research papers
Symmetry relations in wurtzite nitrides and oxide nitrides and the curious case of Pmc21
aStructure and Dynamics of Energy Materials, Helmholtz-Zentrum Berlin für Materialien und Energie, Hahn-Meitner-Platz 1, 14109 Berlin, Germany, bInstitute for Chemistry, Universität Potsdam, Karl-Liebknecht-Strasse 24/25, 14476 Potsdam, Germany, and cGeosciences, Freie Universität Berlin, Malteserstrasse 74-100, D-12249 Berlin, Germany
*Correspondence e-mail: joachim.breternitz@helmholtz-berlin.de
Binary III–V nitrides such as AlN, GaN and InN in the wurtzite-type structure have long been considered as potent semiconducting materials because of their optoelectronic properties, amongst others. With rising concerns over the utilization of scarce elements, a replacement of the trivalent cations by others in ternary and multinary nitrides has led to the development of different variants of nitrides and oxide nitrides crystallizing in lower-symmetry variants of wurtzite. This work presents the symmetry relationships between these structural types specific to nitrides and oxide nitrides and updates some prior work on this matter. The non-existence of compounds crystallizing in Pmc21, formally the highest of the wurtzite type fulfilling Pauling's rules for 1:1:2 stoichiometries, has been puzzling scientists for a while; a rationalization is given, from a of why this is unlikely to be adopted.
Keywords: group–subgroup relationships; nitride materials; wurtzite type.
1. Introduction
GaN and InN in particular are probably some of the most prominent and most important semiconductor materials; indeed, the 2014 physics Nobel Prize was awarded for the invention of blue-emitting GaN LEDs (Nanishi, 2014). As well as LEDs, other optoelectronic semiconductor devices, such as solar cells, have been realized using alloys of InN and GaN (e.g. Aliberti et al., 2010). However, In, which is needed for bandgap values suitable for visible-light absorption, is a very scarce element, accounting only for 0.16 p.p.m. of the earth's crust (Webelements, 2020). While the binary system is limited to trivalent cations to account for the triple negative nitride anion, variations on the cation charges can be realized through more complex substitutions of the cations, such that the overall charge is maintained. In the simplest case, two trivalent cations can, for instance, be replaced by one divalent and one tetravalent cation. This is, for instance, realized in the Zn–IV–V2 nitride materials ZnSiN2, ZnGeN2 and ZnSnN2 (Punya et al., 2011). However, more complex substitutions are also observed in compounds such as Li(+1)Al(+3)Si2(+4)N4(−3) (Ischenko et al., 2002) or Zn3(+2)Mo(+6)N4(−3) (Arca et al., 2018) to name just two. The situation can get even more complex when introducing O2− anions, and can lead to complex structure–composition relationships. The crystal structures of many of these materials, however, can be clearly linked to the wurtzite-type structure (Baur & McLarnan, 1982) and this symmetry relationship can, in turn, be used to rationalize some of the electronic properties of these materials, as the electronic structure is linked to the atomic structure for obvious reasons.
Relationships between ). With the wurtzite type being the i.e. the with the highest symmetry in the system, the lower-symmetry variants, the hettotypes, can be accessed in cascades of group–subgroup descents. While it is beyond the scope of this article to give a conclusive overview of crystallographic group theory, one point is eminently important: when lowering the symmetry from a group to a symmetry operations are only lost, and no other symmetry operations are added. This means that a in the can also be expressed in a but a that genuinely crystallizes in the type cannot be expressed in the group. The has higher which permits shifting of atoms (for instance, out of the centre of tetrahedra), or splitting of crystallographic sites, allowing occupation of different atom types with discrete ordering. One can also make the distinction between subgroups within the same i.e. without a change in point-symmetry operations, known as klassengleiche subgroups (abbreviated by k), or isomorphic subgroups (abbreviated by i) if group and belong to the same space-group type. These group–subgroup transitions are accompanied by an enlargement of the or a loss of unit-cell centring. If the translational symmetry is kept and only point-symmetry operations are lost, however, the subgroups are called translationengleich (abbreviated by t).1 The International Tables for Crystallography Volumes A1 and A (Wondratschek & Müller, 2004; Hahn, 2005) are a comprehensive tool for the establishment of relationships between space groups. A wide-ranging discussion of the group–subgroup relationships in wurtzite variants was published some time ago by Baur & McLarnan (1982), but it does, unfortunately, bear a few inaccuracies at some crucial points. Therefore, we set out to redevelop the symmetry relationships in the system specific to nitrides and oxide nitrides together with barely complete tables of nitrides and oxide nitrides in the different structural types as they appear in the Inorganic Database (ICSD); we will outline some of the difficulties and pitfalls that can arise in the analysis of the symmetry relations in this system.
