New 10H perovskites Ba₅Ln_1−xMn_4+yO_15−δ with spin glass behaviour

Abstract

New 10H perovskite phases of composition Ba₅Ln_1−xMn_4+yO_15−δ (Ln = Sm–Lu) have been synthesised by solid-state reactions at 1300 °C. Their crystal structures have been investigated using X-ray and neutron diffraction. The 10H perovskite structure has an ordered arrangement of Ln and Mn ions on the corner-sharing octahedral centres and the face-sharing octahedral centres respectively. MnO₆ octahedral distortions and short Mn–Mn distances have been identified in these structures, and oxygen vacancies are found in the c-BaO₃ layers. Additionally, magnetic measurements and neutron diffraction identify spin glass transitions in these new materials. These transitions are related to the Mn ions on the corner sharing octahedral sites, which mediate magnetic interactions between tetramers of MnO₆ octahedra.

---

Introduction

Interest in the physics of manganites has motivated investigation of the hexagonal perovskite BaMnO_3−δ system, which shows a variety of crystal structures and magnetic properties.^1–5 A series of polytypes (2H → 15R → 8H → 6H → 10H → 4H) have been identified in the BaMnO_3−δ system as the oxygen vacancy δ increases.^4,5 Further perovskite polytypes may be generated when Mn is replaced by other cations.

Many doped BaM_xMn_1−xO₃ systems have been explored, for example, with M = Ca,^6,7 Ti,^8,9 Fe,^10,11 Co,^12,13 Ru,^14,15 Ir,¹⁶ In,^17,18 and rare earths Ln.^19–23 This led to the discovery of a range of M-cation ordered 10H perovskites Ba₅In_0.93Mn₄O_14.40,¹⁷ Ba₅Sn_1.1Mn_3.9O₁₅,²⁴ Ba₅Ce_0.83Mn_4.17O₁₅,²⁵ and Ba₅Sb_1−xMn_4+xO_15−δ.²⁶ The Ba₅MMn₄O₁₅ 10H perovskite structure consists of close packed BaO₃ layers in a (cchhh)₂ sequence. The large M ions occupy the corner-sharing octahedral centre M1, and the Mn ions sit on the face-sharing octahedral centres M2 and M3, as shown in Fig. 1. Polymorphs of BaMnO_3−δ have long range antiferromagnetic (AFM) order of all the Mn spins, with T_N = 220–270 K.^4,5 However, the long-range magnetic order is suppressed in the 10H perovskites with low-level doping, and the metal occupying the corner-sharing M1 site plays a key role in the magnetic behaviour of these materials. Ba₅In_0.93Mn₄O_14.40 has diamagnetic In³⁺ ions on the M1 site and weak AFM interactions between the tetramers of face-sharing MnO₆ octahedra, and no magnetic order is observed down to 2 K. Ba₅Sn_1.1Mn_3.9O₁₅ (ref. 24) and Ba₅Ce_0.83Mn_4.17O₁₅ (ref. 25) have Mn⁴⁺ ions occupying 10–20% of the M1 sites and undergo spin glass transitions below 10 K. Ba₅Sb_1−xMn_4+xO_15−δ²⁶ phases have 24–36% M1 site occupancy by Mn³⁺ ions, and long range AFM order of partial Mn spins (moment ≈ 0.5 μ_B) below 129–167 K. The presence of Mn³⁺ ions on the M1 site causes ferro- or ferri-magnetic clusters to freeze below a second magnetic transition, in the temperature range 13–22 K.

Fig. 1 10H Ba₅MMn₄O_15−δ perovskite structure (a) the unit cell and (b) the octahedral centres. The M ion is on the M1 site, and Mn ions sit on the M2 and M3 sites.

The BaLn_xMn_1−xO₃ system contains several cation-ordered perovskites: 6H Ba₃LnMn₂O₉ and 12R Ba₄LnMn₃O₁₂ (Ln = Er, Ce, Pr, and Nd).^20–23,27 The 6H and 12R polytypes feature dimers and trimers of face-sharing MnO₆ octahedra respectively,^23,27 alternating with LnO₆ octahedra. The 6H and 12R perovskites are closely related to the 10H perovskite. Considering the similarity of Ln elements, we predicted the existence of 10H BaLn_xMn_1−xO₃ series and started an experimental exploration. Herein we report the syntheses, crystal structures and magnetic properties of 10H Ba₅Ln_1−xMn_4+yO_15−δ (Ln = Sm–Lu) perovskites.

