Multiferroic Transitions and Misfit Phase Diagram in Oxygen‐Deficient Epitaxial (111) PbTiO3

Takahiro Shimada,Yasumitsu Araki,Tao Xu,Jie Wang,Takayuki Kitamura

DOI: https://doi.org/10.1002/aelm.201600113

IF: 6.2

2016-01-01

Advanced Electronic Materials

Abstract:DOI: 10.1002/aelm.201600113 vacancies.[20–26] If an oxygen vacancy could act as a source of magnetism in ferroelectric oxides, one could achieve multiferroic and multifunctional oxide materials through engineering defects. Unfortunately, oxygen vacancies remain nonmagnetic in ferroelectrics owing to the unpolarized nature of the defect spin states.[25,26] Thus, instead of the simple concept of defect engineering, an alternative approach is required for the design of multiferroics with conventional ferroelectrics. Here, we demonstrate using the state-of-the-art hybrid Hartree–Fock density functional theories that misfit (epitaxial) strain can rapidly activate (anti)ferromagnetism for oxygen vacancies in epitaxial (111) PbTiO3 and induce unusual multiferroic phase transitions. Electronic structure analysis reveals that local (anti)ferromagnetism of oxygen vacancies arises from the splitand spin-polarization of the localized defect state owing to the orbital and crystal symmetry breaking via the misfit strain. We also predict nonlinear magnetoelectric coupling due to the magnetic phase transition by polarization switching. Thus, the present results suggest a new design approach for magnetoelectric multiferroics that involves engineering of both defects and mechanical strains in conventional nonmagnetic ferroelectrics. All first-principles calculations are based on the hybrid Hartree–Fock density functional theories parameterized by Heyd, Scuseria, and Ernzerhof (HSE06)[18,19] as implemented in the Vienna ab-initio simulation package (VASP) code.[27,28] The HSE06 hybrid functional well reproduces the structural and electronic properties of wide-bandgap oxides and is essential for the correct description of defect electronic structures in ferroelectric oxides.[20–26] The details are shown in the Supporting Information. For the modeling of defects in epitaxial PbTiO3, we use the common supercell setup containing 120 atoms (Figure S1, Supporting Information), roughly corresponding to × × 6 2 2 2 3 unit cells. The in-plane cell vectors of the supercell are set to ( ) = 6 ,0,0 1 sub a a and ( ) = 0,2 2 ,0 2 sub a a , and kept fixed to simulate the pseudoepitaxy of ferroelectric PbTiO3 on a (111) substrate with a lattice parameter of asub, while the rest of the out-of-plane vector a3 and internal atomic positions are fully relaxed until the Hellmann–Feynman forces are less than 0.01 eV Å–1. According to the previous thermodynamic analysis and first-principles calculations,[29,30] we screened out a few space groups compatible with the symmetry of epitaxial (111) PbTiO3 and performed total energy optimization in these structures. Our calculations revealed that the lowest energy structure under the (111) pseudoepitaxy is in the monoclinic (MA) phase with space group Cm at an optimal lattice parameter of asub = 3.943 Å. This structure was then set as the zero misfit strain state of PbTiO3 under such (111) pseudoepitaxy condition. To include the effect of misfit (epitaxial) strain due to choice of substrate, we considered the lattice parameter asub Since the discovery of multiferroic (MF) materials, which exhibit various coexisting different ferroic orders such as ferroelectricity and (anti)ferromagnetism in one phase,[1–3] scientific and technological interest in multiferroics has been increasing dramatically owing to novel physical phenomena that result from the multiple ferroic degrees of freedom and the cross-coupling between different order parameters, such as the colossal magnetoelectric (ME) effect, and technological innovations that might potentially lead to novel functional device paradigms, such as new information-storage devices made of multi ferroic memory cells for writing electrically and reading magnetically.[4,5] However, there are very few multiferroic materials because of the mutually exclusive mechanisms for ferroelectricity and ferromagnetism. For ferroelectricity in conventional ferroelectrics (FEs) such as PbTiO3 and BaTiO3, a formal d0 empty electron configuration of the transition metal (Ti4+) is essential for cation off-centering for the formation of an electric dipole, while what is necessary for ferromagnetism is a partially occupied d state.[6] Despite numerous experimental and theoretical efforts to seek new multiferroic materials through substitution and doping of elements in perovskite and complex oxides,[7–10] the discovery and understanding of new multiferroics remain limited. Thus, a novel material design concept and strategy based on a unique mechanism is required to develop a new route for multiferroics, which would also provide fundamental insight into diverse functionalities in oxides. Oxygen vacancies, as the most common and abundant point defects in oxide materials, often lead to an intriguingly rich variety of physical phenomena and additional functionalities that the perfect host perovskite materials do not intrinsically possess, such as n-type conductivity,[11,12] characteristic light emission,[13] high 2D carrier density,[14,15] and paraelectric-to-ferroelectric transition.[16,17] Recent advances in firstprinciples theories, such as the hybrid Hartree–Fock density functional,[18,19] have enabled an accurate description of complex defect electronic structures and revealed that the abovementioned variety of additional functionalities arise from the localized d-orbital electronic structure characteristically formed within the band gap by local nonstoichiometry of oxygen www.MaterialsViews.com www.advelectronicmat.de