The Ba2LnFeNb4O15 “Tetragonal Tungsten Bronze”:

towards RT composite multiferroics

M. Jossea,*, O. Bidaultb, F. Roullanda, E. Castela, A. Simona, D. Michaua, R. Von der
Mühlla, O. Nguyena and M. Maglionea
a
ICMCB-CNRS, Université Bordeaux 1, 33608 Pessac, France
b
I.C.B., CNRS - Université de Bourgogne, 21078 Dijon, France

*
To whom correspondence should be addressed: [email protected]

1
 Abstract
Several Niobium oxides of formula Ba2LnFeNb4O15 (Ln = La, Pr, Nd, Sm, Eu, Gd)
with the “Tetragonal Tungsten Bronze” (TTB) structure have been synthesised by
conventional solid-state methods. The Neodymium, Samarium and Europium compounds are
ferroelectric with Curie temperature ranging from 320 to 440K. The Praseodymium and
Gadolinium compounds behave as relaxors below 170 and 300 K respectively. The
Praseodymium, Neodymium, Samarium, Europium and Gadolinium compounds exhibit
magnetic hysteresis loops at room temperature originating from traces of a barium ferrite
secondary phase. The presence of both ferroelectric and magnetic hysteresis loops at room
temperature allows considering these materials as composites multiferroic. Based on crystal-
chemical analysis we propose some relationships between the introduction of Ln3+ ions in the
TTB framework and the chemical, structural and physical properties of these materials.

2
 Introduction

Much attention has been paid in the recent years to multiferroic materials, and
particularly to those presenting both ferromagnetism and ferroelectricity. Magnetoelectric
multiferroics, which will be referred to as multiferroics in the following, could be used, for
example, for the design of new generation of Random Access Memories (RAMs) in which
both the electric and magnetic polarisation are used for data storage. Aside from the
processing of multiferroic thin films and composites [1], the search for bulk multiferroic is a
great challenge.
However up to now the candidate multiferroic materials for room temperature
applications are very few. A large majority of the “multiferroic” publications of the recent
years deal with perovskite-related materials [2,3]. From this abundant literature, two
limitations are widely accepted. First, the electronic structure of the A an B cations in ABO3
multiferroics is hardly compatible with a magnetic ordering at elevated temperatures [4].
Second, a coupling between the ferroelectric and magnetic order may be obtained when the
material exhibit a non standard, e.g. spiral, magnetic structure [5], or a toroidal order [6]. So
far, only the Bismuth ferrite BiFeO3 has demonstrated room temperature multiferroic
behaviour, and yet the complex antiferromagnetic ordering yields to a very small remnant
magnetisation [7].
To face both these limitations, we decided to investigate completely different
networks, namely the Tetragonal Tungsten Bronze (TTB) structural family. At the crystal
chemistry level, this framework has two main advantages. First, the number of cationic sites is
much higher than in perovskites (5 instead of 2) thus enabling extended substitutions
opportunities to foster magnetic interactions. Next, TTB compounds are known to display
incommensurate polar state [8,9], that may favour a coupling between magnetic and
ferroelectric order. On investigating an extended family of TTB compounds, this paper is an
experimental attempt towards this search for alternative multiferroics.
The “Tetragonal Tungsten Bronze” (TTB) structural type is related to the Potassium
tungstate K0.475WO3, the structure of which was elucidated by Magneli [10]. The
characteristic feature of the TTB crystal structure is the three types of open channels that
develop within its octahedral framework (fig. 1). These features allowed solid state chemists
to perform a wide range of substitutions, either in these channels or within the octahedral
framework itself, which led to the discovery of Ba2NaNb5O15 [11], one of the first and most
significant representative of TTB niobates. The physical properties of Ba2NaNb5O15

