Structural and magnetic characterisation of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ and its relationship to “Bi₂Sr₂Nb₂MnO₁₂”

Abstract

A new Aurivillius phase (generic formula M₂A_n−1B_nO_3n+3) has been synthesized with n = 3 and containing manganese, Bi₂Sr_1.4La_0.6Nb₂MnO₁₂. The structure has been investigated by X-ray and neutron powder diffraction and found to be tetragonal (I4/mmm) at temperatures down to 2 K, with a = 3.89970(7) Å, c = 32.8073(9) Å at 2 K. There is significant cation disorder between Bi³⁺ (predominantly on the M sites) and Sr²⁺ and La³⁺ which prefer the A sites: 19(2)% of Bi³⁺ occupy the A sites. This disorder, leading to occupancy of M sites by Sr²⁺, is thought to relieve strain due to size-mismatch between the fluorite-like and perovskite-like blocks. A high level of order exists between Mn and Nb on the B sites, with Mn located predominantly (76.1(6)%) in the central B site whilst Nb preferentially occupies the lower symmetry, outer B site, where it undergoes an out-of-centre displacement towards the fluorite-like blocks. Magnetic measurements indicate that this material displays spin-glass behaviour on cooling. Synthesis of the Mn⁴⁺ analogue Bi₂Sr₂Nb₂MnO₁₂ was unsuccessful, possibly due to the small size of the Mn⁴⁺ cation.

---

Introduction

The Aurivillius phases are a class of layered bismuth oxide materials¹ which have been the focus of much research due to their ferroelectric properties,^2,3 particularly in the light of recent reports of their fatigue-free behaviour in thin film form. This has resulted in their use as non-volatile ferroelectric random access memories (FeRAMs).⁴ Aurivillius phases, of general formula M₂A_n−1B_nO_3n+3 can be described as layered intergrowth structures composed of alternating fluorite like [M₂O₂]²⁺ and perovskite like [A_n−1B_nO_3n+1]²⁻ blocks, where M is generally Bi³⁺, A is generally a large cation such as a lanthanide or group II metal and B is generally a smaller d⁰ transition metal such as Nb⁵⁺, Ti⁴⁺ or W⁶⁺.

It has been suggested by Yee et al.⁵ that these materials might display interesting electromagnetic properties if they could contain transition metals with partially filled d orbitals, but there are very few reports of such materials. A well characterized example is Bi₅Ti₃FeO₁₅, with Fe³⁺ and Ti⁴⁺ statistically distributed over the B sites. At temperatures above the ferroelectric Curie temperature T_c, the material is of ideal, tetragonal symmetry, with space group I4/mmm. On cooling, it undergoes a single step, crystallographic transition, coincident with the paraelectric–ferroelectric transition at 730 °C, to orthorhombic symmetry (space group A2₁am), in which it exhibits a polar displacement of the A site cations and cooperative tilting of the BO₆ octahedra.^6,7

There are striking structural similarities between the Aurivillius structure and that of the low-dimensional oxides that display colossal magnetoresistance (CMR), such as the Ruddlesden–Popper phase La_1.2Sr_1.8Mn₂O₇,⁸ as both contain low dimensional perovskite blocks. With this in mind, we were keen to investigate the possibility of incorporating manganese into the Aurivillius B site.

Yu et al.⁹ recently reported the synthesis and characterisation of a novel manganese-containing n = 3 Aurivillius phase Bi₂Sr₂Nb₂MnO_12−δ. Structure refinement using X-ray powder diffraction (XRPD) data suggested orthorhombic symmetry (Fmmm, a = 5.5243(4) Å, b = 5.5246(3) Å, c = 32.9996(9) Å). The authors suggested that this material exhibits several unusual structural features: oxygen vacancies in the perovskite blocks (equatorial site in the central octahedra); manganese–niobium ordering over the perovskite B sites and complete bismuth–strontium ordering over the M and A sites. However, the structural model provides an unrealistically short Mn–O_(ap) bond length (1.57 Å) and the determinations of the manganese oxidation state were inconclusive.