types are based on the relationships of the underlying symmetries through the use of crystallographic group theory (Müller, 20132. An overview of the wurtzite-related structure types
The most powerful tool for the graphical representation of group–subgroup relationships was developed by Bärnighausen (1980), where the relationships are represented in the form of a tree diagram with the highest-symmetry structure standing at the top. Essentially, four subgroups of the wurtzite type are found amongst the wurtzite-derived nitrides and oxide nitrides besides the (Fig. 1). For the sake of completeness, Fig. 1 also contains the relationship between the wurtzite type and the Lonsdaleite type (hexagonal diamond), which can be understood as the prototype for the atom stacking in these materials.
The first complication of this system arises from the transition of the hexagonal wurtzite type in P63mc to its maximal translationengleiche orthorhombic Cmc21. This is because one needs to transition from the hexagonal coordinate system with angles α = β = 90°, γ = 120° to an orthogonal coordinate system with α = β = γ = 90°. This is symbolized as a, a + 2b, c in the Bärnighausen tree (Fig. 1), which relates the basis vectors of the to those of the group. It also corresponds to a complex transformation of atom coordinates between the two space groups which becomes necessary. Although no structure is observed in the maximal Cmc21, it forms an important link, as it is an intermediate to all lower-symmetry variants. If the atomic positions are given in decimals rather than fractions, they are no longer bound by symmetry to specific positions. The second complication arises because all subgroups are in orthorhombic space-group types, where one particular axis setting has been defined as the standard setting based on the symmetry operations. Müller (2013) advocates for the use of non-standard settings of space groups to avoid unit-cell transformations, but we use the standard settings of the space groups herein to facilitate the use of the work for the wider community. An example of this is the transition from Cmc21 to Pca21, where a and b are swapped and c is inverted.
While the intermediate Cmc21 allows for higher it still contains only two independent crystallographic sites, one for the anions and one for the cations. To accommodate different cations or anions on distinct crystallographic sites, the symmetry needs to be lowered even further. Lowering the symmetry, the cation and anion sites split in different ways to accommodate different ratios of cations and anions, but also to form different ordering patterns. It has been outlined before that Pauling's rules are a decisive factor in the way these materials are built (Baur & McLarnan, 1982; Quayle et al., 2015). It is interesting to note that the nominally highest-symmetry for such an octet-rule-obeying arrangement, Pmc21, is not found in any existing material.
3. The different types
3.1. Wurtzite type
The binary III–V nitrides of main group 3 (apart from BN), namely AlN, GaN, InN and TlN, crystallize in the wurtzite-type , Table 1). It is worth noting that the wurtzite-type structure is non-centrosymmetric, i.e. it lacks a centre of inversion as Consequently, all its hettotypes are non-centrosymmetric too. Besides the pure binaries, a vast number of binary alloy compounds exist, where the cations are disordered on the single crystallographic cation site. This feature is used for the effective bandgap tuning in these compounds, for instance in the system Ga(1−x)InxN (Jakkala & Kordesch, 2017). Not only do III–V nitrides form wurtzite-type crystal structures, but nominally ternary systems, such as BeSiN2 (Schneider et al., 1979) and ZnGeN2 (Larson et al., 1974), have been reported to adopt the disordered wurtzite type. In fact, we recently showed that introduction of oxygen on the anion sites leads to the formation of Zn1+xGe1−xN1−yOy compounds in the wurtzite type with disordered cations and anions (Breternitz et al., 2019). This is in line with multi-cation oxide nitrides that still adopt the wurtzite such as Cd0.25Zn1.13Ge0.62ON (Capitán et al., 2000).