Experimental

Polycrystalline Ba₅Ln_1−xMn_4+yO_15−δ samples were synthesised from BaCO₃, MnCO₃, and Ln₂O₃ (Ln = Sm–Gd and Dy–Lu) or Tb₄O₇. For each sample, the starting materials were thoroughly mixed in an agate mortar and pestle and heated in an alumina crucible at 1000 °C for 10 h to decompose the carbonates. Then the material was reground, pressed into pellets (20 ton per cm²), and heated at 1300 °C for a total of 30–50 h with slow cooling to room temperature. At several intermediate steps the material was re-ground and re-pressed, and X-ray powder diffraction (XRPD) measurements were made at each step to track the reaction to completion.

The phase purity of the samples was checked by XRPD using a Rigaku D/Max-2000 diffractometer with graphite monochromator. This generated Cu Kα radiation at 40 kV and 100 mA. Data were collected over the 2θ range 6–120° in step scanning mode. Neutron powder diffraction data were collected using the BT-1 32 detector neutron powder diffractometer at the NIST Center for Neutron Research reactor, NBSR. A Cu (311) monochromator with a 90° take-off angle gave incident neutrons with λ = 1.5403(2) Å, and in-pile collimation of 15 minutes of arc was used. Data were collected over the 2θ range 3–168° with a step size of 0.05°. The diffraction data were analysed with the GSAS program.²⁸ The Rietveld refinement was carried out assuming a pseudo-Voigt function for the peak shape and a calculated background using linear interpolation between a set of fixed points. Selected-area electron diffraction (SAED) were carried out on Tecnai G2 F30 with an accelerating voltage of 300 kV. The magnetic properties were investigated using a Quantum Design MPMS-SS superconducting quantum interference device (SQUID) magnetometer. Data were recorded in a 20 Oe or 1000 Oe applied field while warming the sample from 2–300 K, following both zero-field cooling (ZFC) and field cooling (FC).

Results and discussion

Chemical composition and 10H phase formation

The initial Ba₅LnMn₄O_15−δ samples consisted of mixed 10H and 12R phases. To eliminate the 12R phase, we first varied and found the Ln content that gave the lowest amount of a 12R phase. Then we fixed the Ln content at these values, and varied and optimised the Mn fractions. By repeating these optimisation steps for several cycles, 10H phase-pure samples were obtained with the Ba : Ln : Mn ratios listed in Table 1. Their general chemical formula is Ba₅Ln_1−xMn_4+yO_15−δ (Ln = Sm–Lu).

Table 1 The optimised Ba : Ln : Mn ratios of the 10H Ba₅LnMn₄O_15−δ samples

Ln	Ba : Ln : Mn	Ln	Ba : Ln : Mn
Sm	5.00 : 0.98 : 4.01	Ho	5.00 : 0.86 : 4.12
Eu	5.00 : 0.87 : 4.02	Er	5.00 : 0.82 : 4.17
Gd	5.00 : 0.82 : 4.17	Tm	5.00 : 0.80 : 4.14
Tb	5.00 : 0.85 : 4.15	Yb	5.00 : 0.83 : 4.17
Dy	5.00 : 0.86 : 4.12	Lu	5.00 : 0.78 : 4.21

---

Fig. 2 shows the XRPD patterns of the pure Ba₅Ln_1−xMn_4+yO_15−δ samples, which are similar to that of Ba₅Sb_1−xMn_4+xO₁₅. The XRPD patterns of all our materials can be indexed using a hexagonal unit cell, with lattice parameters a = 5.76–5.80 Å and c = 23.75–24.04 Å. The Selected-Area Electron Diffraction (SAED) patterns collected on the Ba₅Sm_0.98Mn_4.01O_15−δ sample along [010], [02̄1] and [11̄0] zone axes are shown in Fig. 3. No superstructure reflections are seen, and all the observation reflections agree with the indexing results. The systematic absence of the reflections hhl: l = 2n + 1 is consistent with the space group P6₃/mmc.

