The calcium oxidotellurates Ca₂(Te^IVTe^VIO₇), Ca₂(Te^IVO₃)Cl₂ and Ca₅(Te^IVO₃)₄Cl₂ obtained from salt melts

2.1 Synthesis

Single crystals of mixed-valent Ca₂(Te₂O₇) were obtained from an NaCl/CaCl₂ flux. Te(OH)₆, CaCl₂⋅6H₂O and NaCl were thoroughly ground in an agate mortar in the molar ratio 1:0.6:10 (sample weight: 0.196 g, 0.091 g, 0.5 g, respectively). The mixture was placed in an alumina crucible, which was heated in a muffle furnace under atmospheric conditions with a heating rate of 225 °C h⁻¹ to 820 °C (holding time 5 h) and cooled to 200 °C with a cooling rate of 100 °C h⁻¹ before cooling the crucible to room temperature within 30 min. Optical inspection of the product under a polarizing microscope revealed very small numbers of crystals with a lath-like form and lengths in the range 5–80 µm that were manually separated for single crystal X-ray diffraction. Powder X-ray diffraction (PXRD) of the bulk product showed only the reflections of NaCl.

For an alternative preparation of Ca₂(Te₂O₇), solid state reactions of CaO, TeO₂ and TeO₃ in the molar ratio 2:1:1 were carried out in silica ampoules to maintain a constant oxygen partial pressure. CaO was freshly prepared by heating calcium oxalate dihydrate at 1000 °C for 12 h; TeO₃ was synthesized by heating telluric acid for one day at 400 °C. The educts (0.075 g CaO, 0.107 g TeO₂, 0.118 g TeO₃) were intimately mixed and sealed under vacuum in silica ampoules with an approximate volume of 8 cm³. The ampoules were heated in a muffle furnace for three days at 650 °C and 850 °C, respectively.

Single crystals of Ca₂(TeO₃)Cl₂ and Ca₅(TeO₃)₄Cl₂ were grown from CsCl or CaCl₂ fluxes using similar procedures as described above. Experimental details of selected batches are compiled in Table 1.

2.2 X-ray diffraction and crystal structure analysis

Powder X-ray diffraction (PXRD) measurements were conducted on a PANalytical X´Pert II Pro PW 3040/60 diffractometer using CuK_α1,2-radiation and an X’Celerator detector. For phase analysis, the Highscore+ software suite was used [6].

Intensity data were collected using MoKα radiation on a Stoe Stadivari system or a Bruker Kappa Apex-II diffractometer system in a dry stream of nitrogen. Measurement strategies were optimized with X-area [7] or Apex-4 [8], frame data were reduced to intensity values with X-area or Saint-plus [8] and scaled using Lana [9] or Sadabs [10]. The crystal structures were solved with Shelxt [11] and refined with Shelxl [12]. All atoms were refined with anisotropic displacement parameters. Ca₂(Te₂O₇) crystallized as a twin with four orientation states showing partial lattice overlap. The reflection spots were assigned to four distinct domains (refined values 0.4983(12), 0.1568(6), 0.2466(9) and 0.0983(6)) and integrated with overlap information (HKLF5 style reflection file). For better comparability with the closely related crystal structure of Cd₂(Te₂O₇) [13], a non-reduced cell setting was used (lattice basis (−a, −b, c) with respect to the reduced basis). Moreover, the atom names and coordinates were chosen to be as close as possible to the reported parameters of the Cd analogue.

Crystal structure and refinement data are gathered in Table 2. Further details of the crystal structure investigations may be obtained from the joint CCDC/FIZ Karlsruhe online deposition service: https://www.ccdc.cam.ac.uk/structures/ by quoting the deposition numbers specified at the end of Table 2. Bond-valence sums (BVS) [14] for the three structures were computed on basis of the parameters by Brese and O’Keeffe [15]. Selected bond length, angles and BVS values for the three structures are compiled in Table 3.

