Rare-Earth Pnictide Oxides - American Chemical Society


Rare-Earth Pnictide Oxides - American Chemical Societypubs.acs.org/doi/pdf/10.1021/acs.chemmater.7b03996by S Forbes - â€...

7 downloads 151 Views 3MB Size

Review Cite This: Chem. Mater. 2017, 29, 9605-9612

pubs.acs.org/cm

Rare-Earth Pnictide Oxides (RE,Ca)mPnnOm (Pn = Sb, Bi): A Review of Crystal Structures, Chemistry, Compositions, and Physical Properties Scott Forbes‡ and Yurij Mozharivskyj*,† ‡

Department of Chemistry, Princeton University, Princeton, New Jersey 08544, United States Department of Chemistry and Chemical Biology, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4M1, Canada



ABSTRACT: Rare-earth-based pnictide oxides (RE,Ca)mPnnOm (Pn = Sb, Bi) adopt a variety of interrelated structures based on specific stacking sequences of rare-earth oxide sublattices. These phases are all prepared via high temperature reactions and may be selectively produced in high purity. X-ray single crystal studies also revealed the disorder of pnictogen atoms in the non-charge balanced structures, which gives rise to unexpected physical properties. It is theorized that Anderson localization of the Pn p states results in an activation energy which forces semiconductortype behavior in these phases. Interestingly, the extent of the localization of states is directly dependent on the magnitude of the disorder of the pnictogen atom, which also depends on pnictogen and/or the rare-earth atom that is used. By altering the composition of the sample with respect to the choice of Pn and/or RE, the physical properties of these select phases may be tuned to the desirable level without a change to the overall structure. The progression of this series of phases with respect to chemistry, structure, and physical properties is reviewed and discussed.



variety of the A/RE−M−Pn−O pnictide oxides21−28 when compared to the RE−Pn−O pnictide oxides. The A/RE−M− Pn−O phases can be divided into two groups, with or without fluorite-type [MPn] layers.29 The one with the [MPn] layers is much larger and features five structure types (vs four for the other group), with the ZrCuSiAs-type oxides having the most representatives. Since the RE−Pn−O phases lack a d metal, their structures are different from those of A/RE−M−Pn−O. Still, some similarities are present; e.g. the RE2PnO2 phases (anti-ThCr2Si2 structure) contain antifluorite-type [REO] layers, which are also present in the A/RE−M−Pn−O phases with the ZrCuSiAs-, Th2 Ni 3−x P 3 O-, La 3 Cu 4 P 4 O 2 -, and U2Cu2As3O-type structures. The Eu4Pn2O oxides crystallize with the anti-K2NiF4-type structure, whose ordered version is adopted by Na2Ti2As2O and Na2Ti2Sb2O.30 In the past few years, a family of new rare-earth pnictide oxides with the general formula (RE,Ca)mPnnOm has been discovered, triggering an interest in studying their structures as well as their relationships between one another in terms of chemistry and physical properties. The RE3SbO3 and RE8Sb3O8 phases were the first examples reported,1 and they are a part of a series that is structurally related to the RE2PnO2 phases. Experiments involving the substitution of rare-earth with heteroatoms also led to the discovery of the (RE′RE″)3SbO3

INTRODUCTION Rare-earth pnictide oxides (RE,Ca)mPnnOm (Pn = Sb, Bi) are an interesting class of materials which have been discovered to possess a series of different phases based on stacking sequences and site mixing.1−3 Such trends are well documented for other series, such as perovskites or Ruddlesden−Popper phases.4−8 The (RE,Ca)mPnnOm phases can be imagined as chemical fusions of the rare-earth pnictides (and some fraction of calcium pnictide) with rare-earth oxide, resulting in atomic frameworks with unique structural features and charge transport properties. For our original research, it was hoped that these new phases would yield physical properties suitable for thermoelectric applications, but unfortunately most rareearth pnictide oxides possess high electrical resistivities. Nevertheless, these phases are fascinating from a structural perspective, and the trends discovered from their relation to one another may be used to predict the formation of other phases. Many examples of RE−Pn−O phases are known, but most of them contain the pnictogen atom in the +3 or +5 oxidation state, e.g., RESb3O9,9 RE3Sb5O12,9 RE3SbO7,10 RESbO4,11 REBiO3,12 and RE10Bi8O27.13 Since the energy separation between the occupied O2− p states and empty Pn states is large, these materials are wide band gap insulators. Until recently, there were only three cases of phases containing Pn3−: Eu4Pn2O,14,15 RE2PnO216−19 (Pn = Sb, Bi), and RE9Sb5O520 phases. Incorporation of a d metal, M, and alkaline or alkalineearth metal, A, significantly expands the number and structural © 2017 American Chemical Society

