- © 2016 The Mineralogical Society
Relative to its extremely low abundance in the Earth's crust, tellurium is the most mineralogically diverse chemical element, with over 160 mineral species known that contain essential Te, many of them with unique crystal structures. We review the crystal structures of 703 tellurium oxysalts for which good refinements exist, including 55 that are known to occur as minerals. The dataset is restricted to compounds where oxygen is the only ligand that is strongly bound to Te, but most of the Periodic Table is represented in the compounds that are reviewed. The dataset contains 375 structures that contain only Te4+ cations and 302 with only Te6+, with 26 of the compounds containing Te in both valence states. Te6+ was almost exclusively in rather regular octahedral coordination by oxygen ligands, with only two instances each of 4- and 5-coordination. Conversely, the lone-pair cation Te4+ displayed irregular coordination, with a broad range of coordination numbers and bond distances. A threshold was applied for Te4+–O links of ∼2.45 Å or 0.3 valence units with some flexibility, as a criterion to define strongly bound Te–O polymers and larger structural units. Using this criterion, Te4+ cations display one-sided 3-, 4- or 5-coordination by oxygen (with rare examples of coordination numbers 2 and 6). For both valence states of Te, examples are known of TemOn complexes which are monomeric (m = 1; neso), noncyclic finite oligomers (soro), rings (cyclo), infinite chains (ino), layers (phyllo) and frameworks (tecto tellurates). There is a clear analogy to the polymerization classes that are known for silicate anions, but the behaviour of Te is much richer than that of Si for several reasons: (1) the existence of two cationic valence states for Te; (2) the occurrence of multiple coordination numbers; (3) the possibility of edge-sharing by TeOn polyhedra; (4) the possibility for oxygen ligands to be 3-coordinated by Te; and (5) the occurrence of TemOn polymers that are cationic, as well as neutral or anionic. While most compounds contain only one or two symmetrically distinct types of Te atom, Pauling's Fifth Rule is frequently violated, and stoichiometrically simple compounds such as CaTeO3 can have polymorphs with up to 18 distinct Te sites. There is a tendency for local symmetry features such as the threefold axis of a TeO6 octahedron or the acentric symmetry of a Te4+On polyhedron to be inherited by the host structure; the latter in particular can lead to useful physical properties such as nonlinear optical behaviour. We develop for the first time a hierarchical taxonomy of Te-oxysalt structures, based upon (1) valence state of Te; (2) polymerization state of TemOn complexes; (3) polymerization state of larger strongly-bound structural units that include non-Te cations. Structures are readily located and compared within this classification.
Tellurium (Te) is an unusual element in that its cosmic abundance is greater than that of any other element with an atomic number >40, as measured by relative number of atoms in C1 chondrite (Anders and Ebihara, 1982). Nevertheless, Te is one of the rarest elements in the Earth's crust (0.4‒10 ppb; Parker, 1967; Levinson, 1974; Govett, 1983; McDonough and Sun, 1995; Reimann and de Caritat, 1998) and also in seawater (up to 0.0009 ppb; Andreae, 1984; Lee and Edmond, 1985). It is thus 3 to 5 orders of magnitude less abundant than other even-number elements that are nearby in the periodic table, such as tin and barium, and is in fact rarer than platinum or gold.
The extreme depletion of Te in the Earth's crust is probably due to its strongly siderophile character at high pressure, which resulted in much primeval Te being sequestered in the core, and the small amounts of Te in the outer layers of the Earth arriving after core formation in a “late veneer” (Wang and Becker, 2013). The extreme scarcity of Te makes it all the more remarkable that there are ∼160 Te minerals described from Nature: ∼3% of all known species. Christy (2016) showed that most chemical elements show a power-law dependence between their abundance in the Earth's crust and the number of mineral species in which they are essential constituents. Other elements that are major constituents of 150–200 species are much more abundant, such as Ce and Ni, present in the crust at 33 and 105 ppm, respectively, according to Taylor and McLennan (1985). Conversely, if Te followed the typical trend, there would be only seven Te minerals. Tellurium is, in fact, the most extreme example of an element that forms an anomalously large number of distinct species in the Earth's crust. Telluride minerals, containing Te as an anion, are probably best known, and are well studied due to their association with gold in epithermal Au–Te deposits (cf. Cook and Ciobanu, 2005; Ciobanu et al., 2006), often related to alkaline magmatism (e.g. Jensen and Barton, 2000). Rare sulfosalts are also known in which cationic Te4+ plays a role analogous to As3+, such as the tetrahedrite-group mineral goldfieldite, ideally Cu102(TeS3)4S (Trudu and Knittel, 1998). Hence, Te can adopt either anionic or cationic roles as a chalcophile element, like As and Sb. Also, like those elements, it oxidizes readily to form secondary oxycompounds under near-surface conditions. About half of the known Te minerals are such tellurites and tellurates.
The recent explosion of new secondary mineral species, particularly from Otto Mountain, California, has seen publication of descriptions for 14 new Te minerals from 2010 up to September, 2015 (Kampf et al., 2010a; Back et al., 2011; Housley et al., 2011; Pekov et al., 2010; Christy et al., 2016). This represents the greatest flurry of activity in the study of Te secondary minerals since the 1970s. The majority of these new minerals are also compounds new to inorganic chemistry, and possess new crystal-structure types. Crystal structures are now known for 55 of the ∼80 Te oxyminerals. It is of particular interest that the Te oxyanions show a wide range of polymerization, somewhat analogous to silicates: examples range from isolated [Te4+O3]2– and [Te6+O6]6– anions to complex three-dimensional frameworks. The analogy to rock-forming silicates is strengthened by the observation that, in a locality with an unusually large number of tellurate species, there appears to be a correlation between polymerization state and both the early or late position of a mineral in the local paragenetic sequence, and the abundance of ‘network-modifying’ species such as Cu2+ (Christy et al., 2016).
A search of the Inorganic Crystal Structure Database (ICSD) and recent literature has found good-quality crystal structures for 703 compounds, in all. The number of structures referenced per year for the present study suggests that the rate of synthesis and structure refinement has been increasing through time, and that 40 new compounds and structures per year may now be typical (Fig. 1). Thus, the current interest in both synthetic and natural Te oxycompounds, along with the anomalously large number of the latter, justifies a review of their structural chemistry. It should be noted that new compounds appear in the literature constantly, but we had to stop updating our working list in mid-2015, in order to avoid repeated shuffling of the database and the associated risk of introducing errors.
Examination of the known structures of Te oxycompounds reveals extraordinary diversity due to a combination of factors, namely: (1) Te may occur as Te4+ or Te6+, which are of comparable stability under atmospheric conditions, so compounds also occur with both oxidation states coexisting. (2) Te6+ is almost invariably octahedrally coordinated by oxygen (Christy and Mills, 2013). The Te6+O6 group is strongly bound, in that the average Te–O bond valence is unity. In contrast, Te4+ has a stereoactive lone electron pair, and may adopt a wide range of coordination geometries. Usually, three to four oxygens are strongly bound to form an asymmetric coordination polyhedron, but there may also be several other neighbours at longer distances (Christy and Mills, 2013). (3) As noted above, TeOn polyhedra polymerize readily to form oligomers, chains, layers and frameworks. These units also link readily to other strongly-bonding cations to form heteropoly structural building units. (4) The geometries of TeOn polymers are even more flexible than those of silicates, in that the polymers may contain Te with various coordination numbers, and may carry not just negative or zero net charge, but may also be positively charged in Te ‘salts’. For example, a [Te24+O3OH]+ infinite layer cation can be identified in (Te2O3OH)(NO3) (Anderson et al., 1980), while the [Te4+(OH)3]+ ion has been recently identified in the structure of Na11H[Te(OH)3]8[SO4]10(H2O)13 by Mills et al. (2016). (5)Te polyhedra readily share edges, as well as corners, in contrast to SiO4 tetrahedra. (6) Oxygen may be coordinated by three Te4+, as in winstanleyite, TiTe3O8 (Bindi and Cipriani, 2003). This possibility does not arise in conventional silicates because the short Si–O distance causes strong Si…Si repulsion, although edge-sharing of non-silicon tetrahedra and 3-coordination of oxygen atoms are seen in beryllosilicates and zincosilicates, where the lower cation valence decreases repulsion, and gives a small effective non-bonded radius relative to bond distances (O'Keeffe and Hyde, 1981). The longer bond distances make such geometries possible for Te–O polyhedra as well.
A structural hierarchy for silicates (Bragg, 1930; Zoltai, 1960; Liebau, 1985) is used widely to organize classic textbooks such as Deer et al. (1966). More recent schemes, such as those for borate (Hawthorne et al., 1996) and sulfate minerals (Hawthorne et al., 2000), render intelligible the diversity of these large, complex classes, highlight structure-composition-property relationships and facilitate comparison between species, and also aid in applying group nomenclature (Mills et al., 2009b). A major objective of the present study is to create such a structural hierarchy for Te oxycompounds.
In NMR spectroscopy, a concise ‘Q notation’ to describe polymerization states of silicate species, in which ‘Qn’ (n = 0–4) designates silicate tetrahedra with n bridging oxygen atoms and, by implication, 4-n non-bridging oxygen atoms (cf. Lippmaa et al., 1980). It would be convenient to use a similar notation in the present study for TeOn polyhedra. However, the original symbology makes the assumptions that (1) the coordination of Si is always 4; (2) the coordination of O by Si is either 1 (non-bridging) or 2 (bridging) and, concomitantly; (3) the number of non-bridging oxygen atoms is the same as the number of next-nearest neighbour Si atoms. For Te, all three of these assumptions are invalid, as they are violated as a result of variable Te coordination number (CN), plus the possibilities of edge-sharing and CN3 oxygen. More information is needed to fully specify the polymerization state of a Te cation, including the numbers of oxygen ligands connected to 1, 2 or 3 Te cations, and the number of edges (i.e. 4-rings, Te–O–Te–O) shared between Te polyhedra. An extended notation Qabcz can do this using four single-digit integers: a = number of CN1 oxygen atoms, b = number of CN2 oxygen atoms, c = number of CN3 oxygen atoms and z = number of shared edges. The total Te CN = a + b + c, the corresponding number of oxygen ligands per Te is a + ½ b + ⅓ c, and the number of next-nearest Te atoms is b + 2c – z. The original silicate Qn would be written Q(4–n)n00 in the extended notation (Q0 ≡ Q4000, Q1 ≡ Q3100···Q4 ≡ Q0400). Note that 0 ≤ z ≤ ½ p(p–1), where p = b + c.
A very large number of different Qabcz states can occur. For 4-coordinate Te alone, there are 80 possibilities, and 17 of these are found in the structures of the present study. In order to illustrate the value of the notation, the corresponding topologies are shown very diagrammatically in Fig. 2.
Te–O bond length and polyhedral geometry
‘Bond valence’ is a parameter that expresses the strength of a chemical bond between a cation and an anion in terms of the effective number of electron pairs involved in bonding. It is thus a generalization of the concept of ‘bond order’, well entrenched in organic chemistry (IUPAC, 1997), and of the “electrostatic bond strength” of Pauling (1929). The bond-valence model relates bond distance, r, to bond valence, s, for a given cation‒anion pair via a smoothly varying function of two parameters. The equation most often used is r = r0 – blns, where: r0 is the distance at unit bond valence, and b, a ‘softness’ parameter (Brown and Altermatt, 1985; Brese and O'Keeffe, 1991; Brown, 2002). Bond valences and their sums on a central atom are powerful crystallographic tools for distinguishing species of similar scattering factor, but different valence, and for identifying species such as O2–, OH– and H2O and hydrogen bonds when H cannot be located in structure refinements.
While a universal softness value b = 0.37 Å is often assumed (e.g. Brown and Altermatt, 1985; Brese and O'Keeffe, 1991), it has become apparent that this does not model the bonding behaviour well for many heavier atoms. Several alternative parametrizations for particular species have been published in recent years [e.g. for Pb2+ by Krivovichev and Brown (2001); U6+ by Burns et al. (1997); Tl1+ by Locock and Burns (2004); and Sb3+ and Sb5+ by Palenik et al. (2005), Sidey et al. (2009) and Mills et al. (2009a)], and we examined the available structural data for Te4+–O, Te4+–Cl and Te6+–O bonds in a recent paper (Mills and Christy, 2013). For Te4+‒O, we obtained the parameters r0 = 1.9605 Å and b = 0.41 Å, while for Te6+‒O, we obtained the parameters r0 = 1.921 Å and b = 0.56 Å. We considered all Te‒O distances out to 3.5 Å as at least weakly bonded. Conversely, Te4+ showed a broad distribution of coordination numbers from 3 to 12, with two modes at CN6 and CN8 (fig. 1 in Mills and Christy, 2013). The distribution of Te4+‒O bond distances is also bimodal (Fig. 3). Thus, the oxygen neighbours of Te4+ separate into two groups: strongly-bound ligands on the opposite side of the Te from its stereoactive lone electron pair, with bond valences typically 0.3–1.3 valence units (vu) (corresponding to a distance of 1.85–2.45 Å), and more distant ligands, with bond valences usually < 0.15 vu (2.74 Å). When three short Te4+‒O bonds are present, the Te4+O3 geometry is invariably a rather symmetrical trigonal pyramid, with oxygen atoms at three of the four corners of a tetrahedron, and the lone pair of the Te directed towards the fourth vertex (Fig. 4a); any more distant oxygen atoms are on the same side of the Te as the lone pair. An apparently unique alternative to this geometry for 3-coordination is seen in Nd[Te2O5]Br, where the three Te‒O bonds are coplanar, forming a ‘T’-shape (Tarasov et al., 1996; described as structure #285 below). Te4+O4 may be either square pyramidal, with four O‒Te‒O angles approximately equal (Fig. 4b), or have the oxygen atoms at the apices and two equatorial positions of a trigonal bipyramid, with the lone pair replacing the missing ligand (Fig. 4c). The rarer examples of Te4+O5 approximate octahedra with the lone pair replacing one ligand (Fig. 4d).
Subsequent investigation of the Te4+O6 subset of these data by Christy and Mills (2013) showed that, although the Te–O distances within a given polyhedron could show a large variance, the oxygen atoms of a Te4+O6 polyhedron (including long bonds) generally fall very nearly on the surface of a sphere, not centred on the Te, but on a point ∼1 Å away from the Te atom, which is consistent with the centre of lone-pair electron density. However, the Te–O distances within a given polyhedron could show a large variance, and the radius of the sphere of oxygen atoms increased linearly with the tellurium lone-pair distance (Christy and Mills, 2013). The volume of the Te4+O6 polyhedron varied, depending on both the sphere radius and the uniformity of the distribution of oxygen atoms over the sphere's surface. The polyhedra ranged in volume from nearly twice that predicted for a regular Te4+O6 octahedron, when the oxygen atoms are uniformly distributed and the lone pair is highly stereoactive, to 20% less than that of a regular octahedra, when the oxygen atoms are crowded on one side of the sphere. Extreme flexibility in the volume of coordination polyhedron is, thus, another feature of lone-pair cations, such as Te4+, which can contribute to their accommodation in a wide range of crystal structures.
A further unusual feature, occasionally noted, is that bonding interactions of lone-pair cations are not limited to those of the conventional cation‒anion type. Christy and Mills (2013) found that the most compressed Te4+O6 polyhedra also showed short distances between the lone pair and large cations, such as K or Ba and/or other Te4+ cations, suggesting that weak monopole‒dipole (K,Ba)‒lone-pair and dipole‒dipole Te4+‒lone-pair interactions also help to stablize the relevant structures. These non-classical ‘bonds’ complicate the application of the bond-valence model to structures containing stereoactive lone pairs.
Mills and Christy (2013) verified the strong preference of Te6+ for octahedral geometry: that dataset included 100 examples of Te6+O6 octahedra, plus another five where six additional oxygen atoms lay near the cut-off distance, but only 13 examples of other coordinations. Examples of polyhedra with CN4‒6 are shown in Fig. 4e‒g.
The chemical and structural diversity of Te oxycompounds
In the current study, we examine the 703 Te oxycompounds for which good structure refinements are available. Where multiple refinements were available for a compound, one of the better ones was selected. Structures with obvious errors or that were of solid-solution variants of a pure end-member were usually rejected. The dataset includes 55 mineral species, about two-thirds of those described to date.
The frequency of occurrence of specific elements as essential constituents in these 703 compounds is shown on periodic tables in Fig. 5. Apart from Te and O, the most common constituents in the mineral species are Cu and Pb (22 out of 55), H (21), Cl (9), Fe (8) and Zn (6). While H, Cu, Cl and Pb are also important in the dataset as a whole, including synthetic compounds (223, 70, 68 and 42 compounds respectively out of 703), many of the latter also include Mo (77), Na (73), K (65), N (49), Ba (48) and P (38). The alkali metals Na and K are common as counteranions in many laboratory-crystallized Te oxysalts; while the number of N compounds is boosted by the analogous use of the ammonium ion, NH4+. The anomalously large number of Mo compounds is due to the large number of salts of the tellurohexamolybdate anion, [Te6+Mo66+O24]6–, that have been prepared, while the majority of P compounds are hydrogen-bonded adducts of Te(OH)6 with alkalis and various phosphate anions.
The large proportion of Cu, Pb and H minerals is consistent with these elements, like Te, showing unusually high mineral diversity (Christy, 2015), and the common association of primary telluride minerals with sulfides of Cu and Pb. It is surprising that so few secondary Te minerals containing As, Sb and Bi are known, as these are also mineralogically diverse chalcophile elements. Syntheses of many Cu and Pb tellurates were probably attempted because of the importance of such phases as minerals.
Definition of Te4+ coordination and structural unit
Mills and Christy (2013) chose 3.5 Å as a cutoff distance for inclusion of weak bonds to oxygen in the Te4+ coordination sphere. The corresponding bond valence is ∼0.023 vu. In contrast, the current study is concerned primarily with the strongest bonds of a structure, which define structural building units. For this study, we divide the ‘primary’ and ‘secondary’ Te4+–O bonds at the minimum in the probability distribution between the two modes of Fig. 2. The threshold bond distance is thus 2.40–2.45 Å, corresponding to a bond valence of 0.34–0.30 vu, using the bond-valence parameters of Mills and Christy (2013). Note that this division is consistent with Hawthorne (2014) and references cited therein, who use a bond balance of ∼0.30 vu to differentiate, in crystal structures, between the more strongly bound ‘structural unit’ and weakly bound ‘interstitial complex’. The bond-valence threshold identifying bonds that form the structural unit is employed with some flexibility. The smaller divalent octahedral cations Mg, Zn, Fe2+ and Mn2+, with bond-valence close to 0.33 vu, are part of the structural unit if they bond to tellurate oxygen atoms. However, for Cu2+O4+2 octahedra, elongated due to Jahn-Teller distortion, it was usually the case that only the four shortest bonds were strong enough to be included. Weaker bonds to these small cations were sometimes included, if needed to preserve the integrity of a well-defined coordination polyhedron. Conversely, the larger divalent cations Ca, Sr, Ba and Cd typically occurred with CN > 6 or a mixture of sixfold and higher coordination numbers, and were not generally included, unless they occurred on sites that were occupied by small cations in isostructural compounds. Other large cations with CN ≥ 7 (e.g. REE3+, Zr4+ and Th4+) or highly irregular coordination (Pb2+ or Bi3+) were similarly excluded from the structural unit, except for U6+On polyhedra (n = 6–8). The dimensionality of the heteropoly structural unit was often higher than that of the Te oxyanion alone, as is apparent below.
When long bonds are excluded, the ranges of coordination numbers for the Te–O bonds included in the present study were between 2 and 6 for Te4+ and between 4 and 6 for Te6+. The ‘2-coordinate’ Te4+ of Bi2(TeO3)2O has additional ligands at just over 2.5 Å (Mercurio et al., 1998), while the 6-coordinate examples have the pyrochlore structure type, with Te4+ in octahedral coordination (Loopstra and Goubitz, 1986; Weber and Schleid, 2000). The other polyhedra are all of the types shown in Fig. 3 above. For 428 symmetrically distinct Te6+On polyhedra, frequencies were 2, 2 and 424 for n = 4, 5 and 6, respectively. For 846 symmetrically distinct Te4+On polyhedra, frequencies were 1, 535, 271, 37 and 2 for n = 2, 3, 4, 5 and 6, respectively, although it should be noted that, while the distribution is little changed for Te4+-only compounds (for which the numbers are, respectively, 1, 530, 257, 24, 1), the small sample of mixed-valence compounds show a much greater preference for 4- and 5-coordination (frequencies for CN = 2, 3, 4, 5 and 6 are 0, 5, 14, 13 and 1, respectively).
We have classified the diverse Te-bearing moieties using a set of nested criteria, as follows: (1) Structures are separated into three oxidation-state taxa: those that contain only Te4+ as an essential major constituent, those that contain only Te6+, and those that necessarily contain both Te4+ and Te6+. (2) Within each oxidation-state taxon, we consider only the Te and its strongly-bound oxygen atoms. The next level of subdivision is on the basis of dimensionality of the TemOn species. By analogy with the silicates (e.g. Deer et al., 1966), we use the categories (dimensionality taxa): neso (m = 1), soro (non-cyclic finite groups with m > 1), cyclo (finite groups containing a ring of at least 3 Te), ino (infinite chains), phyllo (infinite layers) and tecto (infinite frameworks). When more than one distinct type of TemOn group is present, the highest-dimensional group with largest m and n determines the classification. (3) Within each dimensionality taxon, species are arranged in an order that facilitates further subdivison, if justified. Cyclo-, ino- and phyllotellurates are first separated depending on whether there is a single or multiple ring/chain/layer. They, along with neso/soro/tecto cases, are then ordered according to the number of Te and anions in the finite complex or, for infinite polymers, the translational repeat unit. (4) Finally, we consider linkage to non-tellurium cations to make larger heteropolymeric ‘structural units’.
Note that consistent focus on Te oxyanions sometimes leads to rather counterintuitive divisions between the ‘Te oxyanion’ and the ‘rest of the structural unit’. For example, on the basis of highest-valence bonds, the structure of mroseite, Ca2Te24+O4(CO3)2, can be divided into two weakly-bonding Ca2+ cations, two carbonate groups and a neutral [Te2O4]0 residual complex that consists of a pair of edge-sharing TeO3 pyramids (cf. Fischer et al., 1975). The formula as written above emphasizes this analysis. However, one oxygen atom of each carbonate triangle also links to a Te via a bond that is strong enough to fall within the bond-valence threshold, to make a larger structural unit that is a finite carbonatotellurite cluster [Te2C2O10]4–. This can be written hierarchically so as to emphasize the carbonate groups, while not showing the full Te coordination, as [(Te2O4)(CO3)2]4–, or alternatively, so as to show the Te coordination, but breaking up the carbonate groups, [(CO2)2(Te2O6)]4–. In the Tables below, mroseite is classified as having an edge-sharing [Te2O6]4– dimer, but both versions of the structural formula are used in the tables and text below, depending on context. Other compounds, in which oxygen atoms are shared by Te and other high bond-valence cations, are treated similarly, that is, with more intuitive or compact versions of formulae alongside structural formulae that emphasize Te environments.
Because the C‒O links in mroseite have very high bond valence (∼1.33 vu), in order to avoid overbonding of the oxygen, the bond from Te to the carbonate oxygen atom is longer and weaker than the other Te‒O bonds: 2.313 Å = 0.42 vu using the parameters of Mills and Christy (2013), as opposed to 0.80‒1.32 vu for the other Te‒O bonds. Similar weak bonding is observed when Te shares an oxygen atom with other high bond-valence cations, and the need to reduce bond valence can increase the Te coordination number. Out of the 24 examples of TeO5 polyhedra in Te4+-only compounds, 16 (67%) have Te4+‒O‒P5+, Te4+‒O‒As5+ or Te4+‒O‒Se4+ links, where the non-Te cation makes a bond of 1.25‒1.33 vu. Interestingly, Te4+ compounds with S6+ and V5+ do not show the same trend, because they tend instead to have very strongly bonded CN1 oxygen atoms on the non-Te cation, thus reducing the valence of the bond to the bridging oxygen atom.
Crystal structure symmetry and complexity
The Te oxysalts in the present study show a nearly even split between Te valences: the dataset contains 375 structures with Te4+ only, 302 with Te6+ only, and 26 with Te in both valence states. Interestingly, the distribution of structures between different crystal systems is quite distinct for the different valence states. As shown in Fig. 6, structures with only Te4+ are significantly more likely than average to be monoclinic or orthorhombic, and less likely to be trigonal, while the converse is true for structures that have only Te6+. Structures that include both valences are particularly likely to be orthorhombic, while having surprisingly few triclinic examples. These differences suggest that, to a degree, the symmetry of the overall structure inherits (or at least is influenced by) the point symmetry of the Te oxyanion. The low symmetry of coordination polyhedra such as those of Figs 4c‒d may make low-symmetry Te4+ structures more numerous, while TeO6 octahedra (Fig. 4h) are likely to have at least one threefold rotation axis, which enhances the number of trigonal and cubic Te6+ phases.
The polarity due to lone-pair stereoactivity in Te4+, in combination with the capacity for local symmetry inheritance by the structure, suggested that there might be a dependence of centrosymmetry on Te valence. However, the percentages of Te4+ and Te6+ structures lacking a centre of inversion were, respectively, 18.9% (71 out of 375) and 17.5% (53 out of 302), not significantly different from each other or the overall average of 18.6%. A higher proportion of acentric structures did occur for the mixed-valence structures (7 out of 26 = 26.9%), but this is also insignificant, given the small sample size. Further subdivision of the dataset by Te polymerization and coordination number did reveal two small groups with significantly high proportions of acentric structures. This was the case for eight out of 24 of the structures with isolated Te4+O4–5 polyhedra and five out of nine structures with mixed-valence layer anions, suggesting that there is a slight tendency to inheritance of polarity.
The structures in the present study markedly violate ‘Pauling's Fifth Rule’ that “the number of essentially different kinds of constituents in a crystal tends to be small” (Pauling, 1929), although Burdett and McLarnan (1984) noted that there is no a priori reason for such parsimony, except as an indirect corollary of some of Pauling's other rules. For both Te4+ and Te6+, the average number of symmetrically distinct polyhedra per structure is greater than unity, there being a total of 846 + 428 = 1274 distinct polyhedra for the 703 structures. Fig. 7 shows the percentages of the 375 Te4+-only, 302 Te6+-only and 26 mixed-valence structures that have different numbers of symmetrically distinct Te sites. Structures with larger numbers of distinct sites are generally less numerous, although a quarter of Te4+-only compounds still have between 3 and 5 distinct Te sites. The Te4+-only compounds also include one example each of structures with 9, 10 and 18 distinct Te sites. The last of these is a polymorph of CaTeO3 (Stöger et al., 2009), dramatically demonstrating that simplicity of formula does not imply simplicity of structure.
Detailed tabulation and descriptions of Te oxysalt structures
The diversity of TemOn polymers is summarized in Tables 1 through 7, which order the different Te‒O topologies according to the hierarchical principles given above. These tables serve, additionally, as an index for the listings of individual structures that follow in Tables 8–26. These tables have been deposited with the Principal Editor of Mineralogical Magazine and are available from www.minersoc.org/pages/e_journals/dep_mat_mm.html. Note that the tables separate Te6+X6 monomers into two groups: compounds that contain anionic [TeO6–x(OH)x](6–x)– groups and those that contain neutral Te(OH)6 molecules which form hydrogen-bonded structures with cations, anions, H2O and polar organic molecules. For conciseness, specific structures are referenced below by the unique ordinal number that they are assigned in Tables 8–26 (deposited), where the corresponding literature reference is cited. These structure numbers will be prefixed with ‘#’ and highlighted in boldface.
The finite oligomeric (soro and cyclo) Te4+ oxyanions of Table 1 have the topologies shown in Fig. 8. The numbers of Te atoms in these complexes range from 2 to 8, although the structure number ranges of Table 1 make it clear that some configurations are strongly preferred: we have 20 examples of the trimer Te3X8 (Fig. 8g), 16 of the simple dimer Te2X5 (Fig. 8a) and 14 of Te4X11 (Fig. 8k). All of these groups are finite linear chains of corner-sharing TeOn polyhedra, but while the dimer has Te in only 3-coordination, the other common anions show a tendency to alternate between 3-coordinated and 4-coordinated Te, which is also widespread among the less usual polymers (cf. Figs 8c,g,j,k,m,o). Tellurium in five-fold coordination is rare, and seems to be a characteristic of compounds that contain other highly electronegative cations such as P, As and Se (Figs 8e,f,i). Corner-linkage of TeOn polyhedra through a 2-coordinate bridging oxygen atom is by far the most common polymerization mechanism, but there are also examples of edge-sharing through two such oxygen atoms (Figs 8b,e,j,o) and linkage through 3-coordinate oxygen atoms (Fig. 8h). The wide range of possibilities available allows formation of isomers with the same composition, but different topologies – Figs 8b and 8c provide an example. Most of the polymers are unbranched soro chains, although Fig. 8l shows an open-branched pentamer, Fig. 8n is a cyclo 6-ring, and Fig. 8i defies classification in the scheme of Liebau (1985), because its three CN5 Te atoms are Q3111 in our extended Q notation, joined through a mutually shared CN3 oxygen atom, as well as through additional bridging oxygen atoms of the conventional CN2 type.
The increased diversity of TemXn polymer topologies relative to silicates is further evidenced by the chain structures collected in the present study. Single-chain topologies are listed in Table 2 and depicted in Fig. 9. Note that although the first entry in Table 2 appears to be a simple einer chain TeX3 in the terminology of Liebau (1985), with all Te atoms translationally equivalent (#281), the bridging oxygen atom is split between two half-occupied positions, suggesting that the crystal structure as published shows an average of disordered zweier chains Te2X6. The stoichiometrically simplest chain type that occurs is a zweier edge-sharing chain of CN4 Te, Te2X4. Note that if all cations are Te4+ and all X are O2–, then this is a neutral complex [Te2O4]0 rather than a chain anion, as is the case in the example Ag(TeO2)(NO3) of Fig. 2a (#282). The mineral telluroperite, Pb3TeO4Cl2 = Pb2(PbTeO4)Cl2, contains topologically similar chains in which the cations Pb2+ and Te4+ are disordered in a 1 : 1 ratio to give an anionic chain [PbTeO4]– (#283).
The most common coordination of Te in the chains is 4. However, CN3 also occurs in Figs 7b,e,g,h,j,k,m,o and CN5 in Figs 9d,j,m,n. Although polymerization is usually achieved through CN2 bridging oxygen atoms, the chains of Figs 9f and 9l also feature edge-sharing. The repeat unit along the chain backbone is most often 2 (zweier), although the chain of Fig. 9h is dreier, and others are vierer (Figs 9e,f,j,m,n) or sechser (Figs 9k,l,o). The chains in Figs 9i,m,n,o have open branches, which attach to a Te cation of the chain backbone through CN2 oxygen in most cases, but via a CN3 bridging oxygen for the chain of Fig. 9i. Figure 9j shows a loop-branched chain, in which a succession of 4-membered rings are united through common vertices.
Overall, the most common chain configurations are corner-sharing types with alternating CN3 and CN4 (eight instances of the zweier chain of Fig. 9b, four of the vierer chain of Fig. 9e) and CN4 chains with regularly-spaced edge-sharing links (four examples of the denningite-type vierer chain of Fig. 9f and six of the spiroffite-type sechser chain of Fig. 9l). Topographic isomers are common: the four different configurations Figs 9e‒h all have the same Te4X10 stoichiometry.
A chain is defined as multiple if it is possible to selectively remove some Te‒O links so as to break it into two or more subchains that themselves remain continuous. The dataset of this study contains several types of double chain, as well as a triple chain and even a quadruple chain. These are listed in Table 3, and depicted in Fig. 10.
The simplest double chain found in this study is the uncharged einer double chain Te2O4 of Bi3(Te2O4)(TeO3)2Cl5 (#326). The bridging oxygen atoms of each subchain make a third Te‒O bond, linking the two subchains together (Fig. 10a), to make a chain of Q1032 Te polyhedra that is almost an infinitely extended homologue of the finite trimer in Fig. 8i. All but one of the other multiple chains have zweier periodicity along the chain length, but show a remarkable range of complexity in the connection patterns between chains. One of the two inequivalent subchains of Fig. 10b can be regarded as loop-branched: backbone Te are Q1300, but connect to additional Q2110 Te, making 3-rings, which, in turn, link to the unbranched subchain of Q2110 Te via the CN3 bridging oxygen. The zigzag pattern of 6-rings in Fig. 10c can be formed through conventional corner-linkage of open-branched subchains. This is also the case for the isomeric chain of Fig. 10d, except that the branches there do not link directly to the other subchain, but instead share edges to form a (Te = Te) pair that is not part of either backbone. The only dreier double chain (Fig. 10e) has unbranched subchains that corner-link directly to form a ribbon of 5-rings, reminiscent of the 6-rings of Fig. 10c. Both of these structures occur for Fe tellurates. The most complex double chain occurs for the chemically simple compound Na2Te4O9 (#332; Fig. 10f). Each zweier subchain backbone can be regarded as loop-branched, so that the subchains are each made of linked 5-rings. However, the loops of the subchains join via a shared edge and two corner-linkages to form a cluster Te4On in which two 3-rings are united at a common shared edge. An isolated cluster with the same topology occurs for Te6+ (Fig. 13d). It will be seen below that this ‘double-triangle’ moiety appears to be unusually stable, recurring as part of several larger polymers. The topology of the zweier triple chain of Fig. 10g is similar to that found in silicates for jimthompsonite and related ‘biopyriboles’ (Veblen and Burnham, 1978), but with additional side branches. Finally, Fig. 10h shows a quadruple chain which exhibits almost every complexity known in Te oxyanions. It contains Te with CN3, 4 and 5. While the two outer chains are zweier, the two central chains are dreier. Furthermore, the central chains join to each other through Te‒O polyhedra sharing edges and corners, to form again the ‘double-triangle’ cluster seen in Fig. 10f. Conversely, the central chains join to the outer chains less conventionally, via oxygen atoms which are CN3, as they also join to additional Te cations, so that the outer chains can be regarded as sequences of trimeric clusters resembling those of Fig. 8i.
Polyhedra Te4+Xn polymerize further to form layers, which may attain considerable complexity. The simplest single layers have TeX4 square pyramids that link via corners to form a sheet TeX2 (Fig. 11a), in which all Te are equivalent. All or half of these polyhedra may be capped by an additional ligand to form sheets with either TeX3 (Fig. 11c) or Te2X5 (Fig. 11f) stoichiometry. A very different type of layer with Te:X = 2 : 5 is seen in Fig. 11b. This compound, ideally Bi(Te2O5)Cl shows considerable structural disorder, and has all Te equivalent in its average structure. The Te shows short distances to one capping ligand (O1), three ∼75%-occupied CN3 oxygen atoms (O2) and six ∼25%-occupied CN2 oxygen atoms (O3). Short distances mean that O2 cannot be occupied simultaneously with its three nearest O3 sites, and O3 cannot be occupied simultaneously with its nearest O2 or its two nearest O3 sites. Figure 11b shows the most symmetrical way of satisfying these short-range order constraints, with ¾ of the Te in distorted 5-coordination (Q1220) and ¼ in the capped pyramidal coordination of Fig. 3d (Q1030). Figure 11d shows a (Te0.5Sb0.5)2X3 sheet in which double chains of the type shown in Fig. 10a, featuring CN3 oxygen atoms, are connected through additional Te–O–Te links to produce a sheet with Q0132 cations [(SbTe)O3]+ that is positively charged, rather than anionic. The layer in Fig. 11e has Te2X4 stoichiometry. All Te are CN4 and all X are CN2, but each TeX4 polyhedron shares one edge, so that it links to only three others (Q0401 configuration). Figure 11e shows the [Te2O3OH]+ complex from (Te2O3OH)(NO3), which again is cationic rather than anionic. However, the neutral sheet [Te2O4]0 of tellurite, TeO2 (#341) has the same topology, but is much more deeply corrugated. When TeX4 polyhedra share three corners only (Q1300) to form 6-rings, a Te2X5 sheet such as Fig. 11g is obtained, topologically similar to the silicate sheet of micas, but less regular geometrically. As is the case for phyllosilicates, the non-bridging oxygens can be distinguished between those that point ‘up’ and those that point ‘down’ relative to the overall plane of the layer, and different up/down ordering patterns of apical oxygen atoms may occur. In the present study, most examples (including that of Fig. 11g) show alternation of pairs of ‘up’ and pairs of ‘down’ polyhedral apices. However, one of the dimorphs of Li2Te2O5 (#351) has all apices oriented in the same direction, similar to the micas.