3.1 Synthesis

In the ternary system Ca–Te–O, the existence and crystal structures of several phases have been reported previously. With respect to oxidotellurates(IV), five crystallographically characterized polymorphs of Ca(TeO₃) [16, 17], one modification of Ca(Te₂O₅) [18, 19], Ca₂(Te₃O₈) [20], Ca₄(Te₅O₁₄) in its ambient [13] and its high-pressure modification [21], are reported. Furthermore, for Ca(Te₂O₅) other polymorphs (one room-temperature modification [22] and three metastable high-temperature modifications) are claimed to exist [23]. With respect to oxidotellurates(VI), Ca(TeO₄) [24], Ca₃(TeO₆) [25, 26], as well as Ca₅(Te₃O₁₄) [27] have been structurally characterized. In addition to the above mentioned phases with a single oxidation state of Te, the mixed-valent Te^IV/VI compound Ca(Te₃O₈), also known as the mineral carlfriesite [28], is part of the Ca–Te–O system.

In the present study, mixed-valent Ca₂(Te₂O₇) has initially formed in a flux synthesis from a Te^VI source, viz. Te(OH)₆, under partial reduction, indicating a complex redox chemistry for tellurium under these conditions. This peculiar behavior has also been observed previously, e.g. in phase formation studies in the Cd–Te–O system [29, 30], or for the preparation of CaTeO₃ and CaTe₂O₅ [31]. Our experiments to repeat the flux synthesis of Ca₂(Te₂O₇) as described in 2.1 were unsuccessful. Nevertheless, Ca₂(Te₂O₇) could be prepared by solid state reactions carried out at 650 °C, but only as part of a product mixture consisting of 75%_wt Ca₂(Te₂O₇) and 25%_wt Ca₄Te₅O₁₄ as revealed by phase analysis of PXRD data. Increasing the temperature to 850 °C did not result in the target phase Ca₂(Te₂O₇), but in a product mixture consisting of 89%_wt α-CaTeO₃ and 11%_wt Ca₄Te₅O₁₄. Hence, the higher temperature of 850 °C appears to be above the thermal stability range of Ca₂(Te₂O₇). Comparison of simulated XRD data on basis of the Ca₂(Te₂O₇) structure model with entry #00-056-0685 (“CaTeO₃” [32]) deposited with the powder diffraction file (PDF [33]) revealed a very good match. From the 22 reflections given for #00-056-0685, 17 could clearly be assigned to Ca₂(Te₂O₇). The presence of the five non-assigned reflections allows the conclusion that in fact #00-056-0685 consisted of a mixture of Ca₂(Te₂O₇) and at least one additional phase.

Under the given conditions (maximum temperature of 820 °C), an oxidation of Te^IV was not observed for flux synthesis batches 2–6 (Table 1). However, the formation of Ca₅(TeO₃)₄Cl₂ as a single phase product (batch 1) could not be achieved in subsequent experiments. Based on the data detailed in Table 1, the amount of CaCl₂⋅6H₂O seems to influence whether Ca₂(TeO₃)Cl₂ or Ca₅(TeO₃)₄Cl₂ are preferentially formed as reaction products. Whereas higher amounts of CaCl₂⋅6H₂O (batches 2 and 3 with a 1.5:1, and batch 5 with a 2:1 ratio relative to Te) led to higher amounts of Ca₂(TeO₃)Cl₂, lower amounts of CaCl₂⋅6H₂O (batches 4 and 6 with ratios of 0.72:1 and 1:1, respectively) were found to favor the formation of Ca₅(TeO₃)₄Cl₂. Other by-products identified in batches 2–6 are α-CaTeO₃, Ca₃TeO₆, and Ca₅Te₃O₁₄ in variable amounts (Table 1).

3.2 Crystal structure and twinning of Ca₂(Te₂O₇)

Ca₂(Te₂O₇) augments our knowledge on phases with a mixed Te^IV/VI valency in the Ca–Te–O system. In the Robin-Day classification of mixed-valent compounds [34] it belongs to class I, i.e. the different valence states each are clearly located on a single atomic site. Other mixed-valent Te^IV/VI compounds with formula type M^II₂(Te^IVTe^VIO₇) or (M,M′)(Te^IVTe^VIO₇) include Cd₂(Te₂O₇) (space group P 1 ‾ ) [13], two polymorphs (Aba2, C2/c) of Hg₂(Te₂O₇) [35], the three isotypic phases (space group Ama2) BaMg(Te₂O₇), BaZn(Te₂O₇) [36] and BaCu(Te₂O₇) [37], as well as the two isotypic phases (Pbcm) PbCu(Te₂O₇) and SrCu(Te₂O₇) [38].