Received: September 20, 2017 Revised: October 30, 2017 Published: October 31, 2017 9605

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials Table 1. Optimal Heating Trends for the Formation of the (RE,Ca)−Pn−O Phases system

loading composition

T, °C

time, h

product

RE3SbO3 Gd3BiO3 RE8Sb3O8 Gd8Bi3O8 (RE′RE″)3SbO3 RE2SbO2 RE2BiO2 CaRE3SbO4 Ca2RE8Sb3O10 RE9Sb5O5 Ca2RE7Sb5O5 Ca2RE7Bi5O5

RESb + RE2O3 41/3GdBi + 11/3Gd2O3a 3RESb + 3RE2O3 or 22/3RESb + 3Sb + 22/3RE2O3 3GdBi + 3Gd2O3b 1 /2RE′Sb + 1/2RE″Sb + 1/2RE′2O3 + 1/2RE″2O3 2RESb + Sb + 2RE2O3 2REBi + Bi + 2RE2O3 CaO + RESb + RE2O3 Ca + CaO + 2RESb + Sb + 3RE2O3 15RESb + 5RE2O3 + 2RE 2CaO + 5RESb + RE2O3 2CaO + 5RESb + RE2O3

1500−1600 1300 1350−1500 1300 1550 1500 1500 1300 for bulk, 1600 for crystals 1300 ∼1500 1350 1275

2 24 6 24 10 16 16 2 48 48 18 18

black, solid unknowna black, solid gray, solid black, molten black, solid black, solid black, solid at 1300 °C, molten at 1600 °C black, solid black, molten gray, molten gray, molten

a Can only be formed in trace amounts with an excess of GdBi. As a result, the color is unknown. bCan be prepared in good purity through extended heat treatment via deposition of impurities.

(RE′ and RE″ are two different rare earth with the same 3+ oxidation state),3 CaRE3PnO4, and Ca2RE8Pn3O10 phases,31 which were found to belong to the same structural family. Furthermore, the role of calcium in RE−Pn−O phases was investigated further by substitution in RE9Pn5O5, yielding the Ca2RE7Pn5O5 series.32 Clearly there is potential for a wide variety of phases with different RE−O and RE−Pn frameworks to exist.



SYNTHESIS OVERVIEW Due to the fact that the (RE,Ca)mPnnOm phases incorporate very stable RE−O and RE−Pn bonds, high temperatures are required for their formation (T ≥ 1000 °C). Furthermore, these phases contain antimony and bismuth, both of which have high vapor pressure at elevated temperatures. As such, the synthetic approaches for the production of rare-earth pnictide oxides must take into consideration the need for high temperature without the risk of damaging equipment due to sublimation of the elements. For this reason, the rare-earth pnictide (REPn) binaries are used whenever possible to minimize any antimony or bismuth lost. These binaries are simple to produce and can be made by simply mixing the elements and heating in a carefully controlled manner. To avoid unwanted reactions with the equipment used, all samples are generally sealed inside tantalum ampules using an arc melter. Tantalum is chosen due to its high melting point and relative chemical inertness. Once suitably prepared, samples are heated using an induction furnace at the required temperature and time until a pure phase is produced. The choice of composition and reaction conditions is crucial as subtle deviations may lead to the formation of the undesired product, due to the similarities in structures between rare-earth pnictide oxides. A summary of ideal conditions for preparing each of the (RE,Ca)mPnnOm series is presented in Table 1. Synthetic conditions used to prepare RE9Sb5O5 and Ca2RE7Pn5O5 are also provided. Figure 1 is a visual presentation of the synthetic conditions from Table 1, and it also provides additional information. For example, (RE′RE″)3SbO3 phases can be prepared with either the monoclinic RE3SbO3 structure when the RE′/RE″ ratio deviates from 1 or the tetragonal structure when the ratio is close to 1. Both the monoclinic RE3PnO3 and tetragonal (RE′RE″) 3 SbO 3 oxides will transform into monoclinic RE8Pn3O8 upon extended annealing at the same or lower temperature via some loss of RE and O. In fact, the RE8Pn3O8

Figure 1. Schematic representation of the synthetic conditions for the (RE,Ca)-Pn-O phases.

phases can only be obtained starting from the RE3PnO3 loading compositions.1,33 Similarly, the CaRE3SbO4 phases will transform into the structurally related Ca2RE8Sb3O10 phases after a longer heat treatment. In most cases, rare-earth pnictide oxides may be prepared for any rare earth up to dysprosium, with the exception of europium due to its different chemistry. Cases of rare-earth oxides with RE = Tm, Yb, or Lu are practically unknown. The reason for this may be twofold: the smaller rare-earth atoms present result in a smaller unit cell with increased strain on the bonding geometry (i.e., deviations in rigid RE−O and RE−Pn bonds) as well as the higher temperature stability of the corresponding rare-earth pnictide and oxide precursors. Reactions involving these rare-earth atoms tend to be sluggish at best and generally do not yield pure phases. As such, they were generally not considered for investigation.