Fig. 2 X-ray diffraction patterns of our Ba₅Ln_1−xMn_4+yO_15−δ samples, each showing a pure 10H perovskite phase.

Fig. 3 SAED patterns of Ba₅Sm_0.98Mn_4.01O_15−δ along (a) [010], (b) [02̄1], and (c) [11̄0]. No superstructure is observed.

The X-ray structures of Ba₅Ln_1−xMn_4+yO_15−δ

Rietveld refinement against the XRPD data was performed using the 10H Ba₅Sb_1−xMn_4+xO₁₅ structural model. Ln ions and Mn ions occupy the M1 site, whilst the M2 and M3 sites are solely occupied by Mn. The ratio of Ln and Mn on the M1 site was fixed according to the ratio of reagents used. The refinements were insensitive to occupancies and isotropic thermal parameters of the oxygen sites, which were held at unity and 0.025 Ų respectively. Good fits were obtained with R_wp (weighted-profile residual) factors of 4.1–6.5% and χ² (goodness-of-fit) values in the range 1.7–3.0. The refined plots for all the sample are shown in ESI (Fig. S1–S10†). The refined crystallographic parameters and selected bond lengths are shown for all samples in Table S1.†

Across the Ba₅Ln_1−xMn_4+yO_15−δ series, the lattice parameters and cell volume increase monotonically, in line with the increasing ionic radius of the Ln³⁺ cation used. The sole exception is the Tb compound, which has considerably smaller lattice parameters and cell volume, as shown in Fig. 4. The refined Tb–O bond length (2.10 Å) is close to the calculated Tb⁴⁺–O²⁻ length, indicating that the Tb ion is tetravalent. Tb⁴⁺ ions have also been found in the 12R perovskite series Ba₄LnIr₃O₁₂ and Ba₄LnRu₃O₁₂.^29,30

Fig. 4 The lattice parameters and unit cell volumes of all our Ba₅Ln_1−xMn_4+yO_15−δ phases vary according to the radius of the Ln ion in each.

All of our 10H perovskites have an ordered arrangement of Ln and Mn cations. Like the M = In³⁺ and Sb⁵⁺ 10H analogues, Mn ions alone occupy the octahedral centres that form face-sharing tetramers. These MnO₆ octahedra are distorted, with two sets of three Mn–O distances. For all our samples the Mn1 and Mn2 sites have very similar average Mn–O bond distances, of about 1.93 Å as shown in Fig. 5. This is the same as the Mn⁴⁺–O²⁻ distance calculated from Shannon's radii,³¹ indicating that the Mn ions in the sites forming tetramers of MnO₆ octahedra in the Ba₅Ln_1−xMn_4+yO_15−δ phases is tetravalent. The Mn–Mn distances inside each tetramer range from 2.35 Å to 2.52 Å, close to the Mn–Mn distance in elemental α-Mn.³² This observation may suggest a degree of d-orbital overlap between neighbouring Mn ions within each tetramer. The distortion of the MnO₆ octahedra forming the tetramers is evidenced by the fact that the Mn1–Mn2 distance is around 0.1 Å longer than the Mn2–Mn2 distance in all of our Ba₅Ln_1−xMn_4+yO_15−δ samples.

Fig. 5 Variations of the Ln–O, Mn–O and Mn–Mn distances in our Ba₅Ln_1−xMn_4+yO_15−δ samples. d_Mn1–O and d_Mn2–O are average values. The solid and dashed lines indicate the trend for d_Ln–O and d_Mn–O values as r(Ln) increases.

The corner-sharing octahedral centre has mixed occupancy, with Ln ions in majority (more than 78%) and a minority of secondary Mn ions (less than 21%). Notably, the Ba₅Ln_1−xMn_4+yO_15−δ samples are non-stoichiometric, with (Ln + Mn)/Ba values of 0.98–1. This non-stoichiometric associated with the vacancies at the M1 site (less than 10%), which has also been observed in Ba₅In_0.94Mn₄O_14.40.¹⁷ As expected, the Ln–O distances increase with the size of Ln³⁺ except for the Tb-containing sample. The values are 10% smaller than the calculated Ln³⁺–O²⁻ distances for all the samples, due to the partial occupation the M1 site by Mn.