From the listed M^II₂(Te^IVTe^VIO₇) or (M,M′)(Te^IVTe^VIO₇) phases, only Cd₂(Te₂O₇) shows a closer relation to Ca₂(Te₂O₇). In fact, in a projection along [100], both structures appear practically identical (Figure 1). Yet, they cannot be considered as isopointal, and in consequence isotypic, since the atoms occupy a different set of Wyckoff positions [39]. In particular, the Cd4 and Cd5 atoms in Cd₂(Te₂O₇) are located on centers of inversion, whereas the corresponding Ca4 atom is located on a general position. Moreover, even though the space group type ( P 1 ‾ ) and the number of formula units (Z = 4) are the same, and the reduced cell parameters [40] of both structures are very close (Ca₂(Te₂O₇): a = 7.4607(3), b = 8.4428(3), c = 10.1847(3) Å, α = 85.554(3), β = 79.500(3), γ = 77.307(3)°; Cd₂(Te₂O₇): a = 7.4328(7) b = 8.3346(6), c = 9.9898(8) Å, α = 87.005(6), β = 78.843(8), γ = 77.210(8)°), the relative atom positions with respect to the reduced basis differ in both structures. A non-reduced setting with the basis (−a, −b, c) with respect to the primitive basis was chosen for Ca₂(Te₂O₇), which simplifies the comparison of the structures. Using this alternative setting, the a and b basis vectors correspond well, but the c vector differs for the reasons described below.

Figure 1:

The crystal structures of (left) Ca₂(Te₂O₇) and (right) Cd₂(Te₂O₇) viewed down [100]. Trigonal-pyramidal [Te^IVO₃] units are given in orange, [Te^VIO₆] octahedra in red, M atoms in blue and O atoms in white. A subscript ‘p’ indicates out-of-plane axes projected on the drawing plane. Displacement ellipsoids are given at the 90% probability level.

In both crystal structures, a ¹_∞[Te^VIO_4/1O_2/2] chain of corner-sharing [TeO₆] octahedra (Te2, Te4) is flanked by one of the [Te^IVO₃] groups (Te3) on one side. The other [Te^IVO₃] group (Te1) is connected to the chain on the opposite side via secondary Te–O⋯Te contacts >2.6 Å (Table 3), forming a double chain extending parallel to [100] (Figure 2). The double chains are embedded into a framework of [MO _x ] polyhedra (M = Ca, Cd, x = 6–8) to leave space for oval channels (short and long diameter ≈3.7 and ≈6.8 Å) running parallel to the chain direction into which the non-bonding 5s² electron lone pairs are directed. In the crystal structure of Ca₂(Te₂O₇), the mean Ca−O distances of the four [CaO _x ] polyhedra (Table 3) are in agreement with reference values, viz. 2.448(133) Å for coordination number 7, and 2.498(151) Å for coordination number 8 [41].

Figure 2:

The oxidotellurate(IV,VI) double chain in the crystal structure of Ca₂(Te₂O₇). Secondary Te–O⋯Te contacts >2.6 Å are indicated by dotted lines. Color code as in Figure 1.

The structural differences are best analyzed by considering two distinct crystal-chemical layers parallel to (011) (extending horizontally in Figure 1). The first kind of layer is shown in Figure 3. In both structures, slightly corrugated chains of vertex-connected [TeO₆] octahedra extend along [001]. The chains are separated by the M atoms. Whereas the Cd atoms are coordinated octahedrally, the Ca atoms additionally coordinate to one of the bridging O atoms of the [TeO₆] chains, which leads to a stronger tilting of the [TeO₆] octahedra. Two of the [CdO₆] octahedra (Cd4 and Cd5) are located on centers of inversion and form a rod of edge-connected octahedra extending along [100]. In the second rod, a single crystallographically unique [CdO₆] octahedron (Cd2) is located on a general position and adjacent polyhedra are related by an inversion operation. The [CaO₇] polyhedron in Ca₂(Te₂O₇) is not symmetric by inversion and therefore both rods are of the ‘second kind’, i.e. built of a single crystallographically unique [CaO₇] polyhedron, whereby pairs of polyhedra are generated by inversion.