CHALLENGES WITH RARE-EARTH BISMUTHIDE OXIDES Compared to rare-earth antimonide phases, the bismuthide analogues are far more difficult to prepare, and fewer of them have been discovered. The reason for this is simple: bismuth is far larger than antimony (r(Sb) = 1.45 Å vs r(Bi) = 1.60 Å).34 This size difference means it is far more difficult to incorporate Bi atoms into rare-earth oxygen frameworks since RE−O bonds are quite rigid. As such, not all rare-earth pnictide phases can be formed with bismuth, but there is a way to predict if a bismuth analogue of a rare-earth antimonide oxide phase will exist. Table 2 lists the known rare-earth oxides with respect to the 9606

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials Table 2. Coordination Polyhedron Volume for Sb or Bi in (Gd,Ca)mPnnOm phase

smallest Sb site volume

largest Sb site volume

smallest Bi site volume

largest Bi site volume

Gd2PnO2 Gd8Pn3O8 Ca2Gd7Pn5O5 Gd3PnO3 Ca2Gd8Pn3O10 CaGd3PnO4

33.3(2) Å3 29.01(4) Å3 28.88(3) Å3 28.39(5) Å3 26.78(3) Å3 27.22(8) Å3

33.3(2) Å3 29.47(4) Å3 29.47(3) Å3 28.39(5) Å3 28.26(3) Å3 27.22(8) Å3

34.2(2) Å3 29.68(4) Å3 29.87(3) Å3 29.50(5) Å3 -

34.2(2) Å3 30.29(4) Å3 30.48(3) Å3 29.50(5) Å3 -

coordination polyhedron size of its pnictogen atoms (Gdcontaining phases were chosen for consistency). An obvious trend is noted from this data: only phases with a certain minimum coordination polyhedron size for the Sb atoms can yield the corresponding rare-earth bismuthide oxides. Our experimental results confirm the conclusions from Table 2: RE2BiO2 (large Pn site) are relatively simple to prepare, while RE8Bi3O8 (smaller Pn site) require highly specific conditions for synthesis. Continuing this trend, the RE3BiO3 phase (even smaller Pn sites) may only be formed in trace amounts, while the CaRE3BiO4, Ca2RE8Bi3O10, and (RE′RE″)3BiO3 are not known to exist. This would indicate there is a minimum Sb-site size of ca. 28 Å3 for the corresponding Bicontaining phase to exist, assuming no structural modifications. The larger size of bismuth also results in another problem: the overwhelming stability of the RE2BiO2 phase in comparison to other RE−Bi−O phases. Due to the natural superlattice-type structure of the tetragonal RE2PnO2 phase, the RE−O bonds within the 2D layers (ab plane) do not need to stretch considerably to accommodate bismuth; instead the lattice expands along the c direction of the unit cell. For example, in the Ho2PnO2 series, the expansion within the Ho−O layer is only 0.013 Å while the interlayer separation increases by 0.033 Å when Sb is fully replaced by Bi.35 Furthermore, since the Pn site in the RE2PnO2 series is far larger than in other RE−Pn−O phases, the RE2BiO2 phases are very thermodynamically stable as a result. This is evidenced by higher temperatures used to prepare RE8Bi3O8 yielding the RE2BiO2 phase instead.33

Figure 2. Two dimensional illustration of the RE4O tetrahedra connectivity in rare-earth pnictide oxides. As RE4O tetrahedra (A) are connected to each other by edge sharing, they form basic building blocks (B). These building blocks are connected to each other by corner sharing, resulting in the creation of empty channels, which may be occupied by pnictogen atoms (C). Further buildup results in a three-dimensional, periodic network that defines many of the rareearth pnictide oxide phases, such as CaRE3SbO4 (D).

