Ternary perovskite titanates with chemical composition ATiO3, where A is a divalent alkaline earth element Ca, Sr, or Ba exhibit unusual and interesting physical properties. The most prominent example is BaTiO3 which undergoes several structural phase transitions from the cubic paraelectric phase at high temperatures to a series of new lower symmetry phases near or below room temperature. The structural distortions involved with these phase transitions induce a ferroelectric polarization, making BaTiO3 a technologically very important high-k material. The isostructural compound SrTiO3 has a great potential application as an active material in oxygen sensor devices but less as a room temperature dielectric material. In contrast to BaTiO3, SrTiO3 which undergoes a cubic-to-tetragonal structural phase transition at 105 K, remains paralelectric to lowest temperatures, however with a remarkably high dielectric constant.
Müller and Burkard have demonstrated that the intrinsic paralectric behavior observed to lowest temperatures is a consequence of a quantum-mechanical stabilization of large q = 0 ferroelectric fluctuations. In order to modify the paraelectric properties and to drive SrTiO3 into the ferroelectric state, numerous experiments such as application of stress or strain, Ca substitution, oxygen isotope substitution or introduction of oxygen vacancies have been carried out.
A different approach undertaken by Katsufuji and Takagi was to replace the Sr constituents by the considerably heavier Eu atoms. EuTiO3 and SrTiO3 share a number of commonalities: At room temperature they both crystallize with the cubic perovskite structure with almost identical lattice parameters. Sr and Eu are both divalent implying Ti to be in the oxidation state +4 with a 3d0 diamagnetic electronic configuration.
Below room temperature both systems exhibit a strong temperature dependence of the dielectric constant concomitant with a large softening of a TO zone center phonon mode. However, in contrast to Sr2+, Eu2+ carries a rather large spin-only magnetic moment arising from seven unpaired electrons in the 4f electronic shell. These moments order with a G-type antiferromagnetic structure at about 5.5 K. Antiferromagnetic ordering causes a 4% drop of the dielectric constant, the origin of which has not been understood yet. Bussmann-Holder et al. have shown that the TO-mode softening in EuTiO3 can be successfully modeled within the scope of the polarizability model using identical coupling parameters as derived for SrTiO3.
The significantly larger atom mass of Eu (152 a.m.u.) as compared to Sr (45 a.m.u.) induces a marked upshift of the cubic-to-tetragonal phase transition which was theoretically predicted to take place close to room temperature. In a series of heat capacity and x-ray diffraction experiments the phase transition was subsequently located at 282(2) K . Here we report thermal dilatometry experiments carried out in order to investigate in more detail a) the cubic-to-tetragonal phase transition near room temperature and b) to study the magnetoelastic coupling in the antiferromagnetic phase and its implications for the low-temperature dielectric anomaly.
Figure 1 displays the lattice contraction, DL(T), of EuTiO3 and for comparison of SrTiO3. Anomalies at the structural phase transitions are clearly visible for both compounds. For EuTiO3, the linear coefficient of thermal expansion, a(T ) = d/dT DL(T), exhibits a sharp anomaly centered at 281(2) K confirming our earlier heat capacity and the x-ray diffraction experiments . These results settle a controversial dispute in the literature as to the exact critical temperature of the structural phase transition in EuTiO3. A theoretical modeling of the lattice anomalies in terms of the thermal displacement correlation function confirms the abrupt change in DL(T) seen at the structural phase transition . The inset in Fig. 1 reveals a pronounced magnetic field dependence on the magnetoelastic contraction below the Néel temperature. Whereas for vanishing magnetic field a peak in a(T) occurs at 5.5 K, application of a magnetic field readily suppresses the lattice contraction.
Figure 2 shows in detail the magnetization and the linear coefficient of thermal expansion in the regime of antiferromagnetic ordering. Below the Néel temperature, saturation of the magnetic moment of 7 mBohr per Eu atom is achieved already for magnetic fields above ≈1.3 T, apparently of sufficient strength to destroy antiferromagnetic order. As seen from Fig. 2(b), a(T ) exhibits a l-shaped peak at the Néel temperature. With increasing field the magnitude is decreased and the peak moves to lower temperatures until it is completely suppressed above the saturation field of ≈1.3 T indicating a lattice expansion induced by a magnetic field application.
Figure 3 outlines the field dependent lattice parameter at 2 K and compares it with the measured and calculated field dependence of the dielectric constant . These findings allow for an understanding of the surprising drop of the dielectric constants, a consequence of magnetoelastic lattice stiffening in the antiferromagnetic phase. Breaking antiferromagnetic order by an external field, in contrast, expands the lattice. Modeling of the phonon properties within the framework of the polarizability model extended with a magnetoelastic coupling term demonstrates that the lattice stiffening stabilizes the relevant TO phonon mode. The concomitant decrease of the dielectric constant below the Néel temperature is a consequence of a 100 ppm decrease of the lattice parameter induced by magnetic ordering.