3
(Ferroelectricity, non-linear optics…) gave rise to extended investigations on TTB niobates,
and to the discovery of many new ferroelectrics and more recently relaxors [12] compositions.
Besides this literature - and forty five years before the “revival of the magnetoelectric
effect” [7] - Fang & Roth reported, in a short paper without experimental data [13], some of
the title compositions as ferroelectric and ferrimagnetic materials, but did not report further
investigations in later articles. A few years later, Krainik & al. [14] synthesized several rare-
earth/iron substituted TTB niobates, and found some of them to be ferroelectrics. Schmidt, in
his “magnetoelectric classification of materials” [15], referenced two of these TTB niobates.
Ishihara & al. also reported on the TTB fluoride K3Fe5F15 as a potential multiferroic [16].
On the basis of these works and of our own knowledge of TTB niobates, we decided
to reinvestigate the Ba2LnFeNb4O15 (Ln = La, Pr, Nd, Sm, Eu, Gd) system and started an in-
depth study of rare-earth/iron substituted TTB niobates in order to confirm and understand
their multiferroic properties.
In this paper, we report the synthesis and physical-chemical characterisation of several
TTB niobates, with emphasis on their dielectric and magnetic properties. We observed that
several compositions display both ferroelectric and magnetic hysteresis loops at room
temperature. X-Ray microprobe experiments revealed that traces of barium ferrite BaFe12O19
are responsible for room temperature magnetic hysteresis. We also discuss the chemical,
structural and physical properties of this family of TTB niobates in terms of crystal-chemical
behaviour.

4
 Experiments

Ba2LnFeNb4O15 (Ln = La, Pr, Nd, Sm, Eu, Gd) compounds have been obtained by
conventional solid state route, from stoichiometric mixtures of BaCO3, Fe2O3, Nb2O5 and
Ln2O3, all reagents with 99.9% minimal purity grade. Ceramics were sintered at 1300°C and
exhibited densities ranging from 90 to 96% of the nominal density of the products.
X-ray diffraction data were recorded on a Philips XPert pro diffractometer (CuKα1, λ
= 1.54056 Å) with 10 < 2θ < 130° and step = 0.008°. Rietveld refinements were performed
using the program FULLPROF [17]. Experimental diffractograms were fitted using a
Thompson-Cox-Hastings function (profile n°7), without any additional constraints.
Microstructural characterisations were performed on a JEOL 6560 SEM equipped with
an EDS analyser.
X-Ray Microprobe mappings were obtained from a CAMECA SX-100 apparatus (15
kV, 10 µA).
Dielectric measurements were performed on a Wayne–Kerr 6425 component analyser
under dry helium, using gold electrodes, at frequencies ranging from 102 to 2.105 Hz.
Ferroelectric hysteresis loops were performed using an analogical Sawyer Tower
circuit with compensation of stray capacitances and resistances.
Magnetic hysteresis loops were obtained on a SQUID magnetometer (MPMS,
Quantum Design Inc.).

5
 Results and Discussion

After standard solid state processing of powders and ceramics, X-Ray Diffraction
(XRD) evidenced the successful stabilisation of the tetragonal tungsten bronze structure for
all the compositions.
However the XRD diagrams of the Sm, Eu and Gd compounds displayed a few
additional lines consistent with the presence of a Fergusonite phase of formula LnNbO4.
Electron microscopy analysis confirmed the XRD results. We investigated various elaboration
procedures in order to avoid this secondary phase formation. This allowed us to observe the
limited thermal stability of TTB ceramics, which start decomposing at about 1360°C with
formation of Hexagonal Tungsten Bronze (HTB) and Barium Ferrite, both identified during
SEM experiments. However thanks to improved synthesis and sintering steps during the
processing, the amount of Fergusonite could be decreased in Sm, Eu and Gd samples. HTB
and Barium Ferrite secondary phases remained undetectable from XRD and SEM analysis. In
addition we point out that Fergusonite phases have no particular dielectric and magnetic
properties in the investigated temperature range.
Rietveld refinements were performed using an averaged model with a “pseudo-
tetragonal” symmetry (space group Pba2, n°32). The refinements converged to satisfying
agreement factors (Table 1, RB ≈ 6, χ² ≈ 2) and confirmed the stabilisation of the TTB crystal-
structure. The cell parameters are reported in table 1, the b parameter is not reported because a
and b refined values were undifferentiated. The Rietveld plot for the Sm representative, in
which the SmNbO4 secondary phase is taken into account, is displayed as an example in
figure 2 (corresponding atomic coordinates are gathered in table 2). One can notice significant
residues in the difference pattern, which could not be taken into account by any alternative
structural model. There are also some discrepancies of the Biso atomic displacement factors
for the octahedral sites (Nb5+ and Fe3+ ions). These features, and the undifferentiated values of
a and b orthorhombic cell parameters, suggest that in these TTBs the crystal structure may be
modulated. However the “pseudo-tetragonal” structures, even if they are averaged ones,
provide a significant amount of reliable information that allows us to extract the trends
governing the crystal-chemistry of these TTB multiferroics.
In these materials the octahedral framework is statistically occupied by Fe3+ and Nb5+
ions while the Ba2+ ions occupy the pentagonal channels. The rare earths occupy the square
channels. It should be noted that in the case of the gadolinium representative, the Rietveld