those of the (RE,Ca)mPnnOm series, and thus they are not discussed in this paper (their structures are analyzed in refs 20 and 32). There are always two types of channels in the (RE,Ca)mPnnOm structures; the empty square ones and the occupied ones that are either square or rectangular. The square channels have one Sb atom, while rectangular ones contain two or three Sb atoms. The (RE,Ca)mPnnOm structures can be classified by the type of the REO building blocks present around the Sbfilled channels. The CaRE3SbO4 phases contain the same two blocks that are mutually orthogonal. Since each block consists of two RE4O tetrahedra, we call this structural arrangement 2 × 2 (Figure 3). The building blocks in tetragonal (RE′RE″)3SbO3 are enlarged by one RE4O tetrahedron, and thus the notation is 3 × 3. In the RE3PnO3 phases, one building block is double in length when compared to CaRE3SbO4, and thus the motif is 4 × 2. The RE8Pn3O8 6 × 2 structure is derived from that of RE3PnO3 by extending the larger block further by two RE4O tetrahedra. The RE2PnO2 phases can be seen as a final member of the series, in which the building block is infinite in size. It is worth mentioning that the [REO] layer in RE2PnO2 is of an antifluorite type and is found in some A/RE−M−Pn−O pnictide oxides (see the Introduction). The Ca2RE7Pn5O5 structure is a combination of the CaRE3SbO4 and RE3PnO3 structures, and thus it is designated as 4 × 2 + 2 × 2.



BUILD-UP PRINCIPLE OF (RE,Ca)mPnnOm Interestingly, many of the known rare-earth pnictide oxide phases are structurally related to each other in terms of stacking sequences of rare-earth oxide “building blocks” and empty channels occupied by pnictogen atoms. To visualize this, one may imagine the simplest type of building block: a single RE4O tetrahedron. Alone it does not form a phase, but when two or more are connected to each other via edge sharing, they form larger blocks that can create a structure-specific pattern by sharing the corners within a plane (Figure 2). These blocks stack on the top of each other in a third direction via edge sharing and thereby create 1D channels in the same direction (2D layers in RE2PnO2). Some of these channels are occupied by pnictogen atoms. The most basic of all rare-earth pnictide oxide structures is the CaRE3SbO4 phase, which must be stabilized by calcium in order to maintain its charge neutrality. By altering the size of the building block(s), different arrangements are formed, which are represented by several rare-earth pnictide oxides (Figure 3 and Table 3). The similarities between structures also form the basis of the synthetic pathways between RE 3 PnO 3 /RE 8 Pn 3 O 8 and CaRE3SbO4/Ca2RE8Sb3O10, as well as atomic disorder of pnictogen atoms. The structures of RE 9 Sb 5 O 5 and Ca2RE7Pn5O5 are identical but completely different from



DISORDERED PNICTOGEN ATOMS, PHYSICAL PROPERTIES, AND ANDERSON LOCALIZATION Except for some of RE 2BiO2 ,18 all other investigated (RE,Ca)mPnnOm phases display semiconducting behavior experimentally.1,2,31,33,35 Assuming 3− for Pn, 2− for O, 2+ for Ca, and a 3+ oxidation state for RE (RE3+ are confirmed by magnetic susceptibility measurements3,31,33), the RE3PnO3 and CaRE3PnO4 oxides are expected to be charge balanced and therefore semiconducting in agreement with the experimental data. However, the RE8Pn3O8 ((RE3+)8(Sb3−)3(O2−)8(h+)), Ca 2 RE 8 Sb 3 O 10 ((Ca 2+ ) 2 (RE 3+ ) 8 (Sb 3−) 3 (O 2− ) 10 (h + )), and 9607

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials

Figure 3. Buildup sequence of the (RE,Ca)mPnnOm phases. Extension of the major RE−O building blocks creates a series of phases related by similar structural features. The RE9Pn5O5/Ca2RE7Pn5O5 phases do not follow this trend, however.

Table 3. Summary of Relative Structural Data for (RE,Ca)mPnnOm Phases

a

system

space group

RE−Sb−O

RE−Bi−O

building block motif

fraction of disordered Pn

RE3PnO3 RE8Pn3O8 RE′1.5RE″1.5PnO3 RE2PnO2 CaRE3PnO4 Ca2RE8Pn3O10 RE9Pn5O5 Ca2RE7Pn5O5

C2/m C2/m P42/mnm I4/mmm I4/m C2/m P4/n P4/n

La, Sm, Gd, Ho La, Sm, Gd, Ho La/Dy, La/Ho, Ce/Ho La−Nd, Sm, Gd, Ho, Er Ce−Nd, Sm, Gd Pr, Nd, Sm, Gd, Tb La, Ce, Pr, Sm, Tb, Dy Pr, Sm, Gd, Dy

Gd Gd none Y, La−Nd, Sm, Gd−Er none none nonea Gd, Dy

4×2 6×2 3×3 ∞ 2×2 4×2+2×2 N/A N/A

none 1 /3 none all in Sb analogues none 2 /3 none none

Known to exist, but no structure has been solved for any RE.