(Fig. 23.2. The Na2SiO3 type
When the cation site is occupied by more than one atom type, an ordered case is normally energetically and geometrically favourable. The latter is easy to quantify through the ionic size of the different cations. If the cations are disordered on one crystallographic site, the difference in coordination environment between the different atom types is only possible within rather strict borders. Therefore, a tendency of the cations to order reduces the strain on the
combined with an energetic advantage.To allow for an ordering of different cation types on different crystallographic sites, the positions need to be split in agreement with the ratio of the cations in the material. Take Zn2PN3, for instance (Fig. 3). The Zn:P ratio is 2:1 and hence a splitting of the crystallographic sites into a different ratio would unavoidably cause disorder on some positions. The simplest 2:1 splitting of the cation 4a site in Cmc21, as the maximal of the wurtzite type, is through an isomorphic symmetry descent of index three (Fig. 1). Thereby, the 4a site is split into one special 4a position with x = 0, i.e. lying in the bc plane, and one 8b general position (x, y, z) that imposes no restrictions on the atom positions. Given the multiplicity of the two different sites, there are only half as many atoms on the 4a positions as on the 8b positions in this arrangement, which are being filled 2:1 in the Na2SiO3 type (Table 2). The Zn and P atoms in Zn2PN3 occupy the Wyckoff positions 8b and 4a, respectively. It is worth mentioning that the cation site splitting goes along with an anion site splitting. In pure nitrides, these two anion positions are both occupied by nitrogen, but the site splitting allows for a deviation of the coordination environment for both cation sites.
|
Taking Zn2PN3 as an example case for the Na2SiO3 type (Fig. 3), it can be easily depicted that the PN4 tetrahedra form strands along the b directions, which is an effect of the special 4a positions on which phosphorus resides, whereas the ZnN4 tetrahedra are interconnecting the strands in three dimensions. It is interesting to note that all nitrogen atoms are connected to phosphorus atoms as well as to zinc atoms, with the nitrogen on 4a being connected to two P5+ and two Zn2+ and the one on 8b to one P5+ and three Zn2+. Therefore, both positions do not strictly fulfil the octet rule as discussed previously, but only approximate Pauling's rules (George et al., 2020) with bond strengths of +3.5 and +2.75, respectively.
3.2.1. The ordered defect variant Si2N2O
The Si2N2O type (Fig. 4 and Table 3) can be seen as a special case of the Na2SiO3 type, since both crystallize in the same – and would be located at the same position in the Bärnighausen tree – but are different in the occupation of the crystallographic sites. The 4a position remains unoccupied, as compared with the Na2SiO3 type. However, since this class of compounds bears two distinct anions, the anion site splitting mentioned above plays an important role here in that the two crystallographically independent anion sites are occupied by nitrogen (8b) and oxygen (4a). In this particular arrangement, every oxygen atom is neighbouring two silicon atoms, whereas every nitrogen atom is neighbouring three silicon atoms and thereby obeying Pauling's rules with formal bond strengths of 2 and 3, respectively.
|
Finally, SiPN3, which should be more correctly written as (Si0.5P0.5)2N3 since Si and P are sharing sites, can be viewed as an intermediate between the Na2SiO3 type and the Si2N2O type, since it only contains one sort of anion, which it shares with the Na2SiO3 type, but exhibits an unoccupied 4a cation position like the Si2N2O type. In fact, both cations share the general 8b position and show no particular order as observed by Baldus et al. (1993). However, this result was determined on the basis of X-ray and electron diffraction only, but P5+ and Si4+ are isoelectronic ions and would hence scatter in much the same way and therefore not allow a truly reliable determination.
3.3. The β-NaFeO2 type
When it comes to a 1:1 ratio of cations, there has been much discussion of the formally highest-symmetry Pmc21. This is a maximal klassengleiche of Cmc21 and is a in the transition to the enargite-type structure (see also Figs. 1 and 6). However, this has not been observed and nitrides with a cation ratio of 1:1 are reported to crystallize in the β-NaFeO2 type in the Pna21, which has a twice as large as that of the hypothetical Pmc21 structure. Since this phenomenon has led to some confusion, we will discuss it in more detail in Section 4.
