Morphology and conducting properties of domain walls in uniaxial ferroelectrics

Student thesis: Doctoral ThesisDoctor of Philosophy

Abstract

Ferroelectric materials possess a spontaneous electrical polarisation that can be reoriented between various stable states by application of an electric field. This property, and the tendency of a ferroelectric to split into domains, have been the at the core of ferroelectrics research for the better part of a century. Recently, focus has shifted away from polarisation reversal and domains, towards the behaviour of domain walls. These are the nanoscale, interfacial regions that separate domains. In the last decade, domain walls have been shown to behave as functional entities in their own right: enhanced electrical conductivity at ferroelectric domain walls has been demonstrated in many systems. Domain walls can be created, erased and moved by application of external fields, which has led researchers to dream of a novel kind of reconfigurable nanocircuitry, based on the dynamic deployment of these ephemeral, conducting interfaces. While much progress has been made in prototype device demonstration, knowledge of the fundamental characteristics associated with transport at domain walls remains lacking, mostly due to the difficulty of adapting standard electronic transport measurements to fit complex domain wall geometries.

The first part of this thesis attempts to address this issue. We focus on the most exciting conducting domain wall system discovered to date, lithium niobate, where an estimated 13 order of magnitude increase in the conductivity at domain walls (compared to the bulk domains) has been established. Single crystal 500nm thick z-cut lithium niobate is partially poled, resulting in the injection of inclined, conducting domain walls. The domain microstructures are characterised by piezoresponse force microscopy and cross sectional transmission electron microscopy, which reveal that the wall morphology is that of a truncated cone (otherwise known as a conical frustum). Using an iterative current density procedure, we develop a framework for analysing current flow in non trivial geometries. We show that this conical domain wall geometry can be likened to a "Corbino disc": the prototypical sample geometry used for measurements of geometric magnetoresistance. This effect has the same physical origin as the Hall effect, with the only difference being the sample geometry that is utilised during the measurement. Geometric magnetoresistance measurements can therefore provide experimental access to the carrier mobility associated with conduction. We perform magnetotransport measurements on the domain wall current and measure an upper value of carrier mobility at room temperature of approximately 3700 cm^2V^-1s^-1. Such a value at room temperature is unparalleled, in oxide semiconductors or interfaces, and provides the first insight into the transport physics in this exceptional system. Extensions to the magnetoresistance measurement and possible interpretations for transport are discussed.


Broadly speaking, charged domain walls in uniaxial ferroelectrics exhibit enhanced conduction due to the non-zero polar divergence that exists at the wall. In this respect, the uniaxial ferroelectric lead germanate stands out as an exception. It exhibits fully head-to-head and tail-to-tail domain walls, but shows no sign of enhanced conduction, nor any trace of bound charge at the domain walls. Phase field models, performed by collaborators, suggest that this is because the polarisation rotates away from the polar axis close to these sections of domain wall, maintaining a net zero divergence of polarisation intrinsically. The resulting domain network, which consists of a series of mutually bifurcating domains, hosts a local polarisation field with genuine topological character. To date, the observation of topological textures based on polar rotation (such as polar skyrmions and vortices) have been limited to multiaxial ferroelectrics (e.g. lead titanate). The models suggest lead germanate is the first example of polar topology in a uniaxial ferroelectric.

In the second part of this thesis, we set out to verify the above notions. Again, the domain wall microstructure is crucial. Due to the bifurcation of abutting domains, the domain wall is shaped like a saddle point, around which dipolar rotation results in a polar texture with topological character. Using tomographic piezoresponse force microscopy, we examine the three dimensional domain microstructure, up to depths of several microns. We extract the coordinates of the domain wall surface, and produce surface renders which demonstrate an abundance of domain bifurcation and saddle point structures. The measured domain wall morphology agrees qualitatively with the models and verifies, to some degree, the topological nature of the polarisation field. We then set out to examine another uniaxial ferroelectric, triglycine sulphate, to test whether the saddle point morphology is general, or a peculiar quirk of lead germanate. Our tomographic mapping confirms there are regions where the domain wall crosses the polar axis (and are hence notionally charged), and moreover confirms that, at these sections, the same saddle point domain wall morphology is present. We also perform conductive AFM mapping combined with tomography and find the same distinct lack of enhanced conduction at these domain wall sections. This suggests the same dipolar rotation mechanism as in lead germanate is at play in triglycine sulphate, and that the saddle point topology is a general mechanism by which uniaxial ferroelectrics can avoid domain wall bound charge.

In chapter 6, the main conclusions of this thesis are drawn, and avenues for future work are discussed. With regards to conducting domain walls, it is crucial to develop four-probe capability for domain wall transport measurements. This will allow low temperature transport measurements, perhaps allowing access to the quantum transport regime. Work on 2D heterointerfaces provides direction on this front, if 4-probe geometries can be realised. With regards to polar topologies, nanoscale Kelvin probe force microscopy, or Oersted field mapping via an NV-center microscope, are discussed as potential methods to provide novel ways to detect topological polar patterns.
Date of AwardDec 2023
Original languageEnglish
Awarding Institution
  • Queen's University Belfast
SponsorsEngineering and Physical Sciences Research Council
SupervisorRaymond McQuaid (Supervisor) & Marty Gregg (Supervisor)

Keywords

  • Ferroelectrics
  • domain walls
  • magnetotransport
  • nanotechnology
  • interfaces
  • lithium niobate
  • condensed matter physics

Cite this

'