RE2SbO2 ((RE3+)2(Sb3−)(O2−)2(h+)) phases are all deficient by one electron and thus should be metallic, which contrasts with the resistivity measurements. An inherent feature of these noncharge balanced phases is the presence of one distinct Pn site with a large displacement parameter. In RE8Pn3O8, the Pn atoms in the middle of the 1D channels display substantial displacements along the channel direction, while the two terminal Pn atoms behave normally; in Ca2RE8Sb3O10, both Sb atoms in the rectangular channel are affected; and in RE2SbO2, all Sb atoms move significantly within the 2D layers (Figure 4).

This atomic displacements appear to be static as they did not diminish upon cooling from room temperature to 100 K; as a result, they were treated as a temperature-independent atomic disorder.1,33 The possibility of superstructures being formed was considered, but no satellite peaks supporting an extended ordering were observed in any of the X-ray data. Even with the disorder, the Sb−Sb distances are at least 3.2 Å and the Bi−Bi distances are above 3.5 Å in the RE8Pn3O8, Ca2RE8Sb3O10, and RE2SbO2.2,31,33 Such Pn-Pn distances are too large to support Pn24− dimer formation and consequently to yield a charge balance formula. This conclusion was further verified by the electronic structure calculations performed for the structures with the Pn atoms at the ideal positions and with disordered Pn atoms.1,31 The results suggest that Pn disorder and formation of the weak Pn−Pn interactions is not sufficient to open a band gap and render the RE8Pn3O8, Ca2RE8Sb3O10, and RE2SbO2 phases semiconducting. Figure 5 represents DOS curves for Gd2SbO2 with two structural models calculated with the TB-LMTO-ASA method36 as implemented in the Stuttgart program.37 In both cases the Fermi level for Gd2SbO2 resides in the valence band, suggesting a metallic conductivity, which contradicts the observed semiconducting behavior.2 In the superstructure model, the band gap even disappeared due to stronger Sb−Sb interactions. It is worth mentioning the valence band is dominated by the Sb states and thus any perturbations to the Sb layers should affect the charge transport properties. It is proposed that Sb atomic disorder leads to the Anderson

Figure 4. (left) Gd2SbO2 structure with disordered Sb atoms. (middle) Electron density maps calculated from the experimental intensities for the Sb and Gd atoms within the ac plane. (right) Electron density map of the Sb atoms within the 2D layer. 9608

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials

Figure 5. Density of states for Gd2SbO2 with two structural models. (left) The Sb atoms sit at the ideal positions of (0, 0, 0). (right) The Sb atoms are shifted to (0.0844, 0, 0) and form Sb−Sb bonds of 3.23 Å. A superstructure was created to allow Sb−Sb dimer formation.

localization38 of charge carriers due to the nonperiodic Sb potential.1 In terms of the band structure, the Sb atomic disorder results in the localization of the crystal orbitals at the bottom and top of the band dominated by the Sb states. This localization is schematically shown in Figure 6. As a result of the localization, some energy is required to move electrons across the crystal structure, and the electrical conductivity displays a semiconducting behavior.

Additional verification of the Anderson localization comes from the Ho2Sb1−xBixO2 system.35 The two end members of the series display different electrical behaviors: Ho2SbO2 with disordered Sb atoms is a semiconductor, while Ho2BiO2 with well localized Bi atoms is a metal. If the Anderson localization is a primary mechanism for charge carrier transport, then localization of the states and activation energy should decrease with an increase in the Bi amount in Ho2Sb1−xBixO2. Figure 7 shows the electrical resistivity and Seebeck coefficient for different Sb/Bi ratios. As predicted, the Bi-richer phases become more conductive with lower activation energy. The activation energy (effective band gap), EA, was estimated from EA = eSmaxTmax (Smax is the maximum Seebeck coefficient at Tmax) and is plotted in Figure 8 in conjunction with the relative atomic displacements of the Sb/Bi atoms. The relative displacement parameters, which are the ratios between the shifts within the ab plane to the shifts in the perpendicular direction, are used to quantify the disorder within the 2D Sb/Bi layer. The Sb/Bi atomic displacements become smaller for larger Bi amounts since the Bi atoms are significantly larger and thus spatially more constrained by rigid Ho−O layers. There is a clear correlation between the level of disorder and activation energy: the smaller the disorder the lower the activation energy. It has to be mentioned that displacement parameters are for single crystals and activation energies for the bulk samples;

Figure 6. Schematic representation of charge carrier localization in Gd2SbO2 due to the Sb disorder. The Fermi level sits at the top of the valence band and in the region of the localized states.