The trigonal Te3X7 sheet of Fig. 11h has CN2 oxygen atoms linking Te into 3-rings and additional CN3 oxygen atoms forming 6-rings (Q1210 configuration); note that the hybrid double chain of Fig. 10b is actually a slice of this structure. Figure 11i shows a Te4X10 sheet with 4- and 8-rings, which again is strongly analogous to a well-known silicate structure, apophyllite (Colville et al., 1971). The isomeric Te4X10 structures of Figs 11j and 11k both have all Te 4-coordinated, but in 10-rings only, which requires some Te to link to only two others rather than three, and hence, for some edges to be shared. In Fig. 11j, the edge-sharing Te are Q1301 and the others Q1300, while in Fig. 11k, the edges are shared by Q0401 polyhedra and the others are Q2200. The isomeric pair of Te8X18 alkali tellurite sheets in Figs 11l and 11m have no shared edges, but both have ¼ of the Te in CN3 rather than CN4. The Te form 6-rings which contain one or two CN3 cations in Fig. 11l, but zero or two CN3 cations in the pseudohexagonal sheet of Fig. 11m. The much more complex and highly convoluted Te16X36 sheet of Fig. 11n has Te in 3-, 4- and 5-coordination, making 3-, 4- and 12-rings. The CN5 Te occur in edge-sharing pairs. Finally, Fig. 11o shows an extraordinarily complex sheet made by Q2200 Te cross-linking elliptical tubes which have 4-, 7- and 8-rings of CN4 Te.
Analogously to the case for chains, a phyllotellurate has a double layer if deletion of selected Te–O bonds can separate it into two distinct sublayers which themselves remain continuous. Our dataset contains two types of double layer, as seen in Table 5 and Fig. 12. The Te6X13 double layer of Fig. 12a has all Te CN = 4, but half of them are Q1300, sharing corners to form 3-rings, while the other half of the Te are branches off these rings, which share edges (Q0311) to link the two sublayers. Oxygen atoms with CN3 link the edge-sharing dimers to complete each of the sublayers. Conversely, the Te6X14 double layer of Fig. 12b has no edge-sharing or CN3 oxygen atoms, but has Te in three different coordination states (Q0300, Q1300 and Q1400 configurations). Like many Te‒O polymers containing CN5 Te, this thick double layer is braced by additional polyhedra containing other high-charge, low-CN cations (Se4+O3 in this case). ⇓
Te4+On polyhedra also form a range of infinite three-dimensional frameworks. Figure 12c shows the electrically neutral tetragonal framework of paratellurite, TeO2, and its metastable orthorhombic distorted variant γ-TeO2; these are polymorphs of tellurite, which has a layered structure of the type seen in Fig. 11e. The paratellurite framework is of interest in that it is isopuntal with the low-cristobalite form of SiO2 (Dollase, 1965), and yet can also be derived from the structure of rutile (and the dense stishovite form of silica) by deformation of coordination octahedra TeO6 → TeO4+2 → TeO4. Note that the Q0400 Te polyhedra are much less symmetrical than SiO4 tetrahedra. The paratellurite structure is thus a shared hettotype structure that could act as a transition state for diffusionless phase transformations between the low-density/high-temperature structure of high-cristobalite on the one hand, and the high-pressure stishovite structure on the other, analogous to the transformation mechanisms described in Christy (1993).
Figure 10d shows a rare example of Te4+ in Q0600 polyhedra that are nearly regular octahedra, with no lone-pair stereoactivity, and which link to form the pyrochlore framework. The frameworks of Figs 12e‒g are all closely related, and like that of Fig. 12d, have cubic unit cells with a ≈ 10 Å; all can ultimately be derived from 2 × 2 × 2 superstructures of the fluorite type. The Te3X7 framework of KGa(Te6O14) (Fig. 10e) is formed when ¼ of the ‘fluorite’ cations are replaced by non-Te species and ⅛ of the anions omitted, to make a framework in which Te are in a Q1210 configuration, linked through 2 × CN2 oxygen atoms and one CN3 oxygen atom. The isomeric cliffordite framework (Fig. 12f) has a similar range of oxygen CN and the same Pa space-group symmetry, but the topology of linkage of the CN2 and CN3 oxygen atoms is different. The structure of the winstanleyite group, M4+(Te3O8), is a slightly distorted fluorite superstructure in which the Te framework has Q2020 Te linked through two CN3 oxygen atoms only; it can thus be represented as a 3-connected net with CN3 oxygen at the nodes and (TeO2) groups decorating the links (Fig. 12g).
Figure 12h shows a more open tetragonal Te4X9 framework in which half the Te cations are Q0400, forming Te4O12 rings which are arranged on a D lattice complex (Fischer and Koch, 2006), analogous to the Ti atoms in anatase (TiO2; Howard et al., 1991) or Ca in scheelite (CaWO4; Zalkin and Templeton, 1964). These Te atoms are linked to form a framework through pairs of Q2200 Te atoms, making additional 4-rings. The Te5X11 framework of Pb(Te5X11) is even more open (Fig. 12i). This structure has five distinct types of Te, all CN4, but in four distinct Q states. Te1 (Q0222) and Te3 (Q0401) form edge-sharing tetramers Te3 = Te1 = Te1 = Te3, while Q1210 Te2 links to Te3 of one tetramer and through CN3 oxygen to both Te1 of the next, so that Te1, Te2 and Te3 form continuous chains of 6-rings which run || x and lie in layers || (001). Between these layers and cross-connecting them are corner-sharing chains running || y of Q1300 Te4 and Te5, where the chain backbones ‒Te4‒O‒Te5‒O‒ have an asymmetrical crankshaft geometry, very similar to that of the Pb‒O chains in massicot (Hill, 1985). Connections are so sparse that the smallest rings to include Te4 or Te5 have eight and ten members.
As noted above, the stereochemistry of Te6+ is much less diverse than that of Te4+, so the range of polymeric complexes is also more restricted. Table 6 shows that the tetrahedral TeO42– anion (Fig. 4e) and bipyramidal TeO54– (Fig. 4f) occur in only three structures altogether, one of which contains both of them. Conversely, octahedral complexes TeX6 (Fig. 4h) are extremely common. The neutral ‘orthotelluric acid’ molecule Te(OH)6, with its ability to make a profusion of hydrogen bonds, is the defining Te species in 59 structures, while no less than 182 have less protonated octahedral anions as their most complex Te complex. Thus, isolated TeX6 octahedra are the most complex Te complex in about one third of the total database. Only five types of finite oligomer are documented. Octahedral dimers may share faces (Q3303 Te2X9, Fig. 13a), edges (Q4202 Te2X10, Fig. 13b) or corners (Q5101 Te2X11, Fig. 13c). The only larger oligomers are a bicyclic ‘double triangle’ tetramer Te4X18 (Fig. 13d), which is a Te6+ analogue of the Te4+ clusters that occur as parts of more complex polymers in Figs 10f and 10h, and a 6-ring Te6X27 with alternating corner-sharing and edge-sharing of octahedra (Fig. 13e).
Table 6 shows that the variety of infinite Te6+ chains is similarly limited. Only single chains are known. Zweier chains may be edge-sharing Te2X8 (Fig. 13f) or corner-sharing Te2X10 (Fig. 13g). Edge-sharing dreier chains Te3X12 occur (Fig. 13h), as do edge-sharing vierer Te4X16 (Fig. 13j) and corner-sharing Te4X20 (Fig. 13k). Note that the vierer periodicity of the latter is determined by having alternation of Q4200 octahedra with cis and trans bridging oxygen atoms, a type of variability that does not occur with coordination numbers below 5. Note that all the corner-sharing zweier chains of this study have trans bridging oxygen atoms. The most complex chain is a loop-branched dreier Te3X12 isomer (Fig. 13i), which is made by polymerization of the ‘double triangle’ cluster of Fig. 13d.
The Te6+ polyhedra also link to form layers and frameworks, but again, the range of polymer types is restricted relative to that seen for Te4+. Q0603 octahedra share edges to form TeX3 sheets with the same topology as those of the Al(OH)3 polymorphs, such as gibbsite (Saalfeld and Wedde, 1974; Fig. 14a). However, the Te in these octahedra appears always to be in solid solution with another cation of lower valence (Fe3+, Ti4+, Mn4+ or Ge4+), which gives the layer an overall negative charge. The same is true for the single case where a sheet is formed through edge-sharing of trigonal prisms, rather than octahedra (Fig. 14b). Such charge reduction is not necessary when Q2400 octahedra share corners to form TeX4 sheets with 4-rings (Fig. 14c). The most complex tellurate layer type has corner-sharing of Q2400 and Q4200 octahedra to form a layer with overall stoichiometry Te3X14 (Fig. 14d), with the same topology as that seen in chiolite, Na5(Al3F14) (Jacoboni et al., 1981). ⇓
Long-range disorder is shown by Te6+ with lower-valence cations again in MTeO4 tellurates with a monoclinically distorted (Te,M)X2 rutile framework, containg Q0062 octahedra (Fig. 14e). Here, disordered substitution with low-valence cations results in electroneutrality of the framework. Further analogies between Te‒O and Al‒F frameworks are provided by the rhombohedrally distorted ReO3 structure of TeO3 itself (Fig. 14f), collapsed so that the oxygen atoms approximate hexagonal close packing, which is shared with a polymorph of AlF3 (Daniel et al., 1990), and the Te2X7 framework of Fig. 14g, which is that of weberite, Na2(MgAlF7) (Knop et al., 1982). The weberite framework contains many 3- and 6-rings, similar to the pyrochlore framework of Fig. 12d, but half of the octahedra are not Q0600, but are instead partly depolymerized to Q2400, which allows extra anions to be included. The unique Te4X12 framework of Fig. 14h also contains many 3- and 6-rings (as well as 4- and 8-rings), but there is also some edge-sharing: half the octahedra are Q0601 and the other half Q0602. In this framework, dense zigzag columns || z share polyhedral edges in the y direction to define layers, which in turn are linked into a framework via relatively sparse Te‒O‒Te bridges. Again, this particular structure has Te mixed with another cation (Fe3+) on all the octahedral sites, in order to give it an overall negative charge.
Last, we consider the 26 structures of the present study which contain both Te4+ and Te6+. Only one of these structures (#675) has separate Te4+ and Te6+ polyhedra which are not linked by strong bonds. In all the rest, ino, phyllo or tecto polymers contain Te in both valence states. Given their different stereochemical preferences, Te4+ and Te6+ are always ordered on distinct sites. The three types of chains are all only zweier in backbone periodicity, but nevertheless, display other complexities. In Fig. 15a, Te4+ and Bi3+ are disordered in pairs of edge-sharing Q0401 polyhedra, which are linked through Q2400 Te6+ to make a chain of ‘double triangles’. Figures 15b and c show chains in which Q2400 Te6+ backbones are decorated by Q2200 Te4+ to make 3-rings. Again, Cd demonstrates a tendency to be associated with structurally complex anions (cf. Fig. 10b and h). Figure 15d shows a relatively common type of Te4+Te6+X6 layer (four examples known) in which Q1400 Te4+ and Q0600 Te6+ share corners to 3- and 4-rings. Other isomeric layers with the same stoichiometry are seen in Figs 13e‒g. The layer of Fig. 15e features zweier chains of Q2400 Te6+O6. Unlike the zweier Te6+ chains of Fig. 13g, these strongly zigzag chains are cis-bridged. They are linked through massicot-like chains (cf. Hill, 1985) of Q0400 Te4+O4, to form 3- and 5-rings; the 3-rings on the Te6+ backbone form a loop-branched chain resembling that of Fig. 15c. The topologies of Figs 15f and 15g both feature edge-sharing between Te6+ octahedra to produce ‘double triangle’ clusters, but differ in coordination number of Te4+, numbers of non-bridging oxygen atoms on Te6+, and ring sizes in the layers. Note that the positions of Te6+ and Te4+ in the ‘double triangles’ are reversed here, relative to the chain of Fig. 15a. Figure 15f has Q1300 Te4+ and Q1501 Te6+ with 8-rings between the clusters, while Fig. 15g has Q1400 Te4+, Q0601 Te6+ and 4-rings. One compound, Ag2(Te4+Te6+O6), has three polymorphs ‘I’, ‘II’ and ‘III’, displaying, respectively, the topologies of Fig. 15f, d and g (#686, 684 and 687). The most popular type of mixed-valence layer, with seven examples, is the Te4+Te6+X7 topology of Fig. 15h. Here again, zigzag cis-bridged chains of Te6+ (Q2400) are connected through Te4+ (Q2200), but the two non-bridging ligands on both types of cation give a greater X/Te ratio. The strong nonplanarity of the layer and small O‒Te‒O angles allow four 6-rings to meet at each Te6+, in contrast to the three 6-rings meeting at each node in the more familiar mica-type layer (Fig. 11g). In Fig. 15i, a layer of stoichiometry Te34+Te6+X9 is formed by 6-rings of Q0400 Te4+ linked through Q0600 Te6+ to make a sheet with additional 3- and 4-rings.
All the mixed-valence framework compounds have unique structures, although some of them are closely related to one another. In particular, most of them contain component layers with the well-known ‘hexagonal tungsten bronze’ or ‘kagome’ topology (O'Keeffe and Hyde, 1996), where 3- and 6-rings alternate around each node of the net in the order 18.104.22.168. Figure 16a shows the framework of Te4+Te6+O5, in which corner-sharing layers (similar to Fig. 14c) of Q0600 Te6+ are cross-linked via corner-sharing massicot-like chains of Q0400 Te4+; cross-linkage produces 3-rings, as so often is seen in tellurate polymers. In carlfriesite, Ca(Te24+Te6+O8), Q2400 Te6+ are cross-linked through edge-sharing dimers of Q0401 Te4+ to make a trellis-like, nanoporous framework (Fig. 16b). The isomeric framework of Sr(Te24+Te6+O8) is also zeolite-like, but is formed by cross-linkage of chains similar to those of Fig. 15a, made from ‘double-triangle’ clusters containing Q0600 Te6+ and edge-sharing dimers of Q0501 Te4+ (Fig. 15c). Figure 16d shows another rather open framework of stoichiometry Te24+Te26+X11, in which Q0500 Te4+ and Q0600 Te6+ define layers with a pseudohexagonal kagome net. Te6+‒O‒Te6+ bridges link pairs of such layers, producing ladder-like double chains of Te6+. Layer pairs are, in turn, linked into a framework through pairs of edge-sharing Q0501 polyhedra. Kagome layers are also found in the Te34+Te6+X12 framework of Fig. 16e, which is that of pyrochlore and, thus, contains kagome nets in four different orientations. The framework, ideally cubic in symmetry, is rhombohedrally distorted as a result of Te4+‒Te6+ ordering. Like Fig. 12d, this is a rare example of Te4+ in octahedral coordination, with no evidence of a stereoactive lone pair: all Te cations are Q0600. The K analogue of the Cs compound of Fig. 16e has the modified pyrochlore framework shown in Fig. 16f. Monoclinic shear of the structure is accompanied by breaking of some Te4+‒O‒Te6+ links, so that Te4+ is now Q0500 (with a stereoactive lone pair) and one-sixth of the Te6+ are Q2400. Kagome nets are also a major feature of the Te24+Te36+X14 framework in Fig. 14g, where the Q0600 Te6+ form such layers, which are cross-linked via pairs of edge-sharing Q5001 Te4+. As for the framework of Fig. 16d, the cross-linkage produces additional 3- and 4-rings. The orthorhombic Te34+Te56+X23 framework is yet another modification of the pyrochlore type, in which the lone pairs of Te4+ are accommodated by breaking some Te4+‒O‒Te4+ links, with complete elimination of 1/24 of the oxygens. The ordering pattern of Te4+ and Te6+ is quite different from those of Figs 16d and 16e. Kagome layers are preserved in two orientations, but are not exclusively Q0600 Te6+: alternate layers have 1/6 or 1/3 of their cations Q0500 Te4+.
Descriptions of individual structures
Finite Te4+–O complexes
Our descriptions of individual structures begin with those that contain finite Te4+‒O complexes (neso, soro and cyclo tellurites), #1‒280. The various topologies that occur are summarized in Table 1 and depicted in Figs 4 and 8.
Structures with monomeric Te4+O3, no larger structural unit, and no other anions or water
Structures #1‒24 are listed in Table 8 (deposited), along with their full references. In Li2[TeO3] (#1), helical columns of edge-sharing LiO4 tetrahedra || x are cross-linked into layers || (002), with TeO3 pyramids bracing the layers. Layers are held together only through weak Te···O interactions across the interlayer.
Na2[TeO3] and Ag2[TeO3] (#2‒3) have a monoclinic √2 × 3√2 × 1 superstructure of the rocksalt type, with the Te coordination environment distorted to give only three close neighbours. Tl2[TeO3] (#4) has an orthorhombic 3 × 2 × 1 superstructure of rocksalt with a different cation ordering pattern. Short bonds form ‒Te‒O‒Tl‒O‒Tl‒O‒ chains || x, with a crankshaft shape reminiscent of the Pb‒O chains in the massicot form of PbO; such chains recur frequently in the structures of the present study, as noted above. Both Tl and Te have stereoactive lone pairs and only three close oxygen neighbours, making the structure strongly layered || (020). AgTl[TeO3] (#5) has similar cell dimensions but a different cation ordering pattern and space group. Again, the structure is layered, but this time || (200).
K2[TeO3] and Cs2[TeO3] (#6‒7) have nearly-identical structures that are both oxygen-stuffed derivatives of the Ni2In type; they are therefore TeO32– analogues of the high-temperature K2SO4 structure (O'Keeffe and Hyde, 1985). The alkali cations are in 6‒9 coordination. Rb2[TeO3] (#8) appears to have a slight monoclinic distortion of the same structure, although the refinement is of poor quality.
Ca[TeO3] and Sr[TeO3] both occur in a large number of polymorphs with large unit cells and low symmetry (#9‒15). All structures are packings of (Ca,Sr)O6–8 polyhedra containing tunnels, with TeO3 groups bracing the sides and the tunnels and lone pairs pointing into the central space. The two forms of Ba[TeO3] are quite different. One of them (#16) has the simple monoclinic structure of KClO3 (Bats, 1978); the Ba and Te substructure resembles the CrB/TlI type (Helmholtz, 1936), and Ba is in 7 + 2 coordination by oxygen. BaTeO3 is thus an oxygen-stuffed analogue of TlI, in the same way that baryte, BaSO4, is an oxygen-stuffed derivative of the closely related FeB structure (O'Keeffe and Hyde, 1985). The other dimorph (#17) has an approximately cubic close-packed (ccp) array of Ba + Te but a very complex ordering pattern, with Ba in 8‒10 coordination.
There are also two synthetic polymorphs with known structures for Pb[TeO3]; interestingly, these appear to be distinct from the two mineral species of this composition, triclinic fairbankite (Williams, 1979) and orthorhombic plumbotellurite (Spiridonov and Tananeyva, 1982), both of which remain poorly characterized. The tetragonal form (#18) has a structure very similar to that of scheelite (CaWO4) but with ¼ of the oxygens removed in an ordered fashion. The coordination numbers are 6 and 3 for Pb and Te, as opposed to 8 and 4 for Ca and W. The lone pairs of both Pb2+ and Te4+ are directed into tunnels running || z. The more complex monoclinic structure of #19 has a framework of PbO4–6 and TeO3, again with tunnels (this time || y) which act as micelles to contain the lone pairs.
Cd[TeO3] (#20) has Te filling interstices in an edge-sharing framework of irregular CdO6 polyhedra. Sc2[TeO3]3 (#21) has edge-sharing layers of ScO6–7 polyhedra || (020), which are bridged by Te. In Ce4+[TeO3]2 (#22), zigzag chains of edge-sharing CeO8 are linked into a framework by Te. The Th and Pu analogues (#23‒24) are isostructural.
Structures with monomeric Te4+O3 and no larger structural unit, but with additional anions or water
Structures #25‒46 are listed in Table 9 (deposited), along with their full references. In Li3[TeO3](OH) (#25), LiO3OH tetrahedra and TeO3 pyramids form honeycomb-like double layers || (100), which are held together only by long Te···O and lone-pair interactions. Thus, the structure resembles that of the chemically similar but anhydrous phase #1. Na2[TeO3] · 5H2O (#26) has three types of Na. Face-sharing trimers (H2O)2Na1 ≡ (H2O)3 ≡ Na2 ≡ (H2O)3 ≡ Na1(H2O)2 share corners with each other and with square-planar Na3(H2O)2O2 to form a very open hydrogen-bonded framework in which TeO3 is only loosely held. KNa[TeO3] · 3H2O (#27) has a denser, simpler structure with K(H2O)6O3 and Na(H2O)3O3 polyehdra sharing faces. The arrangement of K, Na and Te is a threefold ordering of the primitive hexagonal net, so this can be regarded as an O/H2O-stuffed derivative of that archetype. The same is also true of K2[TeO3] · 3H2O (#28), although the oxygen positions there are adjusted to give 8‒13 coordination. Note that the anhydrous analogue (#6) is derived by oxygen-stuffing of a different but equally simple hexagonal arrangement of cations.
MgTeO3 · 6H2O ≡ [Mg(H2O)6][TeO3] (#29) has a rhombohedral structure with angle αrh = 97.4°. It can be regarded as a derivative of the CsCl type, in which Cs+ and Cl– anions are replaced by [Mg(H2O)6]2+ and [TeO3]2– complexes. As the lone pairs of the Te cations all point in the same direction along z, the structure is polar and ferroelectric, with point group R3. The structure of Sr[TeO3] · H2O (#30) is centrosymmetric but strongly anisotropic, with edge-sharing SrO6(H2O) polyhedra defining layers || (100). The layers are held together via H-bonds and Te lone-pair/secondary bonding interactions. Ba[TeO3] · H2O (#31) is isostructural, although with x and z directions exchanged.
The structures of A32+[TeO3]2X2 (A = Sr or Ba; X = Cl or Br) have a large (a ≈ 16 Å) cubic unit cell with the same space group Fdm as diamond (#32‒34). Clusters [Sr6Te4O12]4+ can be distinguished, in which Sr are at the vertices of an octahedron, linked by O along the octahedral edges, and braced by Te above four out of the eight octahedral faces. These clusters are arranged in the same fashion as the C atoms of diamond, and Sr3 triangles of neighbouring clusters face each other to define a second type of octahedron between them. Halide anions centre both types of octahedron, and also occur in the largest remaining interstices of the structure with six Sr arranged in an almost planar hexagon at 3.6‒4.4 Å and six Te above and below the plane at 3.3‒3.8 Å. The structure can be regarded as a derivative of the pyrochlore type, with X replacing CN8 and CN6 cations of pyrochlore and A replacing the framework anions of pyrochlore. Pb3[TeO3]Cl4 (#35) is quite different, in that one tellurate oxygen is tetrahedrally coordinated OTePb3 and the others triangularly coordinated OTePb2 to make a rod [Pb3TeO4]4+ running || z. The rod has a rhombic cross-section with the Pb cations (bonded to only 2 or 3 oxygen atoms) on the exterior, and are arranged in a herringbone fashion. They are held together through four crystallographically distinct Cl anions, bonded to 3‒5 Pb. Again, the structure is polar due to tilt of all TeO3 groups in the same sense along the z direction.
In Ho[TeO3]Cl (#36), Ho3+ is in pentagonal bipyramidal coordination by five O and two Cl (one apical and one equatorial). The HoO5Cl2 polyhedra form edge-sharing ribbons || y, which are again arranged in herringbone fashion but are cross-liked through Te cations. The lone pairs of Te point into well-defined micellar channels that run || y, between the ribbons. Nd5[TeO3]2O4Cl3 (#37) has three types of Nd3+ polyhedra: NdO8 cubes, irregular NdO6Cl and NdO5Cl3 square antiprisms. These form thick layers || (001), with Cl– bridging the interlayer regions and Te bracing the outsides of the layers. The TeO3 group is canted so that Te makes four long bonds to Cl at 3.19‒3.40 Å (Fig. 17). Na2Lu3[TeO3]4I3 (#38) has all the atoms except O in an approximately ccp array, with a layer sequence (Na2Te4), Lu6, (Na2Te4), I6 alternating along the x direction. Oxygen atoms define nearly-cubic LuO8 polyhedra which share edges to form sheets || (100). Some oxygen atoms are in tetrahedral coordination (OLu2NaTe), while others are displaced away to tetrahedral sites so as to be in nearly plane triangular coordination by Lu2Te, with a second Te much more distant at 3.15 Å. Thus, the structure can be regarded as a modified superstructure of the fluorite type (the cell parameters correspond to 2 × 1 × 3 fluorite cubes). Alternatively, it may be described as a structure in which thick Lu‒O sheets are braced by Te4+, with I– occupying the interlayer region and NaO4I4 square antiprisms holding the sheets together. Nd4Cu1+[TeO3]5Cl3 (#39) has a superficially similar stoichiometry but a quite different structure. NdO8, NdO7Cl and NdO7 polyhedra share edges to form walls that surround two types of channel running || y; large and small channels alternate in a checkerboard fashion. The small channels contain one out of five distinct types of TeO3, which again render the structure polar by all pointing in the same sense along the y direction. The large channels are lined by the rest of the Te atoms, but still contain enough space to accommodate a zigzag corner-sharing chain [Cu2Cl6]4– of CuCl4 tetrahedra, held in place by Nd3‒Cl3 links as well as each Cl making two to four long bonds to Te. Bi2[TeO3]2O (#40) has a defect fluorite superstructure (cf. #38) with a unit cell corresponding to 4 × 1 × 4 cubes of fluorite, with 1/8 of the anions missing. Overall, Bi + Te form a ccp array. They are ordered into columns || y, with 1 × 1 and rectangular 2 × 3 blocks of Bi separating 2 × 1 and paralleogram-shaped 2 × 3 blocks of Te. If tetrahedral interstices are surrounded by Bi4, then they are occupied by O. Bi2Te2 tetrahedra have oxygens displaced away from one Te or absent, BiTe3 tetrahedra have oxygens linked only to 1 Bi + 1 Te, and Te4 tetrahedra are unoccupied. Tellurium lone pairs point along ±y for the small Te blocks, but are directed into the interiors of the large blocks, which along with long Bi‒O bonds mark the gap between weakly defined thick layers || (200). Te2 in this structure is our unique example of ‘2-coordinate’ Te; however, if all Te‒O distances out to 3.5 Å are considered, a square-pyramidal coordination polyhedron (Fig. 3e) is defined by oxygens at 1.865, 1.911, 2.519, 2.793 and 3.062 Å. It is possible that the O coordinates are not accurate in this structure, and that the true coordination polyhedron has a narrower spread of bond distances. Smirnite, Bi2[TeO3]O2 (#41) has yet another defect fluorite superstructure with a unit cell corresponding to 2 × 3 × 1 fluorite cubes with 1/6 of the anions missing, and many of the rest displaced so as to approximate plane triangular coordination by 3Bi or 2Bi + 1Te. The Te atoms are on the outsides of thick layers || (100), which are linked only through long Bi…O and Te…O bonds.
In Ca6[TeO3]5(NO3)2 (#42), CaO6–8 polyhedra share edges to form undulating layers || (100) and also bridge these layers into a framework which contains large channels || y. TeO3 pyramids decorate the sides of these channels, and the channels act as micelles which contain the Te lone pairs and also NO3– anions. The structure is thus strongly reminiscent of the nitrate-free tellurites #9‒15. Ca5[TeO3]4(NO3)2 · 2H2O (#43) is rather similar, but the Ca layers || (200) remain completely separate, with no bridges connecting them. Tellurium lone pairs and nitrate groups point into continuous interlayer gaps.
Sc2[TeO3](SeO3)(SeO4) (#44) has zigzag chains of edge-sharing ScO7 polyhedra running || x. TeO3 groups connect the chains into pairs, and the SeO3 pyramids connect these further into layers || (002), while SeO4 tetrahedra share oxygens with four different Sc to link the layers into a rather open framework.
La2[Si6O13][TeO3]2 (#45) contains LaO9 and TeO3 polyhedra, forming layers || (100). These intercalate with a silicate sheet in which dreier double chains like those of okenite (Merlino, 1983) link to their neighbours to make a double layer in a disordered fashion, such that ⅔ of the Si are Q4 and ⅓ are Q3, giving an overall composition [Si6O13]2–. A similar intercalation of La‒Te and silicate sheets occurs in the triclinic structure of La4[Si5.2Ge2.8O18][TeO3]4 (#46) but here, the layers are || (010), La3+ cations have CN8‒10, and the silicate‒germanate anion is a loop-branched single sheet which is formed by cross-linking of narsarsukite-like tubes (Peacor and Buerger, 1962) running || x. Half of the (Si,Ge) are Q3 and half are Q4, and Ge > Si in two out of the eight tetrahedral sites.
Structures with monomeric Te4+O3 as part of a larger structural unit that is a finite cluster.
Details for structures #47‒62 are shown in Table 10 (deposited). HgTeO3 ≡ [Hg2(TeO3)2] (#47) has two types of Hg in quite differently distorted coordination polyhedra. Hg1 has two oxygen nearest neighbours at 2.06‒2.12 Å and four more oxygen atoms at 2.54‒2.73 Å, while Hg2 has one close oxygen neighbour at 2.14 Å, one at 2.30 Å and three more at 2.40‒2.46 Å. Using the parameters of Brese and O'Keeffe (1991), the Hg2 distances correspond to bond valences of 0.57, 0.37 and 0.28‒0.24 vu, so a ‘strong bond’ threshold of 0.3 vu would make both Hg atoms 2-coordinate, with O‒Hg2‒O less symmetrical and less linear than O‒Hg1‒O. The strong bonds define a structural unit that is a neutral molecule in which Hg1 and Hg2 form a ring with two TeO3 groups. These molecules lie in layers || (002) and are linked through long Hg…O and Te…O bonds. The complex structure of Cd4V25+Te34+O15 ≡ Cd4[VO3][(VO3)(TeO3)](TeO3)2 (#48) contains several structure-building elements. Pyroxene-like zigzag vanadate chains [V2O6]2– run || x, although these do not contain Te. The most complex structural unit with Te is a cluster [VTeO6]3– formed by corner-sharing of a VO43– tetrahedron and a TeO32– pyramid. The V-bearing structural units act as bridges between layers || (020) of relatively weakly-bound CdO6–8 polyhedra. The remaining TeO32– are attached to the Cd‒O layers, and all Te lone pairs point into channels than run || x, between the vanadate chains.
In Pb2[Pd2+Cl2(TeO3)2] (#49), two TeO3 groups are linked through a PdO2Cl2 square to make an anionic complex trans-[PdCl2(TeO3)2]4–. These complexes form layers || (002) which are linked via irregular PbO5 polyhedra, with the long axes of complexes oriented towards  and  in alternate layers. Bi2WTe2O10 ≡ Bi2[WO4(TeO3)2] (#50) has two TeO3 groups linked through a WO6 octahedron to make a complex cis-[WO4(TeO3)2]6–. These V-shaped anions are linked through irregularly coordinated Bi3+ cations. The six shortest Bi‒O bonds out to 2.58 Å define edge-sharing chains of distorted BiO6 octahedra || y, which with the Te‒W anions make layers || (200). However, longer Bi‒O at 2.83 and 3.27 Å link the Bi chains into continuous sheets || (002), alternating with layers of Te‒W anions. The cluster anion in Nd2W2Te2O13 ≡ Nd2[W2O7(TeO3)2] (#51) has two TeO3 linked to an edge-sharing pair of octahedra, W2O10. However, while Te1 is joined to both W atoms to form a 3-ring of cation-centred polyhedra, Te2 is attached only to W1, giving the overall stoichiometry [W2O7(TeO3)2]6–. These anions lie in layers || (10), and are connected through CN8‒9 Nd3+.
K4[Mo6Te2O24] · 6H2O (#52) and its isostructural Rb analogue (#53) are our first examples of a large family of salts (mainly telluropolymolybdates) in which the structural unit is a modified Anderson‒Evans anion (Anderson, 1937; Evans, 1948, 1974). Six MoO6 octahedra share edges to form a hexagonal ring, but instead of one Te occupying the vacant octahedral site at the ring centre, two pyramidally coordinated Te sit above and below the vacant octahedral position. The hexagonal heteropolyanions lie in layers parallel to (10), with their planes normal to either  or  directions. Water molecules and 8‒9 coordinated alkali cations lie between them (Fig. 17).
Cs6Na2[W6Mo3S4O20(H2O)3(W9TeO33)] · 11.7H2O (#54) and the nearly isostructural Cs7.15Na1.85[W6Mo3S4O20(H2O)2Cl(W9TeO33)] · 11.2H2O (#55) have extremely complex heteropolyanions which can be derived from incomplete fragments of the cuboctahedral TM12O40 Keggin ion (Keggin, 1934), where T = a tetrahedral cation and M = an octahedral cation. The ion is composed of two dissimilar half-cuboctahedral fragments of the Keggin cage. One fragment has composition [W9O30(TeO3)], and has a TeO3 pyramid rather than a TO4 tetrahedron bracing a bowl-shaped cluster of WO6 octahedra (Fig. 17). The other fragment is [W6Mo3S4O26(H2O,Cl)3]. It has no analogue of the central Te4+ cation, and has Mo6+ rather than W6+ as the cations on one triangular face of the cage. The four anions bonded to two to three Mo and no W are S2– rather than O2–, and the three anions bonded to one Mo only are (H2O,Cl). The two half-cuboctahedra link through six oxygen atoms to form an ellipsoidal cluster that is similar but not identical to the Wells‒Dawson cluster (Wells, 1947; Dawson, 1953; Baker and Figgis, 1970). The clusters are held together by additional water molecules and Cs+ and Na+ cations in a wide range of coordination states.
The heteropolyanion in K8Na2[Pd3(W9TeO33)2] · 51H2O (#56) consists of two half-Keggin subunits linked into a single, large dumbell-shaped anion through a set of three square-planar coordinated Pd2+ cations, to make an expanded version of the cluster in #54‒55. These anions are packed around inverse tetrad axes and linked through K(H2O)7–9 and Na(H2O)6 polyhedra. K9Na[Cu3(H2O)3(W9TeO33)2] · 16H2O · (#57) is almost isostructural but has a much lower water content, and each of the bridging Cu2+ ions also has an H2O molecule associated, to give it 5-fold rather than 4-fold coordination. Unusually, water molecules are included in the structural unit here and in the clusters below, when they are necessary to fully define the coordination environment of a cluster cation.
[N(CH3)4]2Na6[Ni(H2O)2(Ni(H2O)3)2(WO2)(W9TeO33)2] · 23H2O (#58) and its analogue with Zn2+ replacing Ni2+ (#59) have slightly more complex clusters in which the W9TeO33 fragments are not half-cuboctahedra but half-anticuboctahedra, in which two square faces share an edge (Fig. 17). These are linked through two M2+O3(H2O)3 octahedra and two octahedra that are 50% M2+O4(H2O)2 and 50% W6+O6. There is also a partially disordered cluster in K14[V125+Mo126+O69(TeO3)2] · 27H2O (#60), in which two half-Keggin units each have an average composition V4.5Mo4.5TeO33 and are linked through a ring of three VO4 tetrahedra alternating with three MoO6 octahedra. K10[V4(V3Mo17)O74](TeO3)] · 15H2O (#61) has a unique cluster that contains two different kinds of half-Keggin unit. One is (V3Mo5)TeO33 with partial V‒Mo disorder; this has the same topology as the half-Keggin units of #54‒57 and #60, with CN3 oxygen atoms centring the three triangular faces around the sides of the ‘bowl’ but not the bottom triangular face. The other half-Keggin unit is Mo9VO35, of the type found in the Wells‒Dawson cluster, in which it is the bottom face that is centred, and with a VO4 tetrahedron instead of a TeO3 pyramid. Instead of the two ‘bowls’ facing towards each other, the vanadate bowl is inverted, so that the VO4 tetrahedron is pointing away from the centre of the cluster. The two bowls are linked through a ring of alternating V and Mo, as for #60. The extraordinarily complex structural unit of (NH4)[H(Ru44+O6(H2O)9)2(Fe3+(H2O)2)2(W9TeO33)2] · 36H2O (#62) has two half-anticuboctahedra W9TeO33 similar to those of #58‒59, linked through a pair of FeO4(H2O)2 octahedra and also through three oxygen atoms each of two tetrahedral clusters or corner-sharing RuX6 octahedra, Ru4O9(H2O)9.