Figure 3:

[TeO₆] octahedra containing layers in (top) Ca₂(Te₂O₇) and (bottom) Cd₂(Te₂O₇) projected on the layer plane (011). Selected pseudo-symmetry operations in the case of Ca₂(Te₂O₇) are indicated using the common graphical symbols. Color code as in Figure 1.

A basis of the translation lattice of the layers is given by (a, b + c). The second basis vector b + c is indicated by a green arrow in Figure 3. Here, a fundamental difference between both structures manifests itself: Whereas Ca₂(Te₂O₇) possesses a pseudo-rectangular primitive layer lattice (angle between a and b + c: 89.996(3)°), the layer lattice in Cd₂(Te₂O₇) is pseudo-rectangularly centered (angle between a and 2(b + c)−a: 89.724(9)°). This means that translationally equivalent [TeO₆] chains are mapped by translations perpendicularly to [100] in Ca₂(Te₂O₇), but in Cd₂(Te₂O₇) an additional shift along 1/2a (one full [TeO₆] octahedron) has to be applied.

The effect of the different kinds of layer lattices is even more clearly observed in the second kind of layers, which contain the [Te^IVO₃] groups (see Figure 4). Again, these layers can be described as being composed of rods of [Te^IVO₃] groups connected by [MO _x ] polyhedra. In the [Te^IVO₃] rods, two distinct Te^IV-positions with markedly different orientation of the [Te^IVO₃] groups alternate along [100]. Again, in Cd₂(Te₂O₇), owing to the different layer lattices, every second rod is shifted along 1/2a with respect to the same rod in Ca₂(Te₂O₇).

Figure 4:

[TeO₃] unit containing layers in (top) Ca₂(Te₂O₇) and (bottom) Cd₂(Te₂O₇) projected on the layer plane (011). Color code as in Figure 1.

Ca₂(Te₂O₇) crystallizes as twins with four different orientation states, typically designated as domains, even though they are not necessarily contiguous. The multiplicity of the twin is therefore four [42]. However, all twin elements (geometrical elements of the operations that relate the twin domains) are of order two. The twin is thus a twofold and not a fourfold twin according to the classification of Grimmer & Nespolo [42]. Since no twin element can generate all four domains, there are independent twin elements. This makes the crystal a rather rare case of a higher-degree twin [42], in this case of degree two.

One twin operation is a twofold rotation about the [100] axis. A second non-equivalent twin operation is a reflection at the (011) plane, which is parallel to the layers discussed above. Note that owing to point symmetry 1 of the twin individuals, every twofold rotation has a corresponding reflection as twin operation and vice-versa. These two independent twin operations generate the four twin domains. The twin operations and the point symmetry of the structure generate the mmm point group, which can be seen as the overall point group of the edifice. Since all twin operations are of order two, they interchange the domains in a pair-wise fashion. The tetrachromatic twin point group [43] is therefore (2^(2,2)/m^(2,2) 2^(2,2)/m^(2,2) 2^(2,2)/m^(2,2))⁽⁴⁾.

Twinning is often due to pseudo-symmetry of distinct parts of the structure. Classical examples are local pseudo-symmetry operations pertaining to disctinct layers (OD-twin [44]) or pseudo-eigensymmetry of crystallographic orbits [45]. In the case of Ca₂(Te₂O₇), we give an OD interpretation. The layers parallel to (011) containing the [TeO₆] octahedra clearly possess a pseudo-reflection plane normal to [100] (indicated by black lines in Figure 3, top), which explains the twofold rotation about [100]. The reflection at the layer plane (011) is harder to argue. One interpretation would be the pseudo rotation axes normal to the layer planes indicated in Figure 3, top. These rotations, when combined with the centers of inversion result in a pseudo-reflection at the layer plane, which explains the observed twinning. Since the actual structure deviates significantly from this pseudo-symmetry, the twin interface in this case features a distinct distortion with respect to the macroscopic domains. This pseudo-symmetry is nevertheless the most likely cause of the observed twinning. Combining the actual pseudo-symmetry operations, the idealized layers possess pmam symmetry, for which the symbol is given with respect to the basis (a, b + c, a × (b + c)), i.e. the basis of the layer lattice and a vector perpendicular to it.