that complies with Pauling's rules inOne complication that makes the direct comparison of this structure sometimes difficult is that all atoms in the β-NaFeO2-type structure lie on general 4a positions. Therefore, the choice of the unit-cell origin is arbitrary in this system and may necessitate a shift of the experimentally determined coordinates to reveal the group–subgroup relationship derived ones, as is illustrated in the table in Fig. 5. This can either be performed manually or through automatic tools, such as the program COMPSTRU (de la Flor et al., 2016). Since this structure obeys Pauling's rules, every anion is surrounded by two cations of every sort to equalize the charges on every anion position [Fig. 5(b)]. The fact that all atoms lie on general positions further allows a high structural flexibility accommodating cations of distinctly different sizes, for instance (Table 4).
|
One particular case of the β-NaFeO2 type is the zinc nitride halides. In principle, there are two possible ways to view them in terms of the structure type: one either regards them as a special case, where the two cation types are the same and the anion sites are occupied by different atoms, or as anti-β-NaFeO2 type, where cations and anions switch sites. Since all atoms lie on a general position and have the same coordination, both ways lead to the structure and are hence interchangeable from a structural point of view. However, the authors would argue for the latter case as anti-β-NaFeO2 type, since one further point needs to be considered: the occupation of the anion positions by two different types with different charges can only obey Pauling's rule if the cation positions are filled by cations of the same charge, in the simplest case the same cation sort. This mutual dependency is the exact opposite for the pure nitrides in the β-NaFeO2 type and the nitride halides should hence be regarded as anti-β-NaFeO2 type.
3.4. α-LiSiON type
As the nitride halides demonstrate, the β-NaFeO2 type does not allow for an occupation of the cation sites and the anion sites with differently charged ions, while obeying Pauling's rules at the same time; instead, a different arrangement of the tetrahedra needs to be achieved. The number of next neighbours of the distinct crystallographic sites needs to be different for the different sites. This is achieved through a different symmetry descent from the common intermediate into the Pca21 in the α-LiSiON-type structure (Fig. 5). The cations form planes in the ac plane and so do the anions. The effect of this is that every oxygen atom has three lithium and one silicon neighbour, while the nitrogen atoms have three silicon and one lithium neighbour. While the formal bond strengths of 1.75 and 3.25, respectively, do not perfectly obey Pauling's rule, they are considerably closer to the expected values than the 2.5 throughout the β-NaFeO2 type. Given the rather special arrangement in this class, only two compounds, namely LiSiON (Laurent et al., 1981) and KGeON (Guyader et al., 1983), have been experimentally observed in this structure type.
3.5. The enargite type
Finally, one further structure type is observed in the nitride wurtzite system. Instead of a simple 1:1, or a 2:1 splitting of the crystallographic sites as observed in the β-NaFeO2 type and the Na2SiO3 type, cation and anion positions are split into three crystallographic sites (Fig. 6) with different multiplicities and thereby allowing a 3:1 occupation of the cations or anions. Taking Na3MoO3N as an example, sodium occupies a 2a and a 4b position, while Mo occupies a 2a position, with the situation for the anions being analogous. Therefore, the MoN4 tetrahedra are isolated within the and completely surrounded by NaN4 tetrahedra (Fig. 7). Taking a simple view with Pauling's rules fails in this structure type, since all anions are surrounded by three sodium cations and one molybdenum cation. This is probably due to the different nature of the Mo—N/O bonding versus the Na—N/O bonding, with the former being distinctly more covalent. Therefore, a careful consideration of the bond lengths in addition to the simple counting of nearest neighbours is necessary to rationalize this ordering. This situation is even more obvious in the compound Li3SO3N, where the S—O/N bonding is clearly covalent, while the Li—O/N bonding is expected to be mostly ionic. Although this structure type shows a considerable degree of complexity, four compounds, Li3SO3N (Kurzman et al., 2013), Na3MoO3N (Arumugam et al., 2003), Na3WO3N (Elder et al., 1994) and Zn3MoN4 (Arca et al., 2018), span the range for this class from oxide nitrides to pure nitrides.
4. The curious case of Pmc21
In the light of the complexity of group–subgroup relationships and the fact that they are probably rarely taught, problems sometimes occur in the literature that are prone to being replicated in work based on it. They are, however, crucially important for a proper understanding of symmetry relations, and we would like to showcase this for the case of Pmc21, a that has been postulated for ternary nitrides for a long time (e.g. Baur & McLarnan, 1982; Quayle et al., 2015; Quayle, 2020).