6
refinement shows that Ba and Gd atoms are distributed other both pentagonal and square
tunnels.
The crystal structures observed in the Ba2LnFeNb4O15 TTBs are related to the
introduction of the rare-earth cations in the square-shaped tunnels (see fig. 1) which normally
define a 12-coordination. The Ln3+ ions are reluctant to adopt such an environment, and this
results in displacements of the O2- ions defining the surrounding octahedra. The cations
occupying the octahedral sites are almost unaffected by this distorsion. In particular, although
the 4c general position allows any kind of displacement, the corresponding (Nb5+, Fe3+)
cationic sublattice remains almost perfectly tetragonal. The accommodation of the rare earth
thus induces distortions in the octahedral framework. If correlated, these distortions may
induce a modulated crystal structure. We note that, even in absence of Ln substitutions, the
TTB network itself is likely to produce modulated structures, as in the case of Ba2NaNb5O15
[8,9].
The influence of the Ln3+ ions on the TTB framework can be illustrated by the
evolution of the cell volume (tab. 1). The ionic radius of lanthanides ions is known to
decrease with increasing atomic number. Thus the cell volume of the TTB matrix should
decrease accordingly, when going from La to Gd. This trend is observed while going from La
to Sm representatives, but the following representatives, Eu and Gd, points out a different
crystal chemical behaviour. In particular the increase in the cell volume of the Gadolinium
composition is probably related to the Ba/Gd statistical distribution over both pentagonal and
square channels suggested by the corresponding Rietveld refinement. Taking into account the
presence of fergusonite secondary phases in Sm, Eu and Gd TTbs, these observations suggest
that the amount of smaller rare-earths that can be accommodated in the TTB framework is
limited. This limitation is probably related to the distortion induced by the accommodation of
the rare earth, depending on its ionic radius. These statements allow separating rare earth
substituted TTBs into two groups, on crystal-chemical basis: La-Nd (total or almost total
accommodation of the rare earths) and Sm-Gd (partial accommodation of the rare earths and
detection of LnNbO4 from XRD)).
From Scanning Electron Microscopy, the ceramics are highly densified (compactness
> 95%) with micronic particles shaped like faceted rods at the surface. EDS analysis indicated
cationic ratios in agreement with the expected 2:1:1:4 Ba:Ln:Fe:Nb ratio associated with the
Ba2LnFeNb4O15 composition, although Ln ratio was slightly lower in Sm, Eu and Gd
samples. Traces of fergusonite phase were detected in the Pr and Nd samples, suggesting that
in both these materials the accommodation of the rare earth is not complete. Thus SEM results

