Session B30: Ferroelectric Walls, Heterostructures and Superlattices

Two-dimensional electron gases at head-to-head and tail-to-tail domain walls in ferroelectric thin films.

Symmetry breaking at ferroelectric domain walls gives rise to new physical properties, offering the opportunity to use the domain walls themselves as a functional separate object in a device. One example is the appearance of an enhanced conductivity at the boundaries between ferroelectric domains in oxides. A realistic first-principles simulation of the domains walls is limited to highly-symmetric cleanly-cut walls in order to keep the number of atoms in the simulation box small. Here we use a recently developed second-principles method that treats all the lattice degrees of freedom and the relevant electronic ones on the same foot with high accuracy at a modest computational cost. We apply it to the demading physical problem of head-to-head (HH) and tail-to-tail (TT) domain walls in ferroelectric PbTiO$_{3}$ thin films. These interfaces present a large and unfavourable electrostatic energy due to the polarization-induced bound charge at the domain wall. An accurate simulation should capture eventual charge transfers between the walls, and the concomitant electron-lattice coupling. We show how the polarization discontinuity in HH and TT domain walls in PbTiO$_{3}$ thin films can be effectively screened by the formation of two-dimensional electron gases of electrons and holes.   

First-principles prediction of a native ferroelectric metal

The possibility that metals may support ferroelectricity is an intriguing open issue. Over the years, various compounds have been referred to as ferroelectric metals, including non-centrosymmetric metals as well as ferroelectrics whose polar distortion survives moderate metallicity induced by doping or proximity. Yet, we think none of these systems embodies a truly ferroelectric metal with native switchable polarization and native metallicity coexisting in a single phase. Here we report a first-principles prediction of such a material. We show that the layered perovskite Bi5Ti5O17 has a non-zero density of states at the Fermi level and metal-like conductivity, as well as a spontaneous polarization in zero field. Further, we predict that the polarization of Bi5Ti5O17 is switchable both in principle (the material complies with the sufficient symmetry requirements) and in practice (in spite of being a metal, Bi5Ti5O17 can sustain a sizable potential drop along the polar direction, as needed to revert its polarization by application of an electric bias). Our results also reveal striking behaviors -- such as the self screening mechanism at work in thin Bi5Ti5O17 layers -- emerging from the intimate interplay between polar distortions and free carriers in such an exotic material.   

Polarization in asymmetrical intermixed interfaces in SrTiO$_3$/PbTiO$_3$ superlattices

We used first principles density functional theory to study the effects on polarization of asymmetrical intermixing. In our systems, one interface has intermixed A-cations and the other one is pure. We analyze both monodomain and polydomain SrTiO$_3$/PbTiO$_3$ (STO/PTO) superlattices with varying periods. We report how the difference in energy and spontaneous polarization, between the two stable polarization states, scales with period, domain size, thickness of the intermixed layer and oxygen vacancies. Our results are used to explain the origin of the intrinsic polarization asymmetry observed in experimental measurements of ferroelectric hysteresis loops.   

Domain Structure and Properties in Inhomogeneously-Strained Ferroelectric Thin Films

Epitaxial thin-film growth and the ability to deterministically apply lattice mismatch strain has enabled dramatic control over the structure and properties of a range of ferroelectric materials. Modern ferroelectric films, including bilayer and superlattice heterostructures, have also provided access to exotic structures and properties not available in the bulk. In this work, we focus on recent advances in our understanding of how strain can be manipulated and controlled to elicit new types of responses and new understandings about response in ferroelectric materials. In particular, we will explore new modalities of strain control of ferroelectric materials that go beyond traditional lattice mismatch effects and how this can be used to enhance performance, independently tune susceptibilities, and provide new insights into the nature of these complex materials. In particular, we will focus on the deterministic production of large strain gradients (on the order of \textgreater 10$^{\mathrm{-5}}$ m$^{\mathrm{-1}})$ via purposeful compositional gradients. We will highlight work on compositionally-graded versions of PbZr$_{\mathrm{1-x}}$Ti$_{\mathrm{x}}$O$_{\mathrm{3}}$ and Ba$_{\mathrm{1-x}}$Sr$_{\mathrm{x}}$TiO$_{\mathrm{3}}$ where careful control of lattice mismatch and chemistry combine to produce large strain gradients, exotic properties, and new approaches to independently control traditionally coupled properties. As part of this discussion, we will explore the evolution of the crystal and domain structure as a function of the end-members of the compositional gradient, thickness of the film, and substrate. Advanced band-excitation piezoresponse force microscopy, switching spectroscopy, and non-linearity studies have been applied. These studies reveal both unexpected crystal and domain structures can be stabilized in these heterostructures and exotic low- and high-field responses can be obtained. Of particular interest will be the results of temperature-dependent probes of susceptibility which reveal large, nearly temperature-independent properties from 25-500\textdegree C and the observation of highly-mobile ferroelastic domain wall structures which can give rise to local enhancement of susceptibilities. These observations could represent a ground-breaking advance in the performance of these materials.   