Combined X-ray and neutron diffraction studies on Ba₅Lu_0.78Mn_4.21O_15−δ

Due to their weak X-ray scattering, oxygen sites are not well described with PXRD. To find the exact oxygen positions and explore possible oxygen vacancies, we studied our sample of Ba₅Lu_0.78Mn_4.21O_15−δ using powder neutron diffraction. Rietveld refinements using X-ray and neutron diffraction data simultaneously were performed, starting from the refined X-ray structure. The oxygen occupancies were refined, and oxygen vacancies were found on the O2 site. The final refinement gave a good fit, with overall R_wp and χ² parameters of 6.4% and 2.3 respectively. The neutron refinement is shown in Fig. 6.

Fig. 6 Experimental and calculated neutron diffraction pattern of Ba₅Lu_0.78Mn_4.21O_15−δ, with the difference plot offset below.

The refined atomic parameters are listed in Table 2. The O2 site in the c-BaO₃ layers is 6% vacant. Due to this oxygen deficiency, the O2 site has a larger thermal displacement factor than the O1 and O3 sites. The nearby sites Ba2 and M1 sites also have relatively large thermal factors.

Table 2 Refined atomic parameters of Ba₅Lu_0.78Mn_4.21O_15−δ in space group P6₃/mmc at 300 K^a

Atom	x, y, z	Occup.	100× U Å^−2
a Lattice parameters: a = 5.7580(1) Å and c = 23.7445(3) Å.
Ba1	2/3, 1/3, 0.25	1	0.93(9)
Ba2	2/3, 1/3, 0.4422(1)	1	1.09(6)
Ba3	1, 0, 0.3437(1)	1	0.82(6)
M1 (Lu/Mn)	1, 0, 0.5	0.78/0/21	1.10(9)
M2 (Mn1)	1/3, 2/3, 0.4065(2)	1	0.83(11)
M3 (Mn2)	1/3, 2/3, 0.3010(2)	1	0.67(10)
O1	0.1836(4), 0.8164(4), 0.25	1	0.88(8)
O2	0.1765(3), 0.3530(6), 0.4486(1)	0.940(9)	1.36(8)
O3	0.4818(2), 0.9636(4), 0.3516(1)	1	0.80(6)

---

From the refined occupancies, this material has the composition Ba₅Lu_0.78Mn_4.21O_14.64. The average valence state of Mn ions is found to be 4.0, and the Mn ions on the M1 site must also be Mn⁴⁺ ions in order to maintain electroneutrality. Our other Ba₅Ln_1−xMn_4+yO_15−δ samples probably have Mn⁴⁺ ions on the M1 site as well, similar to the Ce⁴⁺ and Sn⁴⁺ analogues.^24,27 However, the Ba₅Sb_1−xMn_4+xO_15−δ analogue is different with Mn³⁺ ions on the M1 site.²⁶ For the BaM_xMn_1−xO₃ 10H perovskites, the average valence of the cations occupying the M1 site approaches +4 in keeping with Pauling's principle.

The Ba₅Lu_0.78Mn_4.21O_14.64 sample has oxygen vacancy in the c-BaO₃ layers, which is similar to that in the In-doped analogue.¹⁷ The M1 site in both compounds has a high occupancy of M³⁺ ions, which requires neighbouring oxygen vacancies to compensate the charge imbalance. When M⁴⁺ or M⁵⁺ ions dominate on the M1 site, oxygen vacancies in the c-BaO₃ layers are no longer present, as seen in the Ce⁴⁺, Sn⁴⁺, and Sb⁵⁺ analogues.^24–27

In Ba₅Lu_0.78Mn_4.21O_14.64, the Mn_M2–O2 distance is 0.12 Å less than the Mn_M2–O3 distance, which indicates considerable distortion of the Mn_M2O₆ octahedron. In analogous materials with a small M ion, such as Sb⁵⁺, the two Mn_M2–O distances are nearly equal and the Mn_M2O₆ octahedron is undistorted.²⁶ This indicates that the degree of Mn_M2O₆ distortion in the Ba₅MMn₄O_15−δ 10H perovskites is related to the size of M ion, r_M. The Mn–O distances are plotted against r_M in Fig. 7. The most noticeable trend is that M_M2–O2 distance shows a significant, linear decrease as r_M increases. The M_M2–O3 distance increases slightly with r_M.

