When electrons meet ferroelastic domain walls in Strontium Titanate

TL;DR

SrTiO3 sits at the crossroads of multiple coupled orders: a cubic-to-tetragonal antiferrodistortive (AFD) ferroelastic transition near 105 K and quantum paraelectricity below 35 K set a background where ferroelastic twin walls host emergent nanoscale polarity. The review synthesizes evidence that wall polarity arises from multiple couplings (flexoelectric rotoflexo, rotopolar, and trilinear interactions) and interacts with doped electrons to create static conduction channels, anisotropic transport, and non-universal metal–insulator transition scaling, while wall dynamics induce glassy relaxation and memory effects. It also highlights emergent orders such as magnetism and quasi-one-dimensional superconductivity localized along walls, pointing to domain-wall landscapes as a versatile platform for polar metals, oxide spintronics, and reconfigurable nanoelectronics. Overall, the work argues for a paradigm where dynamic, strongly coupled domain-wall physics fundamentally shapes charge transport in quantum paraelectric oxides, with broad implications for correlated oxide physics and device concepts.

Abstract

Strontium titanate (SrTiO$_3$), famously described by Nobel laureate K. A. Müller as the "drosophila of solid-state physics", has been extensively investigated over the last seventy five years for its intricate coupling of structural, electronic, and dielectric properties and continues to serve as a foundational platform for advancing oxide electronics. In its pristine form, SrTiO$_3$ exhibits quantum paraelectric behavior below 35 K and undergoes an antiferrodistortive phase transition near 105 K. This transition generates ferroelastic twin domains separated by a dense network of domain walls, which function as nanoscale structural defects with far-reaching consequences. While the static influence of ferroelastic domain walls on carrier transport in electron-doped SrTiO$_3$ is well established, recent experimental results show that the emergence of polarity at these walls, combined with strain fields and inherent quantum fluctuations, induces correlated dynamical phenomena such as glass-like relaxations of electrons and memory effects. In this review, we highlight these recent advances, focusing on the subtle interplay between the emergence of nanoscale polar order, quantum fluctuations, and long-range strain fields. We propose that understanding charge carrier dynamics in the background of these complex ferroelastic domain wall landscapes offers a new paradigm for exploring electronic transport in the presence of local polar order and quantum fluctuations, with broad implications for correlated oxides.

Figures (17)