7
confirm that the tetragonal tungsten bronze structure has been successfully stabilized after the
introduction of paramagnetic cations and indicate that the microstructure of the ceramics is
optimal.
Dielectric measurements on the La sample did not reveal any ferroelectric transition
down to 82K, although the evolution of the dielectric permittivity suggests a possible
transition below 80 K, i.e. beyond our experimental conditions.
The Nd, Sm and Eu representatives are ferroelectric below TC = 323 , 405 and 440 K
respectively, as illustrated by figure 3a. This is confirmed by ferroelectric hysteresis loops
which were obtained at room temperature, i.e. below the transition temperature for the three
samples (figure 4).
On the other hand, the Pr and Gd samples are relaxors-ferroelectrics below 170K and
300K respectively (Fig. 3b). This means that both the real and imaginary parts of the
dielectric permittivity display a broad dielectric maximum whose temperature shifts on
sweeping the frequency. Because of this relaxor-ferroelectric behaviour, no ferroelectric
hysteresis loop could be recorded in these latter materials.
The details of the dielectric properties of all samples are gathered in table 2. One can
note that the dielectric transition temperature and the maximum permittivity follow similar
trends, although the Eu TTB displays a quite high permittivity. However the dielectric
permittivities in Tetragonal Tungsten Bronze ferroelectrics can be affected by order/disorder
phenomenon in the different kind of channels of this framework, and we suspect that this is
the case for Eu TTB. The polarisations extracted from ferroelectric hysteresis loops for the
Nd, Sm and Eu TTBs are also rather low, as compared to the usual polarisations observed in
TTB ferroelectrics (10-20 µC.cm-2). This could be related to a modulated structure (suggested
by Rietveld refinements) that induces misalignments of the local dipoles and thus reduces the
overall polarisation.
The dielectric studies show that a ferroelectric behaviour can be obtained at room
temperature and above in some of these magnetically substituted TTBs. Rietveld studies
indicate that the electric dipoles in the Nd, Sm and Eu samples originate from an off-center
position of the Nb5+ ions within the octahedral sites. All these sites display one short and one
long axial bond length, in agreement with a statistical distribution of niobium and iron over
the whole octahedral framework. Off-center positions of the Nb5+ ions are also observed in
the relaxor Ba2GdFeNb4O15, but display inhomogeneous distortions in the different octahedral
sites.

8
 The evolution of the ferroelectric properties in the Nd, Sm and Eu compounds can be
rationalised from a crystal-chemical point of view, which could be resumed as follows:
- The smaller the rare earth, the greater the distortion of the TTB framework
- The greater the distortion of the TTB framework, the greater the permittivity
- The greater the distortion of the TTB framework, the higher the Curie temperature
These considerations indicate that the accommodation of the rare earth in the TTB
framework is the critical parameter impacting the dielectric properties in this family of
materials.
After this description of ferroelectric properties we will now focus on the magnetic
properties observed in these samples at room temperature.
Magnetic hysteresis loops at 300 K have been obtained from all the samples except the
La one (fig 5). The 20 kOersteds magnetisations are comparable, while the coercive fields are
about 500-700 Oersteds for Pr, Eu and Gd samples and about 2000 Oe for Nd and Sm
samples. It should be noted that all the samples, except the La one, are strongly attracted to a
NdFeB magnet at room temperature.
The details of the dielectric and magnetic properties of the studied samples are
gathered in table 3. It is worth mentioning that the room temperature magnetizations of the
TTB samples evolve in a similar way as the dielectric transition temperature. As was observed
for the dielectric properties, the variation of the magnetic properties with respect to the rare
earth shows that the Ln3+ ions play a critical role concerning the magnetic behaviour of our
TTBs samples.
The magnetic study suggests a successful magnetic substitution, as all the samples,
except the La one, exhibit hysteresis loops at room temperature. The existence of a
spontaneous magnetisation at room temperature in these samples, however, is surprising
considering that :
- The Ln3+ ions are not likely to be magnetically ordered at room temperature
- The Fe3+ ions are diluted and disordered in the [NbO6] octahedral framework
This latter statement in particular, is not in favour of a long range magnetic order.
Thus we will now discuss the possible links between the crystal-chemistry and the magnetic
and ferroelectric properties of these niobates, which may account for the origin of the
magnetic properties.
The trends in magnetic properties, as well as the evolution of dielectric properties, can
be correlated to the evolution of cell volume. As shown by figure 6, the molar magnetisation
of the TTB samples, measured at room temperature under a magnetic field of 20 kOe, and the