Fig. 7 The Mn–O distances of the face-sharing octahedra for several Ba₅MMn₄O_15−δ 10H perovskites vary with the radius of the M ion.

For Ba₅Lu_0.78Mn_4.21O_14.64, the Mn_M3O₆ octahedron is almost undistorted, as evidenced by similarity of the two Mn_M3–O distances. This is similar to other 10H perovskites.

Magnetic properties

Low-temperature magnetic susceptibility data for our Ba₅Ln_1−xMn_4+yO_15−δ materials are shown in Fig. 8. All are paramagnetic at high temperatures and show linear M–H variation, without hysteresis, at 300 K. The susceptibilities of the materials with Ln = Sm³⁺ and Eu³⁺ do not follow the Curie–Weiss law, due to the temperature dependent magnetic moments of these ions. The samples with Ln = Gd–Lu have susceptibility maxima at their spin freezing temperatures T_F = 3–34 K, and divergent ZFC and FC magnetic susceptibilities below T_F. In comparison to the samples containing an Ln³⁺ cation, which have around 15% Mn occupancy of the M1 site, the sample containing Tb⁴⁺ has a higher T_F = 34 K, which is due to anti-site dislocation of Mn⁴⁺ and Tb⁴⁺ ions on the M2 sites increasing the fraction of Mn on the M1 site. Mn⁴⁺/M⁴⁺ site mixing has been observed in the other 10H BaM_x⁴⁺Mn_1−xO₃ perovskites.^24,25 Curie–Weiss fits to the susceptibility above T_F give an effective paramagnetic moment μ_eff = 1.5–2.9 μ_B per Mn ion for the Ln = Gd–Lu samples (Table 3), which is smaller than that expected for localized Mn⁴⁺ spins (3.87 μ_B). Negative Weiss temperatures (θ = −2.5 K to −65 K) indicate that AFM interactions are dominant. Similarly small paramagnetic moments of 0.88–1.44 μ_B per Mn⁴⁺ ion, and large negative θ ≈ −2000 K, have reported in the 10H perovskite analogue Ba₅Ce_1−xMn_4+xO₁₅.^25,27 This is probably due to the formation of AFM Mn clusters inside the tetramers, consistent with the observation of very short Mn–Mn distances and the significantly reduced magnetic moments of the Mn⁴⁺ ions.

Fig. 8 Susceptibility χ vs. temperature for Ba₅Ln_1−xMn_4+yO_15−δ (a) Ln = Tm, Lu, Ho, Gd, and Dy at 20 Oe; (b) Ln = Er, Tb, Yb, Eu, and Sm at 1000 Oe. The insets expand the low-temperature regions. Empty and filled symbols show ZFC and FC data, respectively. For clarity, χ are scaled by 10 for the Ln = Sm, Eu and Tm samples, and by 100 for the Lu sample.

Table 3 Magnetic ordering and Weiss temperatures, and effective paramagnetic moments and saturated magnetic moments, from susceptibility measurements of the Ba₅Ln_1−xMn_4+yO_15−δ samples

Ln	Gd^3+	Tb^4+	Dy^3+	Ho^3+	Er^3+	Tm^3+	Yb^3+	Lu^3+
TF/K	8.5	34.0	7.0	11.5	3.0	3.0	5.5	3.5
θ/K	−2.5	−39	−13	−7.2	−28	−52	−15.5	−65
μeff (f.u.)/μB	8.2	8.5	11.4	10.3	10.4	9.0	6.3	3.3
μeff (Ln)/μB	7.9	7.9	10.6	10.6	9.6	7.6	4.5	0
μeff (Mn)/μB	1.9	2.2	2.8	1.5	2.8	2.9	2.4	1.6
Ms (f.u.)/μB	0.3	0.05	1	0.8	0.5	0.06	0.05	0.05