Structures with monomeric Te4+O3 as part of a larger structural unit that is a chain
The next 11 structures (#63‒73, Table 10, deposited) have TeO3 groups incorporated into infinite heteropolymeric anions. In magnolite, [(Hg2)(TeO3)] (#63), each Hg atom is bonded to one other at 2.53 Å to make a (Hg2)2+ dimer, and has no other neighbours apart from one close oxygen at 2.06 Å and three more distant at 2.69‒3.00 Å. The Hg dimers and TeO3 groups form continuous chains ‒O‒Te‒O‒Hg‒Hg‒O‒Te‒O‒ running || y and zigzagging in the (001) plane, with Te atoms at the apices of the bends. The chains are held together only by long Hg…O and Te…O bonds. BaZn(TeO3)Cl2 (#64), better written for our our purposes as Ba2[Zn2Cl3(TeO3)2]Cl, has double chains || y resembling those of the amphibole minerals, in which TeO3 pyramids instead of tetrahedra alternate with Q3 ZnO3Cl tetrahedra and Q2 ZnO2Cl2. The chains lie in double layers || (002) with additional Cl– and 6‒7 coordinate Ba2+ between them. Dy[CuCl(TeO3)2] and its analogues (#65‒67) have a loop-branched structural unit, in which TeO3 pyramids and CuO4Cl pyramids make CuTeCuTe 4-rings, which link into a chain || x through the Cu atoms (Fig. 17). The chains form layers || (002), which are interconnected through zigzag chains of edge-sharing DyO8. The Te lone pairs and Cl are located in channels which run between the Dy chains and the Cu‒Te chains. The compound Yb3[Cu2+Cl2(TeO3)2]2[Cu1+Cl2] (#68) has an open-branched chain || y in which pyramids Cu2+O3Cl2 and TeO3 alternate along the chain backbone, while a second type of TeO3 acts as a branch sharing an oxygen atom with the Cu. These chains attach on either side of a layer || (100) of edge-sharing YbO8 polyhedra, with isolated, linear [Cu1+Cl2]– anions in the interlayer gap (Fig. 17). LaNbTeO6 ≡ La[NbO3(TeO3)] (#69) has NbO6 octahedra sharing trans corners to make zweier chains || y; oxygen atoms are shared with Te so as to make two differently oriented Nb2Te 3-rings, which alternate along the chain. Chains are held together via long Te…O bonds and 8-coordinated La3+. This compound is isostructural with Pb[MoO3(SeO3)] (Oh et al., 2012).
Tl2[(UO2)(TeO3)2]-β (#70) has Te‒U‒Te‒U 4-rings which link through UO6 octahedra to make loop-branched chains || x. These pack in a herringbone fashion, and are linked through ribbons of edge-sharing TlO5–7 polyhedra. The α polymorph has a layered structure (#221, below). Sr3(UO2)(TeO3)4 ≡ Sr3[(UO2)(TeO3)2](TeO3)2 (#71) has similar chains || y, all sharing a common attitude, and bridging layers || (200) of SrO7–8 polyhedra. Additional isolated TeO3 brace the Sr layers. The complex chain in Yb2[Cu3Cl4(TeO3)4] (#72) is again based on linked 4-rings, but the polyhedra are CuO6 octahedra at the nodes and CuO3Cl2 on the loops. One of two types of TeO3 acts as an additional bridge between two Cu, making Cu2Te 3-rings, while the other type of TeO3 shares an edge with CuO6, so that the bridging oxygens in the Cu chain have CN3. These chains run || x and are linked through edge-sharing zigzag chains of YbO8 to make thick layers || (001) (Fig. 18). The layers are held together only by long bonds to Te. Bi2W3Te2O16 ≡ Bi2[W3O10(TeO3)2] (#73) again has loop-branched chains of 4-rings, but the polyhedra are all WO6, giving a chain stoichiometry W3O14. TeO3 groups share nonbridging oxygen atoms to make W2Te 3-rings, similar to the Cu2Te of #72, and giving a final chain composition W3Te2O16. The chains run || y and lie in layers || (002), which cross-link sheets of BiO8 polyhedra.
Structures with monomeric Te4+O3 as part of a larger structural unit that is a layer
Structures #74‒#112 (Table 11, deposited) have infinite two-dimensional structural units. [Cu(NH3)(TeO3)](H2O) (#74) has a 3-connected net with 4- and 8-rings, topologically similar to the ‘apophyllite’ layer of mackayite (Fig. 11i), but the polyhedra at the nodes are alternately TeO3 and square-planar CuO3(NH3) (Fig. 18). The polyhedra are tilted so that a very elongated octahedron around Cu is completed by another TeO3 oxygen at 2.60 Å and the H2O molecule at 3.07 Å. The layers are || (200), and are connected only through H bonds. Bi[Cu3O2(TeO3)3]Cl (#75) has a complex layer in which a hexagonal honeycomb array of CN3 oxygen atoms are linked through the two bridging oxygen atoms of CuO4 squares. The layer is corrugated because one type of Cu has trans bridging oxygen atoms while the other type has cis. A Te atom sits above or below each CN3 oxygen, sharing its own oxygens with two trans Cu and one cis Cu. Elongated octahedra around all Cu are completed by long bonds to Cl. The layers are || (002) and are connected through BiO8 polyhedra and long Te…O bonds (Fig. 18). Juabite, Ca[Cu10(AsO4)4(TeO3)4(OH)2] · 4H2O (#76) has a structure in which double layers || (010) are held together only through long Te…O bonds. The two sublayers contain edge-sharing blocks OTe = Cu = O2 = Cu = O2 = TeO with long axes || , which are held together by sharing corners with continuous chains of more CuO4 squares and AsO4 tetrahedra (Fig. 18). The resulting net is approximately centred-rectangular, and has 8- and 3-rings. Two such sheets are linked by the inward-pointing apical oxygens of AsO4 tetrahedra connecting to additional Cu2O4(OH)2 dimers. Loosely bound H2O and CN6 Ca2+ lie between the two sublayers.
Sr2V45+Te2O16 · H2O ≡ Sr2[V2O5(TeO3)]2 · H2O (#77) has a structure in which CN5 V1 and CN4 V2 polyhedra share corners with TeO3 to make chains ‒V1‒V2‒Te‒V1‒V2‒Te‒ running || z and zigzagging in the (100) plane. Such chains at two different heights along x are linked by edge-sharing of VO5 polyhedra to make a layer || (200) which is strongly corrugated but not topologically a double layer. Thus, the TeO3 groups and can be regarded as linking [V4O14] tetramers (Fig. 18). One oxygen atom of TeO3 is directed at the interlayer gap, while the lone pair points into a deep fold in the layer. Water molecules and CN8 Sr2+ are between the layers.
In Na[Ga(TeO3)2] (#78), edge-sharing pairs of GaO6 octahedra are linked through corner-sharing TeO3 to make layers || (002) consisting exclusively of Ga‒Te‒Ga‒Te 4-rings. A second type of TeO3 shares the remaining oxygen atoms of the Ga octahedra, forming Ga‒Te‒Ga 3-ring loops, which occur in pairs to make a Ga2Te2 unit resembling the ‘double triangle’ tetrameric Te unit that occurs in many Te-only polymers (Figs 10f, 10h, 13d, 13i, 15f, 15g and 16c). Layers are linked through CN7 Na+ ions (Fig. 18). Na[Fe3+(TeO3)2] (#79) is isostructural. Rodalquilarite, Fe23+Te4O9(OH)3Cl≡ [Fe2(TeO2OH)3(TeO3)]Cl (#80) also has Fe2Te2 ‘double triangles’. These share FeO6 edges to form zigzag chains || x, which are cross-linked via a second type of TeX3 into layers || (001). The Cl– lie between the layers, and are weakly bonded to Te (Fig. 19). Oxygen atoms that are not bonded to Fe are 100% OH– for Te1, 50% O2– 50% OH– for Te2 (Kampf and Mills, 2011). [(Fe2+Fe43+)(TeO3)6]Cl2 (#81) has three types of Fe polyhedra. (Fe1)O6 octahedra share two trans corners with (Fe3)O5. (Fe2)O6, share two edges with each other to make zigzag chains || x, and a third edge with Fe3, linking all Fe polyhedra into an undulating layer || (01) with elongated 12-rings. The mean Fe‒O distances are very similar for Fe1 and Fe2 (2.03 vs. 2.01 Å), implying that there is little ordering of Fe2+ and Fe3+. Two types of TeO3 share all oxygen atoms with Fe polyhedra, with lone pairs pointing into the interlayer space. As in rodalquilarite, interlayer Cl– are weakly bonded to Te.
The phase α-V4+TeO4 ≡ [VO(TeO3)] (#82) is isopuntal with the raspite polymorph of PbWO4 (Fujita et al., 1977). Edge-sharing, distorted VO6 octahedra form V2O8 chains || y, analogous to the W2O8 chains of raspite. These are arranged similarly to the corresponding chains in ferberite, FeWO4 (Ulku, 1967) but with an additional lattice shear so that (1) the anions no longer form a continuous hexagonal close-packed substructure, and (2) the Pb/Te polyhedra are no longer octahedra, but are very irregular (Fig. 19). The structure can be regarded as a distortion of the pucherite type, which is orthorhombic and has an anion array that is still hexagonal close-packed (see discussion of #648–650 below). In raspite, Pb2+ has seven neighbours at 2.3–2.9 Å and no more within 3.2 Å, while TeVO4 shows even less regularity: Te4+ has three strongly bound neighbours at 1.75, 2.00 and 2.25 Å, followed by four more within 3.4 Å, at 2.42, 2.59, 3.04 and 3.17 Å. As is typical for V4+, there is one very short bond of 1.73 Å of one of the oxygen atoms that is not linked to Te, although this ligand is also only 2.04 Å from a second V atom. Note that Te6+ plays the V/W role in the raspite structure in #650, below. The β polymorph of VTeO4 has a very different, layered structure (#222, below).
[InCl(TeO3)] (#83) and its Br analogue (#84) have edge-sharing chains of InO4X2 octahedra (X = Cl or Br) running || z, sharing oxygen atoms with TeO3 groups to make In2Te 3-rings. The resulting loop-branched chains (similar in topology to the Te chain of Fig. 15c), at two different x heights, are linked through the third oxygen ligand of Te to make a corrugated layer || (100), with layers connected only via long Te…X interactions (Fig. 19). [BiI(TeO3)] (#85) has the same space group and similar unit-cell parameters, but is not quite isostructural. The TeO3 groups are differently oriented, so that they cross-link chains at the same x coordinate, and the structural unit can be described as a double layer rather than a highly corrugated single layer. The Bi coordination polyhedra are BiO5I2 rather than octahedra, and share corners along the chain direction z rather than edges, while making new shared-edge connections between sublayers.
YV5+Te2O8 ≡ Y[VO2(TeO3)](TeO3) (#86) has two topological types of TeO3. The first type (Te2 and Te3) share corners with VO6 distorted octahedra to complete V2Te 3-rings above and below a corner-sharing VO4 layer with a square-net topology similar to those of Fig. 11a,f. Layers are || (002), and connect through TeO3 sharing oxygen atoms with edge-sharing sheets of YO8 polyhedra. The remaining Te1 and Te4 act as additional braces on the Y layer, and are not part of the larger structural unit.
BaMo26+TeO9 ≡ Ba[Mo2O6(TeO3)] (#87) has MoO6 octahedra sharing corners to form zigzag zweier double chains || y, which are cross-linked via TeO3 into double layers || (002). BaO10–11 polyhedra lie between the layers. LiV35+Te2O12 ≡ Li[(V5+O2)3(TeO3)2] (#88) has strips of distorted VO6 octahedra sharing edges and CN3 oxygen atoms to form chains || z similar to the Te chain of Fig. 10a. Te1 shares three oxygen atoms with three V of one such chain, making clusters which include a Te2V2 ‘double triangle’ motif, while Te2 shares oxygen ligands with V atoms of two chains, to link them into a double layer || (010). CN6 Li+ cations sit within corrugations of the layer, and layers are held together through weak Li‒O and Te‒O bonds (Fig. 19). (NH4)V4+V5+TeO7 ≡ (NH4)[(V4+O)V5+O3(TeO3)] (#89) has extremely corrugated layers || (200) in which alternating V4+O6 octahedra and V5+O4 tetrahedra each share three corners to form 6-rings. TeO3 shares the otherwise non-bridging oxygen ligand of V5+ and two of V4+; the additional connectivity means that the structural unit can be regarded as a double layer. The remaining ligand of V4+ is very close (1.61 Å), making a well-defined vanadyl group, [V = O]2+. (NH4)+ ions are in the interlayer gap. Cs3[(V4+O)V5+O3(TeO3)]2Cl (#90) has a closely related structure with similar unit-cell dimensions, in which chains (V4+V5+O8) || z of alternating V4+O6 and V5+O4 do not form continuous layers, but are linked into a single layer normal to x through TeO3. There are four such layers per unit cell, which alternate in their facing direction. The apical oxygen atoms of V polyhedra point towards interlayers that contain Cs+ ions only, while Te lone pairs point towards another type of interlayer, which contains both Cs+ and Cl–.
[Al2(TeO3)(SO4)(OH)2] (#91) has zigzag edge-sharing ribbons of AlX6 octahedra || z, which in turn share corners with each other to make continuous sheets Al2O5(OH)2 || (100). The sheets are braced by SO4 tetrahedra sharing two oxygen atoms and TeO3 pyramids sharing all three, and successive sheets are held together through very long Te…O bonds (Fig. 19). The compounds [M5X2(TeO3)4] with M2+ = (Ni, Co or Mg), X– = Cl and Br, (#92‒96) all have a structure with rather dense layers of edge-sharing MO6 and MO5X octahedra || (200), braced by TeO3 groups. Long Te…X bonds are important in holding the layers together (Fig. 19).
The compounds [PTX2(TeO3)] with P2+ = (Co, Cu or Zn), T2+ = Zn and X– = Cl and also [ZnZnBr2(TeO3)] and [CoCoBr2(TeO3)] (#97‒101) have an orthorhombic structure in which PO5X and TO2X2 polyhedra share corners to make layers || (020) (Fig. 20). Although these compounds are not known as minerals, they are isostructural with sophiite, [Zn2Cl2(SeO3)] (Semenova et al., 1992). [Co2Cl2(TeO3)] (#102), the Cl analogue of #101, has a different, monoclinic structure with layers of edge-sharing CoO4Cl2 and CoO3Cl3 octahedra || (001). In both cases, layers are again braced by TeO3 groups and held together through long Te…X bonds (Fig. 20). [Cu3Br2(TeO3)2] (#103) has CuO4 squares sharing all corners with edge-sharing pairs of CuO4Br pyramids to make loop-branched chains || y which feature a Cu analogue of the ‘double triangle’ motif; these chains are bridged by TeO3 groups to make layers || (001). Long Cu···Br bonds complete an elongated octahedron of ligands around square-coordinated Cu1, and long Te…Br bonds hold the layers together.
(NH4)2Mo3TeO12 ≡ (NH4)2[Mo3O9(TeO3)] and its Cs analogue (#104‒105) have a structure in which Mo6+O6 octahedra share four corners to form Mo3O12 layers with the kagome net of 3- and 6-rings. Tellurium atoms share apical oxygen atoms of the Mo octahedra around each 3-ring, to make [Mo3O9(TeO3)]2– layers || (002). Layers are connected through longer Te…O bonds (2.92‒2.95 Å) which complete a very distorted octahedron around Te, as well as through interlayer NH4+ or Cs+ ions (Fig. 20). The positions of Mo and Te atoms correspond to those of O atoms in the tridymite structure. Alternatively, the MoO6 and TeO3+3 octahedra can be regarded as forming a hexagonal relative of the pyrochlore framework, in which half of the Mo3 triangles link to Te above and below the centre of the triangle. The structure is polar (space group P63) as the TeO3 pyramids always point in the same sense along z. Rb2[W3O9(TeO3)] (#106) is almost isostructural, but with the symmetry reduced to P31c due to slight collapse of the layers. Ag6W3Te4O16 ≡ Ag6[W2O6(TeO3)2][WO2(TeO3)2] (#107) has two types of infinite, strongly-bound structural unit: a layer [W2Te2O12]4– and a chain [WTe2O8]2–; for classification purposes, the layer takes precedence. The chains are loop-branched, with WO6 octahedra sharing four corners with TeO3 groups to make W‒Te‒W‒Te 4-rings. They run || z and are stacked to make layers || (200), which alternate with the continuous W‒Te sheets. The latter have WO6 and TeO3 polyhedra alternating in a 3-connected net of 4- and 8-rings with the ‘apophyllite’ topology (cf. Fig. 11i). Layers are held together through long Te…O links and also three types of interlayer Ag+, in very irregular 5-coordination (Fig. 20).
Pb2[(UO2)(TeO3)3] (#108) has a structure in which UO7 pentagonal dipyramids share four equatorial oxygen atoms and two types of TeO3 share two ligands to form layers || (020) of crumpled 8-rings in which Te and U cations alternate. A third type of TeO3 shares the remaining equatorial oxygen ligand of U, and projects into the interlayer gap, where Pb2+ cations in irregular 7‒8 coordination hold the structure together (Fig. 20). In the compounds A2[(UO2)3O2(TeO3)2] (A = K, Rb and Cs: #109‒111), edge-sharing pairs of UO7 polyhedra share both CN2 and CN3 oxygen atoms with UO6 octahedra to form ribbons || y, which in turn are bridged by TeO3 groups to form layers || (10), which are held together by CN7‒8 interlayer A+ cations (Fig. 20). K4[(UO2)5O5(TeO3)2] (#112) has a similar structure in which broader edge-sharing ribbons (4 × UO7 and 1 × UO6 polyhedra per asymmetric unit) are bridged by TeO3 to form layers which are also oriented || (10), with 8-coordinated K+ in the interlayer.
Structures with monomeric Te4+O3 as part of a larger structural unit that is a framework
The TeO32– pyramid is incorporated into an infinite three-dimensional structural unit in compounds #113‒194 (Table 12, deposited). The first two examples are Ga2(TeO3)3-α, which in this context is more appropriately written [(Ga2.670.33)(TeO3)4], and [(Ga2Zn)(TeO3)4] (#113‒114), which both have the structure of eulytine, Bi4(SiO4)3 ≡ Si3(BiO3)4 (Menzer, 1931). This framework is made by CN2 oxygen atoms linking one CN4 (Ga,Zn) atom and one CN3 Te atom (Fig. 21). The cation sublattice (Ga,Zn)3Te4 has the same arrangement as the Th3P4 structure (Meisel, 1939; O'Keeffe and Andersson, 1977; Hyde and Andersson, 1989). The much less dense β polymorph of Ga2(TeO3)3 is described below (#136).
The mineral choloalite, ideally Pb3[(Cu2.672+ Sb0.335+)(TeO3)6]Cl is almost isostructural with SrCu(TeO3)2 = Sr3[Cu3(TeO3)6] (#115‒116). Again, the structure is cubic in symmetry, but rather complex. CuO4 squares share all corners with Te, and TeO3 groups share two corners with Cu, to form a framework with a large unit cell (a ≈ 12.5 Å) and chiral symmetry (P4132). The Cl– anion, if present, is shared by three (Cu,Sb) atoms as a fifth ligand, while (Pb,Sr) is located in large interstices in irregular 9‒12 coordination (Sr) or 6‒9 (Pb) (Fig. 21). Balyakinite, [Cu(TeO3)] (#117) has a structure in which edge-sharing pairs CuO5 square pyramids link corners to form zigzag chains || x, which are cross-linked in the y and z directions into a rather open framework by TeO3 groups (Fig. 21). Structures #142 and #297 are polymorphs. Although the unit-cell parameters and stoichiometry of [Zn(TeO3)] (#118) are similar to those of balyakinite, the structure is different. It features edge-sharing pairs of ZnO5, but they link with four neighbouring dimers through corners to form continuous corrugated layers || (002), with TeO3 providing bridges between layers in the z direction. Cu(SeO3) is not isostructural with balyakinite, but one of its four polymorphs has the structure of Zn(TeO3) (Effenberger, 1986; Hawthorne et al., 1986), and another has the perovskite-type structure of high-pressure Cu(TeO3) (#142) (Kohn et al., 1976). Teineite, Cu(TeO3) · 2H2O ≡ [Cu(H2O)2(TeO3)] (#119), the dihydrate of balyakinite, also features square pyramids, although these are CuO3(H2O)2, and do not link to each other. The Cu polyhedra share oxygens with Te so that Cu and Te define an open 3-connected framework in which their locations correspond respectively to Ca and half of the Cl of the CaCl2 (orthorhombically distorted rutile) structure (van Bever, 1935; Haines et al., 2000). The structure is intrinsically enantiomorphic (space group P212121) and has rather large channels || y, which are braced by hydrogen bonds between water molecules and TeO3 groups, and into which point the lone pairs of Te atoms (Fig. 21). Chalcomenite, Cu(SeO3) · 2H2O, is isomorphic (Pasero and Perchiazzi, 1989). Less obviously, the structural unit of LiV5+TeO5 ≡ Li[(VO2)(TeO3)] (#120) is topologically the same as [Cu(H2O)2(TeO3)] of teineite, although change in axial ratios and atomic coordinates close the channels (which would be || x if they existed). Additional LiO5 polyhedra share two oxygen ligands with (Te + V), two with (Li + V) and one with a single V atom only.
[(Cu62+Cl4)(Mo26+O8)(TeO3)2] · H2O (#121) has a framework made out of two chemically dissimilar components. Square-planar CuX4, CuO4 and Cu2O3Cl, link corners and some edges to form thick layers || (001), which are braced by TeO3 groups and additional long Cu…Cl bonds. There are no tetrahedral molybate complexes: instead, MoO5 square pyramids share edges to form [Mo2O8]4– dimers in the interlayer region. These share the four non-apical oxygens with Cu of the layers on either side. Half-occupied H2O sites complete a very distorted octahedron around Mo: these are at 2.50 Å from the cation, compared to 1.70 Å for the apical oxygen opposite (Fig. 21). [Cu7(TeO3)2(SO4)(OH)6] (#122) similarly has a framework in which relatively dense layers are connected through sparse bridges. There are five kinds of Cu2+, all in square-planar coordination except that Cu5 has a fifth oxygen ligand if the threshold is set at 2.3 Å. Edge-sharing trimers Cu4 = Cu3 = Cu4 and Cu5 = Cu2 = Cu5 link corners to form complex chains || , with additional corners shared between Cu4 and bridging Cu1, which further condense the chains into layers || (11). Layers are linked into a framework through TeO3 and SO4 groups, which show some orientational disorder as indicated by splitting of oxygen sites.
[Ge(TeO3)2] (#123) has a framework in which GeO6 octahedra and TeO3 pyramids share corners. Interestingly, the GeTe2 substructure is a slight monoclinic distortion of the rutile structure, with x as the pseudotetrad axis. The Ge‒O bonds of argutite (rutile-type GeO2; cf. Haines et al., 2000) are replaced by Ge‒O‒Te links in this compound.
HLi2[Ga3(TeO3)6] · 6H2O (#124) has Ga‒O‒Te links between GaO6 octahedra and TeO3 pyramids, making a rhombohedral structure with alternating layers of 2 × Ga surrounded by Te in an approximate trigonal prism, and 1 × Ga surrounded octahedrally by Te. LiO3(H2O)3 octahedra share faces with the former. There are three of each type of Ga layer per cell. Na3[Ga3(TeO3)6] · 7.2H2O (#125) has a closely related structure which retains the alternation of sparser and denser Ga layers, but in which Ga is always surrounded by six Te in a trigonal prismatic fashion, and there are only two of each type of Ga layer per unit cell. Na3(H2O)5 clusters with each Na bonded to four H2O and two tellurite oxygen atoms also lie in the sparse Ga layers, and another water site in the denser layers is 36% occupied. K[Ga(TeO3)2] · 1.8H2O (#126) has Ga of the sparse layers surrounded octahedrally by Te, as for #124. The layers are || (001), but the structure is triclinically distorted. K+ ions in 7‒8 coordination and H2O molecules occupy interstices in the sparse Ga layers (Fig. 21). Li6[Ga8.67(H2O)2(TeO3)14] (#127) also has alternating sparse and dense Ga layers, but these are quite different to those of #124‒126. The GaO6 octahedra of the sparse layer have a pseudo-diad axis || z rather than a true triad axis, and those of the denser layers occur in edge-sharing dimers Ga2O10, which link into a honeycomb-like layer through CN3 water molecules. Out of three types of Te, Te2 and Te3 link the two different types of Ga layer, along with CN5 Li+, while Te1 sit on triad axes and connect Ga octahedra within the sparse layers.
One dimorph of [Fe23+(TeO3)3] (#128) has FeO6 octahedra in an approximately face-centred orthorhombic array of face-sharing Fe2O9 dimers, all with Fe ≡ Fe vectors || y. TeO3 groups connect an upper oxygen atom of one dimer, lower oxygen atom of a second dimer and middle oxygen atom of a third, to make a continuous framework. The other dimorph (#129) has the same Pnma space group but a topologically quite different structure in which FeO6 octahedra share four corners to form Fe2O8 layers || (020). As the unshared ligands are cis to one another, the layers are highly corrugated. One of two types of Te braces the Fe layer, with the lone pair pointing into the interlayer gap, while the other type of Te bridges two Fe layers to make a continuous framework, with its third ligand not bonded to Fe. [In2(TeO3)3] (#130) is isostructural, but was described in a different axial setting.
Structures #131‒136 all have the microporous zemannite framework, which has zeolitic ion-exchange properties. Na2[Zn2(TeO3)3] (#131) has face-sharing octahedral dimers Zn2O9 which are cross-linked through TeO3 in a fashion very similar to #128, except that the dimers are arranged in a very open hexagonal honeycomb pattern (Fig. 22). Thus, the resulting framework has hexagonal channels || z that are very large (∼10 Å across). Sodium cations occur in the channels in #131, and are accompanied by water molecules in NaH[Zn2(TeO3)3] · 2.67H2O (#132) and Na2[Zn2(TeO3)3] · 2.97H2O (#133). Cobalt replaces Zn in #134. The negative charge on the framework can be modified by substituting trivalent cations for Zn2+, and other channel cations may substitute for Na+. The channel cation is Mg2+ in the mineral zemannite itself, Mg0.45[(Fe1.123+ Zn0.80Mn0.08)(TeO3)3] · 4.08H2O (#135), which can be idealized as Mg0.5[(Zn2+Fe3+)(TeO3)3] · 4.5H2O, although the data of Miletich (1995a) show that the mean charge on M can vary over the range 2.33‒2.56. The channel contents are arranged as chains [Mg(H2O)6]2+…(H2O)3…[Mg(H2O)6]2+…(H2O)3 that have local trigonal symmetry, but are orientationally and translationally disordered. Kinichilite, ideally Mg0.5[(Mn2+Fe3+)(TeO3)3] · 4.5H2O, and keystoneite, Mg0.5[(Ni2+Fe3+)(TeO3)3] · 4.5H2O, have similar unit-cell parameters to zemannite, but have not had their structures refined (Miletich, 1995a). The new mineral ilirneyite, Mg0.5[(ZnMn3+)(TeO3)3] · 4.5H2O, is also isostructural (Pekov et al., 2015). The channels are completely empty and the framework electrostatically neutral in [Ga2(TeO3)3]-β (#136), much less dense than the α polymorph with the eulytine structure described above (#113). Synthetic selenite analogues of zemannite, K2M2[SeO3]3 · 2H2O (M = Co or Ni) are also known (Wildner, 1993).
Emmonsite, [Fe23+(H2O)(TeO3)3] (#137) is triclinic but has pseudotetragonal symmetry down the x direction. FeX6 octahedra occur in edge-sharing dimers Fe2O8(H2O)2, but these stack such that they can be derived from a continuous edge-sharing chain || x by deletion of every third Fe atom. The Fe dimers of neighbouring chains are connected through TeO3 pyramids, which define the walls of nearly-square channels || x, which accommodate the Te lone pairs (Fig. 22). There is a marked resemblance to the tetragonal structure of minium, Pb4+Pb22+O4 (Gavarri and Weigel, 1975) or schafarzikite, Fe2+Sb23+O4 (Fischer and Pertlik, 1975). The atomic arrangement of emmonsite can in fact be regarded as a threefold superstructure of the schafarzikite type with ordered vacancies: Fe32+Sb63+O12 = (Fe23+)(Te34+3)(O9(H2O)3). Co62+[Te6+O6][Te4+O3]2Cl2 (#678, below) has a closely related structure. [Ga2(H2O)3(TeO3)3] (#138) has the Ga2Te3 substructure arranged approximately like the atoms of α-Ga2O3, which has the corundum structure (Marezio and Remeika, 1967). However, there are Te‒O‒Ga links from Te to only three out of the four nearest Ga. There are two types of Ga atom, one bonded to only tellurite oxygen atoms, while the other centres a GaO3(H2O)3 octahedron. The structure is polar, as groups Ga1(TeO3)3Ga2(H2O)3 all point in the same sense along z. [Nb3O3(TeO3)4]Cl (#139) has a quite different type of structure, in which linear chains || y of corner-sharing NbO6 octahedra pack in a trellis-like arrangement, and are cross-linked through TeO3. Lone pairs and Cl– anions are accommodated in large square channels || y.
The [M(TeO3)] structures #140–142 (M = Co, Ni or Cu) are all of perovskite type. MO6 octahedra share all corners to form a framework, while Te4+ occupies the cubic cages thus defined. The octahedral tilt system is of a+b–b– type, as in GdFeO3 (Glazer, 1972), producing a unit cell with space group Pnma and cell parameters √2 × 2 × √2 of the aristotypical perovskite cube. Bending of Fe‒O‒Fe links and displacement of Gd in GdFeO3 reduce the Gd coordination from 12 equidistant oxygen atoms to six at 2.26‒2.39 Å, two at 2.82 Å and four effectively non-bonded oxygen atoms at >3.1 Å (Coppens and Eibschuetz, 1965). However, the displacement of Te in Co(TeO3) is much more extreme, giving three O at 1.90‒1.92 Å, five at 2.70‒2.98 Å and four at >3.4 Å. Thus, Te forms only three strong bonds, and acts as a brace on the MO3 framework (Fig. 22). This form of Cu(TeO3) (#142) is a high-pressure polymorph of #117 and #297. Note that Te6+ in octahedral coordination can act as the smaller ‘B’ cation in the perovskite structure: #562‒584 below, are examples. An unusual example of Te4+ in the ‘B’ site of a defect perovskite is provided by #195, below.
[Fe3+F(TeO3)] (#143) has zigzag zweier chains || y of edge-sharing FeO4F2 octahedra (the shared edges are alternately F2 and O2). TeO3 groups link trios of neighbouring chains to make a framework. In one polymorph of V25+Te2O9 ≡ [V2O3(TeO3)2] (#144), alternating VO5 and VO6 polyhedra (V1 and V2 respectively) share corners to make a zigzag vierer chain V2O9, with V2 at the angles in the chain. The V1 polyhedron has a geometry that would be more typical for V4+ than V5+: a square pyramid with four V‒O distances 1.78‒2.00 Å, and a very short distance of 1.58 Å corresponding to an apical V=O double bond. The V2 geometry is an extremely distorted octahedron with four distances 1.97‒2.30 Å to oxygen atoms that are shared with V1, Te or both, and two much shorter cis distances 1.60‒1.72 Å to unshared oxygen atoms. Thus, the structure appears to contain two types of vanadyl (V) complex, [V=O]3+ and [O=V=O]+. The V chains lie in layers || (100). One of two types of Te shares two oxygen atoms with V1 + 2 × V2 of one chain and the other oxygen atom with V2 of an adjacent chain. The other type of Te bridges between layers to complete the framework. A second polymorph containing Te2O5 dimers is described below (#203).
Sonoraite, Fe3+(TeO3)(OH)·H2O ≡ [Fe23+(OH)2(H2O)(TeO3)2]·H2O (#145), has edge-sharing octahedral dimers Fe2O8(OH)2 and Fe2O4(OH)5(H2O) alternating and linked through shared OH– corners to make vierer chains || . The chains are packed in an approximately hexagonal array, and TeO3 groups link trios of neighbouring chains into a framework. The non-framework water molecule is loosely held in a structural cage between TeO3 groups. [Ta2O3(TeO3)2] (#146) has a similar stoichiometry for its structural unit, but a quite different structure in which TaO6 octahedra share three corners to form a layer || (002) of 4- and 8-rings. The Te atoms are in the interlayer regions, and half of the TeO3 groups share all three ligands while the other half share only two, in order to link the Ta layers into a framework. The structure of stoichiometrically analogous [Sb25+O3(TeO3)2] (#147) is again quite different. Half of the SbO6 octahedra share four corners and half share two corners to make [Sb2O9]8– ribbons of zigzagging 4-rings || y. These ribbons pack in a herringbone fashion, and are connected into a framework through four crystallographically distinct types of TeO3 group (Fig. 22). The wide variety of structures possible for M2TenX9 structural units is demonstrated further by Na2[W2O6(TeO3)] (#148), which has a structure in which eight types of WO6 octahedron share two to three corners to form corrugated layers || (200), which have bands of 4-rings || y alternating with bands of 8-rings. Two of the four types of Te act as braces on particular W layers, while the other two types link the layers into a framework. Na+ cations lie between the layers, in irregular 7‒8 coordination. K3[(V4+O)4(V5+O4)(TeO3)4] · 4H2O (#149) has clusters of five VOn polyhedra: four very distorted octahedra V4+O6 (V‒O distances are 1 × 1.61, 4 × 1.96‒2.07 and 1 × 2.26 Å) and a V5+O4 distorted tetrahedron (2 × 1.63 and 2 × 1.82 Å). The octahedra form two face-sharing dimers V2O9, which each share one of their bridging oxygen atoms with the tetrahedron to make a mixed-valence pentameric anion [V44+V5+O20]19–. All oxygens which are not part of the VO4 tetrahedron or the [V = O]2+ cation are shared with TeO3 groups, which again connect trios of neighbouring vanadate units into an open framework which has large channels (7‒8 Å diameter) running || y and z. Water molecules and CN 8‒10 K+ ions are in the channels (Fig. 22).
[Ni11(TeO3)10Cl2] (#150) has five types of NiO6 octahedra forming thick layers || (001), which are linked into a framework through edge-sharing pairs of NiO5Cl octahedra. Five types of TeO3 brace the structure, two of which show orientational disorder, evidenced by mutually exclusive split positions for oxygen atoms. [Ni7(TeO3)6Cl2] (#151) has NiO5Cl octahedra sharing edges in a very open honeycomb pattern, making very low-density layers || (003). The framework is formed by TeO3 and additional NiO6 between the layers and sharing edges with them. Ni3(TeO3)2(OH)2 (#152) is more informatively written [Ni6(TeO3)4(OH)3](OH). It, and its Co analogue (#153), have an unusual structure in which face-sharing dimers of octahedral M2O7(OH)2 share additional edges to form zigzag chains M4O10(OH)2 || z. These chains in turn share corners to act as walls surrounding large channels (9 Å diameter) along the 63 screw axis of the structure, and small channels along the triad axes (Fig. 22). The overall composition of the resulting nanoporous framework is M12O24(OH)6 per unit cell. Two Te occupy the small channels and another six brace the large channels, which contain two more very loosely bound OH– anions to complete the unit-cell content, M12(TeO3)8(OH)8.
In [Ga2Mo6+O4(TeO3)2] (#154), edge-sharing chains || y of GaO6 octahedra are linked through chains [MoO4(TeO3)] of alternating TeO3 and distorted MoO6 octahedra (4 × 1.71‒1.99 and 2 × 2.37 Å) to form undulating layers || (002). A second type of Te connects these layers into a framework. K[Nb3O6(TeO3)2], its Ta analogue and the corresponding Rb compounds (#155‒158) have octahedral MO6 (M = Nb or Ta) sharing four corners to make corrugated layers of 4-rings, M3O12 || (020). Interlayer TeO3 shares all three corners to link these layers into a framework. CN12 K+ ions are also in the interlayer gap.