9
dielectric ordering temperature evolve inversely proportional with respect to the cell volume.
This correlation is reproducible and is observed for several sets of samples. Thus the
correlation between the magnetic behaviour of TTBs and their structural and dielectric
properties may support an intrinsic origin for the multiferroic character of the Ba2LnFeNb4O15
TTBs (Ln = Pr, Nd, Sm, Eu, Gd).
However X-ray microprobe analysis, initially performed in order to determine more
precisely the rare-earth composition in fergusonite-containing samples, revealed traces of a
Fe-rich secondary phase. This parasitic phase is not detected within the La TTB (fig. 7b and
7a respectively), which also do not exhibit magnetic hysteresis loops at room temperature.
Moreover the Fe-rich phase contains a small amount of barium, suggesting a barium ferrite
BaFe12O19 composition, i.e a widely used permanent magnet. Thus the magnetic hysteresis
loops observed at room temperature in most of the Ba2LnFeNb4O15 samples originate from
traces of barium ferrite.
The presence of barium ferrite, and its amount, can be related to the accommodation of
the rare earth in the TTB matrix. In the La sample, the La3+ ion is totally accommodated
within the TTB framework, and no fergusonite phase LaNbO4 appear. To maintain the charge
balance of the TTB phase, Fe3+ ions must be totally incorporated in the TTB framework of the
La sample. In the other samples (Ln = Pr, Nd, Sm, Eu, Gd) the rare earths are partly
accommodated, and the excess of lanthanide oxide reacts with niobium oxide to form
LnNbO4 fergusonite. Thus, to maintain electrical neutrality of the TTB framework in these
cases, iron must be only partly accommodated in the TTB framework. The excess of iron
oxide reacts with a small proportion of barium oxide to form barium ferrite related phases. It
appears that the nature of the rare earth, and particularly its size, i.e. its ability to be
accommodated within the TTB framework, controls the appearance of the magnetic barium
ferrite phase in our samples.
This latter correlation indicates that the rare earth introduced in the TTB framework, in
fine, has a direct influence on its chemical stability, its crystal-chemistry and its physical
properties. Beyond the correlation shown in fig. 6, one has to take into account the additional
and highly critical parameter of the chemical stability of these TTB compositions.
Besides this correlation between chemical stability, crystal-chemistry and physical
properties, it is worth mentioning that all the TTB samples (except the La one), display both
spontaneous polarisation and magnetisation and can be considered as composites
multiferroics. The particular interest of these multiferroic composites is that the magnetic
phase is generated in situ, during the elaboration of the sample, which should favour the

10
quality of the interfaces between the ferroelectric and magnetic part of the composite.
Adequate chemical substitutions can also modify the amount and properties of this magnetic
phase. Moreover, four of these composites materials display both ferroelectric and
ferromagnetic hysteresis loops at room temperature. The particular case of the Pr and Gd
samples should be noted, for they are, below 170 and 300 K respectively, two examples of
composite relaxor multiferroics. Last but not least, this work shows that it is possible to
introduce magnetic cations in a ferroelectric framework without loosing the ferroelectric
properties.

11
 Conclusion

The experiments reported in this paper demonstrate that five over the six studied
Tetragonal Tungsten Bronze samples are composite multiferroics below their
ferroelectric/relaxor Curie temperature. The multiferroic properties of these materials are
directly influenced by the nature of the rare earth accommodated in the TTB framework. Thus
the Nd, Sm, Eu and Gd representatives are multiferroic at room temperature and above.
Moreover the Gadolinium and Praseodymium representative are relaxor multiferroics. The
Ba2LnFeNb4O15 niobates (Ln = Nd, Sm, Eu, Gd) constitute a family of room-temperature
composite multiferroics. Both the composite and single-phase approaches are still under
investigation.
We would like to point out again our initial approach, which was to introduce new
crystalline frameworks in the field of multiferroic materials. We hope that this work will
contribute to the widening of the crystal-chemical aspects in multiferroic materials, for such
an approach may provide a substantial support to the design of single-phased multiferroics.