• Figure 1: a. Schematic illustrating the temperature-driven AFD structural transition of STO from the high-temperature cubic phase (above 105 K) to the low-temperature tetragonal phase (below 105 K). This transition is driven by the rotation of TiO$_6$ octahedra, resulting in the formation of a complex network of ferroelastic twin domains (discussed later in detail). b. Overview of the major research frontiers in incipient ferroelectric STO. Primarily, research has focused on two major fields: stabilizing ferroelectricity via external perturbation (e.g., chemical doping, isotope substitution, strain engineering, and ultrafast optical excitations, etc.) and exploring emergent electronic transport and magnetism through carrier doping. More recently, the intrinsic physics of ferroelastic domain walls has emerged as a distinct field of interest. The central theme of this review (bottom curved arrow) addresses the critical intersection of these areas: understanding how the ferroelastic domain wall actively affects and controls the electronic transport properties in doped STO.
• Figure 2: a. Unit cell schematic of STO illustrating transverse optical (TO) soft-mode. b. Temperature dependence of the soft-mode frequencies at the R point, $\Gamma$ point, and antiphase rotation angle of TiO$_6$ octahedra ($\phi$). The R-point mode (TO-R) softens at 105 K, marking the cubic to tetragonal structural transition. Below this transition, the lowering of crystal symmetry lifts the triple degeneracy of the R-point soft mode, causing it to split into two distinct branches: a singlet corresponding to rotation about the unique $c$-axis and a doublet corresponding to rotations about the equivalent $a$- and $b$-axes Cowley:1969p181. Both branches harden with decreasing temperature. In comparison, the $\Gamma$-point TO mode (TO-$\Gamma$) continues to soften upon cooling but tends to saturate below $\sim$ 35 K due to quantum fluctuations, signifying the emergence of the quantum paraelectric state in STO. The bottom panel shows the temperature dependence of the antiphase rotation angle of TiO$_6$ octahedra ($\phi$) in the tetragonal phase. This panel has been reproduced from references. Cowley:1969p181Sirenko:2000p373hayward:1999p501. c. The schematic phase diagram summarizes the nearby phases around the quantum paraelectric regime. The blue and orange regions denote the ferroelectric phase and a polar-acoustic regime, respectively, which can be accessed via tuning parameters such as flexoelectric coupling. The insets on the left and right depict representative phonon dispersions near the transitions to the FE and polar-acoustic phases, respectively. Panel c is taken from ref. Orenstein:2025p961.
• Figure 3: a. A schematic showing structural twin domains of STO below 105 K. Three types of domains with $c$ axis along [100], [010], and [001] have been marked with $X$ (green), $Y$ (blue), and $Z$ (yellow), respectively. The right panel shows the domain geometry in the $(110)$ cross-sectional plane. The twin boundaries separating $Z$ domains from the $X$ (or $Y$) domains form the characteristic angles of either $55^{\circ}$ ($125^{\circ}$) or $145^{\circ}$ ($35^{\circ}$) in the $(110)$ plane. The wall separating the $X$ and $Y$ domains is aligned along the $[001]$ axis in the $(110)$ plane. This panel is taken from the ref. Harsan:2016p257601b. Atomic representation of two types of domain walls distinguished by the sign reversal of the antiferrodistortive (AFD) pseudovector components $\phi$$_r$ or $\phi$$_s$ corresponding to “head-to-head” (HH, left) and “head-to-tail” (HT, right) configurations. The black arrows indicate the local tilt vector. Panels c - e. show how different order parameters change spatially across the domain wall. c. Spatial variation of the pseudovector components $\phi$$_r$ and $\phi$$_s$ across two domain walls, with the shaded region denoting the nominal wall width (2$\xi$). d. Amplitude of the antiferroelectric Ti displacement mode, $u_{Ti}$ across the domain walls. The inset illustrates the antiferroelectric character of the Ti displacements. e. Spatial variation of the spontaneous strain ($\epsilon$) and the in-plane polarization ($P$) across the domain walls. Panels b - d are taken from ref. Schiaffino:2017p137601
• Figure 4: a. Schematic of the experimental configuration for resonant piezoelectric spectroscopy (RPS). The same setup can be adapted for resonant ultrasound spectroscopy (RUS) by applying the ac voltage to the upper piezoelectric transducer instead of across the sample electrodes. The applied ac voltage in both cases was 25 V. b. Low-temperature RPS spectra of STO recorded between 25 kHz and 100 kHz. The spectrum shown in blue was collected at $\sim$ 80 K. Disappearance of two mechanical resonances in RPS spectra near 40 kHz ($\nu$$_1$) and 85 kHz ($\nu$$_2$) with increasing temperature indicates softening of elastic modes. c. Temperature evolution of the resonance frequencies $\nu$$_1$ and $\nu$$_2$ obtained from RPS and RUS. The close agreement between the two methods confirms that the Fano-like features observed in RPS originate from intrinsic mechanical resonances of STO. Crucially, since bulk STO is centrosymmetric and non-piezoelectric, it cannot normally be driven into mechanical resonance by an electric field. Therefore, the appearance of resonances implies that the domain walls themselves are polar; the applied voltage causes the walls to oscillate, generating the strain fields that drive the sample’s mechanical modes. The polarity at the domain wall develops at 80 K and further enhances below 40 K Scott:2012p187601Salje:2013p247603. This figure is taken from ref. Salje:2013p247603
• Figure 5: a. Electric-field-induced topographical changes obtained by subtracting reflection images taken at 400 V/mm from 0 V/mm. Pronounced contrast at 6 K and 20 K indicates active twin motion, whereas changes diminish at 40 K and vanish by 60 K. b. Temperature dependence of the topography changes, quantified as the normalized mean intensity of the subtracted images shown in panel a. The magnitude of topography changes decreases sharply above 40 K, indicating a rapid reduction in twin-wall mobility with increasing temperature. The inset shows the schematic of the experimental geometry used for electric field-dependent optical imaging. This figure is taken from ref. Casals:2019p032025