The twinning entails a complex diffraction pattern (Figure 5). It is best understood by considering the basis (a, b + c, −b) (see green arrow in Figure 3, top). The two-dimensional pseudo-rectangular lattice spanned by (a, b + c) is practically invariant under all twin operations. However, the third vector −b differs for all the four twin domains. Projecting b on a gives approximately a/4 (formally a·b ≈ 1/4a·a), which means that for h even the reflections of domains related by rotation about [100] overlap nearly perfectly, whereas for h odd, they do not (compare Figure 5, left and right). Following an analogous pattern, (b + c)·b ≈ 2/5(b + c)·(b + c) and therefore the reflections of domains related by reflection at (011) overlap when k + l is a multiple of 5. The corresponding rows of reflections are indicated in Figure 5. In a sense, the two distinct twin elements possess twin index 2 and 5, respectively.

Figure 5:

(left) h = 0 and (right) h = 1 sections of reciprocal space reconstructed from intensity data. The reciprocal basis vectors of the four twin domains are indicated by different colors. The origins of the arrows representing the basis vectors is placed at the 100 reflection of the respective twin domain on the right.

3.3 Crystal structures of Ca₂(TeO₃)Cl₂ and Ca₅(TeO₃)₄Cl₂

In the quaternary system Ca–Te–O–X (X = halogen) existing phases have not been reported so far. However, for the higher homologues Sr and Ba, the following M–Te–O–X (M = Sr, Ba; X = Cl, Br) compounds were synthesized and structurally characterized: Sr₃(Te₂O₆)Cl₂ and isotypic Ba₃(Te₂O₆)Cl₂ [46], Sr₄(Te₃O₈)Cl₄ [47], Ba₆(Te₁₀O₂₅)Br₂ and Ba₃(Te₃O₈)Br₂ [48]. Novel Ca₂(TeO₃)Cl₂ and Ca₅(TeO₃)₄Cl₂ obtained during the present study exhibit different formula types and thus are not directly related to the aforementioned quaternary phases.

The asymmetric unit of Ca₂(TeO₃)Cl₂ comprises two Ca, one Te, three O and two Cl sites with all atoms being located on a general site 2 i of space group P 1 ‾ . Three oxygen atoms coordinate the tellurium atom in the form of a trigonal pyramid. In the crystal structure, the [TeO₃] units are isolated from each other and linked by the Ca atoms. The two unique Ca sites show the same type of distorted coordination polyhedron, viz. coordination by four O atoms of the oxidotellurate(IV) units at shorter distances and by three Cl atoms at longer distances. The respective mean bond lengths are similar (d_av.(Ca1–O) = 2.394 Å, d_av.(Ca2–O) = 2.440 Å; d_av.(Ca1–Cl) = 2.826 Å, d_av.(Ca2–Cl) = 2.836 Å). The two Cl atoms in turn bind to three Ca atoms each. The two types of [CaO₄Cl₃] polyhedra share common edges to form a three-periodic framework that is percolated with channels (short and long diameter ≈4.0 and ≈4.8 Å) running along [010]. This arrangement provides the required space for the non-bonding 5s² electron lone pair of the Te^IV atom (Figure 6).

Figure 6:

The crystal structure of Ca₂(TeO₃)Cl₂ in a projection along [010]. [CaO₄Cl₃] polyhedra are given in blue. Displacement ellipsoids are drawn at the 90% probability level. Red dotted lines given for one channel indicate weak interactions between the Cl atoms and the 5s² electron lone pair E.

