Microscopic flexoelectricity in the canonical PMN relaxor Preprint: AIP/123-QED

J. Hlinka · 2026

Read Microscopic flexoelectricity in the canonical PMN relaxor Preprint: AIP/123-QED by J. Hlinka (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-13202.txt.

Introduction

J. Hlinka revisits neutron scattering data on PMN, the canonical relaxor ferroelectric perovskite oxide. The aim is to gauge the role of intrinsic bulk flexoelectricity in this material. The flexoelectric coupling value drawn from the diffuse scattering data stays in a familiar range. It does not exceed values seen in conventional perovskite ferroelectrics. Hlinka links this result to Ginzburg-Landau-Devonshire theory and soft mode theory. He suggests relaxor traits may stem from a shortened transverse correlation length near the Lifshitz-point regime.

Catalog record

CREATORJ. Hlinka
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE29162
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.13202

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Microscopic flexoelectricity in the canonical PMN relaxor Preprint: AIP/123-QED J. Hlinka Abstract Previously reported neutron scattering investigations of the canonical relaxor ferroelectric perovskite oxide with a chemical formula Pb(Mg 1/3 Nb 2/3 )O 3 (PMN) are revisited in order to appreciate the role of the intrinsic bulk flexoelectricity. Despite the outstanding electromechanical properties of lead-based relaxors, the magnitude of the flexoelectric coupling coefficient f 11 − f 12 ≈ − 2 ​ V f_{11}-f_{12}\approx-2V , derived here directly from the PMN neutron diffuse scattering data, does not exceed the range of values typical for conventional perovskite ferroelectrics. We explain how these findings are related in the framework of the Ginzburg-Landau-Devonshire and the ferroelectric soft mode theory. We propose that the relaxor properties of PMN might be related to the suppression of the transverse correlation length of the flexoelectrically hybridized translational-polarization fluctuations due to its closeness to the Lifshitz-point regime. I Introduction Generation of electric polarization in an insulating solid by elastic strain gradients is a challenging engineering task. The linear part of the underlying universal mechanism, known as flexoelectricity, as well as the associated converse effect, has already been the subject of numerous experimental and theoretical investigations, and it promises remarkable possibilities for the design of unique electromechanical materials properties at the nanoscale. At the microscopical level, it often plays the key role in theories of modulated ferroelectrics and antiferroelectrics Axe70 ; Etxebarria92 ; Pottker ; Liubachko ; Tagantsev13 . Recent comprehensive overviews of the flexoelectricity with a focus on perovskite oxides can be found, for example, in Refs. Yudin13 ; Wang19 . To support the current flexoelectric engineering efforts, it would be convenient to employ the most advanced computer-aided approaches, allowing to explore various combinations of nanoscale geometry and distinct materials. Obviously, for that, an extensive registry of materials properties is needed. However, although the flexoelectric effect appears to be easily accessible for direct measurements, in practice, these measurements are often difficult because of the sensitivity, nonlinearity, surface phenomena or other experimental issues. Potentially, the flexoelectric tensor describing the intrinsic bulk properties of crystals can be calculated ab-initio following the theory exposed e.g. in Refs. Hong13 ; Stengel , but these techniques are still facing the computational limits. Consequently, flexoelectric tensor elements are so far established only for a rather limited number of materials Yudin13 ; Wang19 . In this work, we address the intrinsic bulk flexoelectricity in one of the most remarkable perovskite oxide relaxor ferroelectric crystal, the cubic lead magnesium niobate, PMN. The magnitude of its flexoelectric property will be derived here from parameters of its characteristic atomistic structural inhomogeneity, that has been originally extracted from the extensive experimental neutron scattering study of PMN diffuse scattering in Ref. Vakhrushev . The essential ingredient of our analysis relies on the observation, highlighted and emphasized in Hirota02 , that the typical single unit cell pattern (eigenvector) of the frozen structural modulations differs from the eigenvector of the dynamic soft optic phonon by a systematic uniform shift, oriented in the direction and in the positive sense of the local frozen polarization. We explain that this shift, called a "phase shift" there, can be understood as a natural consequence of the converse flexoelectricity, and we use it to estimate the