The structure of [Ni3(MoO4)(TeO3)2] (#159) bears some resemblance to those of #152‒153. Four zigzag chains of edge-sharing NiO6 and two of corner-sharing NiO5, all chains having the composition Ni2O8, share additional corners to form walls around large pseudohexagonal and small pseudotrigonal channels || x, forming a nanoporous framework Ni6O18. The small channels are empty, but the large hexagonal channels are braced by four TeO3 pyramids and two MoO4 tetrahedra. Reduction of some Ni coordination numbers to 5 occurs because of elimination of a bond to an oxygen atom of a neighbouring Ni polyhedron, preventing overbonding of the latter oxygen, which is part of the MoO4 group. [Co7(TeO3)4Br6] (#160) has layers of cis-CoO4Br2 octahedra sharing edges to make layers with 7-rings, || (200). The layers are linked into a framework via trans-CoO2Br4 octahedra, which share faces with octahedra in the layers above and below. Tellurium atoms brace the layers, rather than acting as interlayer bridges. [Fe23+(TeO3)O2] (#161) has corrugated layers of edge-sharing FeO6 octahedra || (100), with Te bridging across the interlayers. [Co2(H2O)(SO4)(TeO3)] and its Mn analogue (#162‒163) have crankshaft chains || z of edge-sharing octahedra MO5(H2O) (M = Co or Mn), which additionally link their H2O corners to make undulating layers with 6-rings || (100). The layers are braced by TeO3 but connected into a framework by SO4 tetrahedra, which share two oxygen atoms with the layer on one side and one oxygen atom with the layer on the other. [Zn2(MoO4)(TeO3)] (#164) has alternating ZnO6 and ZnO4 polyhedra sharing corners to make layers of 6-rings || (001). TeO3 groups span the 6-rings, acting as braces, while interlayer MoO4 tetrahedra share two oxygen atoms with each adjacent Zn layer to make a framework. [Fe33+O(TeO3)3]Cl (#165) has FeO6 octahedra sharing edges to make a helical vierer chain || x that is a thin fragment of a rocksalt-like structure. These chains are linked through two opposite ligands of additional FeO6 octahedra to make a very open nanoporous framework with rhombic channels || x of diameter ∼12 × 8 Å (Fig. 23). TeO3 groups reinforce the cross-links, and have their lone pairs pointing into the large channels. While the non-tellurite O2– ion is part of the structural unit, in the core of the Fe chain, Cl– is only loosely bound, and sits in the channels.
K2[W3O9(TeO3)] (#166) has WO6 octahedra sharing four corners to form layers || (020) with the 3- and 6-rings of the kagome net. Tellurium atoms link these layers into a framework, forming additional 3-rings with two W atoms on one or other side of the interlayer gap. Potassium in the interlayer is 8-coordinated. [Ni6(Mo4O16)(TeO3)2] (#167) has rhomb-shaped tetramolybdate anions [Mo4O16]8– which are held together by two CN3 and four CN2 bridging oxygen atoms. Layers of molybdate anions || (200) alternate with layers containing zigzag sechser chains of edge-sharing NiO6 running || y. The molybdate groups link Ni chains of the same and successive layers to form a framework. Between the molybdate groups, there are channels || z containing the Te, which also cross-link the Ni layers. [(Mo25+Mo36+)O13(TeO3)] (#168) has five types of Mo in very distorted octahedral coordination. There is a very wide spread of Mo‒O distances: all Mo atoms have one very short Mo‒O of 1.69‒1.71 Å, all except Mo1 have one very long (2.38‒2.46 Å), while the rest are 1.78‒2.13 Å. The bond-valence parameters of Brese and O'Keeffe (1991) give a correspondingly wide spread of individual bond valences (0.22‒1.78 vu) but a narrow range of bond-valence sums of 5.82‒6.12 vu for all Mo, implying that there is no ordering of Mo5+ and Mo6+. MoO6 polyhedra share four corners to form layers || (002). The layers are of a modified ‘tungsten bronze’ type, in which bands of 4-rings || y alternate with bands of 3- and 6-rings. Mo2‒Mo5, in the 4-rings, share additional corners with octahedral in the layers above and below, connecting the layers into a [Mo25+Mo36+O16]4– anionic framework. The charge-balancing Te4+ ion sits in the 6-ring of the layer (Fig. 23). The TeO3 group shares an edge with the Mo1 octahedron, where 3- and 6-rings meet, and its remaining ligand is one apical oxygen atom of the Mo1 octahedron of the layer either above or below. A high-temperature polymorph is described as #281 below. Ba2[Nb6O15(TeO3)2] and its Ta analogue (#169‒170) have octahedral MO6 (M = Nb or Ta) sharing corners to make a three-cation wide layer that is a slice of ReO3-type structure || (221) of the cubic ReO3 cell. The layers are connected into a framework by sparse shared edges, which separate the interlayer into channels || y of the resulting monoclinic cell. The channels contain CN11 Ba2+, while Te braces the layers but unusually has its lone pair directed in towards the centre of the layer rather than into the interlayer gap. Na4[(UO2)3(TeO3)5] (#171) has a large (a ≈ 17 Å) unit cell in a low-symmetry cubic space group (I213). UO7 and TeO3 polyhedra form a framework in which every equatorial ligand of U links to one of three types of Te. The Te2 and Te3 sites lie on triad axes, and are arranged like 2 × 2 × 2 unit cubes of the CsCl structure. Three Te1 and three U form a buckled hexagon around each Te2, and the equatorial oxygen atoms of UO7 link to the central Te2, the nearest Te3, the two nearest Te1 of the same hexagon and one Te1 of a neighbouring hexagon (Fig. 23). Three types of Na+ are weakly held in interstices, in irregular 3‒6 coordination.
Structures with neso Te4+X4–5 as part of the structural unit
Table 13 (deposited) shows structures #172‒194, in which TeX4 or TeX5 polyhedra do not link to other Te. However, it is interesting to note that unlike TeX3, these polyhedra always link to some other relatively strongly bonded cation, and thus are always part of a larger structural unit. Nabokoite, K[Cu7(TeO4)(SO4)5]Cl (#172) is the only example of a structure with neso square-pyramidal [TeO4]4– anions (cf. Fig. 4b). Favreauite, Pb[Cu6(BiO4)(SeO3)4(OH)](H2O), is nearly isostructural according to Mills et al. (2014b), who noted that the lone-pair cations (Bi, Te) are partially surrounded by corner-sharing CuO4 squares to make a thick layer || (002), that can be regarded as a slice of the structure of murdochite, Cu6Pb4+O8 (Dubler et al., 1983). The resemblance is emphasized if the arrangement is considered of oxygen-centred tetrahedra OCu3A (A = Te and Pb), according to the approach of Krivovichev et al. (2013). The layers are braced by SO4 tetrahedra, while the interlayers contain Cl–, which is a distant fifth ligand for one of the Cu2+ cations, and CN8 K+ (Fig. 23). Atlasovite, KCu6Fe3+BiO4(SO4)5Cl, is closely associated and appears to be isostructural with nabokoite (Popova et al., 1987).
Four-coordinate Te4+ is rare in the square-pyramidal geometry, but occurs much more often in the ‘trigonal bipyramid – 1’ or ‘folded rhombus/kite’ geometry of Fig. 4c. An example is the pyrosulfate [Te(S2O7)2] (#173). The (S2O7)2– anions are bidentate ligands for the Te4+ cation, so neutral molecules are formed which consist of butterfly-like pairs of Te‒S‒S 3-rings. These are held together through long Te…O bonds (Fig. 24). In In2[Mo6+O3(TeO4)](TeO3), (#174), InO6 and InO8 polyhedra share corners and edges to form a stepped layer || (10), which is not treated as a structural unit here, given the high coordination number of half of the In cations. The layers are braced by isolated (TeO3) pyramids but also by MoO4 tetrahedra and TeO4 polyhedra, which share a corner to form the dimeric anion [MoO3(TeO4)]3–. Layers are held together by long Te…O bonds. A much more complex finite cluster occurs in (NH4)9K[V44+V85+Mo126+O65(TeO4)(TeO3)2] · 27H2O (#175). The clusters are very similar to the expanded/modified Keggin/Dawson-type anions of #54–62, but are included here because it includes not only two TeO3 groups centring its two dissimilar half-cages, but also a TeO4 polyhedron as part of one half-cage (Fig. 24). There is considerable (V, Mo) disorder in the half-cuboctahedral cages. One Keggin half-cuboctahedron of composition [(V4.25Mo4.75)O30(TeO3)] shares edges and CN3 oxygens with the Mo octahedra of a ring of three MoO6 octahedra alternating with three V5+O4 tetrahedra; the Mo and V polyhedra of this ring also share corners with the other half-cuboctahedron [(V4.75Mo4.25)O30(TeO3)], which has the CN4 Te sitting outside one of the square faces, and bonded to the oxygen atoms surrounding that face. The anions pack with their long axes approximately parallel to , and are held together through weak bonds to K+ and NH4+ ions and H2O molecules.
BaV25+TeO8 ≡ Ba2[(VO2)4(TeO4)2] (#176) has VO5 polyhedra forming edge-sharing dimers [V2O8]6– and isolated tetrahedra [VO4]3–. TeO4 polyhedra share an edge with VO5 and the remaining corners with two VO4 groups to form a zweier double chain with 8-rings, [(VO2)4(TeO4)2]4– (Fig. 24). Expressing the V component as vanadyl [VO2]+ complexes is suggested by the bond distances: 2 × 1.65 and 2 × 1.78–1.83 Å for VO4 and 2 × 1.64–1.65 and 3 × 1.89–1.99 Å for VO5. The chains run || y, and are flattened || (103); they are linked through CN10 Ba2+. The Sr analogue (#177) has a much more complex structure in which eight kinds of VOn polyhedron occur as corner-sharing dimers, either O3V‒O‒VO3 or O3V‒O‒VO4, and these are linked through TeO3 and TeO4 to make sechser double chains with 6- and 10-rings, [V8O18(TeO4)2(TeO3)2]8–. The subchains are joined through TeO4 and VO5 sharing an edge. Two topologically similar but crystallographically distinct types of such chain run || , and are connected through CN8–10 Sr2+ ions. (NH4)4[Mo6O16(TeO4)] · 2H2O and its Rb equivalent (#178–179) have clusters [Mo6O22]6– in which the six Mo atoms are arranged as a pair of tetrahedra sharing an edge; they are held together through two CN4 oxygen atoms at the centres of the tetrahedra and eight CN2 oxygen atoms along edges. These clusters are linked through Te into continuous complex chains || z, held together through NH4+/Rb+ and water molecules.
The structural unit of Ba[NbO(PO3)(TeO4)] (#180) is a highly corrugated layer || (002). Corner-sharing NbO6 octahedra form linear chains || y. Each pair of adjacent Nb polyhedra then share additional corners with a TeO4 polyhedron and a PO4 tetrahedron to form a tetrahedral Nb2TeP cluster. The Te atom makes an additional link to Nb in the next chain, thus joining chains into layers. Note that TeO4 and PO4 share one oxygen atom: the formula above has been written to emphasize the Te coordination, but the structural unit could equally be written [NbO(PO4)(TeO3)]2–. CN10 Ba2+ share edges with the phosphate groups, holding the layers together (Fig. 24). CdMo6+TeO6 ≡ [Cd(MoO2)(TeO4)] (#181) has a layered structure that is a highly modified 1 × 1 × 2 superstructure of the fluorite type with ordered cations and also cation and anion vacancies, (CdMoTe)(O62) ≡ Ca4F8. The cations occupy ¾ of the sites in a ccp array, with a four-layer sequence along z in which layers repeat in the order Cd, (Mo+Te), □, (Mo+Te). Only long Te…O bonds hold the structure together across the layers of vacant cation sites. Cadmium is 8-coordinated by oxygen atoms, but half of the oxygen sites that are not adjacent to Cd are vacant. Displacements of the remaining oxygen atoms reduce the coordination number of Te and Mo from 6 to 4, although the geometries are different: respectively ‘trigonal bipyramid – 1’ and tetrahedral. The TeO4 and MoO4 polyhedra share corners to form chains [MoTeO6]2– that run ||  and  in alternate layers along z (Fig. 24). When smaller divalent cations replace Cd, a small lattice strain reduces the coordination number from 8 to 6, and the symmetry from tetragonal to orthorhombic, as seen for [M(MoO2)(TeO4)], (M = Mg, Mn, Co and Zn) (#182–185). Given the close structural relationship and very strong layering, the Cd compound has been grouped with these as having a layer rather than chain structural unit, even though 8-coordination would normally exclude Cd2+ from the unit as too weakly bonded. In BaMo26+TeO9 ≡ Ba[Mo2O5(TeO4)] and its W analogue (#186–187), MO6 octahedra (M = Mo or W) share three fac corners to form undulating layers || (001) of 6-rings that are in chair configuration. The layers are reinforced by TeO4 sharing edges with two Mo polyhedra and a corner with a third octahedron, while the layers are held together by CN11 Ba2+ ions.
The compounds A2Mo26+Te(PO4)2O6 ≡ A2[Mo2P2O10(TeO4)] (A = Rb, Cs and Tl; #188–190) alternating MoO6 octahedra and PO4 tetrahedra share corners to form layers [(MoO3)2(PO4)2]6– || (002) with a 3-connected net of 4- and 8-rings. Interlayer TeO4 polyhedra link these layers into a framework, sharing corners to make Te‒P‒Mo 3-rings. The A+ cations are in 9-coordination in the interlayer. Note that Tl+ shows little sign of lone-pair stereoactivity (Tl–O = 2.77–3.17 Å, compared to 2.83–3.17 Å for Rb). Mn2+V24+TeO7 ≡ [Mn2+V24+O3(TeO4)] (#191) has double chains ||  in which pairs of VO5 alternate with pairs of VO6, sharing edges and CN3 oxygen atoms in a manner similar to the Te chain of Fig. 10a. These chains are linked into double layers || (001) by markedly asymmetric TeO3+1 groups (Te–O = 3 × 1.86–1.94 and 1 × 2.34 Å; with no more until 2.87 Å), which share edges with two VO6 of one chain and a corner with VO5 of an adjacent chain. The layers are then linked into a framework through MnO6 octahedra, which also share edges with VO6 and TeO4. For simplicity, the formula above does not indicate that each V atom has one very close (1.64–1.67 Å), approximately double-bonded oxygen ligand, as is typical for V4+. The remaining non-tellurite oxygen atom is shared by one VO6 and two VO5 polyhedra.
In Zn3V25+TeO10 ≡ [Zn3(VO3)2(TeO4)] (#192), two types of ZnO5 and one of ZnO6 polyhedra share edges and corners to make very thick, but looped and low-density layers || (020). The higher-valence cations are linked into bow-shaped anions [O3V‒O‒TeO2‒O‒VO3]6–, which are embedded in the Zn layers, but also cross-link them into a framework through one of the terminal vanadate groups (Fig. 24). Co62+Te5O16 ≡ [Co6(TeO4)(TeO3)4] (#193) has blocks of edge-sharing tetramers of CoO6 octahedra sharing corners with edge-sharing dimers of CoO5 to make a framework with large channels running || y. Four types of TeOn groups cross-link across the channels and other interstices between the Co blocks. The only example of TeX5 in a framework is Co32+Te2(PO4)2O2(OH)4 ≡ [Co3(PO2)2(TeO3(OH)2)2] (#194), which has trans edge-sharing chains of CoO6 octahedra running || y. These are linked into layers || (200) through corner-sharing chains of alternating PO4 tetrahedra and TeO3(OH)2 square pyramids, [TePO5(OH)2]3–, and then into a framework through interlayer CoO2(OH)4 polyhedra.
Dimeric Te4+ oxyanions [Te2Xn] (n = 5–9)
Compounds with dimeric sorotellurite groups are #195–227, shown in Table 14 (deposited). The simplest example of the [Te2O5]2– anion (Fig. 8a) is provided by Cs2[Te2O5] (#195). Interestingly, the structure can be regarded as a defect perovskite, with 1/6 of the anions missing. The orthorhombic unit cell corresponds to a 2√2 × 2√2 × 2 supercell of the aristotypical perovskite cube, with Cs+ in the large-cation A positions and Te in B sites. The Te2O5 dimers are separated from one another by vacancies on the anion sites and asymmetrical Te–O…Te links, where the short distances are 1.83–1.97 Å and the long ones are 2.68–4.43 Å. The Te–Te vectors of anions point along  or  in a herringbone pattern (Fig. 25). The Cs+ ions are in 9-coordination. Ba3[Te2O5](TeO3)Br2 (#196) has edge-sharing BaO8Br and BaO5Br3 polyhedra forming columns || y, which share edges sparsely to surround elongated micelles. These channels accommodate the Br– anions and also the lone pairs of the Te dimers and monomers, which brace the micelle walls. In Nd2[Te2O5](TeO3)(MoO4) (#197), NdO8 polyhedra share edges to form undulating honeycomb layers || (200). The layers are braced by molybdate and tellurite groups, and are held together only by long Te…O bonds.
TiTeO3F2 ≡ [Ti2OF4(Te2O5)] (#198) has TiO2(O0.5F0.5)2F2 octahedra sharing corners to make zweier double chains of 4-rings, [Ti4O8F8]8– running || x. Te2O5 groups share four oxygen atoms with a square of adjacent Ti atoms in the double chain to make Te2Ti4 triangular prisms. Long (2.73–2.77 Å) Te…O bonds link the resulting loop-branched double chains into double layers || (020), which are then held together through 2.56 Å Te…F bonds (Fig. 25). V24+Te2O7F2 ≡ [(VO)2F2(Te2O5)] (#199) has cis-VO4F2 octahedra sharing alternately O2 and F2 edges to make zigzag zweier chains || x. Te2O5 groups cross-link the V chains into layers || (010) in such a way that Te < (V = V) > Te ‘double triangle’ clusters are formed. As is typical for V4+, the distance to the non-tellurite oxygen ligand is very short (1.595 Å). The layers are held together via long Te…O and Te…F links. Fe8Cu3Te12O32Cl10 ≡ [(Fe2+Fe73+)(Te2O5)4(TeO3)4](Cu21+Cl6)(Cu1+Cl2)Cl2 (#200) has zigzag chains of edge-sharing FeO6 polyhedra || z, which share corners to make layers || (200). These layers are braced by the monomeric and dimeric Te anions. Between the layers lie three types of Cl– anions. One of these makes long bonds to the Te of Te2O5 in the adjacent layers, while the other Cl– anions form chains of edge-sharing tetrahedra || z. The tetrahedral interstices are half occupied by Cu1+, which also occupies linear twofold coordinated sites in the tetrahedron edges that are || y. Thus, a chain is formed of alternating [Cu2Cl6]4– and [CuCl2]– anions (Fig. 25).
The elegant framework structure of [Cu2(Te2O5)Cl2] and its Br analogue (#201–202) has square pyramids CuO4X (X = Cl or Br) that share edges in groups of four to make clusters with a Cu4O4 cube at the core and have ¯4 point symmetry. These clusters sit in columns || z, and are linked to their neighbours by corner-sharing with Te2O5 groups, which act as the walls of ∼6 Å diameter square channels || z. The channels accommodate the X– anions and Te lone pairs (Fig. 25). Our second polymorph of V25+Te2O9 ≡ [(VO2)2(Te2O5)] (#203) has corner-sharing zweier chains of VO5 polyhedra that lie in layers || (400) and run ||  or  in alternate layers. Te2O5 groups bridge two V chains of one layer and two chains of the next to make a framework. A polymorph with monomeric rather than dimeric TeO3 groups was described above (#144). Cr23+Te4O11 ≡ [Cr2(Te2O5)(TeO3)2] (#204) has edge-sharing dimers of CrO6 octahedra linked into a framework through Te; only the central oxygen of Te2O5 does not link to Cr. Ni3.4[Ni30(Te2O5)6(TeO3)20]Br14.8 (#205) and Ni4.5[Ni30(Te2O5)6(TeO3)20]Cl18.45 (#206) are isotypical compounds with large cubic unit cells (Im, a ≈ 17.5 Å). The ordered part of the structure consists of NiO6 and NiO5 polyhedra which shared edges to form large icosahedral cages, braced by the Te anions, [Ni302+(Te2O5)6(TeO3)20]8+. The rest of the structure shows substantial disorder. The large cages contain an icosahedral cluster of 12 Ni positions which cannot be >50% occupied due to short Ni…Ni distances. Partly occupied halide anion sites are at the core of these clusters, surrounding their exteriors so that the additional Ni2+ are in approximate 4-coordination by (Br,Cl)–, and halfway between the Ni clusters in channels || < 100>, making long Te…(Br,Cl) bonds.
Ho11[Te2O6]2(TeO3)12Cl (#207) has layers || (012) of edge-sharing HoO6–8 polyhedra, sparsely connected into a framework with channels running || x. The walls of the channels are braced by six types of monomeric (TeO3)2– and also a dimer in which TeO3 shares a corner with TeO4 to form the structural unit, [Te2O6]4– (Fig. 8b). The shortest Te‒O distance is 2.18 Å; there are no others until 2.47 Å. The Cl– ion is loosely bound in the centre of the channels, into which the Te lone pairs also point. Moctezumite, Pb[(UO2)(Te2O6)] (#208) has similar Te2O6 dimers, fringing zigzag chains of edge-sharing UO7 polyhedra to form broad ribbons || y. The ribbons lie in layers || (10), which are held together by PbO3+4 polyhedra. A high-pressure (5.09 GPa) structure of (NH4)2WTe2O8 ≡ (NH4)2[WO2(Te2O6)] (#209) has layers || (100) in which WO6 octahedra share four corners and each Te of Te2O6 shares two, to form a net of Te‒W‒Te‒Te‒W 5-rings which has the topology of the ‘Cairo tiling’ (Hyde and Andersson, 1989), seen in less crumpled form in the tetrahedral sheet of the melilite group of minerals, e.g. (Ca,Na)2[(Al,Si,Mg)3O7] (Smith, 1953). Layers are held together by NH4+ ions (Fig. 25). InV5+Te2O8 ≡ [In(VO2)(Te2O6)] (#210) has zigzag corner-sharing chains of InO6 octahedra || y, sharing additional corners with VO4 tetrahedra to make In‒In‒In‒V 4-rings. Te2O6 groups link the resulting ribbons to their neighbours to form thick layers || (10), held together only through long Te…O bonds. Showing the V as vanadyl is validated by the asymmetry in V‒O distances (2 × 1.61‒1.69 Å and 2 × 1.81‒1.83 Å). Poughite, Fe23+Te2O6(SO4) · 3H2O ≡ [Fe2(H2O)2(SO4)(Te2O6)] · H2O (#211), has an edge-sharing dimer of FeO5(H2O) octahedra sharing additional corners with an SO4 tetrahedron to form a finite 3-ring cluster [Fe2O6(H2O)2(SO4)]8–. These are linked together in groups of four through Te2O6 to make layers || (020), with the remaining H2O molecule in the interlayer (Fig. 25). Sr[Cu2Cl(Te2O6)]Cl and the isostructural Ba compound (#212‒213) have Te2O6 sharing all six oxygen atoms with Cu2O6Cl dimers of CuO4 and CuO3Cl squares, to make a rather open framework, with channels || x and y which contain the remaining Cl– and also (Sr,Ba)2+ ions, coordinated by 6O + 2Cl.
Pb3Te2O6Cl2 ≡ Pb6[Te2O6](TeO3)2Cl2 and its Br analogue (#214‒215) contain edge-sharing dimers of TeO4 (Fig. 8c). They have edge-sharing layers || (20) of PbO4X4, PbO5X3 and PbO8X polyhedra (X = Cl or Br). The layers are braced by Te2O6 and TeO3 groups, and loosely held into a framework through Pb‒X and Te…X bonds. The Te‒O bonds out to 2.05 Å in mroseite (#216) define a neutral dimer [Te2O4]0, made from an edge-sharing pair of TeO3 pyramids, suggesting that the formula be written Ca2(Te2O4)[CO3]2. However, each Te is only 2.31 Å from a carbonate oxygen atom, so it could also be expressed as Ca[(CO2)2(Te2O6)]. The S-shaped [(CO2)2(Te2O6)]4– heteropolyanions pack in a herringbone pattern to form strongly corrugated layers || (002), which are held together through CN8 Ca2+ ions (Fig. 26). NaV5+TeO5 ≡ Na2[(VO2)2(Te2O6)] is isotypical with its K+ and Ag+ analogues (#217‒219). In these compounds, edge-sharing Te2O6 dimers share corners with VO4 tetrahedra to make chains with 4-rings Te‒V‒Te‒V, which extend || . These pack, herringbone fashion, in layers || (020). Large cations between the layers are in 8-fold coordination, but while this is fairly regular for K+ (K‒O = 2.67‒2.99 Å), it is less so for the smaller cations (Na‒O and Ag‒O both = 2.41‒2.91 Å). Ba2V25+Te2O11 ≡ Ba2[(VO2)(VO3)(Te2O6)] (#220) has Te2O6 groups alternating with VO4 tetrahedra to make a dreier chain || y, which has a second type of VO4 as an open branch on one of the central oxygen atoms of each Te2O6. The chains lie in layers || (200), which have BaO8–10 polyhedra between them. The α form of Tl2(UO2)Te2O6 ≡Tl4[(UO2)2(Te2O6)(TeO3)2] (#221) has a β dimorph with monomeric TeO3 in a heteropoly chain structure, #70 above. The α structure has edge-sharing pairs of UO7 polyhedra sharing further edges with Te2O6 to make ribbons || . The U polyhedra are bridged by additional TeO3 to link the ribbons into layers || (1) with 12-rings and U‒Te‒U‒Te 4-rings. Between the layers are Tl1 showing little lone-pair stereoactivity (8 × O at 2.63–3.13 Å) and Tl2 with much less symmetrical coordination (3 × O at 2.49–2.69 Å and four more at 3.27–3.56 Å) (Fig. 26). The β phase of V4+TeO4 ≡ [(VO)2(Te2O6)], stable at high-temperature, has corner-sharing VO5 making zigzag chains || (#222); the V polyhedra share edges with Te2O6 to make layers || (100) (Fig. 26). The vanadyl oxygen atom is at 1.61 Å from V, but also makes three weak bonds to Te of the next layer (2.96–3.25 Å). There is little resemblance to the raspite structure of the α dimorph (#82). NiV25+Te2O10 ≡ [Ni(VO2)2(Te2O6)] (#223) has distorted VO6 octahedra sharing edges to make zigzag chains [V2O8]6– || x. Te2O6 groups each bridge four such chains, while NiO6 octahedra bridge two V chains and two Te2O6 groups, forming a quite dense framework with many Ni‒V‒V and Ni‒V‒Te 3-rings.
BaMo2Te2O11 · H2O ≡ Ba[Mo2O4(Te2O7)] · H2O (#224) has cis-corner-sharing MoO6 octahedra forming helical vierer chains || y. These are linked into thick layers || (002) via Te dimers which are corner-sharing [Te2O7]6– groups (Fig. 8d), if Te–O distances out to 2.43 Å are included. Water molecules and CN10 Ba2+ ions are between the layers. Polymorph I of TeO(As5+O3OH) ≡ [(AsOH)2(Te2O8)] (#225), like many compounds of Te with other high bond-valence cations, has Te4+ in 5-coordination. Here, two TeO5 pyramids share edges to make a [Te2O8]8– dimer (Fig. 8e). These share corners with [AsO3OH]2– tetrahedra to make As<(Te=Te)>As clusters which are yet another variant of the common ‘double triangle’ motif (Fig. 26). The remaining unprotonated oxygen atom on As is shared with Te of a neighbouring cluster, linking clusters to form double chains || y. Chains are held together through hydrogen bonds and long Te…O bonds. Tellurium polyhedra are condensed further in polymorph II (#299, below). The edge-sharing Te2O8 group also features in Ba2TeO(PO4)2 ≡ Ba4[(PO2)2(PO3)2(Te2O8)] (#226), where such groups are bridged by pairs of corner-sharing PO4 tetrahedra to form dreier chains with [‒Te = Te‒P‒] backbones, not unlike the Te‒V chains of #220. Again, a second type of tetrahedron forms open branches, but this time, the additional PO4 groups attach to a single Te atom rather than to a bridging oxygen of the chain backbone. The resulting ribbons lie in layers || (10), with BaO9–11 between them. Te2O(PO4)2 ≡ [P2(Te2O9)] (#227) again has CN5 Te4+ (cf. # 225), but this time in corner-sharing dimers [Te2O9]10– (Fig. 8f). The dimers are linked into a framework through two types of PO4 tetrahedra: one type shares oxygen atoms with Te atoms of four different dimers, whereas the other type links only three dimers, but forms a Te‒Te‒P 3-ring with both Te atoms of one of them.
Trimeric Te4+ oxyanions [Te3Xn] (n = 8–11)
Trimeric sorotellurite groups occur in structures #228–249 (Table 15, deposited). Sr3[Te3O8](TeO3) (#228) has a triclinic structure in which 2-wide and 4-wide ribbons of edge-sharing SrO7–9 polyhedra run || x and are linked into a framework through a few additional shared edges. Two types of [Te3O8]4– (Fig. 8g) and two types of monomeric [TeO3]2– line channels || x in the Sr–O matrix. Ba3[Te3O8](TeO3) (#229) is nearly isostructural, although with x and z axes exchanged, as published. However, it retains a centre of inversion symmetry that is lost in the Sr compound, concomitant with which all Te3O8 groups are equivalent and Ba2+ has CN = 8–9. Both structures resemble the unexpectedly complex structures of the (Ca,Sr)(TeO3) phases (#9‒15). BaLa2[Te3O8](TeO3)2 (#230) is quite different, in that it has layers || (200) of edge-sharing BaO12 and LaO8 polyhedra, which are braced by TeO3 groups and bridged by Te3O8. The series A2[Te3O8](Te2O5) (A = Dy, Ho, Er, Tm, Yb, Lu and Y; #231–237) show even less polymerization of large-cation polyhedra. They have 4-wide ribbons of edge-sharing AO7–8 polyhedra running || x. Interestingly, the Te3O8 groups do not link the ribbons, but only decorate their edges, while Te2O5 groups link the ribbons into thick double layers || (001) (Fig. 26). Long bonds to both types of tellurite anion hold the layers together. Sr4[Te3O8]Cl4 (#238) has SrO8Cl, SrO4Cl5 and SrO2Cl7 polyhedra. These share O2 edges to form 8-cation-wide ribbons of a fluorite-like structure || y, which connect through the longer Sr–Cl bonds into continuous layers || (20), and more sparsely into a three-dimensional framework. The Te3O8 groups brace the Sr–O ribbons. In La2[Te3O8](MoO4) (#239), LaO9 polyhedra share faces, edges and corners to form an open framework La2O11 that has channels || x and y. Te3O8 groups run down the length of the y channels, sharing all oxygen atoms with La, while MoO4 tetrahedra brace the x channels, sharing only three ligands with La.
Na2Mo3Te3O16 ≡ Na2[Mo3O8(Te3O8)] and its Ag analogue (#240–241) have MoO6 octahedra sharing edges to form V-shaped trimers [Mo2 = Mo1=Mo2], which are linked via Te3O8 groups [Te2‒Te1‒Te2] to form complex loop-branched dreier chains || x. The chain has two equivalent backbones ‒Te1‒Te2‒Mo2‒ and ‒Te1‒Mo2‒Te2‒, which intersect at Te1 and are further braced through Mo1=Mo2 and a CN3 oxygen atom which bonds to Mo1, Mo2 and Te2. Chains are held together via CN6–8 (Na, Ag)+ (Fig. 26). Ca[Co2Cl2(Te3O8)] and its isostructural Sr‒Co and Sr‒Ni analogues (#242–244) have a structure in which MO5Cl octahedra (M = Co or Ni) share edges to form helical vierer chains || y (cf. #165 and 224). Te3O8 groups bridge these chains to form double layers || (002), the long axes of the Te anions pointing ||  on one side of the layer and ||  on the other side. CN8 (Ca,Sr)2+ ions sit in the cores of the double layers, which are held together through long Te…Cl interactions. Nb2Te3O11 ≡ [Nb2O3(Te3O8)] (#245) has ladder-like einer double chains of 4-rings || z of corner-sharing NbO6 octahedra, arranged in a herringbone pattern. Te3O8 groups share two corners with octahedra in each of the Nb chains in the ±x directions and one corner with each of the Nb chains in the ±y directions, forming a framework (Fig. 27). NaNb3Te4O16 ≡ Na[Nb3O5(Te3O8)(TeO3)] (#246) has broader zweier triple chains of 4-rings || y of NbO6 octahedra, which are decorated by the TeO3 monomer, making Nb‒Nb‒Te 3-rings. The Te3O8 groups link each Nb ribbon to three of its neighbours in a framework. The CN8 Na+ ion lies between the TeO3 monomers of neighbouring Nb chains. Na1.4Nb3Te4.9O18 ≡ Na1.4[Nb3O4(Te2.9O8)(TeO3)2] (#247) has a very similar composition and two very similar unit-cell parameters, but the topology is quite different. The NbO6 octahedra form single and double zweier chains || y, that alternate in the x direction. Te3O8 groups are slightly defective, in that the central Te is only 90% occupied. They link Nb single chains with each other and with neighbouring double chains, forming an open framework with numerous Nb‒Nb‒Te and Nb‒Te‒Te 3-rings, while two types of TeO3 monomer form additional Nb‒Nb‒Te rings on the double chains. The CN8 Na+ ions partially occupy sites that lie in channels || y (Fig. 27).
Dy2(TeO3)3 ≡ Dy4[Te3O9](TeO3) (#248) has a surprisingly complex structure with thick, loop-branched layers of edge-sharing DyO7–8 polyhedra || (100). These are bridged by V-shaped corner-sharing trimers [Te3O9]6–, including Te–O distances out to 2.44 Å (Fig. 8h), and three types of independent (TeO3)2– ion (Fig. 27). The high As5+–O bond valence (∼1.25 vu) leads to Te adopting 5-coordination in Te3O3(AsO4)2 ≡ [As2(Te3O11)] (#249). Three TeO5 pyramids share a common CN3 oxygen atom and two CN2 oxygen atoms to form [Te3O11]10– trimers (Fig. 8i). AsO4 tetrahedra share two corners with one such trimer and one each with two neighbouring trimers to form continuous layers || (010), with Te‒Te‒As 3-rings, Te‒As‒Te‒As 4-rings and Te‒Te‒As‒Te‒Te‒As 6-rings. Layers are held together only through long Te…O bonds (Fig. 27).
Finite Te4+ oxyanions [TemXn] with m ≥ 4
Na2Te2O5 · 2H2O ≡ Na4[Te4O10] · 4H2O (#250) has linear [Te4O10]4– tetramers with a central shared edge (Fig. 8j). These lie in layers || (001) with their long axes all || . Water molecules and CN5–6 Na+ ions lie between the layers. The analogous NH4+ compound (#251) has Te4O10 groups lining up nose-to-tail to form rods || , which pack in a hexagonal array, with NH4+ ions and H2O molecules between them. Sc2Te5O13 ≡ Sc4[Te4O10](Te3O8)2 (#252) has ScO6 and ScO7 polyhedra sharing alternately corners and edges to form zigzag chains || x. Linear Te4O11 groups pointing ||  connect the Sc chains to make double layers || (001), while C-shaped Te3O8 groups wrap around individual Sc chains. The double layers contain slot-like micelles || x which accommodate most of the Te lone pairs; layers are held together through long Te…O bonds. Ca2[Cu(Te4O10)]Cl2 (#253) has Te4O10 groups pointing || , sharing corners with CuO4 squares to make a continuous layer || (01). The Ca2+ ions form zigzag edge-sharing chains of CaO7 polyhedra || x between the Cu‒Te layers. Cl– ions also lie in the interlayer gap, and complete a very elongated octahedron around Cu2+ (Cu–Cl = 2.74 Å). Sm2Mn2+Te5O13Cl2 ≡ Sm4[Mn2(Te4O10)(Te3O8)2]Cl4 and isostructural Dy4[Cu2(Te4O10)(Te3O8)2]Br4 (#254–255) have Te anions pointing ||  and arranged in layers || (201); each layer contains both Te4O10 and Te3O8 groups. The anions are held together by (Mn,Cu)O6 octahedra between the layers, which are considered part of the structural unit, and edge-sharing ribbons of (Sm,Dy)O8 polyhedra, which are not. A very open framework results, in which there remain elongated channels || y which accommodate (Cl,Br)– anions, which are weakly bonded to Te. Unusually, the CuO6 octahedron is not much less regular than MnO6 (Cu–O = 1.99–2.32 Å, vs. Mn–O = 2.12–2.28 Å).