Epitaxial strain effects on layered polar oxides from first-principles

Epitaxial strain is a powerful tool to generate ferroelectric phases in thin films owing to polarization-strain coupling. The coupling of the oxygen rotations to strain can also be exploited to realize oxygen rotation-sensitive properties such as metal-insulator transitions and magnetic reconstructions. Here, we use electronic structure calculations to investigate the effects of biaxial strain on (001) thin films of the hybrid-improper ferroelectric Ca$_{\mathrm{3}}$Ti$_{\mathrm{2}}$O$_{\mathrm{7}}$. Besides the bulk \textit{Cmc}2$_{\mathrm{1}}$ phase, we also find a new phase emerges under both experimentally accessible biaxial compressive and tensile strains. Furthermore, a large change in the dielectric anisotropy of the film is found at the tensile phase boundary, which we propose could be electric field tunable. Our results may offer a route to search for new functionalities in layered-perovskite oxides.   

First-principles study of charge-order-driven ferroelectricity in LaVO$_3$/SrVO$_3$ superlattices

We investigate the structure and electronic properties of the 1:1 superlattice composed of LaVO$_3$ and SrVO$_3$ using the density functional theory plus U (DFT+U) method. We find two low-energy antiferromagnetic Mott-insulating phases with distinct charge ordering patterns. In one of these phases, spontaneous polarization normal to the interface is nonzero due to a layered charge-ordering. The polarization calculated by the Berry phase method is 32 $\mu$C/cm$^2$; we have identified a possible switching path on which the system remains insulating. When fully relaxed, the energy per 5-atom-unit-cell of the polar state is only 3 meV higher than the non-polar state and we find that the energy difference can be reduced to zero by tensile strain. This suggests that that the polar state could be induced by applied electric field, and, depending on the switching process, a ferroelectric hysteresis loop could be observed.   

\textbf{Microwave conductance of ferroelectric domain walls in lead titanate}

Numerous theoretical works predicted electronically conducting domain walls in otherwise insulating ferroelectric crystals. A number of recent experiments reported conducting walls, although conductivity itself and a conclusive proof of conductance mechanism remain elusive, largely due to the electrical contact problem. The latter can be overcome using high-frequency AC voltage. Here we will present our successful measurements of microwave conductance at 180$^{\mathrm{o}}$ domain walls in lead titanate using microwave microscopy. AC conducting domain walls can be repeatably reconfigured and have extraordinary stability in time and temperature. AC conductivity is detected even when DC is not. Quantitative modeling reveals that the conductance of domain walls is comparable to doped silicon. We will also present a new and robust mechanism to create charged domain walls in any ferroelectric lattice. Overall, this sets the stage for a new generation of local experiments on conducting domain walls, and furthers the prospects of their application in fast electronic devices. AT, YC, SVK, PM supported by Division of Materials Sciences and Engineering, Office of Science, Basic Energy Sciences, U. S. DOE. PY supported by the National Basic Research Program of China (2015CB921700).   

Thickness dependence of ferroelectric stability in SrRuO$_{\mathrm{3}}$/BaTiO$_{\mathrm{3}}$/La$_{\mathrm{0.7}}$Sr$_{\mathrm{0.3}}$MnO$_{\mathrm{3}}$ ferroelectric tunnel junctions

Ferroelectric tunnel junctions (FTJs) must meet key requirements in order to become viable device structures. One factor which can limit functionality is the thickness of the ferroelectric layer. The ferroelectric must be thin enough that a detectable tunnel current can flow through it, but if it is too thin, screening of the depolarization field by the electrodes will be insufficient for polarization stability. One mechanism to produce a large change in the tunneling resistance, i.e. a large tunneling electro resistance (TER), is to use asymmetric electrodes. This is disadvantageous from the point of view of switchability, however: the electric field due to the mismatch between metal work functions leads to a preference for one polarization state over the other and, in thinner FTJs, may render one of the polarization states unstable. To explore this effect we perform first-principles density functional calculations on SrRuO$_{\mathrm{3}}$/BaTiO$_{\mathrm{3}}$/La$_{\mathrm{0.7}}$Sr$_{\mathrm{0.3}}$MnO$_{\mathrm{3}}$ (SRO/BTO/LSMO) FTJs with varying BTO thicknesses. We find an energetic preference for polarization to point away from the LSMO electrode. FTJs with BTO thicknesses at or below 4 unit-cells polarization pointing toward LSMO is unstable, and therefore are unswitchable. Analysis reveals that, in addition to the work function mismatch, the difference in screening lengths and the intrinsic layer-by-layer polar nature of LSMO play a significant role in this instability. We will also present an analysis of these effects on the tunneling barrier profile as well as on the TER effect.   