We have therefore re-examined this composition, and extended the investigation to the related material Bi₂Sr_1.4La_0.6Nb₂MnO₁₂. It was hoped that the use of neutron powder diffraction (NPD) would allow the precise location and fractional occupancies of oxygen sites to be determined and would be very sensitive to ordering of Mn and Nb ions over the B sites, owing to the significant difference in their neutron scattering lengths. Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ is essentially single phase and we report here its structural and magnetic properties.

Experimental

Polycrystalline samples were prepared by standard solid state methods. Stoichiometric quantities of Bi₂O₃, SrCO₃, La₂O₃, Nb₂O₅ and Mn₂O₃ (Aldrich, >99%) were intimately ground together and heated in air. Typical reaction conditions were 12 hours at 820 °C, 12 hours at 880 °C followed by 24 hours at 920 °C, with intermittent grindings. A final firing at 1150 °C was required for the synthesis of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ in order to decompose the impurity phase Bi₂SrNb₂O₉. XRPD data were collected over a period of 10 hours on a Siemens D5000 diffractometer (operating in transmission mode with a monochromated Cu K_α1 radiation source with a step size of 0.0204° and a position sensitive detector). NPD data were collected at 2 K on the high resolution powder diffractometer D2B at the Institut Laue-Langevin in Grenoble (λ = 1.59432 Å, Ge monochromator). Rietveld structural refinements were carried out using the GSAS suite of programs.¹⁰ Magnetic measurements were performed using a Physical Properties Measurements System (Quantum Design). Both field cooled and zero-field cooled measurements were carried out using the DC extraction method, in an applied field of 1000 Oersted. The AC susceptibility was measured with an amplitude of 10 Oe. Iodometric titrations were not performed due to the low solubility of the sample.

Results and discussion

Bi₂Sr₂Nb₂MnO₁₂

Our attempts to prepare this phase using the reported synthetic method were unsuccessful. A main phase, isostructural with the reported material, was formed after heating at 820–880 °C, but impurity phases (Bi₂O₃, Bi₂SrNb₂O₉ and a perovskite phase Sr(Mn,Nb)O₃) were also present. Further heating at temperatures up to 960 °C decreased the Bi₂O₃ and Bi₂SrNb₂O₉ content, but the perovskite content remained significant and increased on further heating. The synthetic conditions were modified (for example, synthesis in air, in oxygen atmosphere, heating at lower or higher temperatures) in an attempt to obtain a single phase, but we were unable to prepare a sample free from the perovskite impurity. This perovskite phase is not immediately apparent on consideration of the XRPD data which is dominated by the heavy scattering bismuth-containing phases; the most intense reflection of the perovskite phase is almost coincident with the (110) peak of the main phase and only a small shoulder is noticeable, Fig. 1. Refinement of NPD data revealed that the content of this phase was significant (Sr(Mn,Nb)O₃ weight fraction 14%, Bi₂O₃ weight fraction 0.7%, main phase weight fraction 85.2%). This refinement suggested that the main phase was bismuth rich with approximate composition Bi_2.8Sr_1.2Nb_2.34Mn_0.66O₁₂, although this has not been confirmed from refinement of this model.

Fig. 1 a) Observed (+), calculated and difference XRPD profiles of Bi₂Sr₂Nb₂MnO₁₂ at 298 K. The reflection positions of Bi₂O₃, Sr(Mn,Nb)O₃ and the main phase (from top to bottom) are marked (|) and inset b) comparison of XRPD profiles in range 25 ≤ 2θ ≤ 40° for Bi₂Sr₂Nb₂MnO₁₂ and Bi₂Sr_1.4La_0.6Nb₂MnO₁₂, vertical tick marks indicating reflection positions of Bi₂O₃, perovskite phase Sr(Mn,Nb)O₃ and main Aurivillius phase.