As highlighted above, a number of research articles have pointed out that the highest-symmetry ABN2 nitride could crystallize and that obeys Pauling's rules is not the β-NaFeO2 type, but a in Pmc21. This postulated can be found as an intermediate in the descent to the enargite type (Figs. 1 and 6). The difference between the hypothetical in Pmc21 and the β-NaFeO2 type is, indeed, only found in the relative arrangement of the different cations to each other, which led Baur and McLarnan to speculate in 1982 that the energy difference between the two conformations may be small. Quayle et al. (2015) calculated the energy difference between the observed β-NaFeO2-type structure and the hypothetical Pmc21 structure to be only 13 meV per formula unit in the case of ZnSnN2. From this point of view, it is considered interesting why the Pmc21 structure has not been observed (and in fact has not been observed for any of the wurtzite series of materials).
of the wurtzite type in which anFrom a crystallographic and crystal chemistry point of view, it is not quite as surprising to find that the structure in Pmc21 is not observed. Regarding the group–subgroup relationship as outlined in Figs. 1 and 6, it is evident that the crystallographic sites of the cations and anions split from 2b Wyckoff sites in the hexagonal wurtzite into 2a and 2b sites in Pmc21. It is important to remember that Wyckoff sites not only show the multiplicity, but also indicate the (which is m for both Wyckoff sites). In fact, the 2a and 2b sites in Pmc21 are special positions with the coordinates (0, y, z) and (½, y, z), respectively. This essentially means that the two sites can accommodate different atom types, but that they are bound to lie in the bc plane (Fig. 8). This restriction does not only apply to the cations, but is true for the anions too. The restriction to the bc plane means that the M—N distances (M, cations) outside the bc plane are critically dependent on the a-axis length and this is the same for both types of cations. In essence, the tetrahedral coordination can either strongly distort into a disphenoid,2 or will have to remain very similar for both ions. This arrangement cannot be favourable, since a distortion would create a strongly anisotropic bonding situation, which is hardly energetically favourable, and identical tetrahedron sizes for cations of different size would be peculiar, as this is one of the main drivers for ordering. This is in line with the observation of Quayle et al. (2015) that the energy difference between the β-NaFeO2-type structure and the hypothetical Pmc21 structure is larger for ZnGeN2 than for ZnSnN2, as the Shannon radii of Zn2+ (0.6 Å) and Ge4+ (0.39 Å) are more different than those of Zn2+ and Sn4+ (0.55 Å) (Shannon, 1976).
One further thing should be considered. It is not the atoms that follow the symmetry of the Pmc21 is not a likely choice, as outlined above, the particular arrangement of tetrahedra in this hypothetical structure could well exist. To rectify the symmetry restriction, one needs to lose the mirror planes perpendicular to the a axis. This could be done in a klassengleiche descent of index two into Pca21 with a doubling of the b axis, or through translationengleiche descents into one of the monoclinic space groups Pc or P21. However, no that would correspond to one of these subgroups has been reported in the ICSD.
but the that reflects the symmetry of the atomic arrangement. While the5. Conclusions
The structural variability of nitrides and oxide nitrides in the wurtzite type and its subgroups is rich and gives rise to many different properties that can be attained in particular arrangements. Combining and comparing these different arrangements and putting them into their group–subgroup relationship can greatly aid the interpretation, but it needs to be performed with great care. In particular, group–subgroup relationships, as they exist in this family, between the hexagonal and the orthorhombic Pmc21, although much speculation has been devoted to this point. A thorough understanding of the relationship between electronic and atomic structures must, from our perspective, be preceded by a thorough understanding of the atomic structures themselves.
can be difficult due to the change of basis vectors, as well as those within the orthorhombic due to different unit-cell settings. We have developed the relationship between the most important structural types that are found for wurtzite and wurtzite-derived nitrides and explained the differences between them, which mostly reside in the cation arrangements relative to each other. Finally, we showcased why ternary nitrides with a 1:1:2 stoichiometry are unlikely to adopt a structure in theFootnotes
1There is a degree of overlap in the nomenclature between types t, k, i and indices (Müller, 2013; p. 135) and the corresponding symmetry reductions (e.g. Müller, 2013; p. 134). Since both are so closely related to each other, we argue that a term like t3 could be synonymously used for a translationengleiche of index three as well as a translationengleiche symmetry descent of index three.
