interest for a number of applications, such as solar cells (Ito & Nakazawa, 1988;Heppke et al., 2020;Liu et al., 2018), batteries (Brant, Devlin et al., 2015;Kaib et al., 2013) and magnetic devices (Wintenberger, 1979;Greenwood & Whitfield, 1968). Furthermore, owing to their inherently non-centrosymmetric structures, DLSs are attractive candidates for infrared nonlinear optical (IR-NLO) devices that make use of secondharmonic generation (SHG) crystals (Ohmer & Pandey 1998): only crystals that lack an inversion center can exhibit SHG.
IR-NLO materials are used to shift the radiation of lasers to more suitable wavelengths for use in military (Hopkins 1998), medical (Stoeppler et al., 2012) and industrial applications (Bamford et al., 2007). Currently, ternary DLSs, most of which are sulfides, dominate the market of SHG crystals for use in the infrared (Ohmer & Pandey, 1998). Yet the current commercially available IR-NLO materials suffer from serious drawbacks, such as low laser-induced damage thresholds (LIDTs) and multi-photon absorption (Schunemann, 2007). Turning attention to the discovery of new quaternary DLSs provides a reliable route to next-generation IR-NLO materials that allows for greater control of the material's properties. Compounds such as Li 2 CdGeS 4 , Li 2 MnGeS 4 (Brant, Clark et al., 2015), and Li 4 HgGe 2 S 7  have shown potential to outperform currently used ternary IR-NLO crystals. These DLSs have shown promising SHG capabilities, as well as resilience to high powered lasers, a necessity to broaden future usage (Hopkins, 1998). For these reasons, we were motivated to investigate the Cu-Fe-Ge-S and Cu-Co-Ge-S systems for new DLSs.

Structural commentary
The title compounds, Cu 4 FeGe 2 S 7 (I) and Cu 4 CoGe 2 S 7 (II), are isostructural and crystallize in the non-centrosymmetric, monoclinic space group C2 (No. 5) with the Cu 4 NiSi 2 S 7 structure type (Schä fer et al., 1980). The structure contains two crystallographically unique Cu + ions, one divalent metal (Fe or Co) sited on a crystallographic twofold axis, one Ge 4+ cation and four S 2À anions (one with site symmetry 2) (Fig. 1). The sulfide anions create a 'cubic' close-packed array and the cations reside in one-half of the tetrahedral holes; these tetrahedra share corners to form a three-dimensional network. Two GeS 4 tetrahedra share corners to form (Ge 2 S 7 ) 6À subunits that are isolated from each other (Fig. 1). These subunits are separated by isolated FeS 4 tetrahedra and surrounded by a snaking, three-dimensional network of corner-sharing CuS 4 tetrahedra that serve to link the (Ge 2 S 7 ) 6À and FeS 4 subunits. All of the tetrahedra are aligned along one crystallographic direction, rendering the structure non-centrosymmetric (Fig. 2). All DLSs exhibit a honeycomb pattern in their crystal structure (Fig. 3); the various resulting space groups arise from the different possible cation-ordering patterns.
Selected geometrical data for (I) and (II) are given in Tables 1 and 2, respectively. The average Fe-S (I) and Co-S (II) bond distances are 2.334 (6) and 2.317 (6) Å , respectively. These values align well with other compounds containing iron or cobalt tetrahedrally coordinated by sulfur. For example, the average Fe-S distance found in Li 2 FeGeS 4 is 2.34 (2)  Polyhedral view of Cu 4 FeGe 2 S 7 showing the polar nature of the structure. Table 1 Selected bond lengths (Å ) for (I).

Figure 1
The structure of Cu 4 FeGe 2 S 7 with crystallographically unique ions labeled. Ellipsoids are shown at 99% probability. Cu 4 CoGe 2 S 7 is isostructural and has similar atomic displacement parameters. (Brant, dela Cruz et al., 2014), while the average Co-S distance found in Li 2 CoGeS 4 is 2.31 (3) Å (Brant, Devlin et al., 2015). The average Ge-S distances are 2.240 (4) and 2.244 (5) Å for (I) and (II), respectively. These distances are also close to those of the lithium-containing DLSs: Li 2 FeGeS 4 (Brant, Devlin et al., 2015) and Li  Both S2 and S4 are coordinated by two copper, one germanium and one iron or cobalt cation. S1 is connected to two germanium and two copper cations, while S3 is surrounded by one germanium and three copper cations.