It is unclear why our attempts to prepare Bi₂Sr₂Nb₂MnO₁₂ were unsuccessful. A possibility is that the structure is unable to support a significant Mn⁴⁺ content. Armstrong and Newnham¹¹ report the inflexibility of the B site in Aurivillius phases, which allows accommodation of metals with radii only in the range 0.58 Å (W⁶⁺) to 0.645 Å (Fe³⁺). The ionic radius of Mn⁴⁺ (0.53 Å)¹² lies outside this range and this could explain our difficulties in synthesizing the pure Mn⁴⁺ phase and the formation of a bismuth rich material. To the best of our knowledge, there are no Aurivillius phases with n > 1 that contain oxygen vacancies in the perovskite blocks, and Snedden et al.¹³ suggested that oxygen vacancies are not tolerated by the structure. It therefore seems likely that the formation of an oxygen deficient phase, Bi₂Sr₂Nb₂MnO_12−δ, containing a larger manganese cation in a lower oxidation state, would be difficult. Synthesis of the Mn^3.4+ phase Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ was therefore attempted, in which charge compensation requires no oxygen vacancies.

Structure refinement of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂

It is common for Aurivillius phases to adopt slightly distorted structures of orthorhombic symmetry, as was reported for Bi₂Sr₂Nb₂MnO₁₂;⁹ however, preliminary refinements of the XRPD data gave no evidence of any distortions from the ideal tetragonal structure. The pattern was indexed in a body centred tetragonal cell (space group I4/mmm, a ≈ 3.907 Å, c ≈ 32.85 Å) and the atomic coordinates of Bi₄Ti₃O₁₂¹⁴ were used as a starting model. This preliminary refinement was carried out assuming complete ordering of bismuth and strontium over the M and A sites, and complete ordering of manganese and niobium over the inner and outer B sites respectively, as suggested by Yu et al.⁹ Although the calculated profile was in reasonable agreement with the observed data (R_wp = 6.42%, R_p = 5.11%, χ² = 8.724), there were inconsistencies in the intensities of some reflections particularly the high d spacing reflections (002), (004) and (006). The model did not seem chemically feasible, containing a very short Mn–O_(ap) bond. Also, the isotropic thermal factors observed for the A site were significantly lower than those for the M site. This is reminiscent of the model described by Blake et al.¹⁵ for Bi₂BaNb₂O₉ and suggested the presence of cation disorder between the M and A sites. The model was noticeably improved when this disorder was included in the refinement.

The NPD data collected at 2 K revealed no evidence of distortions from the ideal tetragonal I4/mmm symmetry. The initial model used in this second refinement was based on Bi₄Ti₃O₁₂¹⁴ with complete cation ordering over the M and A sites and fully ordered manganese and niobium over the inner and outer B sites, respectively. It was of interest to refine the anti-site defects of strontium and lanthanum as independently as possible but a unique solution is impossible using a single data set.¹⁰ However, the neutron scattering lengths of bismuth and lanthanum are almost identical: b Bi = 8.53 fm, La = 8.24 fm, and so the stoichiometry was modelled as Bi*_2.6Sr_1.4Nb₂MnO₁₂, with Bi* assigned a scattering length equal to the weighted average for bismuth and lanthanum (8.46 fm). The Rietveld refinement was carried out with background parameters (linear interpolation function), histogram scale factor, diffractometer zero point, lattice parameters, peak shape (pseudo-Voigt), atomic coordinates and anisotropic thermal factors refined. Constraints were applied to the thermal factors of atoms on the same site, and also to site occupancies to ensure that the overall stoichiometry remained constant and that each site was fully occupied. The ordering of Bi* and Sr over the M and A sites was refined and gave a noticeable improvement in the agreement between the observed and calculated data sets. NPD is very sensitive to manganese–niobium ordering due to the large contrast in their scattering lengths (Mn = −3.73 fm, Nb = 7.054 fm) and refinement of the B site occupancies indicated significant disorder. Refinement of the fractional occupancies of the oxygen sites indicated no significant deviation from full occupancy, and these parameters were subsequently constrained at unity. As suggested by Hervoches and Lightfoot,¹⁶ it would be reasonable to expect bismuth and strontium to occupy slightly different positions on the M and A sites. However, allowing their atomic coordinates to refine independently provided no improvement. Likewise, a single set of coordinates was assigned to cations at the same B site. The Rietveld profile of this final model is shown in Fig. 2. This final model from the NPD refinement was then used in the refinement of the XRPD data in order to determine the fractional occupancies of bismuth and lanthanum over the A and M sites. The coordinates and fractional occupancies of each oxygen site were fixed, as were the fractional occupancies of strontium on both the A and M sites and of manganese and niobium on the two B sites. Isotropic thermal factors were refined for each atom, with constraints on all atoms on the same site, for the oxygen sites and for both B sites. The final atomic parameters are given in Table 1 and selected bond lengths in Table 2; the structure is shown in Fig. 3.