2We note that, in the orthorhombic the tetrahedra are all strictly speaking distorted into disphenoids, since they are no longer bound by symmetry to maintain equilateral triangle faces. As the distortions are generally small in these compounds, and in order to allow a direct comparison of the higher-symmetry structures with the lower-symmetry ones, we use the term tetrahedron throughout this article, whether it is a regular or irregular polyhedron.
Acknowledgements
The authors would like to thank Dr Dennis Wiedemann and Dr Paul C. Quayle for helpful discussion. Open access funding enabled and organized by Projekt DEAL.
References
Aliberti, P., Feng, Y., Takeda, Y., Shrestha, S. K., Green, M. A. & Conibeer, G. (2010). J. Appl. Phys. 108, 094507. Web of Science CrossRef Google Scholar
Arca, E., Lany, S., Perkins, J. D., Bartel, C., Mangum, J., Sun, W., Holder, A., Ceder, G., Gorman, B., Teeter, G., Tumas, W. & Zakutayev, A. (2018). J. Am. Chem. Soc. 140, 4293–4301. Web of Science CrossRef ICSD CAS PubMed Google Scholar
Aroyo, M. I., Kirov, A., Capillas, C., Perez-Mato, J. M. & Wondratschek, H. (2006). Acta Cryst. A62, 115–128. Web of Science CrossRef CAS IUCr Journals Google Scholar
Aroyo, M. I., Perez-Mato, J., Orobengoa, D., Tasci, E., de la Flor, G. & Kirov, A. (2011). Bulg. Chem. Commun. 43, 183–197. CAS Google Scholar
Aroyo, M. I., Perez-Mato, J. M., Capillas, C., Kroumova, E., Ivantchev, S., Madariaga, G., Kirov, A. & Wondratschek, H. (2006). Z. Kristallogr. Cryst. Mater. 221, 15–27. Web of Science CrossRef CAS Google Scholar
Arumugam, N., Hönnerscheid, A. & Jansen, M. (2003). Z. Anorg. Allg. Chem. 629, 939–941. Web of Science CrossRef ICSD CAS Google Scholar
Bacher, P., Roult, G., Ghers, M., Merdrignac, O., Guyader, J. & Laurent, Y. (1989). Mater. Chem. Phys. 21, 223–235. CrossRef ICSD CAS Web of Science Google Scholar
Baldus, H. P., Schnick, W., Luecke, J., Wannagat, U. & Bogedain, G. (1993). Chem. Mater. 5, 845–850. CrossRef ICSD CAS Web of Science Google Scholar
Bärnighausen, H. (1980). MATCH Commun. Math. Chem. 9, 139–175. Google Scholar
Baur, W. H. & McLarnan, T. J. (1982). J. Solid State Chem. 42, 300–321. CrossRef CAS Web of Science Google Scholar
Belousov, A., Katrych, S., Hametner, K., Günther, D., Karpinski, J. & Batlogg, B. (2010). J. Cryst. Growth, 312, 2585–2592. Web of Science CrossRef ICSD CAS Google Scholar
Breternitz, J., Wang, Z., Glibo, A., Franz, A., Tovar, M., Berendts, S., Lerch, M. & Schorr, S. (2019). Phys. Status Solidi A, 216, 1800885. Web of Science CrossRef Google Scholar
Bruls, R. J., Hintzen, H. T., Metselaar, R. & Loong, C.-K. (2000). J. Phys. Chem. Solids, 61, 1285–1293. Web of Science CrossRef ICSD CAS Google Scholar
Capitán, M. J., Picard, C. L., Laurent, Y. & Odriozola, J. A. (2000). Mater. Sci. Forum, 325–326, 25–30. Google Scholar
Chen, D. P., Neuefeind, J. C., Koczkur, K. M., Bish, D. L. & Skrabalak, S. E. (2017). Chem. Mater. 29, 6525–6535. Web of Science CrossRef ICSD CAS Google Scholar
Eckerlin, P. (1967). Z. Anorg. Allg. Chem. 353, 225–235. CrossRef ICSD CAS Web of Science Google Scholar