X-ray powder diffraction and thermal analysis
The calculated and observed X-ray powder diffraction patterns match well (Fig. 4), indicating that the title compounds are the major phases of the respective reactions. An optimization of the synthetic protocol is needed to isolate the desired phases in phase-pure form.
Differential thermal analysis (DTA) reveals that Cu 4 FeGe 2 S 7 and Cu 4 CoGe 2 S 7 show relatively high thermal stability and melting and recrystallization events with appropriate hysteresis around 1000 C (Fig. 5). Multiple heatingcooling cycles for each sample were consistent, suggesting that the thermal events are reversible. X-ray powder diffraction of the DTA residues indicated that the samples were not changed by the thermal analyses, implying that they melt congruently. DTA also suggests that neither compound is a single phase, as Comparison of experimental X-ray powder diffraction patterns before and after DTA with those calculated using the single-crystal structures for Cu 4 FeGe 2 S 7 (left) and Cu 4 CoGe 2 S 7 (right).   The 'honeycomb' pattern found in Cu 4 FeGe 2 S 7 , a characteristic of DLSs. Highlighted in the bottom left corner is one of the [Ge 2 S 7 ] 6À subunits that forms in this structure.
there are some small shoulders on the peaks indicative of the thermal events.

Materials and methods
All powdered elements were acquired from commercial suppliers and used as obtained with the exception of germanium metal, which was purchased as chunks and ground to a fine powder using a Diamonite 2 mortar and pestle prior to use. Powder X-ray diffraction data were recorded from 10-100 2 using a PANalytical X'Pert Pro MPD powder X-ray diffractometer operating with Cu K radiation ( = 1.541871 Å ), a tube power of 45 kV and 40 mA and a step size of 0.017 . DTA data were obtained using a Shimadzu DTA50 thermal analyzer. Each sample was vacuum-sealed in a fusedsilica ampoule, placed alongside an ampoule containing an Al 2 O 3 reference of comparable mass, heated from room temperature to 1050 C at a rate of 10 C min À1 and subsequently cooled to room temperature at the same rate. A second heating-cooling cycle was conducted in order to determine the reproducibility of the thermal events.

Synthesis and crystallization
Cu 4 FeGe 2 S 7 and Cu 4 CoGe 2 S 7 were synthesized by combining stoichiometric amounts of Cu (99.999%), Fe (99.99%) or Co (99.99%), Ge (99.999%) and S (99.5%, sublimed) powders. The powders were mixed and placed into 12 mm o.d. fused-silica tubes that were subsequently attached to a vacuum line, evacuated and flame sealed. The reaction vessels were placed upright into ceramic containers inside programmable furnaces, where they were heated to 1000 C in 24 h, held there for 48 h, cooled to 900 C over the course of 50 h, and held there for 96 h, before being allowed to cool to room temperature over a 24 h period. Subsequently, the reaction vessels were cut open and the contents were examined under a light microscope. The products consisted of loose silvery gray microcrystalline powders from which small single crystals were selected for single-crystal X-ray diffraction.

Refinement
Crystal data, data collection parameters, and structure refinement details are summarized in Table 3. Extinction parameters were refined for each compound. After the final refinement, the Flack parameter for both structures refined to 0.06 (3), indicating that the absolute structure is correct. In Cu 4 FeGe 2 S 7 , the largest difference peak is located 1.15 Å from Cu2 while the deepest difference hole is 1.50 Å from S3. For Cu 4 CoGe 2 S 7 , the largest difference peak is 0.67 Å from Co while the deepest difference hole is 0.83 Å from Ge.   (3) Computer programs: SMART and SAINT (Bruker, 1998), SHELXS97 (Sheldrick, 2008), SHELXL2018/3 (Sheldrick, 2015) and CrystalMaker (Palmer, 2014).
Environmental Sciences at Duquesne University, for keeping the diffractometers running.

Crystal data
Cu 4  Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes. Refinement. Refined as a two-component inversion twin

Special details
Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.