Fig. 2 Observed (+), calculated and difference NPD profiles of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ at 2 K, the reflection positions are marked (|).

Fig. 3 The structure of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ showing Mn/NbO₆ octahedra.

Table 1 Atomic parameters of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ from refinement of NPD data collected at 2 K^a

Atom	x	y	z	Frac.	U 11 × 100	U 22 × 100	U 33 × 100
a I4/mmm, a = 3.89970(7) Å, c = 32.8073(9) Å, χ^2 = 3.610, Rwp = 4.82%, Rp = 3.69%. b Denotes occupancy (±0.02) determined from XRPD refinement, see text. c Denotes fractional occupancies fixed, see text.
Sr(1)	0.5	0.5	0.05886(9)	0.58(3)	1.89(11)	1.89(11)	1.77(22)
La(1)	0.5	0.5	0.05886(9)	0.22^b	1.89(11)	1.89(11)	1.77(22)
Bi(1)	0.5	0.5	0.05886(9)	0.19^b	1.89(11)	1.89(11)	1.77(22)
Sr(2)	0.5	0.5	0.21319(9)	0.12(3)	1.62(10)	1.62(10)	3.79(22)
La(2)	0.5	0.5	0.21319(9)	0.08^b	1.62(10)	1.62(10)	3.79(22)
Bi(2)	0.5	0.5	0.21319(9)	0.81^b	1.62(10)	1.62(10)	3.79(22)
Mn(1)	0.5	0.5	0.5	0.759(6)	0.32(54)	0.32(54)	0.32(54)
Nb(1)	0.5	0.5	0.5	0.241(6)	0.32(54)	0.32(54)	0.32(54)
Mn(2)	0.5	0.5	0.37430(9)	0.121(3)	0.55(9)	0.55(9)	1.11(24)
Nb(2)	0.5	0.5	0.37430(9)	0.879(3)	0.55(9)	0.55(9)	1.11(24)
O(1)	0.5	0.0	0.0	1^c	0.44(16)	5.49(28)	4.81(35)
O(2)	0.5	0.0	0.25	1^c	0.94(10)	0.94(10)	1.59(19)
O(3)	0.5	0.5	0.44141(11)	1^c	4.41(16)	4.41(16)	0.12(24)
O(4)	0.5	0.5	0.32002(12)	1^c	5.74(20)	5.74(20)	0.76(23)
O(5)	0.5	0.0	0.11759(9)	1^c	1.21(13)	3.18(16)	4.10(20)

---

Table 2 Selected bond lengths in Å from NPD data collected at 2 K

Sr/Bi(1)–O(1) × 4	2.7441(22)	Mn/Nb(1)–O(1) × 4	1.94985(4)
Sr/Bi(1)–O(3) × 4	2.75752(5)	Mn/Nb(1)–O(3) × 2	1.922(4)
Sr/Bi(1)–O(5) × 4	2.7413(29)	Mn/Nb(2)–O(3) × 1	2.202(5)
Sr/Bi(2)–O(2) × 4	2.2934(16)	Mn/Nb(2)–O(4) × 1	1.781(5)
Sr/Bi(2)–O(4) × 4	2.9649(20)	Mn/Nb(2)–O(5) × 4	1.9679(6)