The compounds A2[Te4O11] (A = La–Nd, Sm–Lu and Y; #256–270) show the progressive, monotonic change in parameters and properties with atomic number that is characteristic of the lanthanide elements. In particular, the fourth-shortest Te2–O distance changes so much along the series, from 2.515 Å for the La compound to 2.354 Å for Lu, that the classification of the Te anion(s) changes if the usual ‘strong bond’ threshold of 2.40–2.45 Å is used. For such a threshold, the compounds of larger A cations such as La would be regarded as having dimeric [Te2O5] + two monomeric TeO3 groups. In order to keep this family of isostructural compounds together, a threshold of 2.53 Å is used for all of them, and they all classify as having corner-linked tetramers [Te4O11]6– (Fig. 8k), with the long Te–O bonds linking the terminal Te1 atoms to the core Te2–Te2 dimer. The third-shortest Te2–O distance increases slightly from 1.989 to 2.022 Å as the A cation decreases in size from La to Lu. The structure is relatively simple, with layers || (002) of edge-sharing AO8 polyhedra linked through linear Te4O11 groups that all point ||  (Fig. 27). Ba2[Cu2(Te4O11)]Br2 (#271) has layers in which Te4O11 groups with long axes || y share corners with CuO4 squares to make sheets with Cu‒Te‒Te 3-rings, Cu‒Te‒Cu‒Te 4-rings and Cu‒Te‒Te‒Cu‒Te‒Te 6-rings. The sheets are parallel to (001) but have a polarity in the z direction; a pair of such sheets occurs back-to-back in every c repeat. BaO10 polyhedra occur in the interlayer gaps that that are faced by Cu, while Br– anions occur in the gap that is lined by Te atoms (Fig. 28). A very similar Cu–Te sheet occurs in Ba4[Cu22+(Te4O11)]2(Cu41+Cl8) (#272). The sheets again occur in back-to-back pairs, but this time they are parallel to (010), with Te4O11 groups trending || . Again, the Cu2+ ions face an interlayer unit of BaO10 polyhedra. However, the Te side of the sheets faces a wide interlayer space that contains [Cu41+Cl8]4– clusters. These consist of a central pair of edge-sharing CuCl4 tetrahedra, linked through corners to two CuOCl2 triangles. The long axis of the cluster is || , approximately perpendicular to that of the Te4O11 groups. The terminal Cu1+ atoms share the middle oxygen atoms of the Te‒O anions, thus providing additional bridges between them, but are not counted as part of the structural unit due to the low valence of Cu. [Co52+(Te4O11)Cl4] (#273) has thick layers || (100) of corner- and edge-sharing CoO4Cl pyramids, CoO5Cl trigonal prisms and octahedra, and cis-CoO4Cl2 octahedra. The layers are braced by Te4O11 groups with their long axes || , and are held together though long Te…O and Te…Cl bonds.
[Cu4(Te5O12)Cl4] (#274) has an elegant tetragonal structure in which the tellurite anion is the pinwheel-like Te5O12 pentamer of Fig. 8l. CuO4Cl square pyramids occur in clusters of four, where the oxygen atoms of the cube-shaped Cu4O4 cluster core are tetrahedrally coordinated by 3 Cu + 1 Te, thus linking Te and Cu polyhedra into an open framework. The Cl– anions are in channels || z, and make additional long bonds to Cu and Te (Fig. 28). Nd5MoTe7O23Cl3 ≡ Nd5(MoO4)[Te5O13](TeO3)2Cl3, its W analogue and the corresponding Pr analogues (#275–278) have edge-sharing columns of (Nd,Pr)O7–8 polyhedra || x, which share a few additional edges to make pillared double layers || (001) with very large channels (14 × 9 Å) running between the sublayers || x. These are further linked through (Mo,W)O4 tetrahedra into a very open framework, leaving small interlayer channels containing monomeric TeO3 groups, while the large channels are lined by V-shaped Te5O13 pentamers (Fig. 8m). The Cl– anions are held loosely, in the centres of large and small channels. Nb2Te4O13 ≡ [Nb8O8(Te10O26)(TeO3)6] (#279) has zweier chains of corner-sharing NbO6 octahedra || x. The pseudomonoclinic unit cell contains four distinct such chains, but despite the apparent simplicity of this arrangement, these are connected together through a remarkably complex arrangement of Te atoms. There are eight distinct Te sites. Sites Te2, Te3 and Te5 are in monomeric TeO3 groups that share corners with Nb to make 4-cation-wide ribbons || x with Nb‒Nb‒Te 3-rings (alternately on each side of the Nb‒O backbones, similar to the chain of Fig. 15c), Nb‒Te‒Nb‒Te 4-rings and Nb‒Nb‒Te‒Nb‒Nb‒Nb‒Te 7-rings (Fig. 28). The ribbons lie in layers || (01), which are then bridged by the open-branched decameric anion [Te10O26]12– of Fig. 8o, in which a central edge-sharing pair of Te8 are linked through Te6 to Te1, which is then connected to both Te4 and Te7. The long axis of the anion Te7…Te8 = Te8…Te7 is approximately ||  in the interlayer gap, while the Te4 branches complete the Nb‒Te layers || (01). Note that this description is valid if the strong bonding threshold for Te is chosen in the range 2.36–2.44 Å; however, two slightly longer Te–O bonds (2.44–2.47 Å) connect the finite decamers into infinite open-branched achter chains, in which the additional bridging oxygens are CN3, shared by Te7 of two different decamers and Te2 as a new branch. (NH4)6Mo8Te8O43 · H2O ≡ (NH4)6[(MoO3)6(Te6O12)(TeO3)2](Mo2O7) · H2O (#280) has the only true cyclo Te4+ complex of the present study. Hexagonal rings [Te6O12]0 (Fig. 8n) and monomeric [TeO3]2– anions are linked into a layer || (001) through edge-sharing dimers of MoO6 octahedra. Additional corner-sharing dimers of MoO4 tetrahedra [Mo2O7]2–, with point symmetry –3, lie in the centres of the hexagonal rings. The NH4+ cations and water molecules lie in the interlayer gap (Fig. 28).
Structures with Te4+ complexes that are infinite chains
Structures #281–334 are inotellurites (Table 16, deposited). The various topological types of chain are summarized in Tables 2 (single chains) and 3 (multiple chains), and depicted in Fig. 9. Our first example is the structure of a high-temperature phase of Mo5TeO16 ≡ [(Mo25+Mo36+)O13(TeO3)] (#281). As in the polymorph #168, MoO6 octahedra share four corners to form layers of 3-, 4- and 6-rings || (010), which then link through the remaining octahedral vertices to form a framework, and there is no ordering of Mo5+ and Mo6+. As before, Te is in the centre of the hexagonal ring, but this time it is 4-coordinated, and shares oxygen atoms with Te of layers above and below to form a corner-sharing chain || y. In the long-range average crystal structure the chain is einer, but the bridging oxygen is on a twofold split site, which implies that the Te–O–Te configuration is nonlinear and the true local periodicity of a chain is at least zweier. AgTeO2(NO3) ≡ Ag2[Te2O4](NO3)2 (#282) has TeO4 polyhedra of the type shown in Fig. 4c, sharing edges to make an electrostatically neutral zigzag chain [Te2O4]0 || z (Fig. 9a). These lie in layers || (020), with interstitial NO3– and CN6 Ag+ ions, which themselves form a weakly bonded layer in which Ag+ and NO3– are arranged respectively like the Pb2+ and O2– in the litharge form of PbO (Boher et al., 1985). Telluroperite, Pb3TeO4Cl2, is more precisely written as Pb2[(Pb0.5Te0.5)2O4]Cl2 (#283). The Te–O complex is a chain topologically similar to Fig. 9a, but with square-pyramidal polyhedra (Fig. 4b) and a net negative charge due to random substitution of 50% of the Te4+ by Pb2+. The (Pb,Te)–O chains and the remaining Pb2+ cations form litharge-type layers || (002), with Cl– anions between the layers (Fig. 29). The resulting structure is isostructural with perite, Pb2[Bi23+O4]Cl2, and its Sb3+ analogue nadorite (Kampf et al., 2010f). Rajite, [Cu(Te2O5)], has the corner-sharing chain of alternating CN3 and CN4 Te4+ seen in Fig. 9b (#284). The chains run || x, and are made from finite dimers by the rather long fourth bond to Te2 (2.30 Å). Chains are cross-linked into a framework through CuO4 squares, which do not link to each other unless a fifth Cu‒O distance (also 2.30 Å) is included, in which case the resulting CuO5 pyramids form edge-sharing dimers. Nd[Te2O5]Br (#285) has layers || (002) of edge-sharing NdO8–12 polyhedra, braced above and below by [Te2O5]2– chains || y which are topologically similar to those of rajite, but have an unusual planar T-shaped geometry for the CN3 polyhedron. This may be due to the presence of four interlayer Br– ions at a relatively short distance from the Te cation (3.35–3.47 Å). The structures of the Cl analogue (#336) and the Ho analogue of that compound (#307) are different (see below). Nd2[Te2O5](Te2O5)(MoO4) and its Pr analogue (#286–287) have edge-sharing zigzag double chains of (Nd,Pr)O8–9 || x, which are bridged via MoO4 tetrahedra to form looped layers || (001). Dimeric [Te2O5]2– ions (Fig. 8a) brace the layers, while zweier chains of the same composition (Fig. 9b) run || x, and hold the layers together. The chains are the structural unit, for the purposes of this classification. In [Ga2(Te2O5)(Te2O6)] (#288), the Te2O5 chains are || . These, and dimeric groups [Te2O6]4–, which consist of two TeO3 units linked by a long (2.35 Å) fourth bond to Te4, cross-link otherwise isolated GaO5 and GaO4-polyhedra into a framework. [InF(Te2O5)] (#289) has trans-InO4F2 octahedra sharing F atoms to form a helical vierer chain || z. These are cross-linked into a framework by Te2O5 chains, which occur in four layers per unit cell along the z direction, and run ||  and ||  in alternate layers.
La4Ta2Te6O23 ≡ La4[Ta2O6(Te2O5.4)(TeO3)2][Te2O5.6] (#290) has a somewhat disordered structure with three fundamentally distinct types of Te. One type (Te3) approximates a chain of the zweier CN4 type [Te2O6]4– (Fig. 9c) running || z. However, the bridging oxygen atoms are only 77–80% occupied. Another type (Te1+Te2) ostensibly forms a similar chain, but the Te atoms are actually on a twofold split site and again the bridging oxygen sites are only partly occupied (71%), implying that the Te atoms are locally in a very strongly asymmetrical 3+1 coordination environment, and that the ‘chain’ is actually a sequence of Te2O5 dimers (58%) and orientationally disordered TeO3 monomers (42%). The Te atoms of the second chain share corners with a parallel zweier chain of corner-sharing TaO6 octahedra, which are decorated by additional TeO3 groups (Te4) to form a second one-dimensional structural unit. All the Te lone pairs point in towards the centres of large rhomb-shaped channels || z of a trellis-like matrix of edge-sharing LaO9 polyhedra (Fig. 29). K[(UO2)(Te2O5(OH))] (#291) has chains || x of Fig. 9c type, which share corners with UO6 octahedra to form pleated layers || (020). The layers have Te‒Te‒U 3-rings and Te‒Te‒U‒Te‒Te‒U 6-rings. The layers are held together by CN10 K+ ions, which sit at the centres of the hexagonal rings.There are two types of bridging oxygen atom in the Te chain, with different Te–O distances of 2.07 and 2.28 Å; the latter oxygen atom probably accommodates the H+. Schmitterite, UO2TeO3 ≡ [(UO2)2(Te2O6)] (#292), also has its TeO3 groups linked into a [Te2O6]4– chain (|| z), and these are linked through U into a pleated sheet || (010). However, the U6+ forms edge-sharing chains of UO7 polyhedra, all Te–U links are through CN3 oxygen atoms bonding to either 2Te + U or to Te + 2U, and layers are held together only through long Te…O bonds (Fig. 29). Tl3[(UO2)2(Te2O5(OH))(Te2O6)] · 2H2O (#293) has two symmetrically distinct Te2X6 chains running || z; as in #291, the hydrogen atom can be located by noting the unusually long Te2–O7 distance: 2.21 Å as opposed to 2.04–2.07 Å for the other bridging oxygen atoms. Isolated UO7 polyhedra link the Te chains into corrugated layers || (040), held together through three kinds of Tl+ ion (CN = 5–9) and water molecules in the interlayers (Fig. 29).
InNbTe2O8 ≡ In2[NbO2(TeO3)]2[Te2O6] (#294) has NbO6 octahedra sharing four corners to form a square-net layer || (010), resembling the (Nb,Ta)nO4n layers of #155–158. Half of the Te are in monomeric TeO3 groups that share additional corners to form Nb‒Nb‒Te 3-rings that are alternately above and below the Nb layer along the x direction, while the other half of the Te form zweier chains Te2O6 || z that do not connect directly to Nb, but brace a layer of edge-sharing InO7 polyhedra that holds together the Nb‒Te layers. Thus, there are two separate structural units, a Te chain and a Nb‒Te heteropoly layer (Fig. 29). The related compound BiNbTe2O8 ≡ Bi2[NbO2(TeO3)]2[Te2O6] (#295) also has corner-sharing NbnO4n sheets, but while the unshared corners are trans in #294 so that the sheets are planar, in the Bi compound the unshared corners are cis, and the sheets || (002) are strongly pleated. Again, half of the TeO3 link to the Nb sheets, completing Nb‒Nb‒Nb‒Te 4-rings rather than 3-rings, this time, while the rest of the Te form Te2O6 chains between the layers, running || y. The BiO4+3 polyhedra share edges to form a corrugated layer that is braced by the Te chains and also links to the Te of the Nb‒Te units. Cs3[Nb9O20(Te2O6)(TeO3)2] (#296) has very thick layers of NbO6 octahedra || (200) that are slices of a pyrochlore framework. Kagome layers of 3- and 6-rings || (011) and (01) are very prominent in projection down z. The pyrochlore framework is displaced by c on planes || (200), which breaks otherwise infinite chains of octahedra running || x after only four Nb atoms. These chains are terminated by half of the Te rather than Nb, and the offset of the pyrochlore framework allows Te to retain 3-fold rather than 6-fold coordination. Thus, the layers are not condensed any further by strong bonds. The other half of the Te atoms form [Te2O6] chains which run || z in notches in the sides of the thick layers. The Cs+ ions are in 6–8 coordination in large interstices within the layers, which are held together only through long-distance Te…O interactions (Fig. 29). CuTeO3 ≡ [Cu2(Te2O6)] (#297) is a polymorph of balyakinite (#117) and the high-pressure phase #142. In this structure, CuO4 squares form edge-sharing dimers, condensed into chains || x through a fifth Cu–O bond at 2.43 Å. The Cu chains lie in layers || (020), which are cross-linked into a framework by Te2O6 chains running || . Like Cu, Te has a very irregular coordination environment, with three Te–O distances of 1.87–1.92 and the fourth at 2.43 Å; it has no other oxygen atoms within 2.8 Å. TlV5+TeO5 ≡ Tl2[(VO2)2(Te2O6)] (#298) has trans corner sharing chains of distorted VO6 octahedra in a centred-rectangular array, running || x. These are cross-linked by Te2O6 chains running || z into a framework with large channels || x and z. The resulting three-dimensional net has intersecting layers of kagome topology || (011) and (01), like the weberite structure (Knop et al., 1982, Fig. 14g, #672 below). In fact, the framework can be derived from that of weberite by a displacement of slices || (002), thus condensing into chains what would otherwise be isolated Te polyhedra, and adjusting the Te‒O‒V bonding pattern slightly. In #298, oxygen atoms are removed so that Te has CN = 4 rather than 6, while the environment of V is changed from Q0600 to Q1411, with V making a bond to a CN3 bridging oxygen of the Te chain rather than to the missing anion. The V–O–V bridges are very asymmetrical, one distance being 2.24 Å while the other is 1.66 Å, almost identical to the distance between V and the CN1 ligand. Thus, nonlinear [VO2]+ complexes can be recognized in the structure. The Tl+ ions are in very one-sided 7-coordination in the channels.
Polymorph II of TeO(As5+O3OH) ≡ [(AsOH)2(Te2O8)] (#299), like polymorph I (#225 above), has Te in 5-fold coordination, but the TeO5 polyhedra now form corner-sharing zweier chains Te2O8 || y (Fig. 9d) rather than edge-sharing dimers. [AsO3OH]2– tetrahedra share the three unprotonated ligands with the Te chains to form a corrugated layer || (002) with Te‒Te‒As 3-rings and Te‒Te‒As‒Te‒As 5-rings. Hydrogen bonds brace individual layers, which are held together through long Te···O bonds. The triclinic polymorph of TeO(Se4+O3) ≡ [Se2(Te2O8)] (#300) has CN5 Te in similar zweier Te2O8 chains || x. The chain backbones have the asymmetrical crankshaft geometry of the Pb‒O chains in massicot (Hill, 1985). SeO3 pyramids share all corners with the Te chains to make a framework with hexagonal channels || y, which accommodate the lone pairs of both Se and Te. The monoclinic (pseudo-orthorhombic) polymorph #301 is topologically very similar, but slight displacements of atoms mean that the two Te atoms of the chain repeat are now related to each other through a glide plane, rather than being symmetrically independent.
Chekhovichite, Bi2Te4O11 ≡ Bi4[Te4O10](TeO3)4 (#302) has the zigzagging vierer Te chain of Fig. 9e, with CN3 Te at the sharp bends. The chains trend || y, and zigzag in layers || (002). BiO7–8 polyhedra share edges to form corrugated sheets with a honeycomb net, between the Te chains, and the monomeric TeO3 groups brace the Bi sheets. Na2MoTe4O12 ≡ Na2[MoO2(Te4O10)], its W analogue and the corresponding Ag‒Mo compound (#303–305) have topologically similar but more contorted vierer Te chains, which are cross-linked in pairs through otherwise isolated (Mo,W)O6 octahedra to make looped heteropoly chains || z. These chains lie in layers || (200), which are loosely held together through CN7 (Na,Ag)+ ions. The Te bond threshold for these structures has been set at 2.38 Å. If an additional bond at 2.39–2.48 Å is included, then Te2 atoms of neighbouring chains share edges to join the chain structural units into a continuous layer with 10-rings || (200). The compound Fe8Te12O32Cl3Br3 ≡ [(Fe22+Fe63+)(Te4O10)(Te2O5)2(TeO3)4]Cl3Br3 (#306) has four types of Fe atom in 5 or 6 coordination, which link in a honeycomb net to form layers || (004). The different linkage patterns for Fe1–Fe4 polyhedra are described concisely by Q states as Q0603, Q1502, Q1501 and Q2300, respectively. Bond-valence sums indicate that Fe2+ is ordered at Fe3. Two types of interlayer gap alternate between Fe layers. One gap type contains two-thirds of the loosely bound, partly-disordered (Cl, Br)– ions, and is lined by Te2O5 groups which brace the Fe layers. Monomeric TeO3 groups decorate the other side of the Fe layers, facing the other type of gap, but on this side, the layers also link to Te4O10 chains running || y, which join the Fe layers in pairs and thus complete the extremely complex pillared double layer structural unit (Fig. 30). The remaining (Cl, Br)– are in channels || y between the Te chains. The chains still have alternation of CN3 and CN4 Te, but are quite different in conformation from those of #302–305: they are helically coiled, with the sharpest bending at the CN4 Te atoms rather than CN3.
In HoTe2O5Cl ≡ Ho2[Te4O10]Cl2 (#307), nearly-cubic HoO8 polyhedra share edges to make square-net layers || (001); half of the Ho atoms have interlayer Cl as a more distant neighbour. The Ho layers are braced by vierer Te chains which are a topological isomer of those in #302–306, with a different sequence of CN3 and CN4 Te atoms (Table 2, Fig. 9f). The chains trend || . Layers are held together only through weak Te…Cl interactions. The compositionally similar Nd compounds #285 and #336 have quite different structures. Na2Te2O5 ≡ Na4[Te4O10] (#308) has yet another isomer of Te4O10 chain, with all Te 4-coordinated, but an alternation of corner- and edge-sharing (Fig. 9g). Chains trend || , and are held together by CN6–7 Na+ ions. The arrangement of (Na,Te) and O atoms can be described as slightly distorted slabs || (100) of rocksalt type, alternating with fluorite-like slabs, with there being two of each slab type per unit cell. K2Te2O5 · 3H2O ≡ K4[Te4O10] · 6H2O (#309) has similar vierer chains which lie in layers || (002). The chains run in  and  directions in alternate layers. Water molecules and CN7–8 K+ ions lie between the layers. The edge-sharing Te4O10 chains of Fig. 9g are also found in the denningite structure of Mn2+Te2O5 ≡ Mn2+[Mn2+(Te4O10)] (#310). This nanoporous tetragonal structure has two different types of Mn site, with coordination numbers 8 and 6 respectively, which alternate in edge-sharing chains || z. The Mn chains are linked to their neighbours by Te chains which also run || z, so delineate square channels with a minimum diameter of ∼5 Å (Fig. 30). Note that the CN6 Mn are considered to be part of the overall framework structural unit, while CN8 Mn is not. The CN6 Mn2+ may be substituted by Cu2+ to produce solid solutions and ultimately the end-member Mn2+[Cu2+(Te4O10)] (#311). Although Mn-dominant synthetic compositions have been called ‘denningite’, it seems likely that the mineral denningite itself has Ca replacing Mn in the 8-coordinated cation site. The type material has composition (Ca0.60Mn0.40)[(Mn0.72Zn0.24Mg0.04)(Te4O10)] (Mandarino et al., 1963), and the ideal composition is given as CaMn2+Te44+O10 in the IMA list of minerals (http://ima-cnmnc.nrm.se/IMA_Master_List_2015-05.pdf). Although the refinements of Walitizi (1964, 1965) constrained occupancies to be the same on CN8 and CN6 sites, Ca would be expected to partition strongly into the larger 8-coordinated site, but the structure needs to be reinvestigated to confirm this. A different polymorph of Mn[Te2O5] is described as #347.
Despite the apparently simple formula, K2Te2O5 (#312) has an open-branched zweier Te4O10 chain (Fig. 9h). The chains have a V-shaped cross-section, run || x and pack in a herringbone fashion. They are cross-linked by undulating layers || (001) of KO7–8 polyhedra. Ba2V5+Te4O12(OH) ≡ Ba2[VO3(Te4O9(OH))] (#313) contains yet another Te4X10 isomer, the open-branched dreier tellurite chain of Fig. 9i. The hydroxide group is located on the CN4 Te atom that does not have the branch (Te4), while a VO4 tetrahedron shares an oxygen atom with the branch Te atom (Te1). Chains run || y and lie in double layers || (10), with the branches directed into the interiors of the double layers. BaO8–9 polyhedra lie between the double layers. (NH4)2WTe2O8 ≡ (NH4)4[(WO2)2(Te4O12)] (#314) has zweier Te chains with branches attached to CN3 bridging oxygen atoms (Fig. 9j). The chains run || y, and are cross-linked by individual WO6 octahedra, which share four corners to form layers || (100) with Te‒Te‒W 3-rings and Te‒Te‒W‒Te‒W 5-rings. NH4+ ions are between the layers. NiTe2O5 ≡ [Ni4(Te6O14)(TeO3)2] (#315) has loop-branched vierer chains Te6X14, with CN5 Te at the nodes and CN3 Te in the loops (Fig. 9k). The chains run || x, and alternate with monomeric TeO3 groups in layers || (002). The tellurite anions are cross-linked into a framework through chains of trans edge-sharing NiO6 octahedra which run || y; micelles between these chains accommodate the Te lone pairs.
Pb2Te3O8 ≡ Pb8[Te6O16](Te3O8)2 (#316) has both zigzag sechser chains Te6O16 of the type shown in Fig. 9l, and the soro Te3O8 groups of Fig. 8g. Note that the chains contain both CN3 and CN4 Te, with the sharp bends at the CN3 Te atoms. The chains and tritellurite long axes all point || x. Chains lie in double layers with Te3O8 groups between the layers, and the resulting thick sheets repeat || (002). Tellurite anions within the thick sheets are held together by Pb2+ ions that lie on the outsides of the sheets, and sheets are held together via long Pb…O bonds. Lead is in distorted 7‒8 coordination with three short bonds to oxygen atoms (Fig. 30). The compounds M22+Te3O8 ≡ [M4(Te6O16)] with M = Mg, Mn, Co, Ni, Cu or Zn (#317‒322) have the spiroffite structure. No structure refinement exists for an Fe compound with this stoichiometry, while those with M = Mn and Zn are the minerals spiroffite and zincospiroffite, respectively. The MO6 octahedra share edges and corners to make honeycomb layers || (200), and are linked into a framework by a different Te6X16 sechser chain in which all Te are CN4, but edge-sharing pairs alternate with a polyhedron that shares only corners, at which the chain bends (Fig. 9m). The chains run ||  and lie in layers || (202). The trellis-like intersection pattern of M and Te layers results in 5 Å diameter channels running || y and z, which accommodate the Te lone pairs (Fig. 30).
Fe3Te4O12 ≡ [Fe22+Fe43+(Te6O18)(TeO3)2] (#323) has edge-sharing dimers of Fe3+O6 octahedra (Fe1 = Fe3, Fe‒O = 1.94‒2.15 Å) and edge-sharing dimers of Fe2+O6 octahedra (Fe2 = Fe2, Fe‒O = 2.05‒2.48 Å), which share CN3 oxygen atoms to make layers || (100). The layers are braced by TeO3 monomers and linked into a framework through open-branched vierer chains Te6O18 (Fig. 9n). The chains run || y, but zigzag || , obliquely to the Fe layers. Te3O3(PO4)2 ≡ [P4(Te6O22)] (#324) also has an open-branched vierer chain, but with CN4 and CN5 Te atoms (Fig. 9o). The chains run || z, with branches extended in the y direction, and lie in layers || (200). Two types of PO4 tetrahedra share all corners with the Te chains to make a framework which, unusually, has no 3-rings, but does have Te‒Te‒Te‒P and Te‒P‒Te‒P 4-rings and Te‒Te‒P‒Te‒P 5-rings. Ca4Te5O14 ≡ Ca8[Te8O22](TeO3)2 (#325) has open-branched sechser chains (Fig. 9p) running || y and zigzagging in the x direction, and lying with TeO3 monomers in layers || (004). The chains wind through a trellis-like framework of CaO7 polyhedra, which has 7 Å channels || z accommodating the chain branches.
The next few structures have multiple chains (Table 3). Interestingly, two of them are compounds with Fe3+, and two are complex compounds with Cd and Cl. Bi3Te4O10Cl5 ≡ Bi3[Te2O4](TeO3)2Cl5 (#326) has the simple einer [Te2O4]0 ribbon of Fig. 10a, in which Q1032 Te atoms are linked through CN3 oxygen atoms. These run || y, and with TeO3 monomers and BiO4Cl2 and BiO4Cl5 polyhedra, define layers || (201), with the Cl– ions in the interlayers. Although the Te double chain is nominally neutral, in reality, Te1 makes weak bonds to Cl–, and the non-bridging oxygen is shared with Bi1. Cd7Te7O17Cl8 ≡ Cd7[Te5O12](Te2O5)Cl8 (#327) has a commensurately modulated structure in which seven types of Cd polyhedra (variously CdO3Cl3, CdO4Cl3 or CdO3Cl4) form edge-sharing ribbons, flattened in the yz plane and running || y. These ribbons are braced on one side by the rather complex zweier double chain Te5O12 of Fig. 10b and also by the relatively simple Te2O5 single chain of Fig. 9b. The Cd‒Te compound layers repeat || (200), with Cl– ions in the interlayer and long Te…Cl bonds holding layers together. However, the facing direction of the pair alternates back and forth along the 28 Å c repeat (Fig. 30).
Fe3+Te3O7Cl ≡ [Fe2(Te6O14)]Cl2 (#328) and the Br analogue (#329) have dreier double chains Te6O14 made out of 5-rings (Fig. 10c), flattended on (10) and running || y. The chains are linked through edge-sharing dimers of FeO5 polyhedra to make very thick double layers || (100), which contain channels || z. Halide ions lie in the interlayer spaces, and the layers are held together only through weak Te…(Cl, Br) and Te…O interactions. Te2 in this structure has coordination that is strongly 3+1: the third- and fourth-nearest oxygen atoms are at ∼1.95 and 2.42 Å. If the fourth ligand is not included, the double chain becomes a serpentine sechser single chain. Tl2Te3O7 ≡ Tl4[Te6O14] (#330) has a very different Te6O14 double chain which is only zweier but has the subchains linked through edge-sharing Te dimers (Fig. 10d). Thus, the subchains are open-branched but are linked via the branches so that the double chain as a whole is not. The resulting ribbons are flattened on approximately (01) and run || x. The Te chains interpenetrate with and cross-link layers || (010) of edge-sharing TlO5 polyhedra (Fig. 30). Fe23+Te4O11 ≡ [Fe4(Te6O16)(Te2O6)] (#331) has a Te6O16 chain whose description is again dependent on the Te–O bonding threshold. Here, we include a fourth Te2–O link at 2.496 Å, which makes a rather simple zweier double chain of 6-rings (Fig. 10e). Without that link to close the rings, the chain is an open-branched vierer single chain. The chains are flattened on (102) and run || y. The chain is topologically the same as a slice of the aluminosilicate sheet of prehnite, Ca2[Al(AlSi3O10)(OH)2], which has a similar alternation of Q2 and Q4 polyhedra in its 6-rings (Papike and Zoltai, 1967). FeO6 and FeO5 polyhedra share a corner to form Fe2O10 dimers, which cross-link the Te chains, leaving channels || y which accommodate additional Te in edge-sharing dimers Te2O6 (Fig. 8c). Na2Te4O9 ≡ Na4[Te8O18] (#332) has a very complex zweier double chain in which the two subchain backbones are linked through the familiar ‘double-triangle’ tetrameric clusters (Fig. 10f). These chains run || y, and their parallelogram-shaped cross-section defines the geometry of the unit cell. The chains are held together through layers || (100) of NaO5–6 polyhedra. The K and NH4 analogues have a quite different layer structure (#354–355).
We have two instances of chains with multiplicity greater than two. Te4O5(PO4)2 ≡ [P2(PO)2(Te8O24)] (#333) has the open-branched zweier triple chain shown in Fig. 10g. As noted earlier, the central ribbon of 6-rings resembles that of ‘biopyribole’ silicate minerals such as jimthompsonite, (Mg,Fe)5(Si6O16)(OH)2 (Veblen and Burnham, 1978), except for the increased coordination number of the Q2300 polyhedra. The Te ribbons are run || y, and are connected to form strongly pleated layers || (201) by PO4 tetrahedra, half of which share all four corners and half of which share only three. Cd4Te6O13Cl6 ≡ Cd2[Te6O13](Cd2Cl6) (#334) contains the extraordinary quadruple chains of Fig. 10h. The outer two backbones are zweier, and contain edge-sharing Te trimers surrounding a CN3 oxygen atom, similar to those of Fig. 8i. Conversely, the central backbones are dreier, and are linked through shared ‘double-triangle’ clusters. The resulting ribbons are flattened on (012) and run || x. They stack en échelon, and are connected into layers || (001) through edge-sharing ribbons of CdO7 and CdO8 polyhedra. Between these layers, CdCl6 octahedra form edge-sharing ribbons, and the structure is held together by weak Te…Cl bonds.
Structures with Te4+ complexes that are infinite sheets
Table 17 (deposited) lists phyllotellurites #335–363, whose various types of Te–O sheet are summarized in Table 4 and shown in Fig. 11 (single layers) or Table 5 and Fig. 12 (double layers). Bi10Te2O17Br4 ≡ Bi10[TeO2]2O13Br4 (#335) has the simple, electrostatically neutral TeO2 layer of Fig. 11a, with square-pyramidal TeO4. It is a pseudotetragonal with a ≈ b ≈ 4 Å. The Te pyramids share edges with BiO8 cubes which have BiO4 pyramids on the other side. The composite Bi2TeO4 layers || (001) that are thus formed can be regarded as slices of a fluorite-like structure. These layers alternate along z with topologically similar but Te-free Bi3O4 layers. In both cases, an additional ‘interstitial’ anion site, in an ‘octahedral’ interstice of the cubic close-packed cation slab, is 25% occupied by O2– to give a ‘stuffed fluorite slab’ stoichiometry (Bi,Te)3O4(O0.25). These interstitial oxygen atoms are weakly bonded to Bi and Te, but are too far from Te (2.51 Å) to be included in the Te–O complex of our classification. It should be noted that short O…O distances of 2.38 Å suggest that accommodation of the additional oxygen requires adjustment to other parts of the structure. Br– anions lie between the layers, which are held together through weak Bi…Br and Te…Br interactions (Fig. 31). NdTe2O5Cl ≡ Nd[TeO2]2OCl (#336) is structurally very similar, although it is truly tetragonal, has all stuffed-fluorite NdTe2O4(O) layers equivalent, and thus has a halved c repeat. The similarity between these compounds is best seen if the formulae are written as [(Bi2Te)O4(O0.25)][Bi3O4(O0.25)]Br2 and [(NdTe2)O4(O)][(NdTe2)O4(O)]Cl2. The fully-occupied interstitial oxygen site has 8 × O at 2.58 Å, 4 × Nd atoms at 2.86 Å and 2 × Te atoms at 2.42 Å. If these oxygen atoms were included as Te ligands, then the Te coordination would increase from square-pyramidal CN4 to CN5 (Fig. 4b,e), and the Te–O complex would be not the single [TeO2]0 layer of Fig. 11a, but a double layer [Te2O5]2–, in which sublayers of the type shown in Fig. 11c are linked by corner-sharing of the additional oxygen atoms. Note the very different chain structures exhibited by the Ho analogue (#307) and the Br analogue (#285). Compound #337 approximates Bi[Te2O5]Cl, but is more accurately written Bi0.87[Te2O4.9]Cl0.87. Again, it has a strongly layered structure in which Bi atoms are at the core of the layers, Te on the outsides, and weak Te…Cl bonds are holding the layers together. However, layers have trigonal rather than (pseudo)tetragonal symmetry, and there is very extensive long-range disorder. Tellurium has one apical ligand O1 at 2.02 Å, but other oxygen sites at 3 × 2.14 Å (O3), 3 × 2.37 Å (O3) and 3 × 2.41 Å (O2). The O2 and O3 sites are only 73% and 24% occupied, respectively, and short O…O distances imply that there must be considerable short-range order. The O3 sites occur in triangles with O3…O3 = 1.27 Å, so only one position out of each triplet can be occupied, and each O2 site has three O3 at 1.70 Å, so either O2 or one of those O3 positions can be occupied. The simplest and most symmetrical occupancy pattern which satisfies these constraints and approximates the refined occupies of the average structure is shown in Fig. 11b. Each Te atom has ligands which are 1 × O1 plus either 3 × O2 (¾ of the time) or 2 × O2 and 2 × O3 (¼ of the time). The resulting layer has 3-rings of Q1030 and Q1220 Te polyhedra with the geometry of Fig. 4d and a distorted variant of Fig. 4e respectively, an ideal stoichiometry Te2O5, and a within-layer repeat that is a 2 × 2 superstructure of the crystallographic unit cell. The long-range disorder reflects at least stacking disorder of that ordering pattern, possibly with an admixture of other short-range ordering schemes. Compound #338 is another Bi tellurite halide with an ostensibly simple stoichiometry concealing structural complexity (Fig. 31). It approximates BiTeO3Br, is more accurately represented by the structural formula Bi2[TeO3][TeO2]OBr2, which if partial occupancies are indicated becomes in turn Bi1.93[TeO3][TeO2]OBr1.8. Like #336, it is a tetragonal structure with a ≈ 4 Å. There is a stuffed-fluorite compound layer (BiTe2)O4(O) similar to the (NdTe2)O4(O) layers of that structure or (BiTe2)O4(O0.25) of #335. Again, there is a fully-occupied ‘interstitial’ oxygen site, but this is now at 2.87 Å from 4 × Bi, 2.59 Å from one Te and only 1.88 Å from the other Te atom. Thus, individual layers are polar, the Te–O complex on one side being the neutral layer [TeO2]0 with CN4 Te (Fig. 11a), while that on the other side is anionic [TeO3]2– with CN5 Te (Fig. 11c). The full stacking sequence includes two such fluorite-like Te1‒Bi‒Te2 slabs of opposing polarity, and also a separate Bi2O2 layer which is thinner, with a geometry more obviously similar to the litharge form of PbO (Boher et al., 1985). Indeed, the overall structure has the same P4/nmm space group as litharge. Bromium forms double layers between the two fluorite-like slabs and also single layers between fluorite and litharge slabs; the structure is held together through weak Te…Br and Bi…Br bonds.