12
 Acknowledgements

The authors want to thanks L. Raison for the X-Ray microprobe experiments, and E.
Lebraud and S. Pechev for the collection of X-Ray diffraction data.
This work is supported by the European Network of Excellence “FAME”
(www.famenoe.net) and the STREP MaCoMuFi (www.macomufi.eu).

13
 References

[1] Zheng, H., Wang, J., Lofland, S.E., Ma, Z., Mohaddes-Ardabilli, L., Zhao, T., Salamanca-
Riba, L., Shinde, S. R., Ogale, S. B., Bai, F., Vielhand, D., Jia, Y., Schlom, D. G., Wuttig, M.,
Roytburd, A. & Ramesh, R. Science (2004), 303, 661-663.
[2] Kimura, T., Goto, T., Shintani, H., Ishizaka, K., Arima, T. & Tokura, Y. Nature (2003),
426, 55-58.
[3] Spaldin, N. A. & Fiebig, M. Science (2005), 309, 391-392.
[4] N. A. Hill, J. Phys. Chem. B (2000), 104, 6694-6709
[5] Mostovoy, M. Phys. Rev. Lett. (2006), 96, 067601
[6] Van Aken, B. B., Rivera J.-P., Schmid H. & Fiebig M. (2007) Nature 449, 702-705
[7] Fiebig, M., J. Phys. D.: Appl. Phys., (2005), 38, R123-R152.
[8] Schneck, J.; Toledano, J. C.; Joffrin, C.; Aubree, J.; Joukoff, B.; Gabelotaud, A. Phys.
Rev. B, (1982), 25(3), 1766-1785.
[9] C Filipic, Z Kutnjak, R Lortz, A Torres-Pardo, M Dawber and J F Scott J. Phys.: Condens.
Matter (2007), 19, 236206
[10] Magneli, A., Arkiv foer Kemi, (1949), 1, 213-221.
[11] Jamieson, P.B.,Abrahams, S.C.,Bernstein, J.L., J. Chem. Phys., (1969), 50, 4352-4363.
[12] Ravez, J., Simon, A., C. R. Chimie, (2002), 5, 143-148.
[13] Fang, P. H., Roth, R. S., J. Appl. Phys., (1960), 31(5), 278S.
[14] Krainik, N. N., Isupov, V. A., Bryzhina, M. F., Agranovskaya, A. I., Kristallografiya,
(1964), 9(3), 352-357.
[15] Schmid, H., Magnetoelectric Interaction Phenomena In Crystals, ed. A.J. Freeman & H.
Schmidt (Gordon & Breach), (1972), 121-146.
[16] Ishihara, S., Rivera, J.-P., Kita, E., Ye, Z.-G., Kubel, F., Schmid, H., Ferroelectrics
(1994), 162(1-4), 51-61.
[17] Rodriguez-Carvajal, J., Physica B.(1993), 192, 55-69

14
Figures

Figure 1

15
 * *

Figure 2

16
Figure 3a

17
Figure 3b

18
Figure 4

19
Figure 5

20
Figure 6

21
a

Figure 7

22
 Figures captions

Figure 1: Crystal structure of Ba2LnFeNb4O15 viewed along the c axis, showing the distorsion
of the square channel.

Figure 2: Rietveld Plot for Ba2SmFeNb4O15 (dots: experimental, continuous line: calculated),
insert shows SmNbO4 contributions (*). Upper vertical bars indicate TTB Bragg positions
(lower bars: SmNbO4).

Figure 3: a) Dielectric properties of Ba2LnFeNb4O15 TTBs, Ln = Nd, Sm, Eu at 10kHz b) Ln

= Pr, Gd, above vertical scale break: real part of the dielectric constant, below scale break :
imaginary part of the dielectric constant (for clarity only the relevant sections of the
imaginary part are displayed).