---

Structure of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂

It was hoped that doping the A site with trivalent lanthanum would reduce the B site, increasing the average size of the B site cations which might facilitate the synthesis of a single phase sample. Refinement of the NPD data revealed this phase to be virtually pure, with just a trace of Sr(Mn,Nb)O₃ (weight fraction < 1%).

It has been predicted that odd layer Aurvillius phases will adopt tetragonal symmetry (described by I4/mmm) and any distortions occurring at lower temperatures will be described by either the orthorhombic space group B2cb¹⁷ or the monoclinic space group B1a1.¹⁸ The orthorhombic space group Fmmm reported for Bi₂Sr₂Nb₂MnO₁₂⁹ is unusual and care was taken in assigning the high symmetry I4/mmm space group to this A site doped analogue, particularly for the refinement of the low temperature NPD data. In order to confirm this space group, preliminary refinements were also carried out in the previously reported space groups B2cb¹⁷ and B1a1,¹⁸ however, the slight improvement in fit was insufficient to justify this lowering of symmetry. This assignment of the centrosymmetric space group precludes the possibility of the material displaying ferroelectric properties.

Strain often occurs in Aurivillius phases, due to the mismatch in a parameters of the smaller fluorite like layer and the larger perovskite like blocks.¹¹ This strain is often relieved by octahedral tilting in the perovskite block, lowering the symmetry of the structure, as exhibited by Bi₅Ti₃FeO₁₅^6,7 and Bi₄Ti₃O₁₂.¹⁶ However, another mechanism for decreasing the size mismatch involves cation disorder over the M and A sites, as is observed here. Originally, the fluorite blocks were thought to be inflexible in terms of composition, only accommodating bismuth cations. Further research demonstrated that bismuth could be substituted for other lone pair cations such as lead and tellurium¹⁹ but it was commonly held that only cations in possession of a stereochemically active lone pair could occupy this square antiprismatic site. The possibility of anti-site defects occurring, involving spherical cations partially occupying the M site, was first suggested by Smolenski et al.² Blake et al.¹⁵ have since confirmed this for the n = 2 phases Bi₂ANb₂O₉ (A = Ba, Sr, Ca) and more recently, Hervoches and Lightfoot¹⁶ and Haluska and Misture²⁰ have reported site mixing in n = 3 Aurivillius materials.

The results of these refinements suggest that 19(2)% of bismuth cations occupy the A site. This degree of disorder is comparable to that reported for other n = 3 Aurivillius phases such as Bi₂Ln₂Ti₃O₁₂ (Ln = La 18.1(1)%; Ln = Pr 14.8(3)%; Ln = Nd 14.0(3)% and Ln = Sm 12.1(3)%).²¹ However, it is interesting to note that our refinements indicate that strontium has a slightly higher preference for the A site over the M site than lanthanum, with 26.66(6)% of lanthanum occupying the M site, compared with only 17.14(4)% of strontium. Hyatt et al.²¹ report that for a range of rare earth cations, the degree of cation disorder decreases as the size of the rare earth decreases. In contrast to what we observe, it might be expected that the smaller La³⁺ cation would be less inclined to occupy the M site than the larger Sr²⁺ cation (twelve coordinate ionic radii La³⁺ = 1.36 Å, Sr²⁺ = 1.44 Å, eight coordinate ionic radii La³⁺ = 1.16 Å, Bi³⁺ = 1.17 Å).¹²