Elder, S. H., DiSalvo, F. J., Parise, J. B., Hriljac, J. A. & Richardson, J. W. (1994). J. Solid State Chem. 108, 73–79. CrossRef ICSD CAS Web of Science Google Scholar
Endo, T., Takizawa, H. & Shimada, M. (1992). Proceedings of the 16th Annual Conference on Composites and Advanced Ceramic Materials, Part 2 of 2: Ceramic Engineering and Science Proceedings, Vol. 13, pp. 844–851. Westerville, Ohio, USA: John Wiley & Sons, Inc. Google Scholar
Esmaeilzadeh, S., Hålenius, U. & Valldor, M. (2006). Chem. Mater. 18, 2713–2718. Web of Science CrossRef ICSD CAS Google Scholar
Flor, G. de la, Orobengoa, D., Tasci, E., Perez-Mato, J. M. & Aroyo, M. I. (2016). J. Appl. Cryst. 49, 653–664. Web of Science CrossRef IUCr Journals Google Scholar
George, J., Waroquiers, D., Di Stefano, D., Petretto, G., Rignanese, G.-M. & Hautier, G. (2020). Angew. Chem. Int. Ed. 59, 7569–7575. Web of Science CrossRef CAS Google Scholar
Guyader, J., Jacquet, R., Malhaire, J. M., Roult, G. & Laurent, Y. (1983). Rev. Chim. Miner. 20, 863–870. CAS Google Scholar
Guyader, J., L'Haridon, P., Laurent, Y., Jacquet, R. & Roult, G. (1984). J. Solid State Chem. 54, 251–255. CrossRef ICSD CAS Web of Science Google Scholar
Hahn, T. (2005). International Tables for Crystallography, Vol. A. Heidelberg: Springer. Google Scholar
Häusler, J., Eisenburger, L., Oeckler, O. & Schnick, W. (2018). Eur. J. Inorg. Chem. 2018, 759–764. Google Scholar
Häusler, J., Neudert, L., Mallmann, M., Niklaus, R., Kimmel, A. L., Alt, N. S. A., Schlücker, E., Oeckler, O. & Schnick, W. (2017). Chem. Eur. J. 23, 2583–2590. Web of Science PubMed Google Scholar
Häusler, J., Niklaus, R., Minár, J. & Schnick, W. (2018). Chem. Eur. J. 24, 1686–1693. Web of Science PubMed Google Scholar
Häusler, J., Schimmel, S., Wellmann, P. & Schnick, W. (2017). Chem. Eur. J. 23, 12275–12282. Web of Science PubMed Google Scholar
Ischenko, V., Kienle, L. & Jansen, M. (2002). J. Mater. Sci. 37, 5305–5317. Web of Science CrossRef ICSD CAS Google Scholar
Jacobs, H. & Mengis, H. (1993). Eur. J. Solid State Inorg. Chem. 30, 45–53. CAS Google Scholar
Jakkala, P. & Kordesch, M. E. (2017). Mater. Res. Express, 4, 016406. Web of Science CrossRef Google Scholar
Jorgensen, J. D., Srinivasa, S. R., Labbe, J. C. & Roult, G. (1979). Acta Cryst. B35, 141–142. CrossRef ICSD CAS IUCr Journals Web of Science Google Scholar
Kurzman, J. A., Jouan, G., Courty, M., Palacín, M. R., Armand, M. & Recham, N. (2013). Solid State Sci. 25, 28–32. Web of Science CrossRef ICSD CAS Google Scholar
Larson, W. L., Maruska, H. P. & Stevenson, D. A. (1974). J. Electrochem. Soc. 121, 1673–1674. CrossRef ICSD CAS Web of Science Google Scholar
Laurent, Y., Guyader, J. & Roult, G. (1981). Acta Cryst. B37, 911–913. CrossRef ICSD CAS Web of Science IUCr Journals Google Scholar
Li, Y., Liu, X. & Dronskowski, R. (2016). Inorganics, 4, 29. Web of Science CrossRef ICSD Google Scholar
Li, Y. Q., Hirosaki, N., Xie, R. J., Takeda, T. & Mitomo, M. (2008). Chem. Mater. 20, 6704–6714. Web of Science CrossRef ICSD CAS Google Scholar
Li, Y. Q., Hirosaki, N., Xie, R. J., Takeka, T. & Mitomo, M. (2009). J. Solid State Chem. 182, 301–311. Web of Science CrossRef ICSD CAS Google Scholar