Figure 7. Electrical resistivity and Seebeck coefficient of the Ho2Sb1−xBixO2 phases. The maxima in the Seebeck coefficient correspond to the temperatures, at which minority charge carriers are excited across the band gap. The values corresponding to these maxima are used to estimate the activation energies. Reproduced with permission from ref 35. Copyright 2013 American Chemical Society. 9609

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials

Figure 8. (left) Ratio between the in-plane and out-of-plane displacement parameters from the single crystal refinements of the Ho2Sb1−xBixO2 phases. This ratio represents the atomic disorder within the Sb/Bi layers. (right) Activation energy derived from the Seebeck data. Reproduced with permission from ref 35. Copyright 2013 American Chemical Society.

Figure 9. Electrical resistivity and Seebeck coefficient for some RE2SbO2 phases. The Nd and Sm oxides display behaviors characteristic of variable range hopping. Their low-temperature resistivities and Seebeck coefficients can be fitted to the corresponding formulas with dimensionality parameter of 3 for Nd and 2 for Sm.

Figure 10. (left) Shifts of the Sb atoms within the 2D layers and the Sb−Sb distances as a function of rare-earth size in RE2SbO2. (right) Schematic representation of localization of the valence band states and charge transport mechanism for different RE2SbO2 phases. Only the valence band dominated by the Sb p states is shown. A rigid band approximation is used to plot the valence band.

therefore, there is some mismatch between the Bi concentrations in the two plots. Thermal excitation of charge carriers due to the localization of states is only one of the possible mechanisms for carrier transport. According to Mott and Davis,39 charge carriers can also tunnel (via variable-range hopping) across the localized energy region, when the energy span between the localized states is large. Such variable range hopping was observed in Nd2SbO2 and Sm2SbO2, while the RE2SbO2 analogues with smaller rare earths (Gd, Ho, and Er) displayed a regular semiconducting behavior (Figure 9).2 The single crystal diffraction studies revealed larger Sb shifts for the Nd and

Sm oxides, which explains stronger localization of the states and the associated variable range hopping (Figure 10). An Sb disorder decreases for the Gd, Ho, and Er phases, and a regular semiconducting behavior emerges. Interestingly, the Sb−Sb distances appear to be rather insensitive to the degree of Sb disorder and are around 3.24 Å.



SUMMARY The (RE,Ca)mPnnOm phases follow a basic buildup principle centered around the elongation of RE−O building blocks. These oxides can be further divided into two classes based on their structures and stoichiometries: charge-balanced with 9610