As previously mentioned, a likely driving force for this cation disorder is relief of strain due to size mismatch between the fluorite like and perovskite like blocks. Armstrong and Newnham¹¹ calculated the ideal a parameter of the [Bi₂O₂]²⁺ blocks to be 3.80 Å (based on consideration of the isostructural red lead II oxide), and derived an expression for the ideal a parameter of the perovskite block [A₂B₃O₁₀]:

a = 1.33r_R + 0.60r_M + 2.36 Å

where r_R is the 6 coordinate ionic radius of the B site cations and r_M is the 8 coordinate radius of the A site cations. Using the following mean ionic radii (calculated from ionic radii published by Shannon¹²) [Sr_1.4La_0.6] = 1.23 Å, [Nb₂Mn^3.4+] = 0.626 Å, the ideal a parameter of the perovskite block [Sr_1.4La_0.6Nb₂MnO₁₀] was calculated as 3.93 Å. This is significantly bigger than that of the ideal fluorite layer. The observed a parameter reflects the compromise reached, as a result of this cation disorder.

Anti-site defects between the larger Sr²⁺ cation and bismuth will have the effect of decreasing the average A site cation radius, therefore reducing the a parameter of the perovskite block and reducing the size mismatch. However, the driving force for disordering of La³⁺ over the two sites is less clear, as La³⁺ and Bi³⁺ have almost identical ionic radii (1.16 Å and 1.17 Å respectively).¹² As suggested by Hervoches and Lightfoot,¹⁶ perhaps it is more surprising that there is relatively little disorder between these two, similarly sized cations. A possible reason might lie in the preference of the spherical lanthanide for the undistorted coordination environment of the A site, rather than the square antiprismatic M site.

It is likely that bismuth is not located in exactly the same environment as strontium and lanthanum on both the M and A sites, and this provides a reason for the slightly elevated thermal factors for these sites. Hervoches and Lightfoot¹⁶ were able to illustrate that in Bi_1.8Sr_2.2Ti_0.8Nb_2.2O₁₂, the strontium cation does indeed adopt a much more regular coordination environment than bismuth on the M site.

High anisotropic thermal factors (U₂₂ and U₃₃) are observed for O(1) and O(5), which seem typical of cation disordered, n = 3 Aurivillius phases such as Bi₂Ln₂Ti₃O₁₂²¹ and Bi_1.8Sr_2.2Ti_0.8Nb_2.2O₁₂.¹⁶ It has been suggested that these reflect subtle differences in the coordination environments of bismuth and spherical A site cations in this site and are consistent with twisting of the inner BO₆ octahedra around the [001] direction. Attempts to move O(1) from the ideal 4c position to the 16l site were unsuccessful, resulting in an unstable refinement. The thermal factors for O(4) in the ab plane are also high. This is expected for this structure, although the magnitude may also reflect the slight variations in the coordination environments of the different cations in the M site.

This material displays a high level of ordering of the B site cations, with 76.1(6)% of the manganese ions located in the central, high symmetry B site, while niobium preferentially occupies the lower symmetry outer B site. This ordering was suggested by Yu⁹ from refinement of XRPD data and is confirmed by this NPD study. Similar ordering was reported for the n = 3 Ruddlesden–Popper material Na₂Sr₂Ti₂RuO₁₀,²² in which 63.7(5)% of the Ru⁶⁺ cations are located in the inner B sites of the perovskite blocks, while the Ti⁴⁺ cations show a preference for the outer B sites. It is usual for d⁰ transition metals in the outer octahedral site to show an out-of centre displacement towards O(4) and the [M₂O₂]²⁺ layers. This distortion is evident in this material with the Nb(2)–O(3) bond being significantly longer than the Nb(2)–O(4) bond (2.202(5) Å and 1.781(5) Å, respectively) as illustrated in Fig. 4. The short bond is typical for the outer octahedra in n = 3 Aurivillius phases: for Bi_1.8Sr_2.2Ti_0.8Nb_2.2O₁₂, for example, the equivalent Nb–O distance is only 1.72(1) Å.¹⁶ The displacement, which is often ascribed to the second order Jahn–Teller effect,^23–25 may also be influenced by electrostatic effects; it would induce instability for a transition metal with a partially filled d shell, and this may be the reason for the manganese preference for the central B site of higher symmetry.