(Cu1+Cl2)[Sb3+TeO3] (#339) has a disordered 50 : 50 mix of Sb3+ and Te4+ cations, which are 4-coordinated and form (Te,Sb)2O3 layers of the type seen in Fig. 11d, which can be generated by condensing the double chains of Fig. 10a through additional corner-sharing. Note that in this compound, the layers are cationic, [SbTeO3]+. The layers are || (20), and between them are intercalated 3-wide ribbons running || y of CuCln polyhedra. The outer two Cu positions form well-ordered chains of corner-linked CuCl4 tetrahedra, while the central Cu atom is delocalized across triangular 3-fold and linear 2-fold coordinated positions. [Te2O3OH](NO3) (#340) has corrugated layers || (200) of the topology shown in Fig. 11e, in which Q0401 polyhedra form a 3-connected net of 6-rings despite having CN4, by virtue of sharing one edge. The Te atoms and O1, which does not participate in the shared edge, form corner-linked chains of massicot-like asymmetrical crankshaft geometry running || z (cf. Hill, 1985). Te–O distances for O1 (1.89 + 2.05 Å), O2 (2 × 1.93 Å) and O3 (2 × 2.17 Å) indicate unambiguously that the H atom is attached to O3. The resulting layers are again cationic, [Te2O3OH]+, and are held together through their electrostatic attraction to interlayer NO3– anions (Fig. 31). A layer of the Fig. 11e topology also occurs in the mineral tellurite, an orthorhomic polymorph of TeO2 (#341), where very tightly corrugated layers of this type are || (200). These layers are electrostatically neutral, and are held together only by long Te…O bonds. It should be noted that the structure is isopuntal with the brookite polymorph of TiO2, but whereas the Ti atoms of brookite have six oxygen neighbours at 1.863–2.052 Å (Meagher and Lager, 1979), the lone-pair stereoactivity of Te4+ distorts the coordination octahedron to give only four neighbours in the range 1.88–2.20 Å, an additional neighbour within the layer at 2.64 Å, and a weak Te–O bond across the interlayer gap at 3.07 Å. A variant of the structure with additional anion-anion bonding occurs for the pararammelsbergite form of NiAs2 (Fleet, 1972) and intermetallic compounds such as AuSn2 (Rodewald et al., 2006). Two other polymorphs of TeO2 are discussed as #364–365 below. At present, there do not appear to be examples of more complex structures that contain uncharged Te4+−O sheets, although analogues are known for other p-block elements. The mineral lucabindiite, ideally K[As4O6]Cl (Garavelli et al., 2013), has planar As4O6 sheets which direct their lone pairs towards interlayers of Cl–, while the oxygens face interlayers of K+.
Bi4Te2O9Br2 ≡ Bi4[Te2O5]O4Br2 (#342) is another Bi tellurite containing fluorite-like Bi‒Te‒O slabs || (001), like #335 and #338. In this case, a and b ≈ √2 × 4 Å ≈ 5.6 Å, as the Te–O component layer contains alternating CN4 and CN5 Te atoms (Fig. 11f), and is in effect an ordered intermediate between the layers of Figs 11a and 11c. It should be noted that the CN4 Te atom is in almost square-planar coordination, which is unusual, and suggests that some atomic coordinates may be incorrect. The (Bi4Te2)O8(O) stuffed-fluorite slabs are polar, with Te on only one side. Single layers of Br– ions lie between slabs, which are linked through weak Te…Br and Bi…Br bonds.
In one form of Li2[Te2O5] (#343; the other polymorph is #351), Q1300 Te polyhedra link to form layers of 6-rings, topologically equivalent to the silicate sheets of the ‘micas’ (Fig. 11g). Resemblance to micas is further enhanced by the fact that such sheets occur in pairs, and their apical oxygen atoms are directed inwards, towards a ‘sandwich filling’ layer of electropositive cations. In this compound, however, the core of the layer is composed of tetrahedrally coordinated Li+ cations, rather than higher-valence species in octahedral coordination. The compound mica-like layers are || (10), with Te lone pairs directed into the interlayer gap and only weak Te…O bonds connecting layers (Fig. 31). (Te2O3)(SO4) ≡ [(SO2)(Te2O5)] (#344) also has phyllotellurite sheets || (010) with the topology of Fig. 11g, but the non-bridging oxygen atoms are not all on one side of the sheet. Hexagonal rings are bent in a boat configuration, and pairs of apical oxygen atoms point alternately up and down along y. All ‘up’ or ‘down’ pairs have Te–Te || x, and individual 6-rings have either four ‘up’ polyhedra and two ‘down’ or vice versa (Fig. 31). This is an analogue of the silicate sheet found in sanbornite, Ba[Si2O5] (Hesse and Liebau, 1980), rather than the mica structure type. Each pair of apical oxygen atoms is shared with a SO4 tetrahedron, thus completing the heteropoly layer structural unit. Layers are held together through long bonds between the Te of one layer and the non-bridging sulfate oxygen atoms of the next. Tilt of the SO4 groups and asymmetry of the Te–O–Te angles make the structure polar || z. The layer || (010) of (Te2O3)(PO3OH) ≡ (POOH)[Te2O5] (#345) also has pairs of ‘up’ or ‘down’ Te polyhedra, but each 6-ring has three of each, and the ‘up’ (or ‘down’) pairs are arranged in a herringbone pattern. The PO3OH tetrahedra again share two corners with adjacent Te polyhedra of a layer. Interatomic distances indicate that the H atom is attached to the non-bridging phosphate oxygen atom O7, and forms hydrogen bonds that brace the layer, rather than connecting between layers. Layers are connected by weak bonds between Te and the other non-bridging phosphate ligand O5, and the overall structure is polar || z for the same reasons as #344. The b parameters of these two structures are very similar, reflecting the similarity of layers and their stacking.
[Mg(Te2O5)] and the β polymorph of [Mn(Te2O5)] (#346–347) have layers || (020) with 6-rings of the ‘four up/two down’ type, like #344, but the 6-rings are distorted so that ‘up’ Te–Te pairs point along  while ‘down’ pairs point || . Layers are cross-linked into a framework through zigzag edge-sharing chains of MO6 octahedra (M = Mg or Mn). The octahedral chains run || z, with small channels between them that can accommodate the Te lone pairs. The ‘denningite’ polymorph of the Mn compound was discussed above (#310). MoTe2O7 ≡ [(MoO2)(Te2O5)] (#348) also has ‘four up/two down’ layers, this time || (002), but the layers are strongly pleated so that the Te–Te pairs point obliquely to the overall trend of the layer, ||  (‘down’ pairs relative to +z) or  (‘up’). Layers contain very obvious massicot-like Te‒O‒Te‒O chains || x. Tellurium layers are linked into a framework through MoO6 octahedra chains, which occur as edge-sharing Mo2O10 dimers which then share four additional corners to make ladder-like double chains || x (Fig. 31).
La2[Te3O7]2(WO4) (#349) has the unique layers of 3- and 6-rings shown in Fig. 11h; note that the node where three 6-rings join is a CN3 oxygen atom. These layers are || (002), and the non-bridging oxygen atoms of the Q1210 Te polyhedra all point inwards from two such layers towards a central sheet of LaO10 polyhedra. Between the 6-rings of the two Te sheets are large interstices which contain WO4 tetrahedra, disordered between ‘up’ and ‘down’ orientations. Mackayite, Fe3+(Te2O5)(OH) (#350) has layers that should be written Te4O10 to reflect the translational periodicity. These have 4- and 8-rings similar to the ‘apophyllite’ type, as seen in Fig. 11i. In the mackayite layer, individual ‘upward’ and ‘downward’ pointing polyhedra alternate, unlike apophyllite, where whole 4-rings of ‘up’ and ‘down’ tetrahedra alternate (Colville et al., 1971). The layers lie || (004), repeated by a screw tetrad axis. The layers are held together by edge-sharing dimers of Fe octahedra, Fe2O8(OH)2 (Fig. 32). Our second polymorph of Li2Te2O5 ≡ Li4[Te4O10] (#351) like its dimorph #343 has Te layers || (020) in which all Te are CN4; however, the polyhedra form 10-rings, and some of them are only 2-connected, occurring as edge-sharing Q1301 pairs rather than the Q1300 polyhedra typical of silicate-like sheets (Fig. 11j). The edge-sharing dimers act as bridges between bands where the other Te tetrahedra point downards and bands where they point upwards. The pointing direction reverses every c. Tellurium sheets are held together through the non-bridging oxygen atoms, which link to LiO4 tetrahedra, which occur as pairs of vierer corner-sharing helices (cf. Li2[TeO3], #1) running || x between Te layers. The regular inversion of pointing direction in the Te layers, and breakup of the Li component into discrete ribbons, gives this structure a resemblance to commensurately modulated phyllosilicates such as sepiolite, Mg4[Si6O15](OH)2 · 6H2O (Post et al., 2007). The Te4O10 layer of Ca[Te2O5] (#352) also has 10-rings and edge-sharing dimers of Te polyhedra; however, the latter are now 3-connected Q0401 type, as seen in Fig. 11k. The layers are crumpled and rather thick, with the edge-sharing Te=Te vectors almost normal to the overall layer plane (100). Layers are linked through sheets of CaO7 polyhedra (Fig. 32). Note that this description requires a slightly longer than usual strong-bonding distance threshold; if a Te–O distance of 2.450 Å is not included, the layer breaks up into finite Te4O10 tetramers of the type seen in Fig. 8j, #250–255. A similar topology of layer, less tightly corrugated, occurs || (10) in Tl2[Te2O5] (#353). Layers are linked through very irregular TlO5–7 polyhedra. K2[Te4O9] (#354) and its NH4+ analogue #355 are isopuntal, but giving them a common description again requires a careful choice of Te–O bonding threshold, which is 2.43 Å here. This excludes additional neighbours at 2.433 and 2.489 Å to respectively Te1 and Te2 of the K compound, which are at 2.713 and 2.615 Å and hence clearly not strongly bonded in #355. The convoluted Te8O18 layers || (100) in these structures have the topology of Fig. 11l, with 6- and very elongated 14-rings containing both CN3 and CN4 Te polyhedra. The layers are held together by large cations, 7–8 coordinated by oxygen in the case of K+. This structure is very different from the double-chain type of the Na analogue (#332, Fig. 10f). Although K2[Te4O9] · 3.2H2O (#356) is triclinic, it is strongly pseudohexagonal (b ≈ c, α ≈ 60°) and has rather symmetrical Te8O18 layers || (100) containing two types of 6-ring (Fig. 11m). One quarter of the rings are regular hexagons containing only CN4 Te, with non-bridging ligands pointing alternately ‘up’ and ‘down’. The rest of the rings are elliptical, and three of these meet at a CN3 Te atom. Between the Te layers are a central plane of water molecules and K+ ions in 7–8 coordination (Fig. 32). (NH4)Rb[Te4O9] · 2H2O and the corresponding compound with Cs replacing Rb (#357–358) have a rather complex Te16O36 layer (Fig. 11n). The layers are || (002), and can be regarded as formed by condensation of Te8O20 clusters. The clusters, in turn, consist of a central edge-sharing pair of Q0501 Te atoms, common to two 4-rings which are each completed by a pair of Q1300 Te atoms, each 4-ring in turn sharing one side with a 3-ring formed by links to a Q0300 Te atom. The remaining corners of the 4-rings and the CN3 Te atom of the 3-rings then link to other clusters to make a continuous layer, which has very elongate 12-rings. Large cations and water molecules are between the layers.
Our final single-layer structure, Ba6[Te10O25]Br2 (#359), also has an extraordinarily complex layer with a translational repeat Te40O100 and ten symmetrically distinct types of Te (Fig. 11o). The layers are || (002) and are rather thick but looped rather than double, with tubular cavities running || x which contain the Br– ions. Ba2+ cations between the layers link to 8–9 O atoms of the tellurite layer and 0–1 Br. Again, the layer is most simply described if an unusually long bonding threshold of 2.52 Å is used. This includes all the moderately strong Te–O bonds: seven out of the ten Te sites have oxygen neighbours in the 2.41–2.52 Å range, while none have any other neighbours within 2.98 Å. With the Te–O bond network thus defined, Te are CN4 (Te1 Te2, Te4 = Q0400, Te9 = Q1300 and Te5–Te8 = Q2200) except Te3 (Q0300). Two types of tube alternate, both with flattened elliptical cross-sections. The denser type contains a ribbon of 4-rings (Te2‒Te9‒Te2‒Te9) and 8-rings (Te2‒Te3‒Te1‒Te9‒Te2‒Te3‒Te1‒Te9), with arches of three additional Te atoms completing 7-rings (Te4‒Te8‒Te7‒Te3‒Te2‒Te9‒Te1). These tubes are linked into a layer through Te10 and bridges ‒Te5‒Te6‒ but the second type of tube defined by those is much more sparsely connected, its smallest rings having 9 and 10 members.
The compounds [M2+(Te6O13)] (M = Zn, Fe and Mg; #360–362) have Te6O13 double layers (Fig. 12a). These stack || (003) and are linked through corner-sharing with MO6 octahedra. The individual sublayers contain triplets of Q0311 Te1 sharing a common CN3 oxygen atom, connecting to 3-rings of Q1300 Te2 so as to form crumpled 12-rings of Te1 and Te2, which surround the M cations. Two such sublayers are linked by their Te1 atoms sharing an edge (Fig. 32). (Te3O5)(Se4+O3) ≡ [(SeO)2(Te6O14)] (#363) has a double layer || (001) with three types of Te polyhedron: Q0300, Q1300 and Q1400 (Fig. 12b). Each sublayer has a net of 3-connected 8-rings, with CN4 and CN5 Te atoms at the nodes and CN3 Te making two links within the sublayer. These links result in smaller 4- and 6-rings that are shared by the sublayers. Two such sublayers are held together by links between CN3 Te of one and CN5 Te of the other. The SeO3 pyramids are on the outside of the double layer, sharing one oxygen atom with CN5 Te and one with CN4 Te. Layers are held together by weak Te…O and Se…O bonds.
Structures with Te4+ complexes that are infinite frameworks
Tectotellurite structures #364–375 are listed in Table 18 (deposited), summarized in Table 5 and have their Te–O frameworks depicted in Fig. 12. The structure of the paratellurite polymorph of TeO2 (#364, Fig. 12c) is derived from that of the rutile form of TiO2 (cf. Meagher and Lager, 1979) in a way analogous to the derivation of the tellurite structure from brookite (#341 above). The lone-pair stereoactivity of the Te4+ changes the coordination environment from the relatively regular octahedron of rutile to a strongly distorted 4 + 2 pattern with 2 × O at 1.88 Å, 2 × O at 2.12 Å and 2 × O at 2.87 Å. However, unlike tellurite, the lengthening and weakening of two bonds per Te atom does not disrupt the framework of the TiO2 aristotype. However, the tetragonal c repeat is doubled, and the 42 screw axis of rutile becomes a 43 (or 41) axis in paratellurite. Paratellurite is ∼4.5% denser than tellurite. Interestingly, not only is the structure a distortion of the rutile structure, but it is isopuntal and topologically equivalent to the metastable low-temperature α-cristobalite form of SiO2 (cf. Downs and Palmer, 1994). This suggests that the paratellurite geometry provides a pathway for diffusionless structural transformation of the type discussed by Christy (1993), between the relatively low-density cristobalite structure type (stable at relatively high temperature and low pressure in the SiO2 system, for instance) and the high-density rutile type (stable at very high pressure for SiO2, as stishovite). At high pressure (∼0.95 GPa), paratellurite undergoes a continuous, displacive phase transition to a topologically similar but orthorhombic variant TeO2-γ, whose structure has been refined at 1.98 GPa (#365). Paratellurite and the high-pressure phase are isostructural with, respectively, the β and γ phases of SnF2, as noted by Denes et al. (1980), who also discuss transformations involving the cristobalite and rutile structures.
Pr2[Te2O6]O (#366) has a structure of the well-known pyrochlore type, which can be derived from a 2 × 2 × 2 block of face-centred cubes of the fluorite structure by slight displacement of ¾ of the anions so that half of the cations are in octahedral coordination by them, forming a continuous framework (Fig. 12d) and omission of half of the remaining anions. The overall stoichiometry is A2[B2X6]Y, where the larger cation type A is coordinated by 6X + 2Y while B bonds to 6X only. In this case, B is Te4+, which unsually is in rather regular octahedral coordination by oxygen, with no lone-pair stereoactivity (Fig. 32). The next few structures are also fluorite derivatives. K[Ga(Te6O14)] (#367) has a similar-sized a ≈ 11 Å cubic unit cell to #366, but 1/8 of the cations are 8-coordinated K+, 1/8 are 6-coordinated Ga3+, and the rest are 4-coordinated Te4+. The K, six Te and Ga are arranged in the LiCa6Ge pattern (Pavlyuk et al., 1993), a superstructure of the common Cu3Au type (Kear and Wilsdorf, 1962). Relative to fluorite, 1/8 of the anions are missing, as for pyrochlore. The remaining anions are of three types, bonded respectively to 2Te + K, Te + Ga and 3Te + K. If we consider only the Te–O substructure, it has Te6O14 stoichiometry and Q1210 Te polyhedra forming the complex network shown in Fig. 12e. The Ga octahedra reinforce this framework. Cliffordite, [(UO2)(Te3O7)] (#368), has Te4+ and U6+ in a 3 : 1 ratio, ordered in the Cu3Au arrangement like Te and (K+Ga) of #367 (Fig. 32). As for that structure, the Te–O framework has Q1210 Te polyhedra and Te6O14 stoichiometry, but the topology is different (Fig. 12f), and whereas TeO4 polyhedra share corners with GaO6 octahedra in #367, they share edges with UO8 bipyramids in cliffordite. Note that the total anion content is higher than that of fluorite: while the uranyl oxygen positions and those of the oxygen atoms that link U to Te can all be derived by small displacements from their counterparts in the fluorite aristotype, this is not true for O5 of cliffordite, which joins 3 Te. Very closely related is the winstanleyite structure type of compounds [A4+(Te3O8)] (#369–373), where A = Ti in the mineral winstanleyite (#372), (Fe0.673+ Te0.336+) in walfordite (#373), and Sn, Zr or Hf in synthetic analogues (Fig. 33). Again, Te and A atoms are in the Cu3Au pattern. The oxygen atoms of the fluorite aristotype are all present, but are displaced so as to form quite regular octahedra around the A cations and the common ‘folded rhombus’ arrangement of Fig. 4c aound Te. The Te polyhedra are Q2020, and form a framework (Fig. 12g) where the non-bridging ligands share corners with A octahedra. Note that for taxonomic purposes, the minority Te6+ content of the A sites of walfordite is ignored. Cs2[Te4O9] (#374) has a tetragonal unit cell with the a parameter similar to that of #366–373 but c about twice as large. The Cs and Te atoms form two cubes of the MgCu2 Friauf-Laves structure arrangement (Friauf, 1927; Hyde and Andersson, 1989). That is, they are geometrically equivalent to, respectively, the D and T cubic lattice complexes of Fischer and Koch (2006). This is also the pattern of Mg and Al atoms in normal spinel, MgAl2O4, and of YB2 atoms in the A2B2X6Y pyrochlore structure. However, whereas the B cations of the pyrochlore framework are linked through X to six B neighbours (cf. #366 above), the Te atoms of #374 are bonded to only a subset of these: half of them are CN4 (Q0400) and the other half are CN3 (Q1300). There are no 3- or 6-rings, as in the pyrochlore framework: every CN3 Te atom is a member of one 4-ring, while the CN4 atoms join two such rings, and the next-smallest rings have 8 members (Fig. 12h). Vacant sites which would be occupied by oxygen atoms in a pyrochlore allow a more open framework, with rather large interstices to accommodate Cs+ ions in 9–10 coordination. Pb[Te5O11] has a very complex, open framework with five symmetrically different types of Te atom (#375). If Te–O distances out to 2.45 Å are included as bonds, then the topology is as seen in Fig. 12i. Te1, Te2 and Te3 (respectively Q0312, Q1210 and Q0401) form complex double chains that run ||  at z = 0 and ||  at z = ½. These are in turn made from Te3=Te1=Te1=Te3 tetramers which are joined by Te2, which shares a CN3 oxygen with 2Te1 and a CN2 oxygen with Te3 (Fig. 33). Between these double chains, Te4 and Te5 (both Q1300) form massicot-like single chains (cf. Hill, 1985) which run || y at z = ¼ and ¾. The component chains are joined in three dimensions byTe4 linking to Te3 on one side of these chains, while Te5 links to Te2 on the other. Lead atoms are in 8-coordination between the massicot-like chains.
Te6+(OH)6 and its adducts
Our first examples of Te6+ compounds are polymorphs and derivatives of orthotelluric acid, Te6+(OH)6. The large valence (∼1 vu) of bonds to O from both Te and H means that any additioanl bonding must be weak, so the compound Te(OH)6 has discrete octahedral molecules which are held together only through hydrogen bonds. Unsurprisingly, it is hygroscopic, extremely water-soluble, and unknown in minerals. As the molecule has six oxygen atoms, each of which carries one donor H atom and is able to accept low-valence bonds from other electron donors outside the molecule, Te(OH)6 is able to co-crystallize with a wide range of other compounds to form adducts. The other components may be polar organic molecules, simple salts of large cations such as alkali halides, or salts with larger and more complex anions such as polyphosphate species. These adducts are physically and crystal-chemically distinctive enough that they are considered separately from other Te6+ compounds. In many cases, small tilts of Te octahedra or changes to the H-bonding pattern may result in very similar molecular arrangements occurring in a range of different space groups and unit-cell shapes. The TeO6 octahedron (Fig. 4h) is always rather regular, with Te–O = 1.90–2.07 Å, corresponding to bond valences of 1.04–0.77 vu using the parameters of Mills and Christy (2013). The individual structures #376–431 are listed in Table 19 (deposited).
The cubic polymorph of [Te(OH)6] itself has the molecules packed in a fcc array, but with the octahedra tilted and hydrogen bonded so that the cell repeat is doubled to 15.71 Å, and the space group is Fdc (#376). There is also a monoclinic polymorph (#377), where Te…Te distances shorter than 6.3 Å define 12 nearest neighbours in a monoclinically sheared face-centred cube, with pseudocube edges || ,  and  of the P21/n cell. (NaF) · [Te(OH)6] (#378) retains a fcc arrangement of Te(OH)6 molecules, albeit with rhombohedral distortion. Na+–F– ion pairs occur in the octahedral interstices between them, so that (NaF) units and Te(OH)6 molecules form a rocksalt arrangement. Similarly, (KF)2 · [Te(OH)6] (#379) has K+–F– ion pairs in the tetrahedral interstices of an orthrohombically distorted fcc array of Te(OH)6 molecules, so that (KF) and Te(OH)6 are arranged like F and Ca of the fluorite structure (Fig. 34). Such ion pairs are almost unknown in mineral structures, although the Ca2+–CO32– pair occurs in a matrix of H-bonded water molecules in ikaite, CaCO3 · 6H2O (Swainson and Hammond, 2003) and hsianghualite, (LiF)2Ca3[Be3Si3O12], can be regarded geometrically as having ion pairs Li+–F– replacing the Cs+ of pollucite, Cs23[Al2Si4O12], or water of analcime, (H2O)2(Na2)[Al2Si4O12] (Rastsvetaeva et al., 1991). However, the bond distances and the parameters of Brese and O'Keeffe (1991) indicate that the Li–F bond is not unusually strong, bond valences being ∼0.25 vu for all bonds from Li to F + 3 O and to F from Li + 3 Ca.
Alkali halide adducts with larger ions do not have ion pairs intercalated into a matrix of fcc Te(OH)6. (CsCl)2 · [Te(OH)6] (#380) has Te and Cl forming an array of monoclinically distorted CaCl2 (or collapsed rutile) type, with Cs located in tetrahedra of four Cl. (RbCl)3 · [Te(OH)6] (#381) has a quite different rhombohedral structure in which Te(OH)6 octahedra alternate with Cl3 triangles to form rods || z which are linked through Rb, with each Rb having five Cl neighbours and vice versa.
Na2(SO4) · [Te(OH)6] (#382) has a unique structure in which Te(OH)6 octahedra form a primitive hexagonal array, and alternate trigonal prisms of such molcules contain either two Na or SO4 tetrahedra. The pseudohexagonal layers of Te are || (002), with layers of Na and SO4 groups between them. One type of Na bonds to only two SO4 while the other type bonds to three SO4, and their arrangement makes the structure polar in the (010) plane. There are a very large number of adducts of the type A2(TO4) · [Te(OH)6], with larger cations A = (K, Rb, Cs, Tl and NH4) and T = (S or Se) (#383–398). Despite the apparent diversity of symmetries and cell parameters, all of these are again based on a fcc array of Te(OH)6 molecules, with the tetrahedral anion occupying octahedral interstices and the A cations occupying tetrahedral interstices. The edges of the pseudocubic cell have different indices depending on the axial setting chosen by authors, as follows. K2(SO4) · [Te(OH)6] (#383) has space group P, with pseudocube edge vectors are ½, ½ and ½. The monoclinic structures with space group C2/c, Cc, P21/a, P21/c, P21/n or Pn all have pseudocube edges ½, ½ and ½ (#384–390 and 392–397) except for (NH4)1.16K0.84(SO4) · [Te(OH)6] (#391), which was published in a different axial setting where the pseudocube edges are ½, ½ and ½. Note that these structures include centrosymmetric and acentric polymorphs of K2(SeO4) · [Te(OH)6] (#384 and 385), structures in which two A species occur in solid solution and others in which they are ordered, such as CsK(SO4) · [Te(OH)6] (#397), and that there may be one or two symmetrically distinct types of Te(OH)6 octahedron per unit cell. Cs2(SO4) · [Te(OH)6] (#398) has a rhombohedrally distorted structure, in which the pseudocube edges are ⅓, ⅓ and ⅓. Hydrogen ions may be quite mobile in these compounds; Dammak et al. (2005) investigated the protonic conductivity behaviour of Cs0.86(NH4)1.14(SO4) · [Te(OH)6], which has the P21/c variant of the structure but has been excluded from this study because of some unrealistically short Te–O distances in the refinement.
K2(NO3)2(H2O)2 · [Te(OH)6] (#399) has a layered structure || (001) not unlike that of #382, in which Te octahedra form a monoclinically sheared primitive-hexagonal array with trigonal prisms containing the other components. Conversely, (Cs3.5Rb0.5)(SeO3)1.7(SO3)0.3 · [Te(OH)6]3 (#400) has the familiar fcc array of Te, although this time, large cations occupy only ⅔ of the tetrahedral interstices, pyramidal anions are in the other ⅓ of the tetrahedral voids, and the octahedral positions are empty. The pseudocubic edge vectors are , ⅙ and ⅙ of the orthorhombic cell. The iodate adducts all have layered structures. K(IO3) · [Te(OH)6] (#401) has K+Te and I layers alternating || (001), while K2(IO3)2 · [Te(OH)6] (#402) has thicker layers I‒K‒Te‒K‒I || (003), with the lone pairs of I5+ cations directed into the interlayer gap (Fig. 34). (NH4)(IO3)(H2O) · [Te(OH)6] (#403) has layers of Te + I alternating with NH4 + H2O || (200).
A very large number of adducts have been made that contain phosphate anions, along with large cations Na, K, Rb, Ag, Tl or NH4. The phosphate groups range from tetrahedral monomers [PO3OH]2– or [PO2(OH)2]– (#404–412) through diphosphates [P2O6OH]3– or [P2O5(OH)2]2– (#413–415), cyclo triphosphates [P3O9]3– (#416–422), [P4O12]4– (#423–424), [P6O18]6– (#425–427), [P8O24]8– (#428–429) and even [P12O36]12– (#430). Most of these compounds have layered structures with large unit cells, mainly of low symmetry, and will not be discussed in detail here. However, we note that K3Na3(P3O9)2 · [Te(OH)6] has monoclinic and rhombohedral dimorphs (#420‒421), with alternation of (pseudo)hexagonal layers of Te + P and Na + K || (200) and (003), respectively. Most of the phosphate adducts are crystal-chemically unique: there are few examples of more than one compound sharing the same stoichiometry, and even when they do, the structures are different. Thus, despite the apparent chemical similarity, (NH4)4(P4O12)(H2O)2 · [Te(OH)6] and its K analogue (#423‒424) are not isostructural, and the same is true for (NH4)8(P8O24)(H2O)2 · [Te(OH)6] and the corresponding K compound (#428‒429). The most complex phosphate adduct, (C(NH2)3)12(P12O36)(H2O)24 · [Te(OH)6]12 (#430), has as a counterion not an alkali metal cation but guanidinium, [C(NH2)3]+. This compound has six Te layers and three polyphosphate layers in its rhombohedral cell with c ≈ 51 Å. The threefold rotational symmetry is inherited from the Te octahedra, triangular planar guanidinium complexes and cyclophosphate rings (Fig. 34). The final adduct described here also has a small organic molecular component, the neutral urea molecule in (CO(NH2)2)2 · [Te(OH)6] (#431). This compound has a rather simple structure in which each Te octahedron has four neighbours at 5.1–6.6 Å, in a monoclinically distorted version of the diamond arrangement. Each Te octahedron also has four nearby urea molecules (Te…C = 4.2–4.5 Å), and with them defines dense layers in the structure || (20). Hydrogen bonds link molecules both within and between these layers (Fig. 34).
A few additional compounds containing Te(OH)6 molecules as hydrogen-bonded adducts along with other Te in other environments are included below at #498–502, 504 and 631.
Monomeric Te6+Xn anions (n = 4–6)
Te6+X4 and Te6+X5
Te6+ almost always occurs in octahedral coordination with oxygen, as noted above and by Mills and Christy (2013). However, we have a very small group of compounds in which the coordination number is 4 or 5. Ligands distribute themselves symmetrically around the closed-shell Te6+ cation, unlike the situation with Te4+, which usually has a strongly stereoactive lone pair of electrons. The coordination polyhedron for CN4 Te6+ is a tetrahedron (Fig. 4f), while that for CN5 Te6+ is a trigonal bipyramid (Fig. 4g). Cs2[TeO4] (#432) has the K2SO4-β structure, in which Te and Cs form the same arrangement as Pb and Cl in cotunnite, PbCl2 (O'Keeffe and Hyde, 1985) (Fig. 35). Cs2K2[TeO5] (#433) has a tetragonal structure in which TeO5 polyhedra are linked through CN6 K+ and CN6–8 Cs+. Rb6[TeO5][TeO4] (#434) contains both types of Te polyhedron. In this compound, Rb and Te together form an approximately cubic close-packed array with pseudocube edge vectors [½00], [0½0] and [¼0½] of the monoclinic cell. Rubidium atoms are in 6–8 coordination by oxygen (Fig. 35). Note that in all these structures, the counterions are large, weakly-bonding alkali metal species. Details for these structures are summarized in Table 20 (deposited).
Monomeric Te6+X6 that are not part of a larger structural unit.
Although Te6+ occurs almost exclusively in one type of coordination polyhedron, and the range of Te–O polymers that it forms is very restricted compared with Te4+, the monomeric tellurate octahedron TeX6 is the single most prolific structure-forming Te–O complex in the present study: there are 172 compounds with such octahedral anions (#435–616 below), in addition to the 56 adducts of neutral molecular Te(OH)6 that were briefly described above. Nesotellurates up to #456, with no larger structural unit including strongly-bonding non-Te cations, are listed in Table 20 and described here.
(NH4)2[TeO2(OH)4] (#435) has a simple structure in which Te octahedra form a centred regular net in layers || (001), with layers of NH4 cations between them. The Te–O distances show that the unprotonated oxygens (O1) are ordered in trans positions in the octahedron, with O‒Te‒O vectors || . K3Na2Li[TeO6] (#436) has layers of K alternating with layers of Na+Li+Te || (020). Within the latter layers, TeO6 octahedra and LiO4 tetrahedra form edge-sharing chains || , alternating with chains of NaO5–6 polyhedra. The three types of K atom are in 6–8 coordination. K3Li3[TeO6] (#437), somewhat similarly, has alternation of K and Te+Li layers || (200). Within the Te+Li layers, chains || y of corner-linked LiO4 tetrahedra alternate with chains of edge-sharing TeO6 octahedra and unusual square-planar LiO4 groups. The Li‒O distances in the distorted tetrahedra are 1.94–2.05 Å; those in the squares are similar, at 2.01–2.11 Å, with no additional O neighbours until two at 3.27–3.29 Å, which complete a very elongated octahedron around Li. Potassium is 8–9 coordinated. K4Na2[TeO6] (#438) has K layers alternating with Te+Na || (001). The Te form a face-centred rectangular net, and NaO6 octahedra share two opposite faces with neighbouring TeO6 octahedra to complete the layer. Potassium coordination is 6–8.
The structure of KNa5[TeO6] (#439) is best described as having cations in a hcp array, with layers of composition KNa3 and Te(Na0.670.33)3 alternating along z, and two of each layer type per cell. Oxygen atoms are all equivalent and in octahedral interstices in the cation array, coordinated by Te + K + 2Na + 2(Na0.67). The coordination environments for all alkali cations are slightly unusual: trigonal prismatic for K (K–O = 2.79 Å), square planar for partially occupied Na1 and square pyramidal for fully occupied Na2 (Na–O = 2.35 Å in both cases), while Te is in very regular octahedral coordination with Te–O = 1.94 Å (Fig. 35). K[TeO(OH)5] · H2O (#440) has KO9 polyhedra sharing edges and faces to make a sheet with a honeycomb net || (100). The TeX6 octahedra are bound to this sheet to make a compound layer, with H2O molecules in the interlayer gaps. The layers are held together only by hydrogen bonds.
Na[TeO(OH)5] (#441) is very different from the above structures. It has a 2 × 2 × 2 cubic superstructure of the ReO3 type with alternation of Na+ and Te6+ in octahedral coordination and no long-range order of O2– and OH–, and is thus isotypical with wickmanite, Mn2+[Sn4+(OH)6] and a family of related hydroxostannates, germanates and antimonates, including several mineral species (Basciano et al., 1998). Because of nonlinear M‒O‒M′ links and orientational order of O‒H groups, these compounds do not have Fmm symmetry but either Pnm (Strunz and Contag, 1960), Pn (Morgenstern Badarau and Michel, 1976; Cohen-Addad, 1977; Basciano et al., 1998) or P42/n (Mikhaylov et al., 2011; Kleppe et al., 2012; Lafuente et al., 2015). Na[TeO(OH)5] may either be orientationally disordered, or may actually crystallize in one of these lower symmetry space groups. The ReO3 and wickmanite structure types are derivatives of the ABX3 perovskite type in which large cations A are absent, leaving only octahedrally coordinated species B. Many other perovskite-related tellurates are described below (#562–584), while here, we include two unusual examples which have large cations in both A and B positions. These are Sr3[TeO6] and Ba3[TeO6] (#442‒443) whose unit cells are very large superstructures of the basic perovskite cube (the cells reported are respectively 4 × 4 × 4 and √20 × √20 × 8 of the ∼4 Å cube), with rather low symmetry: the triclinic structure of #442 has eight distinct Te sites, four ‘B-type’ Sr sites and eight ‘A-type’ Sr, while for #443, which is tetragonal, the corresponding numbers of distinct sites are five, five and 13. Substantial rotations of TeO6 octahedra allow some ‘B-type’ (Sr,Ba) to increase their coordination number from 6 to 7 (Sr) or even 8 (Ba).
Rhombohedral Li6[TeO6] and Tl6[TeO6] (#444‒445) are isopuntal, even though the great difference in size and stereochemistry between Li+ and Tl+ means that their coordination environments are rather different. These compounds have defect superstructures of rocksalt, in which (Li,Tl) and Te are ordered on a ccp array. In #444, oxygen atoms occupy 7/8 of the octahedral interstices. Tellurium has six oxygen neighbours at 1.93 Å, Li has four at 1.94–2.08 Å and one at 2.37 Å, and oxygen is surrounded by Te + 5Li. In the Tl compound, the inter-cation distances and cell parameters are much larger, and the oxygen atoms are in one of the triangular faces shared by an octahedral and a tetrahedral interstice of the ccp array. While the TeO6 octahedron itself is little changed (Te–O = 1.95 Å), Tl is very irregularly coordinated by six oxygen neighbours at 2.11, 2.46, 2.86, 3.60, 3.88 and 3.90 Å. There is a remarkable relationship between the structure of Tl61+[TeO6] and that of [Tl63+(TeO6)O6] (#616, below), which can be derived from it by stuffing with additional oxygen atoms (Fig. 35). Li4Zn[TeO6] (#446) also has a defect rocksalt superstructure, but this time, there are no vacancies. Edge vectors of a face-centred pseudocube are [½0¯½], [½⅓½] and [½¯⅓½] of the true monoclinic cell. Lithium and Zn are partially ordered on three types of octahedral site, although Li is dominant in all cases.
Ag2[TeO2(OH)4] (#447) has TeX6 octahedra arranged on the D lattice complex (Fischer and Koch, 2006) of its Fdd2 space group. The unprotonated oxygen atoms (O1) are ordered in a cis fashion, and their orientation defines the polarity of the structure along the z direction. The Te octahedra are linked by Ag+ ions in irregular tetrahedral coordination (Ag–O = 2.23–2.58 Å). Silver and the hydroxide oxygen O2 form massicot-like chains (cf. Hill, 1985) which lie in layers || (400) between the Te octahedra, and run ||  or  in alternate layers. Pb5[TeO6]O2 (#448) has an unusual, very dense arrangement of Pb and Te atoms. Cations lie in rods || x that are in a pseudohexagonal array (√ = 1.95 ≈ 2), but the x coordinates are such that the cation substructure is not conventionally close-packed. Layers of cations || (040) form nets in which cations form squares and triangles, such that the connectivity is 36, 3342, 3342 and then repeats along the z direction (Fig. 35). The arrangement can be generated from thinned and faulted 2 × 3 blocks of hexagonal close-packing. Each cation has from 11 to 13 cation next-nearest neighbours at 3.6–4.6 Å, while the oxygen coordination number of Pb2+ is 5–8. Oxygen atoms are in interstices such that layers || (040) of tetrahedrally coordinated oxygen alternate with layers where CN = 5–6. Pb6Cd[TeO6]O4 (#449) has an approximately ccp array of Pb, Cd and Te atoms, the unit cell containing a 2 × 2 × 2 array of distorted face-centred cubes, with irregular PbO4–6 polyhedra, CdO6 trigonal prisms and TeO6 octahedra.
