Implications of phase coexistence in VO(2) thin-films across the metal-insulator transition

Abstract

A Metal-insulator transition (MIT) is the ability of some materials to change between metal and insulator electric behaviours as a function of some external stimuli such as temperature, stress, voltage, magnetic field or light. The Fermi level position with respect to the band structure determines one character or the other, and in some materials this band structure is very sensitive to electron-electron correlations. This is the case of some transition metal oxides, which despite having partially filled bands allowing, in theory, metallic conduction, electron- electron interactions split the half-filled band in a lower energy band that is full first and a higher energy sub-band still empty, resembling an insulator.

An archetypical example is vanadium dioxide (VO2), a system with a 3d electronic configuration which leads to a first-order MIT happening near room temperature (~68 ºC) with a change in conductivity of several orders of magnitude, accompanied by a structural phase transition (SPT) that occurs simultaneously. This results in a high-temperature state given by a metallic rutile (tetragonal) phase that turns into a semiconductor with monoclinic M1 structure at the low-temperature state. Thus, the electronic and structural elements of the transition in VO2 are tightly entangled, whereby mechanical stress induces lattice deformations in the crystal that distort the surroundings of V atoms affecting the orbital properties within the unit cell, the lattice’s electrostatic potential field and vibrational modes. This thesis explores in detail the consequences of in-plane tensile strain in epitaxial VO2 thin films grown on rutile (001)-oriented TiO2 substrates by the pulsed laser deposition (PLD) technique, how the strain relaxes as film thickness increases, and how the strain relaxation itself gives raise to new phases and phase coexistence phenomena of interest. What makes this work novel with respect to other strain studies is the focus on the relation between local properties and global behaviour. As the thesis will show, space-averaged measurements of transport properties or X-ray diffraction (XRD) can miss out on a lot of interesting and important physics that happens on a nanoscopic level, including nano-tweeds or phase coexistence in the form of metal-insulator (M-I) phase boundaries. This work has used various microscopy techniques effective at different scales, from optical microscopy (micron scale) to scanning-probe microscopy (nanoscale) to transmission electron microscopy (atomic scale), yielding a very complete picture of the MIT in strained VO2 films across the different scales. After a brief introductory chapter contextualizing the thesis, chapter one describes the

growth and average structural characteristics of the films, including their evolution due to aging. It was grown a thickness set from ~2 nm up to ~150 nm and surprisingly, the films do not relax with thickness neither conventionally nor progressively, resulting in crack formation, M-I phase separation toward the cracks and strong-strain gradients within the film. The microstructure far from crack-edges is presented in chapter two. It shows that the

strained VO2 regions are not stabilized in the standard high-temperature rutile phase but in a metrically tetragonal structure that triples the rutile periodicity with monoclinic (2/m) symmetry. Moreover, the orientational variants of this new phase (x3M) coexist forming tweed patterns of few nanometers in size. Chapter three explores the functional consequences of the coexistence between strain- relaxed insulating phase near the films’ cracks and metallic phase away from them. The peculiar pattern induced by crack formation leads to the concept of “self-pixelation”, whereby each island of VO2 bounded by insulating cracks behaves de facto as a “pixel” whose MIT can be individually triggered independently of the rest of the film. Finally, the last chapter contains the general conclusions and outlook for future work.

Una transició metall-aïllant (MIT) es la habilitat d’alguns materials per canviar el seu comportament elèctric de metall a aïllant en funció d’estímuls externs com la temperatura, estrès, voltatge, camp magnètic o la llum. Aquest és el cas d’alguns òxids de metalls de transició, els quals malgrat tenir bandes d’energia parcialment ocupades permetent, en teoria, conducció metàl·lica, les interaccions electró-electró divideixen aquesta banda parcialment ocupada en una d’energia més baixa que s’omple primer i una altra d’energia major que queda buida, semblant a un aïllant. Un arquetip d’això és el diòxid de vanadi (VO2), un sistema amb configuració electrònica 3d1 i una MIT de primer ordre, que succeeix ~68 ºC amb un canvi de conductivitat de varis ordres de magnitud, produint-se també una transició de fase estructural (SPT) simultàniament. Aquesta dualitat resulta en un estat d’alta temperatura metàl·lic amb estructura rutil (tetragonal) que es transforma en un semiconductor amb estructura monoclínica M1 en el seu estat de baixa temperatura. Per tant, els elements electrònics i estructurals de la transició al VO2 estan estretament entrelligats. L’estrès mecànic indueix deformacions a la xarxa del cristall que distorsiona els voltants dels àtoms de V afectant les propietats orbitals dins de la cella unitat, camp de potencial electrostàtic i modes vibracionals. Aquesta tesi explora en detall les conseqüències de la tensió en el pla, en capes primes epitaxials de VO2 crescudes sobre substrats rutil de TiO2 orientats en (001), per la tècnica de dipòsit de làser polsat (PLD). El que fa que sigui diferent respecte la resta d’estudis amb estrès mecànic és la focalització en la relació que hi ha entre propietats locals i comportament global del material. Com es mostrarà a la tesi, les mesures espacials mitjana poden perdre’s processos físics importants a nivell nanoscòpic, com nano-tweeds o coexistència de fases amb fronteres metall-aïllant (M-I), i per això les mostres s’han caracteritzat amb varies microscòpies a diferents escales (micro, nano i escala atòmica) aconseguint una visió completa de la MIT en capes tensionades de VO2.