Fig. 4 Structure of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ highlighting out-of-centre displacement of cations in outer octahedra towards O4.

Magnetic characterisation of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂

From the low temperature data collected, the divergence of the field cooled (FC) and zero-field cooled (ZFC) datasets at 9.8 K (Fig. 5a) suggests that the material is paramagnetic at higher temperatures, but on cooling becomes a spin glass. Spin glass behaviour is associated with frustrated systems, resulting from competing interactions. Superexchange interactions between Mn³⁺/Mn³⁺ and between Mn⁴⁺/Mn⁴⁺ would be of antiferromagnetic nature, whereas superexchange or double exchange involving Mn³⁺/Mn⁴⁺ would be ferromagnetic. These interactions between disordered Mn³⁺ and Mn⁴⁺ ions could give rise to frustration and spin glass behaviour. Plots of inverse susceptibility versus temperature are not linear (Fig. 5b), indicating that the material does not behave as a Curie–Weiss paramagnet in this temperature range. However, the high temperature behaviour suggests that ferromagnetic exchange is dominant, and the variation of gradient with temperature is consistent with the presence of ferromagnetic clusters linked via antiferromagnetic interactions,²⁶ so that the low temperature state is best described as a ferromagnetic spin-glass. Field sweep measurements at 2 K, Fig. 6, demonstrate significant hysteresis (with a remnant magnetization of ∼0.6 Am² mol⁻¹ and a coercive field of ∼500 Oe) which is also supportive of this magnetic model. Moments calculated from the plots of inverse susceptibility using the Curie–Weiss law are clearly unreliable, but those estimated were similar to those previously reported for Bi₂Sr₂Nb₂MnO₁₂ (5.95 μ_B),⁹ being high and more characteristic of d⁵ Mn²⁺; for example the slope for 100 K < T < 200 K gives a moment of 6.17 μ_B. High moments are again to be expected for a model containing ferromagnetic clusters, and similar behaviour has previously been reported in La_1.2Sr_0.8MnO_4.27, which contains similar corner-linked Mn³⁺/Mn⁴⁺ octahedra.²⁷

Fig. 5 a) Magnetic susceptibility versus temperature plots of Bi₂Sr_1.4La_0.6Nb₂MnO₁₂: ZFC (○), FC (1000 Oe, ●), AC susceptibility (10 Oe, +); b) inverse susceptibility versus temperature plots.

Fig. 6 Magnetic moment against field for Bi₂Sr_1.4La_0.6Nb₂MnO₁₂ measured as a powder at 2 K.

Conclusions

In conclusion, we present here the synthesis and structural and magnetic characterisation of the novel n = 3 Aurivillius phase Bi₂Sr_1.4La_0.6Nb₂MnO₁₂. Rietveld analysis of NPD and XRPD data indicates that this material shows significant disorder between Bi³⁺, Sr²⁺ and La³⁺ cations over the M and A sites. Disorder between the larger Sr²⁺ and bismuth will tend to relieve the strain due to size-mismatch by reducing the a parameter of the perovskite-like blocks and increasing that of the fluorite-like layers. The ordering observed between similarly sized bismuth and lanthanum cations may be a result of the preference of the spherical lanthanide for the more regular A site, rather than the asymmetric M site. A high degree of ordering of the B site cations was observed, with the d⁰ Nb⁵⁺ cations predominantly occupying the outer, lower symmetry B site, where it undergoes an out-of-centre displacement towards the [M₂O₂]²⁺ layers, whereas the manganese preferentially occupies the higher symmetry, central B site. Magnetic measurements indicate that the material shows spin glass behaviour at low temperatures and displays a degree of hysteresis. We were not able to obtain a pure sample of the previously reported Aurivillius phase Bi₂Sr₂Nb₂MnO₁₂, synthetic difficulties may be due to the small size of the Mn⁴⁺ ion.