Figure 4: Ferroelectric hysteresis loops at room temperature for the Nd, Sm and Eu TTBs

Figure 5: Magnetic hysteresis loops at room temperature for the Ba2LnFeNb4O15 (except La)
TTBs

Figure 6: Evolution of unit cell volume, molar magnetization (RT, H = 20kOe) (divided by
2.5 for scale purpose) and dielectric ordering temperature of Ba2LnFeNb4O15 TTBs versus
ionic radius of the rare earth.

Figure 7: X-Ray microprobe mappings of a Ba2LaFeNb4O15 (a) and Ba2EuFeNb4O15 (b) TTB
samples (BSE : backscattered electron image of the mapped zone). White dots in the Eu and
Fe cartography of Ba2EuFeNb4O15 correspond to EuNbO4 and barium ferrite, respectively.

23
Ln a (Å) c (Å) V (Å3) RB B Rp Rwp Rexp χ²

La 12.514(1) 3.941(1) 617.2 5.74 15.1 13.7 10.35 1.77

Pr 12.484(1) 3.928(1) 612.3 5.87 15.4 16.6 10.07 2.73
Nd 12.478(1) 3.925(1) 611.0 6.45 21.9 19.4 12.65 2.39
Sm 12.458(1) 3.928(1) 609.8 5.83 17.0 13.6 11.76 1.35
Eu 12.460(1) 3.930(1) 610.2 7.06 19.8 17.0 11.55 2.17
Gd 12.485(1) 3.936(1) 613.5 6.24 21.7 17.4 13.27 1.74

Table 1

24
Atom Wickoff occupancy x y z Biso (Ų)
B

Ba 4c 1.000 0.1656(2) 0.6768(2) 0.3510(6) 0.77(3)

Sm 2a 1.000 0 0 0.3480(7) 0.81(3)
Nb1 2b 0.800 0 1/2 0.8713(5) 0.45(5)
Fe1 2b 0.200 0 1/2 0.8713(5) 0.45(5)
Nb2 4c 0.800 0.2847(4) 0.4267(3) 0.8555(5) 0.02(2)
Fe2 4c 0.200 0.2847(4) 0.4267(3) 0.8555(5) 0.02(2)
Nb3 4c 0.800 0.5741(3) 0.2878(4) 0.8510(6) 0.02(2)
Fe3 4c 0.200 0.5741(3) 0.2878(4) 0.8510(6) 0.02(2)
O1 4c 1.000 0.163(2) 0.499(2) 0.753(3) 0.70(13)
O2 4c 1.000 0.490(2) 0.157(2) 0.850(2) 0.70(13)
O3 4c 1.000 0.291(1) 0.775(1) 0.813(2) 0.70(13)
O4 4c 1.000 0.351(1) 0.575(2) 0.733(3) 0.70(13)
O5 4c 1.000 0.436(1) 0.378(1) 0.885(1) 0.70(13)
O6 4c 1.000 0.211(1) 0.950(1) 0.358(2) 0.70(13)
O7 4c 1.000 0.092(1) 0.190(1) 0.403(2) 0.70(13)
O8 2b 1.000 0 1/2 0.346(3) 0.70(13)

Table 2

25
 La Pr Nd Sm Eu Gd
Dielectric prop. - Rel. Ferroel. Ferroel. Ferroel. Rel.
TC or Tm (K) - 170 323 405 440 300
ε’max (1 kHz) 178 186 206 266 731 137
PS (µC/cm2) - - 0.53 1.08 0.59 -
Magnetic prop. (RT) - Hyster. Hyster. Hyster. Hyster. Hyster.
Hc (Oe) - 540 1860 2150 580 700
M(2T) (e.m.u. / mol) 192.8 430.0 741.9 916.2 1004.5 554.3

Table 3

26
 Tables captions

Table 1: Cell parameters (b ≈ a) and volume, agreement factors of the Rietveld

refinement.

Table 2: Atomic coordinates and isotropic atomic displacement parameters for

Ba2SmFeNb4O15

Table 3: Dielectric and magnetic properties of Ba2LnFeNb4O15 samples (ε’max taken at

TC or Tm except for La TTB (80K))

27