Lindqvist, O., Sjöberg, J., Hull, S. & Pompe, R. (1991). Acta Cryst. B47, 672–678. CrossRef ICSD CAS Web of Science IUCr Journals Google Scholar
Liu, T.-C., Kominami, H., Greer, H. F., Zhou, W., Nakanishi, Y. & Liu, R.-S. (2012). Chem. Mater. 24, 3486–3492. Web of Science CrossRef ICSD CAS Google Scholar
Liu, X., Wessel, C., Pan, F. & Dronskowski, R. (2013). J. Solid State Chem. 203, 31–36. Web of Science CrossRef ICSD CAS Google Scholar
Müller, U. (2013). Symmetry Relationships between Crystal Structures. Oxford University Press. Google Scholar
Nanishi, Y. (2014). Nat. Photon. 8, 884–886. Web of Science CrossRef CAS Google Scholar
Orth, M. & Schnick, W. (1999). Z. Anorg. Allg. Chem. 625, 1426–1428. CrossRef CAS Google Scholar
Paszkowicz, W., Černý, R. & Krukowski, S. (2003). Powder Diffr. 18, 114–121. Web of Science CrossRef ICSD CAS Google Scholar
Paszkowicz, W., Podsiadło, S. & Minikayev, R. (2004). J. Alloys Compd. 382, 100–106. Web of Science CrossRef ICSD CAS Google Scholar
Pilyankevich, A. N., Shaginyan, L. R. & Andreeva, A. F. (1974). Bull. Acad. Sci. USSR Phys. Ser. (Engl. Transl.), 38, 163–165. Google Scholar
Punya, A., Paudel, T. R. & Lambrecht, W. R. L. (2011). Phys. Status Solidi C, 8, 2492–2499. CrossRef CAS Google Scholar
Quayle, P. C. (2020). Acta Cryst. A76, 410–420. Web of Science CrossRef ICSD IUCr Journals Google Scholar
Quayle, P. C., Blanton, E. W., Punya, A., Junno, G. T., He, K., Han, L., Zhao, H., Shan, J., Lambrecht, W. R. L. & Kash, K. (2015). Phys. Rev. B, 91, 205207. Web of Science CrossRef Google Scholar
Schneider, G., Gauckler, L. J. & Petzow, G. (1979). Ceramurgia Int. 5, 101–104. CrossRef ICSD CAS Google Scholar
Schultz-Coulon, V. & Schnick, W. (1997). Z. Anorg. Allg. Chem. 623, 69–74. CAS Google Scholar
Schulz, H. & Thiemann, K. H. (1977). Solid State Commun. 23, 815–819. CrossRef ICSD CAS Web of Science Google Scholar
Sedlmaier, S. J., Eberspächer, M. & Schnick, W. (2011). Z. Anorg. Allg. Chem. 637, 362–367. Web of Science CrossRef ICSD CAS Google Scholar
Shannon, R. D. (1976). Acta Cryst. A32, 751–767. CrossRef CAS IUCr Journals Web of Science Google Scholar
Sjöberg, J., Helgesson, G. & Idrestedt, I. (1991). Acta Cryst. C47, 2438–2441. CrossRef ICSD Web of Science IUCr Journals Google Scholar
Suehiro, T., Tansho, M. & Shimizu, T. (2017). J. Phys. Chem. C, 121, 27590–27596. Web of Science CrossRef ICSD CAS Google Scholar
Watanabe, H. & Kijima, N. (2009). J. Alloys Compd. 475, 434–439. Web of Science CrossRef ICSD CAS Google Scholar
Webelements (2020). https://www.webelements.com/indium/geology.html, accessed 20/05/2020. Google Scholar
Wondratschek, H. & Müller, U. (2004). International Tables for Crystallography, Vol. A1. Dordrecht: Kluwer Academic Publishers. Google Scholar
Yashima, M., Maeda, K., Teramura, K., Takata, T. & Domen, K. (2005). Chem. Phys. Lett. 416, 225–228. Web of Science CrossRef ICSD CAS Google Scholar
Zhang, Q.-H., Wang, J., Yeh, C.-W., Ke, W.-C., Liu, R.-S., Tang, J.-K., Xie, M.-B., Liang, H.-B. & Su, Q. (2010). Acta Mater. 58, 6728–6735. Web of Science CrossRef ICSD CAS Google Scholar
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.