The asymmetric unit of Ca₅(TeO₃)₄Cl₂ comprises one formula unit, with all Ca, Te and O atoms being located on a general 8 d position, and four Cl atoms on mirror planes (4 c; site symmetry .m.) of space group Pnma. Again, three oxygen atoms coordinate each tellurium atom to form a trigonal pyramid, with all [TeO₃] units isolated from each other in the crystal structure. The five calcium atoms exhibit different coordination numbers and coordination partners. Ca1 is surrounded by five O and two Cl atoms, Ca2 by eight O atoms, Ca3 by seven O atoms, Ca4 by four O and two Cl atoms, and Ca5 by eight O atoms. The corresponding [Ca1O₅Cl₂], [Ca2O₈], [Ca3O₇], [Ca4O₄Cl₂] and [Ca5O₈] polyhedra share common corners and faces and define a three-periodic framework perforated by channels (short and long diameter ≈4.3 and ≈5.5 Å) running parallel to [001], into which the electron lone pairs of the Te^IV atoms are directed. Each Cl atom is only bonded to two Ca atoms. A view of the crystal structure is provided in Figure 7.

Figure 7:

The crystal structure of Ca₅(TeO₃)₄Cl₂ in a projection along [001]. [CaO₈] polyhedra are given in blue, [CaO₅Cl₂] and [CaO₇] polyhedra in turquoise, and [CaO₄Cl₂] polyhedra in light-blue. Red dotted lines given for one channel indicate weak interactions between the Cl atoms and the 5s² electron lone pair E.

In the crystal structures of Ca₂(TeO₃)Cl₂ and Ca₅(TeO₃)₄Cl₂, the large Cl atoms act as spacers and influence the packing motif. In both cases, together with the Te^IV atoms they are part of the atoms that define the border of the channels. As discussed previously [46, 49], weak interactions between the halogen atom and the electron lone pair E of the opposite [TeO₃] unit with Cl⋯Te distances <3.6 Å appear to be a stabilizing factor of the crystal structure (Ca₂TeO₃Cl₂: Cl1⋯Te1 = 3.5064(9), 3.5533(7) Å; Cl2⋯Te1 = 3.3870(8) Å (Figure 6). Ca₅(TeO₃)₄Cl₂: Cl1⋯Te1 = 3.3891(13) Å (2×); Cl2⋯Te2 = 3.5005(12) Å (2×), 3.5337(13) (2×) Å; Cl4⋯Te4 = 3.4768(12) Å (2×) (Figure 7)).

3.4 General aspects of the three crystal structures

The Te–O distances (Table 3) in the three crystal structures are in the expected ranges for threefold coordinated Te^IV or sixfold coordinated Te^VI and are in good agreement with the corresponding averaged literature values [50] of 1.843(15) Å and 1.923(60) Å, respectively. The [Te^IVO₃] unit is the most common coordination polyhedron observed for oxidotellurates(IV) and in almost all cases its shape is that of a trigonal pyramid [51] as also observed for the present examples. However, many examples for coordination numbers of 4 and 5 are also known if Te–O distances are considered up to 2.45 Å. Further weakly ligating O atoms are found in a range up to ≈3.0 Å defining the secondary coordination sphere of Te^IV [51]. This behavior has been reviewed for the first time by Zemann [52] and is also found for the two Te atoms in Ca₂(Te₂O₇) and for one Te atom (Te1) in Ca₅(TeO₃)₄Cl₂ (Table 3). Contrary to the peculiar crystal-chemical features of oxidotellurates(IV), the corresponding oxidotellurates(VI) exhibit almost solely octahedral [TeO₆] groups with a very narrow Te–O bond lengths distribution [51]. Likewise, this finding is also observed for the two [TeO₆] octahedra in Ca(Te₂O₇), however with somehow greater distortions due to condensation with [TeO₃] groups.

The results of bond-valence analyses for the three structures agree with the expected formal charges, with the exception of one Ca atom (Ca4) in Ca₂(Te₂O₇), which exhibits a bond valence sum of 2.42 v.u., significantly higher than expected. Such deviations are not unusual and have been observed for other alkaline earth cations in oxidotellurates, e.g. in Ba₃(Te^IV₃O₈)Br₂ [48] or in Ba₃Te^VIO₆ [53], indicating that the concepts of the bond valence model not always work properly for large cations [54].