magnitude of the flexoelectric coefficient f_{11}-f_{12} , as well as the lower bound for the corresponding gradient coefficient G_{11}-G_{12} . Moreover, we argue that the system is close to the Lifshits point and that this proximity is one of the critical ingredients responsible for the polar nanodomain formation in perovskite relaxors. II "Phase shift" model of Hirota et al. By inspecting neutron diffuse scattering intensities of the PMN relaxor single crystal in multiple inequivalent Brillouin zones, Vakhrushev et al. in Ref. Vakhrushev determined that there must be inhomogeneous distribtuion of atomic displacements superposed over its cubic average structure, which locally, i.e. at the unit cell level, consists in the set of parallel atomic displacements \{\delta_{i}\} , i=1-5 with the following relative amplitudes: \displaystyle\delta_{1}=\delta_{\rm Pb} \displaystyle= \displaystyle 1.00, (1) \displaystyle\delta_{2}=\delta_{\rm MN} \displaystyle= \displaystyle 0.18, \displaystyle\delta_{3}=\delta_{4}=\delta_{5}=\delta_{\rm O} \displaystyle= \displaystyle-0.64. These quantities are normalized by the lead ion displacement d_{\rm abs} . The lead displacements are disordered across the lattice, on average reaching d_{\rm abs}\approx 0.25-0.3 Å at the lowest temperatures. What matters in the following is that these five values \delta_{i} can be considered as a sort of universal "frozen phonon eigenvector" of PMN. Hirota et al. in Ref. Hirota02 pointed out that (i) this pattern does not satisfy the center-of-mass condition for optic phonon modes, or in other words, \sum_{\kappa}\delta_{\kappa}m_{\kappa}\neq 0, (2) where m_{\kappa} is the mass of the atom \kappa and the summation goes over the 5-atom perovskite unit cell; that (ii) there is a shift \tau\approx 0.58 such that the "shifted eigenvector" h_{\kappa}=\delta_{\kappa}-\tau (3) would satisfy the center-of-mass condition, and that (iii) this "shifted eigenvector" \displaystyle h_{1}=h_{\rm Pb} \displaystyle= \displaystyle 0.42, (4) \displaystyle h_{2}=h_{\rm MN} \displaystyle= \displaystyle-0.40, \displaystyle h_{3}-h_{4}=h_{5}=h_{\rm O} \displaystyle= \displaystyle-1.22 is consistent with the dynamical structure factor of the soft phonon mode observed in their inelastic neutron scattering experiments. In agreement with the intuitive picture of the frozen ferroelectric nanodomains or frozen polar nanoregions, the ultimately short range correlations should be of a ferroelectric type. The diffuse scattering is sensitive to the correlations in the distribution of these frozen entities. In particular, the characteristic anisotropy of diffuse scattering in various Brillouin zones suggests that the dominant contribution can be described as a superposition of harmonic polarization waves in the form {\bf P}({\bf r})={\bf P}({\bf q})\sin({\bf q}\cdot{\bf r}+\varphi), (5) where vectors {\bf q} and {\bf P}({\bf q}) fall along one of the twelve \langle 110\rangle -type crystal directions, and at the same time, {\bf q} and {\bf P}({\bf q}) satisfy the transversality condition, that is, {\bf P}({\bf q})\in\langle 110\rangle~\wedge~{\bf q}\in\langle 110\rangle~\wedge~{\bf P}({\bf q})\cdot{\bf q}=0, (6) what is fulfilled for example with {\bf P}({\bf q})\parallel [1-10] and {\bf q}\parallel [110] and obviously all m\bar{3}m -symmetry equivalent pairs. Since the range of |{\bf q}| is broad, a plausible real-space pattern might contain bundles of differently spaced lamellar nanodomains with (110) -type boundaries, at which the [1-10]-component of the frozen polarization fluctuations alternates its sign – similarly as, for example, in the case of electromechanically compatible O180(110) or T90(110) walls of barium titanate Marton . The condensation of {\bf P}({\bf q})\parallel [1-10] and {\bf q}\parallel [110] waves is obviously in the heart of the so-called pancake model Xu2004 , but let us stress that the assumptions of eqs. ( 5 , 6 ) about the frozen wave basis do not necessarily imply that the resulting local polarization vector has to preferentially point along one of the twelve \langle 110\rangle directions. For example, the nominally uncharged R109(110) domain wall of [1\bar{1}1|110\rangle\bar{1}11] type can be obtained as a superposition of nominally uncharged O90(110) [1\bar{1}0|110\rangle\bar{1}10] wall (which satisfies the above geometry requirements) with a homogeneous [001]-polarized background (which does not contribute to the diffuse scattering away from the Bragg reflection position at all). Thus, diffuse scattering by waves described by eqs. ( 5 , 6 ) is well compatible with the usual assumption that the resulting preferential local polarization favors \langle 111\rangle type crystallographic directions. In fact, strictly speaking, the atomistic pattern of eq. (

Read the complete text