Acknowledgements

We thank EPSRC for financial support and the ILL for the provision of neutron diffraction facilities. We are grateful to Dr E. Suard for experimental assistance with the collection of NPD data.

References

1. B. Aurivillius, Ark. Kemi, 1949, 1, 499.
2. G. A. Smolenski, V. A. Isupov and A. I. Agranovskaya, Sov. Phys. Solid State (Engl. Transl.), 1953, 3, 651.
3. E. C. Subbarao, J. Phys. Chem. Solids, 1962, 23, 665.
4. C. A.-Paz de Araujo, J. D. Cuchiaro, L. D. McMillan, M. C. Scott and J. F. Scott, Nature, 1995, 374, 627.
5. K. A. Yee, T. A. Albright, D. Jung and M.-H. Whangbo, Angew. Chem., Int. Ed. Engl., 1989, 28, 750.
6. A. Snedden, C. H. Hervoches and P. Lightfoot, Phys. Rev. B, 2003, 67, 92102.
7. C. H. Hervoches, A. Snedden, R. Riggs, S. H. Kilcoyne, P. Manuel and P. Lightfoot, J. Solid State Chem., 2002, 164, 280.
8. Y. Moritomo, A. Asamitsu, H. Kuwahara and Y. Tokura, Nature, 1996, 380, 141.
9. W. J. Yu, Y. I. Kim, D. H. Ha, J. H. Lee, Y. K. Park, S. Seong and N. H. Hur, Solid State Commun., 1999, 111, 705.
10. A. C. Larson and R. B. von Dreele, Los Alamos National Laboratory Report No. LA-UR-86-748, 1987.
11. R. A. Armstrong and R. E. Newnham, Mater. Res. Bull., 1972, 7, 1025.
12. R. D. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751.
13. A. Snedden, S. M. Blake and P. Lightfoot, Solid State Ionics, 2003, 156, 439.
14. C. H. Hervoches and P. Lightfoot, Chem. Mater., 1999, 11, 3359.
15. S. M. Blake, M. J. Falconer, M. M. McCreedy and P. Lightfoot, J. Mater. Chem., 1997, 7, 1609.
16. C. H. Hervoches and P. Lightfoot, J. Solid State Chem., 2000, 153, 66.
17. R. E. Newnham, R. W. Wolfe and J. F. Dorrian, Mater. Res. Bull., 1971, 6, 1029.
18. A. D. Rae, J. G. Thompson, R. L. Withers and A. C. Willis, Acta Crystallogr., Sect. B, 1990, 46, 474.
19. P. Millán, A. Castro and J. B. Torrance, Mater. Res. Bull., 1993, 28, 117.
20. M. S. Haluska and S. T. Misture, J. Solid State Chem., 2004, 177, 1965.
21. N. C. Hyatt, J. A. Hriljac and T. P. Comyn, Mater. Res. Bull., 2003, 38, 837.
22. J. N. Lalena, A. U. Falster, W. B. Simmons, E. E. Carpenter, J. Wiggins, S. Hariharan and J. B. Wiley, Chem. Mater., 2000, 12, 2418.
23. N. S. P. Bhuvanesh and J. Gopalakrishnan, J. Mater. Chem., 1997, 7, 2297.
24. R. A. Wheeler, M.-H. Whangbo, T. Hughbanks, R. Hoffmann, J. K. Burdett and T. A. Albright, J. Am. Chem. Soc., 1986, 108, 2222.
25. M. Kunz and I. D. Brown, J. Solid State Chem., 1995, 115, 395.
26. K. V. Rao, M. Fähnle, E. Figueroa, O. Beckman and L. Hedman, Phys. Rev. B, 1983, 27, 3104.
27. R. K. Li and C. Greaves, J. Solid State Chem., 2000, 153, 34.

---