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials

praseodymium, neodymium, and samarium). Zh. Neorg. Khim. 1978, 23, 1737−1740. (11) Varfolomeev, M. B.; Toporenskaya, T. A.; Narnov, B. A. Synthesis and some properties of lanthanum antimonate (LaSbO4). Zh. Neorg. Khim. 1979, 24, 244−247. (12) Nasonova, S. N.; Serebrennikov, V. V.; Narnov, G. A. Preparation and some properties of rare earth bismuthites. Zh. Neorg. Khim. 1973, 18, 2350−2353. (13) Michel, C.; Caignaert, V.; Raveau, B. An oxygen deficient fluorite with a tunnel structure: bismuth lanthanum oxide: (Bi8La10O27). J. Solid State Chem. 1991, 90, 296−301. (14) Schaal, H.; Nuss, J.; Hoenle, W.; Grin, Y.; Von Schnering, H. G. Crystal structure of tetraeuropium diantimonide oxide, Eu4Sb2O. Z. Kristallogr. - New Cryst. Struct. 1998, 213, 15. (15) Hoenle, W.; Schaal, H.; Von Schnering, H. G. Crystal structure of tetraeuropium dibismuthide oxide, Eu4Bi2O. Z. Kristallogr. - New Cryst. Struct. 1998, 213, 16. (16) Benz, R. Ce2O2Sb and Ce2O2Bi crystal structure. Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1971, 27 (4), 853−854. (17) Nuss, J.; Jansen, M. Syntheses, structures and properties of the pnictide oxides R2PnO2 (R = Ce, Pr; Pn = Sb, Bi). J. Alloys Compd. 2009, 480, 57−59. (18) Mizoguchi, H.; Hosono, H. A Metal-Insulator Transition in R2O2Bi with an Unusual Bi2‑ Square Net (R = Rare Earth or Y). J. Am. Chem. Soc. 2011, 133, 2394−2397. (19) Muir, S.; Vielma, J.; Schneider, G.; Sleight, A. W.; Subramanian, M. A. The hunt for LaFeSbO: Synthesis of La2SbO2 and a case of mistaken identity. J. Solid State Chem. 2012, 185, 156−159. (20) Nuss, J.; von Schnering, H. G.; Grin, Y. Ü ber die Antimonidoxide La9Sb5O5 und Ce9Sb5O5 sowie die binären Phasen La2Sb und Ce2Sb. Z. Anorg. Allg. Chem. 2004, 630, 2287−2291. (21) Ozawa, T. C.; Kauzlarich, S. M.; Bieringer, M.; Wiebe, C. R.; Greedan, J. E.; Gardner, J. S. The Effect of Interlayer Cations on the Magnetic Properties of the Mixed-Metal Pnictide Oxides: A2MnZn2As2O2 (A = Sr, Ba). Chem. Mater. 2001, 13, 973−980. (22) Ozawa, T. C.; Kauzlarich, S. M.; Richardson, J. W., Jr.; Greedan, J. E. Structure-property relationships and single-crystal growth of a new CDW/SDW compound: Na2Ti2Sb2O. Book of Abstracts, 219th ACS National Meeting, San Francisco, CA, March 26−30, 2000, American Chemical Society: 2000; INOR-743. (23) Ozawa, T.; Olmstead, M. M.; Brock, S. L.; Kauzlarich, S. M.; Young, D. M. Synthesis and Characterization of a New Compound with Alternating MnO22‑ and Zn2As22‑ Layers: Ba2MnZn2As2O2. Chem. Mater. 1998, 10, 392−396. (24) Nientiedt, A. T.; Jeitschko, W. Equiatomic Quaternary Rare Earth Element Zinc Pnictide Oxides RZnPO and RZnAsO. Inorg. Chem. 1998, 37, 386−389. (25) Brock, S. L.; Kauzlarich, S. M. Structure-property relationships in a series of mixed layer pnictide oxide compounds: A2Mn3Pn2O2 (A = Sr, Ba; Pn = P, As, Sb). J. Alloys Compd. 1996, 241, 82−88. (26) Brock, S. L.; Raju, N. P.; Greedan, J. E.; Kauzlarich, S. M. Structure-property relationships in the mixed layer pnictide oxide compounds A2Mn3Pn2O2 (A = Sr, Ba; Pn = P, As, Sb). Book of Abstracts, 210th ACS National Meeting, Chicago, IL, August 20−24, 1995; American Chemical Society: 1995; (Part 1), INOR-006. (27) Brock, S. L.; Hope, H.; Kauzlarich, S. M. Synthesis and structure of a new layered pnictide oxide containing close Mn-Mn interactions: barium manganese arsenide oxide (Ba2Mn2As2O). Inorg. Chem. 1994, 33, 405−406. (28) Stetson, N. T.; Kauzlarich, S. M. A new material with alternating metal-oxide and metal-phosphide layers: barium manganese phosphate (Ba2Mn3P2O2). Inorg. Chem. 1991, 30, 3969−3971. (29) Ozawa, T. C.; Kauzlarich, S. M. Chemistry of layered d-metal pnictide oxides and their potential as candidates for new superconductors. Sci. Technol. Adv. Mater. 2008, 9, 033003. (30) Ozawa, T. C.; Kauzlarich, S. M.; Bieringer, M.; Greedan, J. E. Possible Charge-Density-Wave/Spin-Density-Wave in the Layered

ordered Pn atoms and non-charge balanced disordered. Despite this, all phases except for some RE2BiO2 show semiconductortype behavior. Electronic structure calculations cannot validate this behavior, as the distribution of states in the non-charge balanced phases suggests metallic-type conduction. Instead, it is proposed that the non-charge balanced phases experience varying degrees of Anderson localization, which creates an activation energy that must be overcome for conduction. The magnitude of this localization, based on electrical resistivity data, seems to be directly linked to the disorder of Pn atoms in each structure. Furthermore, the extent of the Pn atom disorder is linked to the RE that is used, allowing the fine-tuning of physical properties without altering the structure or charge carrier concentration. This approach may be applied to other systems with similar degrees of disorder which would allow better physical property optimization.



AUTHOR INFORMATION

Corresponding Author

*(Y.M.) E-mail: [email protected]. ORCID

Yurij Mozharivskyj: 0000-0003-4498-5545 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by a Discovery Grant and a CREATE HEATER Grant from the Natural Sciences and Engineering Research Council of Canada.