---

Magnetic hysteresis and saturation occur at 2 K for the Ln = Gd–Lu samples, and the Gd³⁺, Dy³⁺, and Er³⁺ samples show history dependent magnetisation, as shown in Fig. 9. The Tb⁴⁺ and Yb³⁺ samples show small magnetic anisotropy with coercive fields of 20 mT at 2 K, and the other samples have isotropic magnetization, as shown in Fig. 9 inset. Due to the Mn ions in the tetramers forming AFM clusters, their total moments are almost zero at 2 K. The saturated moment M_s is therefore determined by the moments of the magnetic ions on the M1 site. The early Ln³⁺ samples have large M_s = 0.3–1 μ_B, close to that of ferromagnetic spin freezing Mn⁴⁺ cations on M1 site. The later Ln³⁺, and Tb⁴⁺, samples have a much smaller M_s ≈ 0.05 μ_B, similar to that of the Sn⁴⁺ and In³⁺ analogue materials.^17,24 This probably indicates AFM spin freezing of the Mn⁴⁺ ions on the M1 site. The different behaviours of the samples with early and late Ln cations reflects the changing difference in size between the Ln and Mn ions. A large size difference, as is found for the early Ln materials, tends to introduce oxygen vacancies and thus Mn³⁺ ions in the c-BaO₃ layers, resulting in a Mn³⁺–O²⁻–Mn⁴⁺ local ferromagnetic network and ferromagnetic spin freezing clusters.

Fig. 9 Isothermal M–H curves at 2 K for the Ba₅Ln_1−xMn_4+yO_15−δ (Ln = Gd–Lu) samples. The inset expands the low field region of the plot.

Neutron diffraction data were collected for Ba₅Er_0.82Mn_4.17O_15−δ at 1.5 K, 10 K and 60 K, as shown in Fig. 10. No magnetic scattering is observed below the T_F = 3.0 K transition observed in the susceptibility, suggesting that this transition is due to short-range spin freezing, though further neutron studies, AC magnetic susceptibility measurements and μSR experiments will be useful in characterising this state more fully.

Fig. 10 Neutron diffraction profiles of Ba₅Lu_0.78Mn_4.21O_15−δ at different temperature. Fitting to the 1.5 K pattern is shown.

Conclusions

The new 10H perovskites Ba₅Ln_1−xMn_4+yO_15−δ (Ln = Sm–Lu) have been synthesised by high temperature solid-state reactions. These materials have an ordered arrangement of Ln and Mn cations. The face-sharing octahedral centres are occupied by Mn⁴⁺ ions, whilst the corner-sharing octahedra are occupied by Ln and Mn cations. Oxygen vacancies are found in the c-BaO₃ layers. Within tetramers of MnO₆ octahedra, the two Mn_M2O₆ octahedra are distorted but the two Mn_M3O₆ octahedra are almost undistorted. The Mn_M2O₆ distortion is related to the size of the M ion in the Ba₅MMn₄O_15−δ 10H perovskite structure.

The Ba₅Ln_1−xMn_4+yO_15−δ perovskites (Ln = Sm–Lu) are paramagnetic at high temperatures. The Ln = Gd–Lu samples have spin freezing transitions at T_F = 3.0–34 K, which have been characterised from their magnetic hysteresis at 2 K and low-temperature neutron diffraction. The spin freezing transition is related to the partial occupation of Mn⁴⁺ ions on the M1 sites of the 10H structure. The saturated moment is also related to the magnetic ions on the M1 site.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Grant 21271014, Grant 21301077 and Grant 51662013), and Scientific Research Key Program of Beijing Municipal Commission of Education (KM201010005019). We thank J. P. Attfield for the advice and encouragement and Alexander Browne for English modification. We acknowledge the support of the National Institute of Standards and Technology, U. S. Department of Commerce, in providing the neutron research facilities used in this work.