Ferroelectric switching in epitaxial PbZr$_{\mathrm{0.2}}$Ti$_{\mathrm{0.8}}$O$_{\mathrm{3}}$/ZnO/GaN heterostructures

As a wide-bandgap semiconductor, ZnO has gained substantial interest due to its favorable properties including high electron mobility, strong room-temperature luminescence, etc. The main obstacle of its application is the lack of reproducible and low-resistivity p-type ZnO. P-type doping of ZnO through the interface charge injection, which can be achieved by the polarization switching of ferroelectric films, is a tempting solution. We explored ferroelectric switching behavior of PbZr$_{\mathrm{0.2}}$Ti$_{\mathrm{0.8}}$O$_{\mathrm{3}}$/ZnO/GaN heterostructures epitaxially grown on Sapphire substrates by RF sputtering. The electrical measurements of Pt/PbZr$_{\mathrm{0.2}}$Ti$_{\mathrm{0.8}}$O$_{\mathrm{3}}$/ZnO/GaN ferroelectric-semiconductor capacitors revealed unusual behavior that is a combination of polarization switching and a diode I-V characteristics.   

Room-temperature Ferroelectricity in Uniaxially Strained Single-crystalline SrTiO$_{\mathrm{3}}$ Freestanding Films

Single crystal pure bulk SrTiO$_{\mathrm{3}}$ (STO) is an incipient ferroelectric whose dielectric permittivity rises to high values as temperature is reduced, but remains paraelectric to the lowest observable temperatures. Ferroelectric phases of STO may be stabilized via doping and strain, whose common effect is to split the spatial free energy well of ionic displacements. With epitaxial strain of the order of a few percent, Curie temperatures $T_{\mathrm{C}}$ \textasciitilde 293 K have been observed. By exploiting a highly novel process to exfoliate epitaxial oxide films deposited by pulsed laser deposition, we have isolated sub-100 nm-thick freestanding films of STO which are readily manipulated and mechanically strained to high levels. Measurements of the in-plane dielectric properties for various applied strains reveal a continuously tunable ferroelectric $T_{\mathrm{C}}$. A two-order-of-magnitude enhanced dielectric response is displayed by a 1.2{\%}-strained sample at $T_{\mathrm{C}}$ \textasciitilde 290 K, as compared with the same sample unstrained at the same temperature. This is consistent with a phenomenological Ginzburg-Landau model, and previous studies on anchored films. The functional properties of strained STO have generated intense interest and debate, and have been suggested for device applications due to e.g. high voltage-tunable dielectric properties. Our work exhibits strain as a continuously variable experimental degree of freedom, which can induce numerous functional effects.   

Manipulation of Carrier Density near Ferroelectric/Semiconductor Interfaces

Switchable polarization of a ferroelectric (FE) opens up the opportunity to control the charge density and transport characteristics at the FE/metal and FE/semiconductor (SC) heterointerfaces. Carrier manipulation near such regions can be used in high density non-volatile memories, switchable diodes, and photovoltaic devices. FEs can be utilized as gate oxides in a metal oxide field-effect transistor configuration for non-volatile memory applications with lower gate voltages compared to that of transistors with linear dielectrics. The channel conductance can be modulated reversibly, for instance, by tuning the magnitude and spatial distribution of polarization in the FE. In this study, we show that FE heterostructures can be used to manipulate the conductivity of a FE/SC interface. We employ a non-linear thermodynamic model based on Landau-Ginzburg-Devonshire (LGD) theory to obtain the equilibrium polarization of heterostructures. The carriers along the heterostructures are mapped through coupling the LGD equation with the Maxwell equations and Fermi -- Dirac distribution of charged carriers/ionized dopants in the FE and SC. We consider various configurations including FE/SC/paraelectric and FE/SC/FE stacks to investigate the carrier distribution and band bending near such interfaces. The resulting properties are explained through the phase transition characteristics and domain structure of the stacks.