The compounds A23+[TeO6] (A = Y, La and Gd; #450–452) have the enantiomorphic (P212121) structure of the orthorhombic form of Nd2WO6 (Efremov et al., 1984). The TeA2 substructure is in the cotunnite (PbCl2) arrangement (Léger et al., 1996, and cf. #432 above), with oxygen atoms coordinated by Te + 2A or Te + 3A so as to form two different types of AO7 polyhedron. Conversely, A23+[TeO6] with A = Sc, Yb, In or Tl3+ adopt a different, trigonal structure (#453–456). These compounds have either smaller A–O bonded distances than those above or A cations with relatively large non-bonded radii in the sense of O'Keeffe and Hyde (1981), so all oxygen atoms are 3-coordinate. The structure adopted is shared with malladrite, Na2[SiF6] (Babel, 1967). The oxygen atoms are approximately hcp, with (A, Te) cations filling variously 4/9 or 5/9 of the octahedral interstices between alternate oxygen layers, and the smaller minority cation (Te ≡ Si) forming a substructure with the AlB2 arrangement (Hofmann and Jäniche, 1935). Every occupied cation site alternates with vacancies along z, so as to avoid face-sharing of octahedra (Fig. 35). Note that #453–456 could be considered as examples of [A23+(TeO6)] frameworks, given the relatively low CN and high bond valence for the A cations, but they are included here because of the chemical similarity to #450–452 and the isostructurality with malladrite, in which the A cation Na+ is much more weakly bound.
Monomeric Te6+X6 as part of a larger structural unit that is a finite cluster
Table 21 (deposited) lists compounds #457–502, in which monomeric TeX6 anions are strongly bound to non-Te cations as part of a larger structural unit. The first of these compounds is unusual, in that the structural unit is in part organic: it is tris(tetramethyldisilyl) tellurate, [(CH3)2Si–Si(CH3)2]3(TeO6) (#457). The three tetramethyldisilyl groups each bond to two oxygens of the tellurate octahedron, to form a propellor-shaped neutral molecule. The monoclinic cell contains four such molecules, two of opposite rotational senses with their local pseudotriad axes pointing || ±, and two with their pseudotriad axes || ± (Fig. 36).
Compounds #458–461 have stoichiometry of the form Na5[M3+(TeO4(OH)2)2] · 16H2O, where M = Cu, Ag and Au, except that #459 is a variant of #458 with 13% of the Na+ replaced by H+. All are isostructural. Two TeX6 octahedra share trans edges of a MO4 square in these compounds, the square-planar coordination being typical for M3+ in a low-spin d8 electronic configuration. The hydrogen atoms were located in the refinement of #461, where they are located on the Te ligands that lie out of the principal plane of the cluster. The corresponding Te‒O distances are long, 1.98–1.99 Å, compared to 1.97 Å for the Au=Te bridging oxygens and 1.85–1.86 Å for the unprotonated non-bridging Te ligands opposite the Te=M shared edge. The lath-shaped [Te=M=Te] clusters all have their long axes ||  of the triclinic cell, and their principal plane approximately || (2). These structural units cross-link layers || (10) of NaX6 polyhedra, which contain the additional H2O molecules (Fig. 36).
The next three compounds #462–464 feature similar trimeric [Te=M=Te] heteropoly clusters containing a high-valence noble metal cation, but the central polyhedron is an MX6 octahedron rather than an MX4 square. The three structures are all different. K6Na2[Pt4+(OH)2(TeO5OH)2] · 12H2O (#462) has all [Te=Pt=Te] clusters parallel, with long axes ||  and equatorial planes of octahedra || (101). The clusters are held together through separate Na(H2O)6 octahedra and K+ ions in 6–8 coordination. In Rb2Na4[Os6+O2(TeO4(OH)2)2] · 16H2O (#463), the [Te=Os=Te] complexes have their long axes ||  and equatorial planes || (102). They share the two O2– ligands of one Te (Te2) with similarly oriented edge-sharing tetramers of NaX6 octahedra, Na4O2(H2O)16, to form long structural rods || y, which are held together by 8-coordinated Rb+. Despite the similar stoichiometry, Na6[Ru6+O2(TeO4(OH)2)2] · 16H2O (#464) has a very different triclinic structure with trimer long axes ||  and equatorial planes || (302). These structural units act as bridges between layers || (100) in which NaX6 octahedra share corners, edges and faces with each other. The H atoms were located in this refinement, and are again confirmed to be located on the oxygen atoms away from the principal plane of the cluster (Te–O = 1.98–2.01 Å), rather than the terminal oxygen ligands in that plane (Te–O = 1.85–1.86 Å). K2Na8[Pd24+Te4O18(OH)6] · 20H2O (#465) has a heteropoly hexameric cluster with a central core of two edge-sharing PdX6 octahedra. Each shares its trans edge with a Te1 cation, while the two Te2 octahedra share the remaining Pd ligands to form 3-rings with both Pd atoms, and in addition make a third bond to the oxygen atoms of the Pd = Pd edge. The hydroxide H atoms are not located in the structure. They are most probably located on the non-bridging oxygen atoms of TeX6 (of which there are 14 per cluster), but only two of these have long Te‒O distances (Te1–O4 = 1.99 Å) as opposed to the typical 1.80–1.83 Å. The rhombus-shaped clusters have long axes ||  and are flattened || (11). They lie in layers parallel to that plane with NaX6 octahedra, while additional NaX6 and KX7 lie between the layers (Fig. 36).
A large group of structures feature the [M66+Te6+O24]6– Anderson–Evans heteropolyanion, where M = (Mo or W), and six MO6 octahedra form a hexagonal ring around a central TeO6 octahedron (Anderson, 1937; Evans, 1948, 1974). The shape and size of the unit cell is controlled largely by the stacking of these large, tabular structural units, which are held together principally by hydrogen bonds to hydrated alkali cations or NH4+ (#466–470), hydrated lanthanide cations (#471–487) or both (#488–490), hydrated transition elements (#491–497) or alkalis plus additional molecular Te(OH)6 (#498–502).
Although the alkali tellurohexamolybdates and tungstates are all triclinic with very similar cell dimensions, they differ in detail. Li6[Mo6TeO24] · 18H2O (#466) has the planes of the Mo–Te hexagons || (2), with LiX6 octahedra between them, sharing edges to form rods of rocksalt-like structure || . Na6[Mo6TeO24] · 22H2O and its W analogue (#467–468) have layers of NaX6 polyhedra || (10), between M‒Te hexagons that lie parallel to approximately (63). Rb6[Mo6TeO24] · 10H2O (#469) has hexagons || (11), sitting in voids in a three-dimensional framework of RbX7–9 polyhedra. (NH4)2Na4[Mo6TeO24] · 16H2O (#470) has layers || (001) of NH4+ and Mo‒Te ions, alternating with layers of monomeric NaX6 octahedra and Na2X10 dimers.
(Ce(H2O)4)2[Mo6TeO24] · 3H2O and the Nd analogue (#471–472) have Mo‒Te hexagons that lie in layers || (020), and are canted slightly relative to that plane in opposite senses in alternate layers. Layers are cross-linked by AX9 polyhedra (A = Ce or Nd), while additional H2O molecules lie in the Mo‒Te layers. The next several compounds are all triclinic with similar unit-cell dimensions, with only a single orientation of Mo‒Te hexagon. Again, these lie in layers with some of the water molecules, while AXn polyhedra lie between the layers and bridge them. The coordination number, n, is usually 9, although it is reduced to 8 for the smaller Eu3+, Ho3+ and Yb3+ cations in #485–487. The layers are usually || (100), although they are || (110) for (Sm(H2O)5)2[Mo6TeO24] · 6H2O (#482), which has a different axial setting. The orientation of the Mo‒Te hexagons varies depending on the hydration state and cation size. For (La(H2O)7)2[Mo6TeO24] · 6H2O (#473), the hexagons are || (10), while for the less hydrated (La(H2O)6)2[Mo6TeO24] · 6H2O (#474) and its Ce analogue (#475), they lie flatter, approximately || (411) (Fig. 36). The hexagons also lie flatter for the Nd compounds: || (411) in (Nd(H2O)6)2[Mo6TeO24] · 6H2O (#479), || (611) in (Nd(H2O)7)2[Mo6TeO24] · 5H2O (#480) and || (41) in (Nd(H2O)7)2[Mo6TeO24] · 5H2O (#481), although they are strongly canted || (20) in the Pr analogue of the latter compound (#478). In K6(Eu(H2O)7)2[Mo6TeO24]2 · 16H2O (#488), layers of two differently oriented and symmetrically independent Mo-Te hexagons alternate along the z axis; one type has EuO2(H2O)7 polyhedra connecting them into columns || y, while the other type does not. A matrix of CN9–10 hydrated K+ ions holds the layers together. The Gd analogue (#489) is nearly isostructural, but is in a different axial setting, with the layers of hexagons || (010) not (001), and the Gd‒Mo‒Te columns || x, not y. Also, slight atomic rearrangements lead to doubling of the c parameter relative to the corresponding a parameter of #488.
(Co(H2O)6)3[Mo6TeO24] and its Ni analogue (#490–491) have a simple, highly symmetrical rhombohedral structure in which [Mo6TeO24]6– hexagons alternate with triangles of [M(H2O)6]2+ octahedra (M = Co or Ni) to make columns running along the threefold rotation axes. (NH4)2(M2+(H2O)3)2[Mo6TeO24] · H2O (M = Mn, Co, Ni, Cu and Zn; #492–496) (Fig. 36) and the Ni‒W analogue (#497) are also highly symmetrical (cubic, space group Pa, a ≈ 14 Å). The TeM2 part of the structure forms a distorted fluorite array of the type found as predicted for SiO2 at very high pressure by Park et al. (1988) and found experimentally for SnO2 above 21 GPa by Haines and Léger (1997). Because of the distortion, M2+ bonds to only three of its four nearest Mo‒Te hexagons, as well as to three water molecules. The other components in the structure, NH4+ ions and the remaining H2O molecule (O6) form hydrogen-bonded dumbells (N…O = 2.84 Å) which, together with the Te atoms, are arranged similarly to the covalent [S2]2– dumbells and Fe in pyrite, FeS2.
The remaining compounds of this section are alkali tellurohexamolybdates which also contain Te(OH)6 molecules as adducts. Li6[Mo6TeO24](H2O)18 · [Te(OH)6] (#498) has sinuous chains of LiX4–6 polyhedra and Te(OH)6 || y, which cross-link Mo‒Te hexagons that are oriented || (103). Cs6[Mo6TeO24](H2O)2 · [Te(OH)6]2 (#499) has one Mo‒Te anion per unit cell oriented || (12), embedded in a trellis-like open framework of CsX7–10 and Te(OH)6 polyhedra. Rb6[Mo6TeO24](H2O)6 · [Te(OH)6]2 (#500) has a C-centred monoclinic cell with strong c pseudosymmetry, and thus four Mo‒Te hexagons per cell. These are all oriented || (010) and alternate with Te(OH)6 molecules along the y direction, forming Te-rich rods in a matrix of RbX7–9 polyhedra (Fig. 36). The ordered K‒NH4 compound #501 is nearly isostructural, as is the more highly hydrated NH4 compound #502 (although this is in an A-centred axial setting).
Monomeric Te6+X6 as part of a larger structural unit that is an infinite chain
The next few structures have TeX6 anions that are strongly bound with other cations into one-dimensional structural units. They are listed in Table 22 (deposited). [Hg2+(TeO2(OH)4)] (#503) has TeX6 octahedra linked via their cis unprotonated ligands through linear 2-coordinated Hg2+ into chains || y, which pack in a hexagonal array. Chains are connected only through hydrogen bonds and long, weak Hg…O bonds (within-chain Hg–O distances are 2.02–2.05 Å, while the two next nearest oxygen atoms to Hg are at 2.67 and 2.75 Å). In [(Hg2+2)(TeO2(OH)4)] · [Te(OH)6] · 2H2O (#504), TeX6 octahedra linked through trans unprotonated ligands to linear [Hg–Hg]2+ dimers, forming infinite chains || . These lie in layers || (010), which alternate with layers containing Te(OH)6 and H2O molecules. Mercury is even more strongly 2-coordinate than in the previous compound, with one O neighbour at 2.11 Å, one Hg at 2.50 Å. and the next O neighbours not until 2.91 and 2.93 Å. K2[Cu(TeO4(OH)2)] · H2O (#505) has TeX6 octahedra sharing opposite edges with CuO4 squares to form straight chains, which lie in layers || (002). Chains of alternate layers run || to  or to . Water molecules and CN8–9 K+ ions lie between the layers (Fig. 37). Copper has no additional ligands within 3 Å. Very slight tilts and displacements reduce the symmetry from orthorhombic and centrosymmetric (space group Cccm) to monoclinic and acentric (Cc). The mineral raisaite, (Mg(H2O)6)[Cu(TeO4(OH)2)] (#506) has TeX6 octahedra sharing edges which are not opposite with CuO4 squares (Cu‒O = 1.94–1.98 Å), to form zigzag chains || z. These form a centred-rectangular array, which have [Mg(H2O)6]2+ octahedra lying betwen them. The structure is held together through hydrogen bonds, and also weak Cu…O bonds: Cu has two O atoms of water molecules at 2.78 Å, completing an elongated octahedron of ligands (Fig. 37). Ag4[Cu(TeO6)] (#507) has TeO6 octahedra sharing an edge with one CuO4 square and corners with two others to make a double chain running || z that zigzags in the (100) plane. The double chain consists of Cu=Te‒Cu=Te 4-rings, united by the shared edges. Two additional O ligands are much closer to Cu than in #506 above (Cu–O = 2 × 1.98, 2 × 2.00 and 2 × 2.49 Å). If these are included to complete a CuO4+2 octahedron, the chains become ribbons of edge-sharing Cu and Te octahedra in which Cu atoms form a central zigzag backbone while Te sit on the outside of the ribbon. The chains form a centred-rectangular array, and are held together by three types of Ag+ ions in distorted octahedral coordination (Ag–O = 2.27–2.83 Å). The oxygen atoms approximate ccp, with pseudocube edge vectors || [¼0¾], [¯¼⅓¼] and [¯¼¯⅓¼]. If the structure is considered as a packing of (Ag, Cu, Te) octahedra, then it is a superstructure of the rocksalt type. Tl4[Cu(TeO6)] (#508) has Cu‒Te chains with the same topology as those of #507 and also forming an approximately centred-rectangular array, but this time running || x and zigzagging in the (01) plane of the triclinic cell. The oxygens do not form a continuous close-packed array as in #507, but discrete close-packed ribbons which surround large channels || x. Four types of Tl+ ions in irregular 5–7 coordination hold the chains together, with their lone pairs pointing into the channels (Fig. 37). (NH4)2V25+TeO8(OH)2 ≡ (NH4)2[(VO2)2(TeO4(OH)2)] (#509) has TeX6 octahedra sharing non-opposed edges with edge-sharing dimers V2O8 of VO5 square pyramids. One of the bridging oxygens links to 1 Te + 1 V, while the other connects to 1 Te + 2 V atoms. Small atomic displacements break the potential 2/a symmetry of the chain, so that there are two distinct Te atoms and four distinct V atoms per repeat unit. These chains run || x and lie in layers || (002), with NH4+ ions between the layers.
Monomeric Te6+X6 as part of a larger structural unit that is an infinite layer
Table 22 also includes structures #510–547, where monomeric TeX6 octahedra are linked with non-Te cations to form two-dimensional strucural units. Na2[Cu2(TeO6)] (#510) has a Cu‒Te layer || (001) that can be regarded as an ordered version of a brucite-like trioctahedral sheet, but with strong Jahn-Teller distortion of the CuO6 octahedra, giving Cu a square of four O neighbours at 1.98–2.00 Å and two more distant ligands at 2.53 Å. If all six neighbours are considered, then the structural unit has the same bond topology as those of #538–541 below. However, unlike those structures, the oxygen atoms of #510 are approximately ccp in three dimensions, and Na atoms are ordered in ⅔ of the octahedral sites between [Cu2(TeO6)]2– layers. The structure can be regarded as an ordered defect derivative of the rocksalt type. NaTl31+[Cu42+(TeO6)2] (#511) has similarly distorted Cu‒Te layers || (200), alternating with layers that contain an ordered array of Na+ in octahedral coordination and Tl+ in very irregular 6-fold coordination. The three-dimensional arrangement of oxygen atoms can again be considered a derivative of cubic close-packing, except that the coordination requirements of Tl cause the Cu‒Te‒O layers to undulate, and lead to the Tl atoms being far from the centres of octahedral interstices.
Frankhawthorneite, [Cu2(TeO4(OH)2)], (#512), has oxygen atoms forming a slightly distorted hcp array, with close-packed layers || (002). Copper and Te atoms occupy octahedral interstices to form ribbons in which one Te atom alternates with two Cu along the y direction. Ribbons form a centred rectangular array, and share corners with their neighbours. However, Jahn-Teller distortion of the Cu coordination polyhedron is of a similar degree to that in #507–508 above (Cu–O = 4 × 1.98–2.06 Å and 2 × 2.46–2.52 Å), so only the four shortest bonds are counted when defining the structural unit, which is hence not a framework but a layer || (10). In frankhawthorneite, the component ribbons of the layer have each Te octahedron sharing two opposite edges with CuO4 squares, and ribbons are linked through CN3 oxygen atoms in a stepped pattern. The Jahn-Teller distortion also reduces the symmetry to monoclinic P21/n from orthorhombic Pmnn, and allows ordering of the H atom on O2, as shown by long Te–O bonds (1.99 rather than 1.91–1.92 Å). Layers are held together by long Cu…O bonds and hydrogen bonds (Fig. 37). A similar arrangement of more regular octahedra is found in kotoite, Mg3(BO3)2, which has similar unit-cell dimensions but the orthorhombic space group, and has additional B atoms in triangular coordination, providing strong links between ribbons (Berger, 1988). The octahedral arrangement in the average structure of one form of (H,Li)2[Ti(TeO6)] is also very similar (#585, below). Paratimroseite, Pb[Cu2(TeO6)] · H2O (#513) has stepped layers || (002) composed by condensation of ribbons || x, in exactly the same topology as frankhawthorneite (#512), but the layers are separated widely and also alternate between two orientations related by a screw diad axis, so their anions form disconnected oblique slices of hcp structure rather than constituting a three-dimensionally continuous close-packed substructure. Water molecules and Pb2+ cations in irregular 9-coordination lie between the layers, and the Pb coordination geometry causes shift of the layers in the x direction such that any reflection symmetries are eliminated, and the space group is P212121 rather than the Pbca or Pbcm of hypothetical aristotypes. An elongated coordination octahedron around Cu is completed by an interlayer water molecule at 2.42 Å and an additional tellurate oxygen atom at 2.54 Å (Fig. 37). The structure of timroseite is closely related (#555, below). Sr2[Cu2(TeO6)]Br2 (#514) has edge-sharing ribbons of CuO4+1 square pyramids (Cu–O = 4 × 1.93–2.02 Å and 1 × 2.36 Å) and TeO6 octahedra running || y, which share corners with each other so as to form a continuous layer || (100) in which the oxygen atoms again form the stepped, hcp slice of #512–513. Interlayer SrO4Br3 polyhedra hold the structure together. The Br– anion is also a very distant sixth ligand for Cu2+ (2.97 Å). In bairdite, Pb2[Cu4(TeO5OH)2](SO4) · H2O (#515), two stepped Cu‒Te layers || (100) stack adjacent to each other, related by a screw diad axis, and are linked through long Cu…O bonds so as to form a double layer (Cu2–O4 = 2.36 Å, as opposed to 1.94–2.04 Å for the four shortest Cu2–O bonds, and 2.57 Å for an additional within-layer distance that completes the octahedron). O4 also has another Cu2 at 2.00 Å and an unusually long distance of 2.03 Å to Te, and is where the H is located. Between Cu‒Te double layers are Pb atoms which show some positional disorder, and SO4 tetrahedra. Cu1 has a square of O atoms at 1.91–2.00 Å and one at 2.41 Å within the Cu‒Te layer, and a sulfate oxygen atom at 2.46 Å. A differently oriented oblique slice through an hcp anion array is an element of the ‘tri-harmunite’ structure of #598, below.
The polytypes of khinite (-3T and -4O), Pb[Cu3(OH)2(TeO6)] (#516–517) share a more complex layer type, in which rows of edge-sharing CuO4 squares =Cu1=Cu2= alternate with rows of CuO4 squares and TeO6 octahedra, =Cu3=Te=. The coordination octahedron of the Te is completed by sharing bridging oxygen atoms of the all-Cu chain, thus making a layer containing 5-rings [–Te=Cu3=Te–Cu1=Cu2–]. In projection normal to the layer, the cations form a hexagonal net, but the all-Cu chain is at a different height from the Cu‒Te chain, so the layer has an overall polarity (Fig. 37). In the 3T polytype (originally known as ‘parakhinite’), the subchains of layers point along x, y or –, successive layers rotating by 120°, consistent with a screw triad axis. There are three layers || (003) per P32/P31 unit cell. In the 4O polytype, layers are || (004), alternate layers have subchains ||  or , and the layers are related by the d glides of the space group Fdd2. Note that a4O ≈ a3T, and that a pseudohexagonal metric is retained, as b4O ≈ √3a4O. In both cases, Pb2+ lies between the layers, in 8-fold coordination. Agaite, Pb3[Cu(TeO5OH)](CO3)(OH) (#518) has layers || (020) in which alternating TeX6 and CuX5 polyhedra form a 63 honeycomb net. These polyhedra share non-opposing edges to form zigzag chains trending || x, which are joined in the z direction by sharing of a fifth corner. Between the Cu‒Te layers lie triple layers of CN8 Pb2+ ions, additional OH– ions (bound to 3 Pb) and a central plane of CO32– ions (each oxygen atom bound to 3 Pb as well as to C). The orientation of Cu=Te zigzags in the structural unit and the pointing direction of CO3 triangles both define a polarity in the z direction (Fig. 38).
Na1.8[(Sn0.94+Te0.16+)(TeO6)] and Na2[Ge(TeO6)] (#519–520) both have a trigonal structure in which oxygen atoms approximate hcp (cf. #512). It is a superstructure of one of the TlSbO3 polytypes (Bouchama and Tournoux, 1975). The other such polytype has the structure of ilmenite (Fe2+Ti4+O3) in that ⅔ of the octahedral interstices between each pair of oxygen layers are occupied by cations, and that two types of cation occupy alternate cation layers. However, whereas the vacant octahedral sites in the ilmenite polytype are offset in a sequence ABCABC so as to produce a 6-layer rhombohedral cell (space group R), the vacant sites in the current structures are arranged in a pattern ABAC, giving a 4-layer trigonal cell with space group P1c. In these tellurate compounds, layers of Na cations alternate with layers of (M4+ + Te6+), where M = (Sn or Ge). NaO6 octahedra share faces with TeO6 octahedra above and below along the z direction, while M4+O6 octahedra have vacant sites above and below. Note that such face-sharing relationships would not be achievable in an ilmenite-like polytype, where every occupied octahedron shares one face with another that is occupied. Sr[Ge(TeO6)] (#521) also has hcp oxygen atoms, and a primitive trigonal cell that has the same ∼5 Å a parameter but half the c parameter of #519–520. (Ge+Te) atoms occupy ⅔ of the octahedral sites every other octahedral cation layer, forming a honeycomb pattern with Ge and Te alternating, while Sr atoms are in the intervening layers, above and below the vacant sites in the (Ge+Te layer). This is the P312 structure of NaNi4+I7+O6 (Brown, 1969), which is a cation-ordered superstructure of the P1m type of Li2ZrF6 (Brunton, 1973) or rosiaite, Pb2+Sb25+O6 (Basso et al., 1996). The compounds A3+[Cr3+(Te6+O6)] (A = La, Pr, Nd, Sm–Yb and Y; #522–534) and La[Fe(TeO6)] (#535) have a polytypical relative of the Sr[Ge(TeO6)] structure in which the c repeat is doubled from 5.4 Å to ∼10 Å and the space group changes to P, because the relative positions of (Cr,Fe) and Te are reversed in alternate Cr‒Fe‒Te layers. This structure is also that of colquiriite, CaLi[AlF6] (Yin and Keszler, 1992). Compound #535 shows some Fe‒Te disorder which may imply short-range mixing of the two cations, but given the large difference in charge, more probably implies displacement of layers by stacking faults, while retaining two-dimensional order within layers. Thus, it may have a nanoscale intergrowth of rosiaite and colquiriite structure types.
Ba[Ge(TeO6)] (#536), despite the chemical similarity to its Sr analogue #521, has a different structure with the same P312 space group. Oxygen atoms are double-hexagonal close-packed (AABB stacking), and layers of (Ge+Te) in octahedral coordination alternate with layers of BaO6 trigonal prisms. Backite, Pb2[Al(TeO6)]Cl (#537) has dioctahedral AlTeO6 layers resembling the MTeO6 unit in structures #519–536 (Fig. 38). However, these layers are now widely spaced, and the oxygen atoms are not three-dimensionally close-packed. Instead, the unit cell contains one such layer, separated from the next by layers of Pb, Cl and Pb such that each Pb atom lies above a vacant octahedral site of the Al‒Te layer and is coordinated by 3O + 6Cl. The compounds Na2–x[M2(TeO6)] (M = Ni, Zn and Co, x = 0–0.05; #538–540) are based upon a AABB stacking of oxygen atoms, like #536. M and Te cations are in octahedral coordination, and form a structural unit that is an ordered brucite-like trioctahedral layer, with Na cations partially occupying the trigonal prismatic sites between the M‒Te layers. The Na+ ions are highly mobile, leading to fast-ion conduction in these compounds (Evstigneeva et al., 2011). Small atomic displacements result in the Zn and Co compounds being acentric (P6322), while the Ni compound has space group P63/mcm. The Cu analogue of these compounds was discussed above (#510).
Leisingite, (Mg(H2O)6)[Cu2(TeO6)] (#541) has a layered [Cu2(TeO6)]2– structural unit || (001) which resembles the trioctahedral layers of #538–540, with no evident Jahn-Teller distortion of CuO6 octahedra (Cu–O = 6 × 2.11 Å), unlike the layers of #510–511. Isolated [Mg(H2O)6]2+ octahedra lie between these layers, and connect them through hydrogen bonding. Apart from a 30° rotation of the Mg octahedron and the locations of H atoms, this mineral is isotypical with zincalstibite (Bonaccorsi et al., 2007), a member of the cualstibite group of the hydrotalcite supergroup of Mills et al. (2012). The correspondence can be shown as [Cu2(TeO6)][Mg(H2O)6] (leisingite) ≡ [Zn2Al(OH)6][Sb(OH)6] (zincalstibite). Zincalstibite has a lower-symmetry space group (P as opposed to P1 m), but very similar unit-cell parameters a = 5.321(1) Å and c = 9.786(2) Å (Fig. 38).
Mojaveite, [Cu6(TeO4(OH)2)(OH)7]Cl (#542), has brucite-like octahedral sheets in which 1/7 of the cations are Te and 6/7 are Cu, while 1/14 of the anions are Cl, and thus are not counted as part of the structural unit. The degree of Jahn-Teller distortion is relatively small, half the Cu are regarded as coordinated by five (O,OH) ligands (+1 Cl) while the other half have six O neighbours. The ordering pattern of cations and anions forces the structure to adopt the relative low-symmetry polar space group R3. The layers are held together by hydrogen bonds (Fig. 38). The mineral is isostructural with bluebellite, Cu6[I5+O3(OH)3](OH)7Cl, which is unusual in that Te6+ does not have a lone pair of electrons while I5+ does so (Mills et al., 2014a). The substitution is presumably facilitated by the polar symmetry of the Te/I site, which frees the coordination environment to distort. Fuettererite, Pb3Cu6(TeO6)(OH)7Cl5 ≡ (Pb3(OH)Cl3)2[Cu6(TeO6)(OH)6]2Cl4 (#543), has Cu‒Te layers very similar to those of mojaveite, although all Cu atoms now have one Cl as a sixth ligand, rather than just half of them. There are Cl atoms on both sides of the layer, and the structure retains a centre of inversion symmetry. Pairs of Cu‒Te‒O layers are linked through a shared Cl atom, and these layers are stacked with the other components between them in the sequence [Cu6(TeO6)(OH)6]…Cl…[Cu6(TeO6)(OH)6]…Cl…(Pb3(OH)Cl3)…Cl…(Pb3(OH)Cl3)…Cl (Fig. 38).
Markcooperite and its synthetic analogue, ideally Pb22+[(U6+O2)(TeO6)], have a quite different type of layer (#544–545). TeO6 and UO6 octahedra both share four corners to form a layer with 4-rings, similar to that of Fig. 14c but with alternation of Te and U atoms. These layers lie || (100) and have CN7 Pb2+ ions between them. In the end-member synthetic compound (#545), the Te octahedron is relatively regular (Te–O = 2 × 1.91, 2 × 1.94 and 2 × 1.95 Å) while the U octahedron is strongly compressed (U–O = 2 × 1.84, 2 × 2.24 and 2 × 2.29 Å), consistent with the presence of linear [O = U = O]2+ groups (Fig. 38). The natural mineral shows 25% substitution of Te on the U site, but the pattern of bond-length variation remains similar. The substitution makes the structure transitional towards those with continuous layers of corner-sharing TeO6 octahedra (Fig. 14c), such as #688–670 below.
ThV25+TeO10 · 2H2O ≡ Th[(VO2)2(TeO6)] · 2H2O (#546) has layers || (200) consisting of TeO6 octahedra, VO5 (V1) trigonal bipyranids and VO4 tetrahedra (V2). V1 polyhedra share an edge with Te on one side and a corner with Te on the other side in the z direction, while V2 tetrahedra share a corner with Te on each side in the y direction, so the layer has square-shaped 8-rings of alternating Te and V in which Te is 4-connected while V is always only 2-connected. Each V atom has two non-bridging ligands with short V–O distances (1.63–1.67 as opposed to 1.77–2.04 Å), validating their description as vanadyl groups [VO2]+. Th4+ ions are in 9-coordination between the layers. Ba2Nb2TeO10 ≡ Ba2[Nb2O4(TeO6)] (#547) has corrugated layers || (020) containing zweier chains of corner-sharing NbO6 octahedra which run || x. The TeO6 octahedron shares one edge and one corner with a pair of Nb octahedra in the chain on one side, making an [‒Nb‒Te=Nb‒] 3-ring, and the same with the next chain on the other side, so that the layer has 3- and 6-rings with the topology of the kagome net. Ba2+ ions between the layers are 9-coordinated. Although the layer contains four non-tellurate oxygen atoms per formula unit, only two of these are non-bridging oxygen atoms with short Nb–O distances (1.78 Å; compare 1.90–2.26 Å for all other Nb–O) (Fig. 39).
Monomeric Te6+X6 as part of a larger structural unit that is a framework
A large group of compounds have TeX6 octahedra strongly bound to non-Te cations to form a three-dimensional framework. These are listed as #548–618 in Table 23 (deposited). [Hg32+(TeO6)] (#548) has a highly symmetrical structure with a large cubic unit cell (Ia, a ≈ 13 Å). Two types of Te atom are arranged as a 2 × 2 × 2 block of CsCl unit cubes, while the Hg atoms are positioned so as to form, in combination with Te, a slightly perturbed 2 × 2 × 2 array of cubes with the Cr3Si arrangement (Boren, 1933; Andersson, 1978), analogous to the cations of the garnet structure (O'Keeffe and Andersson, 1977; Grew et al., 2013; #549–553 below), which has the same space group and similar unit-cell parameter. However, Hg2+ has linear two-fold coordination, unlike the CN4/CN8 cations of garnet: two Hg‒O distances are 2.06 Å, while the next neighbours are not until 2.57, 2.59, 2.76 and 2.83 Å. Links Te‒O‒Hg‒O‒Te connect each Te atom to six out of the eight neighbouring Te of the other type (Fig. 39).
The important garnet structure is an ordered and anion-stuffed superstructure of the Cr3Si type (Geller, 1967; O'Keeffe and Hyde, 1985). The Iad cubic cell of typical garnets A3B2C3X12 has a 2 × 2 × 2 array of body-centred cubes of B atoms corresponding to Si of Cr3Si, and non-intersecting rods || <100> directions of alternating A and C atoms corresponding to Cr. Anions X occupy distorted tetrahedral interstices, so that each anion is bonded to 2A + B + C, while coordination numbers of A, B and C are, respectively, 8, 6 and 4. The numerous mineralogical examples of garnets were reviewed by Grew et al. (2013). The B2C3X12 substructure forms a framework in which each X links to one cation of each type. Depending on the relative bond valences, it may be reasonable to identify within this framework either BX6 or CX4 as a principal anionic complex. An example of the former would be cryolithionite, Na3Al2Li3F12 ≡ Na3Al2[LiF4]3 more appropriately than Na3Li3[AlF6]2, while silicate garnets are examples where the strongest-bound cation is in the tetrahedron: grossular, Ca3Al2Si3O12 = Ca3Al2[SiO4]3. Several tellurate garnets are known, in which the B cation is Te6+; these compounds are of the first type, because of the high Te–O bond valence. These include the mineral yafsoanite (Fig. 39), Ca3[Zn3(TeO6)2] (#549) and also synthetic Na3[M33+(TeO6)2] (M = Fe0.5Al0.5 or Ga, #550–551). Nd3Li3.05[(Te0.975Sb0.025)O6]2 (#552) and Nd3Li4[(Te0.5Sb0.5)O6]2 (#553) are included here because they still have the garnet structure, albeit slightly modified. However, the C cation Li is now so low in charge that it no longer forms a structural unit with the octahedral cations: the Nd–O bond valence (0.375) is larger than that of Li–O (0.25). Furthermore, both these compounds have considerably higher Li contents than the 2 atoms per formula unit expected for a normal garnet, due to stuffing of additional Li into normally vacant interstices, and the Li are very mobile in the structure, making the compounds fast-ion conductors (O'Callaghan et al., 2008).
LaV35+TeO12 · 3H2O ≡ La[(VO2)3(TeO6)] · 3H2O (#554) has a framework in which an approximately primitive cubic array of TeO6 octahedra are linked to all six of their neighbours through 2-connected VO4 tetrahedra. La3+ ions are near the centre of each cube. Thus, the LaTeV3 substructure corresponds to the atomic arrangement of an ABX3 perovskite. The La atom is coordinated by six oxygen atoms of the framework but also by three water molecules which all lie on one side of La, reducing the symmetry to polar rhombohedral R3c. Relative to a cubic metric, the structure is slightly stretched along the rhombohedral triad axis (c/a = 2.48 = √6.15, rather than √6). As is typical for vanadyl tellurates, V–O distances are much shorter (1.63–1.64 Å) for the V‒O‒La oxygens than for the V‒O‒Te oxygens (1.80–1.84 Å).
The minerals of the dugganite group include dugganite itself, Pb3[Zn3(TeO6)(AsO4)2] (#555), its phosphate analogue kuksite (#556) and also cheremnykhite (the vanadate analogue), Pb3[Zn3(TeO6)(VO4)2] (Kim et al., 1990), whose structure has not been refined, and joëlbruggerite, ideally Pb3[Zn3(Sb5+O6)(AsO3.5(OH)0.5)2] (Mills et al., 2009c), which has only minor Te substituting for Sb. The dugganite structure has layers || (001) in which 3-connected AsO4 on triad axes and 2-connected ZnO4 tetrahedra share corners to form a net of trefoil-shaped 12-rings in which Zn and As alternate. One third of the triad axes does not have an As tetrahedron, but instead have Te octahedra between two Zn‒As layers and linking the layers. Each TeO6 octahedron links to three Zn tetrahedra below and to three more above. Thus, the three-dimensional framework that results contains [Zn‒Te‒Zn‒Te] 4-rings and [Zn‒As‒Zn‒Te‒Zn‒As] 6-rings (Fig. 39).