Notes and references

1. J. M. Gonzalez-Calbet, M. Parras, J. M. Alonso and M. Vallet-Regi, J. Solid State Chem., 1993, 106, 99.
2. T. Negas and R. S. Roth, J. Solid State Chem., 1971, 3, 323.
3. E. J. Cussen and P. D. Battle, Chem. Mater., 2000, 12, 831.
4. J. J. Adkin and M. A. Hayward, Chem. Mater., 2007, 19, 755.
5. J. J. Adkin and M. A. Hayward, J. Solid State Chem., 2006, 179, 70.
6. N. Floros, C. Michel, M. Hervieu and B. Raveau, J. Solid State Chem., 2002, 168, 11.
7. N. Floros, C. Michel, M. Hervieu and B. Raveau, Chem. Mater., 2000, 12, 3197.
8. G. M. Keith, C. A. Kirk, K. Sarma, N. M. Alford, E. J. Cussen, M. J. Rosseinsky and D. C. Sinclair, Chem. Mater., 2004, 16, 2007.
9. L. Miranda, A. Feteira, D. C. Sinclair, K. Boulahya, M. Hernando, J. Ramírez, A. Varela, J. M. González-Calbet and M. Parras, Chem. Mater., 2009, 21, 1731.
10. L. Miranda, J. Ramirez-Castellanos, M. Hernando, A. Varela, J. M. Gonzalez-Calbet and M. Parras, Eur. J. Inorg. Chem., 2007, 15, 2129.
11. L. Miranda, D. C. Sinclair, M. Hernando, A. Varela, A. Wattiaux, K. Boulahya, J. M. Gonzalez-Calbet and M. Parras, Chem. Mater., 2009, 21, 5272.
12. L. Miranda, A. Feteira, D. C. Sinclair, M. G. Hernandez, K. Boulahya, M. Hernando, A. Varela, J. M. Gonzalez-Calbet and M. Parras, Chem. Mater., 2008, 20, 2818.
13. L. Miranda, J. Ramirez-Castellanos, A. Varela, J. M. Gonzalez-Calbet, M. Parras, M. Hernando, M. T. Fernandez-Diaz and M. G. Hernandez, Chem. Mater., 2007, 19, 1503.
14. J. G. Zhao, L. X. Yang, Y. Yu, F. Y. Li, R. C. Yu and C. Q. Jin, J. Solid State Chem., 2008, 181, 1767.
15. C. L. Yin, G. B. Li, W. A. Kockelmann, J. H. Lin and J. P. Attfield, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 094420.
16. N. A. Jordan and P. D. Battle, J. Mater. Chem., 2003, 13, 2220.
17. C. L. Yin, G. B. Li, T. N. Jin, L. P. You, J. L. Tao, J. W. Richardson, C. K. Loong, J. L. Sun, F. H. Liao and J. H. Lin, Chem. Mater., 2008, 20, 2110.
18. N. Creon, C. Michel, A. Hervieu, A. Maignan and B. Raveau, Solid State Sci., 2003, 5, 243.
19. X. J. Kuang, C. Bridges, M. Allix, J. B. Claridge, H. Hughes and M. J. Rosseinsky, Chem. Mater., 2006, 18, 5130.
20. H. Yang, Y. K. Tang, L. D. Yao, W. Zhang, Q. A. Li, F. Y. Li, C. Q. Jin and R. C. Yu, J. Alloys Compd., 2007, 432, 283.
21. H. Yang, Z. E. Cao, X. Shen, W. J. Feng, J. L. Jiang, J. F. Dai and R. C. Yu, J. Mater. Sci., 2008, 43, 5679.
22. A. F. Fuentes, K. Boulahya and U. Amador, J. Solid State Chem., 2004, 177, 714.
23. C. Rabbow and H. Mullerbuschbaum, Z. Naturforsch., B: J. Chem. Sci., 1994, 49, 1277.
24. C. L. Yin, G. B. Li, T. N. Jin, J. L. Tao, J. W. Richardson, C. K. Loong, F. H. Liao and J. H. Lin, J. Alloys Compd., 2010, 489, 152.
25. A. M. Mario, M. Olivier, C. Silviu, J. C. Gabriel and H. G. Gilles, J. Solid State Chem., 2013, 198, 186.
26. C. L. Yin, G. B. Li, W. A. Kockelmann, F. H. Liao, J. P. Attfield and J. H. Lin, Chem. Mater., 2010, 22, 3269.
27. A. M. Mario, M. Olivier, P. Caroline, R. Pascal, C. Silviu and H. G. Gilles, New J. Chem., 2015, 39, 829.
28. A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory Report, LAUR, 2004, pp. 86–748, unpublished.
29. Y. Shimoda, Y. Doi, Y. Hinatsu and K. Ohoyama, Chem. Mater., 2008, 20, 4512.
30. Y. Shimoda, Y. Doi, M. Wakeshima and Y. Hinatsu, J. Solid State Chem., 2009, 182, 2873.
31. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751.
32. J. A. Oberteuffer and J. A. Ibers, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1970, 26, 1499.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra02183f

---