REFERENCES

(1) Wang, P.; Forbes, S.; Kolodiazhnyi, T.; Kosuda, K.; Mozharivskyj, Y. Synthesis, Crystal and Electronic Structures of New Narrow-BandGap Semiconducting Antimonide Oxides RE3SbO3 and RE8Sb3‑δO8, with RE = La, Sm, Gd, and Ho. J. Am. Chem. Soc. 2010, 132, 8795− 8803. (2) Wang, P. L.; Kolodiazhnyi, T.; Yao, J.; Mozharivskyj, Y. Decoupling the Electrical Conductivity and Seebeck Coefficient in the RE2SbO2 Compounds through Local Structural Perturbations. J. Am. Chem. Soc. 2012, 134, 1426−1429. (3) Forbes, S.; Wang, P.; Yao, J.; Kolodiazhnyi, T.; Mozharivskyj, Y. Synthesis, Crystal Structure, and Electronic Properties of the Tetragonal (REIREII)3SbO3 Phases (REI = La, Ce; REII = Dy, Ho). Inorg. Chem. 2013, 52, 1025−1031. (4) Anderson, M. T.; Vaughey, J. T.; Poeppelmeier, K. R. Structural Similarities among Oxygen-Deficient Perovskites. Chem. Mater. 1993, 5, 151−165. (5) Anderson, M. T.; Greenwood, K. B.; Taylor, G. A.; Poeppelmeier, K. R. B-Cation Arrangements in Dounble Perovskites. Prog. Solid State Chem. 1993, 22, 197−233. (6) Woodward, P. M. Octahedral tilting in perovskites 0.2. Structure stabilizing forces. Acta Crystallogr., Sect. B: Struct. Sci. 1997, 53, 44−66. (7) Battle, P. D.; Rosseinsky, M. J. Synthesis, structure, and magnetic properties of n = 2 Ruddlesden-Popper manganates. Curr. Opin. Solid State Mater. Sci. 1999, 4, 163−170. (8) Greenblatt, M. Ruddlesden-popper Ln(n+1)Ni(n)O(3n+1) nickelates: Structure and properties. Curr. Opin. Solid State Mater. Sci. 1997, 2, 174−183. (9) Ok, K. M.; Gittens, A.; Zhang, L.; Halasyamani, P. S. Synthesis, structure and characterization of two new antimony oxides-LaSb3O9 and LaSb5O12: Formation of LaSb5O12 from the reaction of LaSb3O9 with Sb2O3. J. Mater. Chem. 2004, 14, 116−120. (10) Varfolomeev, M. B.; Narnov, B. A. Synthesis and x-ray diffraction study of Ln 3SbO7 compounds (Ln = lanthanum, 9611

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612

Review

Chemistry of Materials Pnictide-Oxides: Na2Ti2Pn2O (Pn = As, Sb). Chem. Mater. 2001, 13, 1804−1810. (31) Forbes, S.; Yuan, F.; Saparov, B.; Sefat, A. S.; Kosuda, K.; Kolodiazhnyi, T.; Mozharivskyj, Y. Synthesis, Crystal Structure, and Electronic Properties of the CaRE3SbO4 and Ca2RE8Sb3O10 phases (RE = Rare-Earth Metal). Chem. Mater. 2014, 26, 2289−2298. (32) Forbes, S.; Yuan, F.; Kosuda, K.; Kolodiazhnyi, T.; Mozharivskyj, Y. Investigation of the transport properties and compositions of the Ca2RE7Pn5O5 series (RE=Pr, Sm, Gd, Dy; Pn=Sb, Bi). J. Solid State Chem. 2016, 242, 148−154. (33) Forbes, S.; Yuan, F.; Kosuda, K.; Kolodiazhnyi, T.; Mozharivskyj, Y. Synthesis, crystal structure, and physical properties of the Gd3BiO3 and Gd8Bi3O8 phases. J. Solid State Chem. 2016, 233, 252−258. (34) Slater, J. C. Atomic radii in crystals. J. Chem. Phys. 1964, 41, 3199−3204. (35) Wang, P. L.; Kolodiazhnyi, T.; Yao, J.; Mozharivskyj, Y. Disorder-Controlled Electrical Properties in the Ho2Sb1‑xBixO2 Systems. Chem. Mater. 2013, 25, 699−703. (36) Andersen, O. K.; Pawlowska, Z.; Jepsen, O. Illustration of the linear-muffin-tin-orbital tight-binding representation: Compact orbitals and charge density in Si. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 34, 5253−5269. (37) Jepsen, O.; Burkhardt, A.; Andersen, O. K. The TB-LMTO-ASA Program, version 4.7; Max-Planck-Institut für Festkörperforschung: 1999. (38) Anderson, P. W. Absence of Diffusion in Certain Random Lattices. Phys. Rev. 1958, 109, 1492−1505. (39) Mott, N.; Davis, E. A. Electronic Processes in Non-Crystalline Materials. In International Series of Monographs on Physics; Oxford University Press: Oxford, U.K., 1978.

9612

DOI: 10.1021/acs.chemmater.7b03996 Chem. Mater. 2017, 29, 9605−9612