[Be4O(TeO6)] (#557) has a simple face-centred cubic structure in which Be4O6 tetrahedra, with Be atoms at the corners and O1 oxygen atoms along the edges, and additional oxygen atoms (O2) are arranged in the ‘zincblende’ arrangement. Thus, there is a rather open [Be4O7]6– framework in which O2 is at the centre of an anion-centred tetrahedron, where four BeO4 tetrahedra meet. Oxygen atoms occupy 7/8 of the positions of ccp, with the eighth position vacant, and the sixth O1 oxygen atom defines an octahedral interstice that contains Te (Fig. 39). The structure may be compared with that of swedenborgite, Na[Be4O(Sb5+O6)], which has a similar stoichiometry. In swedenborgite (Huminicki and Hawthorne, 2001), the oxygen atoms again occupy 7/8 of the positions of ccp, but the stacking of close-packed layers is now ABAC rather than ABC, and the eighth position contains CN12 Na+ ions. There is also a unique octahedral interstice, which contains Sb, and a Be4O7 framework in which Be4O6 tetrahedra can be identified. However, the additional oxygen atom that connects the tetrahedra is no longer bonded to four Be, but instead to two Be + Sb, while the oxygen that centres an OBe4 tetrahedron is now part of the Be4O6 tetrahedron. Apart from #557 and swedenborgite, isolated OBe4 tetrahedra have also been reported from two polymorphs of (Be4O)(NO3)6 (Haley et al., 1997; Troyanov et al., 2000; Krivovichev et al., 2013).
Pb[Cu3O(TeO6)] (#558) has Cu‒Te‒O rods running || y in which the oxygen atoms approximate small blocks of cubic close-packed structure. Relatively regular TeO6 octahedra alternate with pairs of Cu1 atoms in elongated CuO4+2 polyhedra, while Cu2O4 squares brace the sides of the rods, as well as sharing corners with Te and Cu1 atoms of neighbouring rods to form a framework with large channels || y, which contain CN7 Pb2+ ions. Jensenite, Cu3TeO6 · 2H2O ≡ [Cu3(H2O)2(TeO6)] (#559), has brucite-like Cu2TeO6 layers || (10) in which Cu and Te are ordered in a honeycomb pattern. While the Te octahedral are quite regular (Te–O = 1.89–1.98 Å), the Cu polyhedra show the usual Jahn-Teller distortion, with Cu–O = 4 × 1.94–2.06 Å and 2 × 2.32–2.45 Å. The layers are bridged into a framework by an additional Cu cation in the interlayer gap that is in strict square-planar coordination. This Cu atom (Cu1) has as ligands two interlayer water molecules, plus a tellurate oxygen atom in each layer (Fig. 39).
Timroseite, Pb2[Cu5(TeO6)2](OH)2 (#560) has stepped hcp layers || (002) of the type previously seen in frankhawthorneite, paratimroseite, Sr2[Cu2(TeO6)]Br2 and bairdite (#512–515, above). As in paratimroseite, the layers are in two different orientations which alternate, but in timroseite, they are connected into a three-dimensional framework through additional CuO4 squares (Fig. 40). Large channels remain || x, which contain OH– anions and Pb2+ in 8–10 coordination. All Cu are in square-planar coordination if a bonding cutoff of <2.27 Å is used, but all Cu atoms also have one channel OH– anion at 2.27–2.64 Å, and Cu1 and Cu2 within the hcp layers, also have an additional tellurate oxygen at 2.47–2.71 Å. Quetzalcoatlite, [Zn6Cu3(TeO6)2(OH)6] · (AgxPbyClx+2y) (#561), has a more explicitly nanoporous structure in which hexagonal rings of six corner-sharing ZnO2(OH)2 tetrahedra alternate along z with layers || (001) in which TeO6 octahedra share edges with CuO4 squares to form a honeycomb net with walls Te=Cu=Te surrounding large hexagonal voids. The hexagonal channels contain rods of alternating, partially occupied (Ag+, Pb2+) and Cl– sites (Fig. 40).
A large group of tellurate compounds have a superstructure of the perovskite type (#562–584). An aristotypical ABO3 perovskite has a unit cube with Z = 1, a ≈ 4 Å and space group Pmm. The compounds described here all have two types of octahedrally coordinated B atom, one of which is Te6+. Other perovskite variants have been described earlier, including M2+Te4+O3 (M = Co, Ni and Cu) with Te4+ as the A cation (#140–142), Cs2[Te24+O5] with Te4+ as the B cation and some ordered oxygen vacancies (#195), the wickmanite-like Na[Te6+O(OH)5], with a vacant A site (#441), and the complex superstructures of M3[Te6+O6] (M = Sr and Ba; #442–443). One form of Pb2[Co2+(TeO6)] (#562) has Te alternating with Co in the B sites, a √2 × √2 × 2 superstructure of the basic perovskite type, and tetragonal space group I4/mmm, consistent with slight tetragonal distortion but no octahederal tilt. Howard et al. (2003), in their analysis of possible octahedral tilt systems (Glazer, 1972) and symmetries of ordered perovskites, expressed scepticism about the existence of such structures in the absence of strong Jahn-Teller or other distortion (as occurs in ‘CsAuCl3’ = Cs2[Au1+Cl2][Au3+Cl4]; Tindemans-van Eijndhoven and Verschoor, 1974). It is possible that the oxygen positions in this structure need reinvestigation. A rhombohedral polymorph occurs at high and at low temperature (#582, below). A2[Cu2+(TeO6)] (A = Sr and Ba; #563–564) have similar cell dimensions to #562 but the tetragonal space group I4/m is obtained by octahedra tilt according to the a0a0c– pattern in these compounds, independent of the Jahn-Teller distortion of the CuO4+2 octahedra (Howard et al., 2003; Howard and Carpenter, 2010). The ICSD gives as the archetype of this perovskite structure Sr2NiWO6 (Köhl, 1973). Ba2[Cu(TeO6)] (#564) has another polymorph whose structural topology is polytypically related to that of normal perovskites (#587, below).
NaLa[Mg(TeO6)] (#565) has similar cell dimensions again but in addition to two types of B cation ordered in a ‘rocksalt’ fashion, there are two types of A cation ordered layerwise along the z direction, and the symmetry is monoclinic, P21/m. However, most of the perovskites with √2 × √2 × 2 superstructure have the P21/n space group exhibited by cryolite, Na3AlF6 = Na2[Na(AlF6)] (Hawthorne and Ferguson, 1975). These include compounds A2[B(TeO6)] with A = Ca or Sr and B = Ca or Co (#566, 569–571) (Fig. 40), Cd2[Cd(TeO6)] (#567), Na2[Sn4+(TeO6)] (#568) and A = disordered (Ca0.5Pr0.53+), (Sr0.5Pr0.53+) or (Sr0.5Eu0.53+), B = Li (#572–574).
Sr2[Ni(TeO6)] has a larger 2 × 2 × 2 superstructure than the perovskites above, and C2/m symmetry (#575), although it is nearly cubic, with all three cell parameters within 0.3% of their mean value, and β ≈ 90.4°. The next five compounds have the truly cubic ‘double perovskite’ structure of elpasolite (K2Na[AlF6]: Sabelli, 1987), with a 2 × 2 × 2 supercell but Fmm symmetry. In all these compounds, Te alternates in the B sites in a ‘rocksalt’ fashion with Ni (#575), Ca (#576), Li (#577–578), Mg (#579) or partially occupied Bi (#580), while in #577 and 578, the alkaline earth cation Ba2+ and lanthanides (La,Pr)3+ are disordered in the A site, analogous to the situation in #572–574 above.
Our final group of perovskites have trigonal cells with a ≈ √2 and c ≈ 2√3 times the edge of the fundamental perovskite cube. This is an alternative axial setting for a structure produced by compression or extension of a 2 × 2 × 2 cube along one  direction. The lattice type for such a perovskite is R in most cases, but in Ba3[Bi2O3(TeO6)] (#581), the unusual ordering pattern, with Te in only ⅓ of the B sites, reduces the symmetry to Pc1 (Fig. 40). The space group is R or Rm for the other structures #582–584, which have the standard ‘double perovskite’ cation arrangement, and include the rhombohedral polymorph of Pb2[Co(TeO6)] (cf. #562 above). Three additional compounds in the present study could be regarded as aberrant examples of rhombohedral double perovskites, but are more usefully considered as superstructures of the corundum type. These are [Ni3(TeO6)] and its relatives, #599–601 below.
In the ABX3 cubic perovskite structure, the large cations A and anions X together form a cubic close-packed array, with X3A ordered in the Cu3Au pattern. The structure can thus be described as an ABC stacking of AX3 layers, with B cations filling the X6 octahedral interstices. Other ‘perovskite polytypes’ can be produced by stacking AX3 differently (Mitchell, 2002). These structures require some BX6 octahedra to share faces, implying short B…B distances and the possibility that they may require some B…B bonding interaction to be stable. Most mineralogical perovskites are derived from only the 3C polytype, although some examples with face-sharing octahedra are known for antiperovskite polytypes, with anions O2–, F– or Cl– in the B sites (Krivovichev, 2008). The synthetic compound Ba2[Co(TeO6)] (#585) has very similar unit-cell parameters to #584 but a different space group and structural topology. Unlike its analogues with Ca, Sr and Mn in B sites (#576, 583 and 584 above), it is based on a 6H ABACBC stacking of BaO3 layers: the hkk sequence, where ‘h’ = ‘hcp-like (layers above and below superimposed)’ and ‘k’ = ‘ccp-like (layers above and below not superimposed)’. A very different 6H perovskite is described as #620, below. Because layers of Co2+ and Te6+ cations alternate along z in #585, there are 12 oxygen layers altogether in the unit cell, which has space group Pm1, and two crystallographic types each of Co and Te. Co2 and Te2 octahedra each share one face with each other, while Co1 and Te1 share only corners (Fig. 40). In contrast, Ba2[Ni(TeO6)] (#586) has a 12R (hhkk) layer sequence (space group Rm), in which corner-sharing octahedra (Te1)O6 alternate along z with columns of three face-sharing octahedra, Ni ≡ (Te2) ≡ Ni. Ba2[Cu(TeO6)] (#587) has the same topology, but distortion of the CuO6 octahedra reduces the symmetry to triclinic P. The distortion is not the usual symmetrical elongation, but gives Cu2+ five oxygen neighbours at 1.98–2.11 Å and the sixth at 2.26 Å. The triclinic (001) plane corresponds to (003) of the pseudorhombohedral cell, while the pseudotriad axis is  in the triclinic axial setting. A tetragonally distorted 3C perovskite polymorph of this compound was discussed at #564, above.
The next four closely-related compounds have structures derived from the LiSbO3 type (Genkina, 1992). This structure has a hcp oxygen substructure, as for the corundum superstructures of LiNbO3 and NaSbO3, but a quite different arrangement of cations. LiSbO3 has an orthorhombic structure with space group Pncn, which is permuted into the Pnna axial setting here. Then, cell parameters a, b, c ≈ √3, √(8/3) and 3, measured in units of the mean ‘close-packed’ O…O distance. There are two close-packed anion layers per cell (as opposed to six for corundum/ilmenite) || (020), and between each layer pair, Li and Sb each occupy ⅓ of the octahedral interstices. SbO6 octahedra form edge-sharing chains || x, and these share corners with the chains above and below to form a framework. The vacant octahedral sites of each cation layer lie above and below Sb, so that SbO6 octahedra do not share faces. LiO6 octahedra do not form edge-sharing chains, but do form face-sharing columns || y. Even in the idealized structure, a and b are only ∼6% different, so the structure is metrically pseudotetragonal with pseudotetrad axis || z, and the two types of cation are arranged similarly to Ca and W of the scheelite structure (Hazen et al., 1985), although the disposition of oxygen atoms is quite different. The compounds Li2[M4+(TeO6)] (M = Sn or Ti; #588–589) have this structure, but with M and Te alternating in the zigzag chains, which reduces the symmetry to Pnn2 (Fig. 40). Partial leaching of Li+ from the Ti compound and replacement with H+ indicated no structural change for small degrees of leaching (Crosnier et al., 1992), but (H1.68Li0.32)[Ti(TeO6)] showed considerable redistribution of non-Te cations, while preserving the (Te + O) substructure. A significant proportion of Ti occupied former Li or vacant sites (#590), giving an arrangement with the space group Pnnm, which is in effect a tri-CaCl2 (orthorhombically collapsed trirutile) structure. The structure of frankhawthorneite, [Cu2(TeO4(OH)2)], is closely related (#512 above). Annealing of this compound re-ordered the octahedral site occupancies to give a tetragonal but acentric trirutile structure with space group P42nm (#591). More conventional P42/mnm trirutile structures are covered below (#594–596).
Pb6[Co9(TeO6)5] and its Ni analogue (#592–593) have an unusual structure containing defect brucite-like layers with 1/6 of the cations missing, (M72+Te32)O24 (M = Co or Ni). These are connected into a framework through additional M ≡ Te face-sharing dimers, which share corners with the layers above and below to form a pillared-layer structure with very large interlayer channels, which contain the Pb atoms (Fig. 41). [M23+(TeO6)] with M = Cr, Fe or Ga (#594–596) have the well-known trirutile structure also known for minerals such as the byströmite and tapiolite groups, M2+Sb25+O6 (M = Mg or Zn) and M2+Ta25+O6 (M = Fe or Mn) (Byström et al., 1942). The space group is the same as rutile, P42/mnm, while the c repeat is tripled due to cation ordering. An unsual acentric trirutile phase with M1‒M2‒ ordering along its pseudotetrad direction was discussed above (#586). Li[Mn2+Mn3+(TeO6)] (#597) again has slightly distorted hcp of oxygen atoms, a pseudotetragonal unit-cell metric, and cell dimensions very similar to #583–586 and #594–596 above. However, a greater proportion of the octahedral sites are occupied, although it should be noted that irregularity of the octahedra and off-centring of cations mean that the coordination number is unambiguously 6 only for Te1, Te2 and Mn1–3, while Mn4 and Li2 are CN7 and Li3 is CN8, if all cation–oxygen distances within 3 Å are included. Bond-valence sums using the parameters of Brese and O'Keeffe (1991) indicate that Mn1 and Mn3 are Mn2+, while Mn2 and Mn4 are Mn3+. The approximate close-packed oxygen layers are || (002), with the pseudotetrad direction || y. Cations are arranged such that two out of every three sites are occupied along the y direction. Two types of cation layer alternate. In one of these, Te1 and Mn2 share edges to form a zigzag chain || x, with an adjacent zigzag of Li2 and Li3 on one side. In the other layer, zigzag chains || x are formed by Te2 and Mn4 and by Mn1 and Mn4. The vacancies of the second layer share faces with Te1 and Mn2, while the vacancies of the first layer share faces with Mn1 and Mn3. The refinement indicates some mixing (17%) of Li on Mn4 and of Mn on Li3, but if this is ignored, all ‘Li’ sites excluded from the structural unit and all ‘Mn’ sites included, then we define a framework in which layers of 2-wide and 4-wide octahedral ribbons share edges, with channels between them || x which accommodate Li+ ions. Na3[(Mn32+Mn3+)(TeO6)2] (#598) has a ‘tri-harmunite’ structure which can be derived from the Pnma structure of harmunite, CaFe23+O4 (Gaulskina et al., 2014), one of three closely related structures known to high-pressure researchers as ‘post-spinel’ phases, as they have the same cation:oxygen ratio as spinel but larger coordination numbers, making them potential high-pressure polymorphs (cf. Yamanaka et al., 2013). These structures all have frameworks made by corner sharing between 2-wide edge-sharing ribbons of octahedra, with channels which contain CN8 cations (Fig. 41). In #598, the octahedral ribbons run || y, but ordering of Te and mixed-valence Mn triples the unit cell repeat in this direction. Na occupies the CN8 site. Note that the octahedral ribbons define oblique, stepped slices of cations in an hcp anion array, repeated by twinning on (020), as noted by Hyde and Andersson (1989). The hexagonal close-packed planes are || (210) or (20) in alternate twin lamellae. The oblique hcp slices differ from those of frankhawthorneite etc (#512–515) in that the ribbon direction is parallel to an octahedral edge, rather than perpendicular.
[Ni3(TeO6)] (#599) has a hcp array of oxygen atoms in which ⅔ of the octahedral interstices are filled in the same pattern as corundum, Al2O3. However, alternate cation layers along the z direction are either all Ni or are Ni+Te, ordered so as to reduce the symmetry to the polar space group R3. The resulting structure is a superstructure not just of the corundum type (Rc), but also of its two zellengleich 1 : 1 superstructures: ilmenite (FeTiO3: R) and LiNbO3 (R3c). Ilmenite-structure compounds in which there is alternation of cation layers with Na and with disordered (M4+ + Te6+) are discussed below at #664–665. The compounds Li2[M4+(TeO6)] (M = Zr or Ge; #600–601) are placed here because they are isopuntal, although cation layers of (Li1 + M) and (Li2 + Te) now alternate along z. Lithium octahedra share faces with M or Te octahedra, and the Li cations are displaced strongly away from these neighbours along z, in accord with the polar symmetry of the structure (Fig. 41). If the low-valence Li atoms are excluded from the structural unit, the remaining MTeO6 framework has a rhombohedrally stretched version of the ‘double perovskite’ type. The c/a ratio is 17% larger in #601 than the value of √12 which would correspond to a primitive cubic arrangement of cations. As Li can be regarded as an off-centre ‘A’ cation in the A2BB'X6 double perovskite topology, these compounds could in fact be classified with the rhombohedral double perovskites #582–584 above, which have similar cell dimensions. Thus, the range of cation valences in these compounds, and the resulting choice of whether or not to include cations in the structural unit, highlight a relationship between perovskite and corundum structure families which is not otherwise obvious.
[Mg3(TeO6)] (#602) is an archetype for several isostructural tellurates of Mn, Mn+Cu or Cd+Mn (#603–605). The rather dense structure is not conventionally close-packed, but the key to comprehending it is to note that the rhombohedral cell parameters for Mg are arh = 6.047 Å and αrh = 90.86°, with Te atoms forming an almost perfect body-centred cubic array (more precisely a CsCl-type array, as there are two nonequivalent Te atoms per cell) (Fig. 41). Like [Hg3(TeO6)] (#548) and the tellurate garnets (#549–553) above, these compounds have an oxygen-stuffed Cr3Si structure, with oxygen atoms occupying a different set of tetrahedral interstices than those that they do in garnets, such that each oxygen bonds to 3 Mg (≡Cr) and 1 Te (≡Si), while Mg and Te are all in octahedral coordination. There is no long-range order of Cd and Mn in [(Cd2Mn)(TeO6)] (#605). The Co and Zn analogues have monoclinically distorted superstructures (#606–607). The Co and Zn compounds have space group C2/c, with amon ∼ √3atrig, bmon ≈ btrig, cmon ≈ ctrig, β = 92–95°. There are still two distinct Te sites per cell, but the single octahedrally-coordinated M2+ site of the rhombohedral phases splits into five distinct sites with a wider range of coordination numbers: CN = 6, 6, 5, 6 and 4 for [Co3(TeO6)] (#606) and CN = 4, 4, 5, 5 and 6 for [Zn3(TeO6)] (#607). The ordered Cu–Zn compound [Cu5Zn4(TeO6)3] (#608) has similar cell parameters, but additional displacements of atoms which reduce the symmetry further to C2. There are three Te sites, six Cu sites and four Zn sites. While Te is in fairly regular octahedral coordination, the Cu atoms are in Jahn-Teller distorted 4+2 coordination if the threshold between strong and weak bonding is set at Cu–O = 2.2 Å, except for Cu2, which is 4+3 coordinated, with three oxygen ligands in the range 2.5–2.8 Å. Two Zn atoms are 4-coordinated and two are 5-coordinated.
The mineral mcalpineite has been recently redefined as anhydrous [Cu3(TeO6)] (Carbone et al., 2013). It is interesting to note that its structure (#609) is the explicitly ternary variant of the ‘C-sesquioxide’ structure of heavy rare-earth oxides and the bixbyite group of minerals M23+O3, where M = Mn, Tl, Y and Sc, in respectively, bixbyite (Zachariasen, 1928), avicennite (Radtke et al., 1978), yttriaite-(Y) (Mills et al., 2011) and kangite (Ma et al., 2013) (Fig. 41). This is another structure like the pyrochlore type (cf. #366, 700) which can be regarded as a defect fluorite. Again, the overall cation array is cubic close-packed and there is a cubic unit cell with a ≈ 10 Å, corresponding to 2 × 2 × 2 fluorite unit cubes, but this time, ¼ of the anions are missing, which correspond to O2 in the winstanleyite structure (#369–373), which has similar cell dimensions and the same Ia space group. As in winstanleyite, the cations are split into two distinct types in a 3:1 ratio. All cations are 6-coordinate in mcalpineite, but one type (Te6+, here) has a coordination polyhedron that is close to regular octahedral geometry, while the other (Cu2+) has a less regular geometry that is best described as twisted trigonal prismatic. Up to half the Cu may be substituted by Co or Zn, with no further ordering (#610–612). Some Te-free synthetic isotypes are also rich in Cu, such as the phases Cu2–x2+Fe2x3+Ti2–x4+O6 (Mouron et al., 1985) and Cu22+M3+Sb5+O6 (M = Mn, Fe and Ga; Bazuev et al., 1994).
Ba3[Zn6(Si2O7)2(TeO6)] (#613) has a unique zincotellurosilicate framework, with 9- and 12-coordinate Ba2+ ions in large cages. Paired ZnO6 and single TeO6 octahedra share edges to form ribbons running || z that resemble the Cu2Te octahedral ribbons of frankhawthorneite and related compounds (#512–515). These ribbons are linked into a framework by sharing corners with corrugated layers || (200) of ZnO4 tetrahedra and Si2O7 tetrahedral dimers. The tetrahedral layers contain [Zn‒Si‒Zn‒Si] 4-rings and [Zn‒Si‒Si‒Zn‒Si‒Si] 6-rings (Fig. 41).
[M63+(TeO6)O6] (M = Y, In and Tl: #614–616) again have a defect fluorite structure like the bixbyite isotypes #609–612 above, but the rhombohedral structure is more complex, with approximately cubic close-packed cations ordered into two types in a 6 : 1 ratio, 1/7 of the anions missing, and the rest split into two types, coordinated either by 2 M + 1 Te or by 4 M. All cations are 6-coordinate. This structure is also known for compounds such as Y6(U6+O6)O6 (Bartram, 1966), the meteoritic mineral allendeite, Sc4Zr3O12 ≡ [(Sc4Zr2)(ZrO6)O6] (Thornber et al., 1968; Ma et al., 2014) and Pr7O12 ≡ [(Pr43+Pr24+)(Pr4+O6)O6] (von Dreele et al., 1975). Remarkably, Tl61+[TeO6] (#445, above) has the same space group and cation substructure, and very nearly the same cell parameters as Tl63+(TeO6) (#616). However, all the non-tellurate oxygen atoms are missing in #445, and the Tl–O bonds to the remaining oxygen atoms are longer and less regular.
This section concludes with two Co-rich frameworks. Na5Co15.52+(TeO6)6 ≡ Na2.5(Na2.5Co1.5)[Co14(TeO6)6] (#617) has a nanoporous hexagonal structure in which Co1 and Te octahedra share edges to form 2-wide ribbons running || z. These ribbons link at corners to delineate relatively large hexagonal tunnels (diameter ≈ 6.5 Å) and smaller ditrigonal tunnels, which are internally braced by Co2 atoms in trigonal prismatic coordination, to complete a framework [(Co1)12(Co2)2(TeO6)6]8–. An additional site in the small channels is occupied by mixed Na + Co, while the remaining Na+ ions are in the large channels. [Co82+(TeO6)(AsO4)2O2] (#618) has four types of Co2+ ion (CN = 4, 6, 6 and 6) sharing edges and corners to make thick double layers || (020). TeO6 octahedra act as internal braces in the middle of the double layers, while AsO4 tetrahedra cross-link the layers into a three-dimensional framework (Fig. 42).
Soro or cyclo finite polymers Tem6+Xn
Structures #619–645 contain Te6+O6 octahedra that are linked into finite polymers TemXn (m = 2–6); these are listed in Table 24 (deposited). Our first two examples are closely related to each other. Ba3[Te2O9] (#619) has a simple hexagonal structure containing the face-sharing dimer of Fig. 13a. These groups lie with ⅓ of the Ba2+ ions in layers || (002), which alternate with layers containing the other ⅔ of the Ba2+ ions. Ba3[Fe3+((Fe3+Te6+)O9)] (#620) has the same P63/mmc space group and nearly identical unit-cell parameters. It contains the same dimeric anion, except that 50% of the Te6+ is now randomly substituted by Fe3+, and charge balance is maintained by insertion of additional Fe3+ ions into octahedral interstices which were vacant in #619. The additional Fe octahedra in #620 join corners with the face-sharing dimers to form a three-dimensional framework which is a 6H perovskite polytype, with hkk stacking of (BaO3) close-packed layers. It is thus very closely related to Ba2[Co(TeO6)] (#585), where the different cation-ordering pattern results in lower symmetry. The relationship to these phases shows that #619 can be considered a defect 6H perovskite with ordered B-site vacancies.
Li4TeO5 ≡ Li8[Te2O10] and its Na analogue (#621–622) has the edge-sharing dimeric anion of Fig. 13b. All atoms are 6-coordinated, with each oxygen atom linked to 5 Na + 1 Te or to 4 Na + 2 Te. The cation array is ccp and the structure is actually a superstructure of the rocksalt type, with the pseudocube edge vectors parallel to [0½½] and approximately [ ] and [ ] of the triclinic cell in #622. The Te = Te dimer axes are oriented ||  (Fig. 42). The Ag analogue (#623) has a monoclinic structure in which layers of close-packed cations stack || (004) in an ABAC sequence (alternating h and k layers). All cations and anions again occupy 6-coordinated interstices of the opposite substructure, so while Ag+ and Te6+ of the k cation layers are in octahedral coordination, the Ag+ ions of h layers are in trigonal prisms. Within the k layers, edge-sharing Te=Te dimers alternate with Ag=Ag=Ag triplets along rows || x.
K4[Te2O6(OH)4] · 5H2O (#624) has a simple but incompletely determined structure in which dimers with Te=Te axes || x are linked into layers || (020) through K+ ions. Water molecules were not located, but are presumably in the interlayer gap and complete the coordination polyhedra of the K+ ions. The structure of K4[Te2O6(OH)4] · 8H2O (#625) is more completely determined, with all O and H atoms located, and bears little resemblance to that of its lower hydrate. Again, all Te=Te dimers are parallel (|| ), and lie in layers || (200), which alternate with layers of KO8–9 polyhedra. Rb4[Te2O6(OH)4] · 10H2O (#626) is somewhat similar, but in this triclinic cell, the layering is || (110), Te=Te axes are || , and the Rb atoms are 8–10 coordinated and lie in less convoluted layers. Cs3[Te2O5(OH)5] · 4H2O (#627) has one Te=Te group per triclinic cell, aligned || , in a matrix of water molecules and CN10 Cs+ ions, while Cs4[Te2O4(OH)4] · 8H2O (#628) has a more obviously layered structure in which Te=Te groups pointing ||  and CsX9–10 polyhedra form sheets || (10), which are linked only through hydrogen bonds. K4Na2[Te2O8(OH)2] · 14H2O (#629) has a body-centred array of Te=Te dimers pointing || z, which bridge layers || (10) of water molecules, CN8 K+ and CN6 Na+ ions. K7Na[Te2O6(OH)4]2 · 12H2O (#630) has Te=Te dimers || x, lying in undulating layers || (020) that are connected weakly through CN6–10 K+ and CN6 Na+ ions.
The next two compounds are ditellurates of organic complexes. ((CH3)3Si)8[Te2O10] (#631) has a trimethylsilyl group strongly bonded (1 vu) to each of the eight non-bridging oxygen atoms of the Te=Te anion, to form a large neutral molecule that has only point symmetry 1, with two distinct Te and eight distinct Si atoms. Four such molecules in slightly different orientations pack together per unit cell (Fig. 42). (C(NH2)3)4[Te2O6(OH)4] (#632) is not molecular, having well-defined guanidinium cations [C(NH2)3]+ and tellurate anions [Te2O6(OH)4]4–. However, the structure is surprisingly complex, with four distinct C atoms and two types of Te=Te anion per unit cell. Te1 dimers pointing ||  form layers || (100), which alternate with layers that contain both Te2 dimers pointing ||  and one type of guanidinium group. The other three types of guanidinium lie between the Te layers.
Two structures have edge-sharing Te2X10 (Fig. 13b) in combination with monomeric TeX6 octahedra. K6[Te2O6(OH)4](TeO2(OH)4) · 12H2O (#633) has Te=Te dimers (Te2) aligned || , alternating along z with monomeric octahedra (Te1). The rows of Te anions lie between layers || (10) of CN8–10 K+ ions (Fig. 42). Cs2[Te2O4(OH)6] · (Te(OH)6) (#634) has Te=Te pointing ||  and lying in layers || (001), which alternate with layers of CN9 Cs+ ions and neutral Te(OH)6 molecules. Note that this is a special case of an orthotelluric acid adduct, more conventional examples of which are described as #376–431 above.
Thorneite, Pb6[Te2O10](CO3)Cl2 · H2O (#635, see front cover image), has Te=Te dimers pointing ||  or , lying in sheets || (200). The Te sheets have on each side of them PbO7Cl, PbO6Cl2 and PbO5Cl2 polyhedra which complete thick layers, and carbonate groups and water molecules lie between the layers (Fig. 42). K3[Ga(Te2O8(OH)2)] · H2O (#636) has Te=Te dimers alternating with GaO4 tetrahedra to form loop-branched vierer chains || z. The Te and Ga polyhedra between them form ‘double-triangle’ clusters; for a topologically similar chain made only of Te octahedra, see K2[Te3O8(OH)4] below (#658; Fig. 13i). The chains form an approximately hexagonal rod packing, and are held together through CN7–8 K+ ions and water molecules. Tl61+[Cu2+(Te2O10)] (#637) has chains that are very similar in topology, but the non-Te component is a CuO4 square rather than GaO4 tetrahedron. The four shortest Cu‒O distances are 1.96–1.98 Å, but a fifth neighbour to Cu at 2.52 Å provides an additional brace along the chain. The chains run || x and are arranged in a chequerboard fashion, leaving large square channels. Tl+ ions line the sides of these channels, and are in one-sided 5–8 coordination. (NH4)2V5+TeO6(OH) · H2O ≡ (NH4)4[(V5+O2)2(Te2O8(OH)2)] · 2H2O (#638) has a loop-branched dreier chain made from regular Te octahedra and very distorted V octahedra. Edge-sharing Te=Te and V=V pairs alternate along the chain || z. Every Te atom shares an oxygen ligand with each V of its adjacent dimer, and also a CN3 oxygen atom with both of them. V–O distances are 1.65–1.66 Å (non-bridging oxygen), 1.95–1.96 Å (CN2 bridging oxygen) and 2.16 and 2.47 Å (CN3 bridging oxygen), consistent with its description as the core of a [VO2]+ cation. The chains pack in a centred-rectangular array, with NH4+ ions and H2O molecules between them (Fig. 42).
Eckhardite, Ca2[Cu2(Te2O10)] · 2H2O (#639) has TeO6 and CuO4+2 polyhedra sharing edges to form layers || (101). The Cu ligands form a slightly twisted square at 1.96–2.02 Å, plus two more completing an elongated octahedron at 2.51 and 2.67 Å (Fig. 43). The oxygen atoms associated with all these polyhedra form a stepped, oblique slice through a hcp array, somewhat similar to the layer seen in the frankhawthorneite-related structures #512–515 above. However, while the frankhawthorneite layer ‘steps’ are edge-sharing ribbons which alternate between two octahedra in width (Cu=Cu) and one (Te), the eckhardite ribbons alternate between three octahedra (Cu==Cu) and two (Te=Te) in width. These run || y. Between the layers lie CN7 Ca2+ ions. Ag[(UO2)(Te2O8(OH)2)] (#640) has edge-sharing ribbons of UO7 pentagonal bipyramids running || x, which share CN3 oxygen atoms with Te2O10 dimers to form layers || (020). The layer topology contains ‘double-triangle’ clusters U < (Te=Te) > U. The Ag+ ions are in irregular 5-fold coordination between the layers. In K2[Ga2(Te2O10)] · 2H2O (#641), GaO4 tetrahedra and Te2O10 octahedral pairs share corners to form a three-dimensional framework. Two ligands of each Ga connect it to a neighbouring Te = Te dimer as part of a ‘double triangle’ Ga < (Te = Te) > Ga, while the other two ligands link Ga to two other dimers, as part of a [Ga‒Te‒Ga‒Te] 4-ring. The double triangles and 4-rings alternate, forming crankshaft ribbons || z, which are arranged in a centred-rectangular array and are connected into a framework via the remaining Ga‒O‒Te links. The framework is very open, with 6–8 Å diameter channels || x and z, which contain water molecules and CN9–11 K+ ions, and has very strong pseudotetragonal symmetry in projection down x (Fig. 43).
Schieffelinite, Pb10[Te2O8(OH)3]2(TeO2(OH)4)2(SO4) · 5H2O, and its chromate analogue chromschieffelinite (#642–643), have as their structural unit the corner-sharing Te2X11 dimer of Fig. 13c. The Te–Te dimers lie in undulating layers || (020), with CN8–10 Pb2+ ions on either side. Water molecules, additional TeX6 monomers and orientationally disordered SO42– or CrO42– tetrahedra lie between the layers.
We have two examples of cyclo anions made from Te6+ octahedra. K2[Te4O8(OH)10] (#644) has an edge-sharing pair of octahedra joining the two halves of a Te4X18 ‘double triangle’ tetramer (Fig. 13d). These isolated clusters are arranged in a herringbone pattern in layers || (100), with CN10 K+ ions between the layers. K8.5(H3O)0.5[Te6O18(OH)9] · 17H2O (#645) has ditrigonal rings Te6X27 in which Te octahedra alternately share corners and edges with their neighbours (Fig. 13e). Rings are in two different orientations with the ring plane always normal to z, and form pseudohexagonal columns running || z, which are arranged in a hexagonal rod packing, with hydronium and CN6–11 K+ ions between them. Successive layers of the rods || (200) are shifted by c, which reduces the symmetry to monoclinic.
Infinite polymers Tem6+Xn
In structures #646–677 Te6+ octahedra are linked to form infinite polymers (Table 25, deposited). Our first examples of ino tellurates have edge-sharing zweier chains Te2X8. In all cases, the shared edges of an octahedra are not cis to one another, so the chain zigzags (Fig. 13f). The database includes two polymorphs of Na2TeO4 ≡ Na4[Te2O8]. The Pbcn polymorph (#646) has Te chains || z, arranged in a centred-rectangular array. The chains are flanked by edge-sharing ribbons of NaO6 polyhedra which hold the structure together. Overall, the oxygen arrangement approximates hcp, with close-packed planes || (200) and (Na+Te) occupying ¾ of the octahedral interstices between each pair of layers. The P21/c polymorph (#647) has a slightly sheared version of the same sructure, with xmon = [ 0]orth and zmon = orth. The isostructural pair of compounds CaTeO4 and SrTeO4 (#648–649) have the same space group as #646 (Pbcn, although strongly pseudo-Cmnm), very similar cell parameters, and the same type of Te chain. However, only half of the Na sites of #646 are occupied by alkaline earth cations. (Ca,Sr)O6 and TeO6 octahedra together form open-branched zweier chains of edge-sharing octhedra running || z between each pair of close-packed anion layers, thus avoiding shared faces between (Ca,Sr) octahedra. This structure is shared with the pucherite polymorph of BiVO4 (Qurashi and Barnes, 1953) and alumotantite, AlTaO4 (Ercit et al., 1992) as well as synthetic MUO4 (M = Cr, Fe and Ni; Hoekstra and Marshall, 1967). A Te-rich variety of raspite, Pb[(W0.56Te0.44)O4] (#650) is included in this review, as the W:Te ratio is close to 1:1. Raspite is ideally PbWO4 (Fujita et al., 1977), and was discussed earlier, as one of the forms of Te4+V4+O4 is isostructural (#82, above). While Te4+ may occupy large, irregularly-coordinated sites like Pb2+, also a cation with a stereoactive lone pair, Te6+ readily enters octahedral sites, and so behaves analogously to W in raspite or V in TeVO4. The raspite structure is a monoclinically sheared derivative of the pucherite type, in which the anion arrangement is strongly perturbed away from hexagonal close-packing and the larger cation is irregularly 7-coordinated. The P21/a axial setting can be related back to the Pbcn setting of pucherite through the relations xpuch = zrasp, ypuch = rasp and zpuch = yrasp; a pucherite-like cell for raspite would have space group P1121/n, a = 5.59 Å, b = 13.03 Å, c = 5.02 Å and β = 96.1° (Fig. 43). KTeO3(OH) ≡ K2[Te2O6(OH)2] (#651) has similar cell parameters to #646–649 and a very similar arrangement of Te chains, although