Metal–insulator transition tuned by oxygen vacancy migration across TiO2/VO2 interface

Oxygen defects are essential building blocks for designing functional oxides with remarkable properties, ranging from electrical and ionic conductivity to magnetism and ferroelectricity. Oxygen defects, despite being spatially localized, can profoundly alter global properties such as the crystal symmetry and electronic structure, thereby enabling emergent phenomena. In this work, we achieved tunable metal–insulator transitions (MIT) in oxide heterostructures by inducing interfacial oxygen vacancy migration. We chose the non-stoichiometric VO2-δ as a model system due to its near room temperature MIT temperature. We found that depositing a TiO2 capping layer on an epitaxial VO2 thin film can effectively reduce the resistance of the insulating phase in VO2, yielding a significantly reduced ROFF/RON ratio. We systematically studied the TiO2/VO2 heterostructures by structural and transport measurements, X-ray photoelectron spectroscopy, and ab initio calculations and found that oxygen vacancy migration from TiO2 to VO2 is responsible for the suppression of the MIT. Our findings underscore the importance of the interfacial oxygen vacancy migration and redistribution in controlling the electronic structure and emergent functionality of the heterostructure, thereby providing a new approach to designing oxide heterostructures for novel ionotronics and neuromorphic-computing devices.

The importance of oxygen point defects in dictating physical properties has been more and more widely recognized by the functional oxide community [1][2][3][4] . Oxygen defects, disguised by the name, can be actually used to advantageously enhance functionalities and device performances, ranging from electronic [5][6][7] , magnetic [8][9][10] and multiferroic properties 11 to energy storage and conversion applications 2,12,13 . Especially, there has been an increasing interest in utilizing oxygen point defects for tailoring electronic structures of oxides, due to the designability and reversibility of this approach 14 . Electrical switching enabled by oxygen defects has been proven promising for applications in neuromorphic computing due to the fact that oxygen defects allow actively tuning electrical conductivity on demand 14,15 . In the pursuit of a defect-tuning functionality in oxides, it was found that the effects of oxygen defects can go beyond what can be explained within the simple rigid band model and band-filling picture. In correlated oxide systems, changes in oxygen content can profoundly alter electronic structure and even trigger a metal-insulator transition (MIT). VO 2-δ (δ denotes oxygen non-stoichiometry) is such a material system, wherein a small change in composition can lead to a large modulation in correlation effects [16][17][18][19][20] . Nearstoichiometric VO 2 has 3d 1 electron configuration and shows a MIT with a transition temperature T c ≈ 340 K, accompanied by a structural transition from low-temperature M1 (monoclinic) phase to high-temperature R (rutile) phase. The MIT transition temperature can be lowered by 60 K using epitaxial strain imposed by substrates 21,22 . Interestingly, the MIT has been shown to be completely suppressed by inducing oxygen vacancy formation, either chemically 17,23 or electrochemically 16,24 . Oxygen-deficient VO 2-δ was shown to remain in the metallic R phase when cooled down to below the MIT transition temperature. This makes oxygen stoichiometry a knob for reversibly controlling the MIT in VO 2 .
Oxygen non-stoichiometries in functional oxide thin films are most commonly manipulated by changing the oxygen electrochemical potential via annealing in a certain oxygen partial pressure (pO 2 ) or by applying electrical bias 25 . In contrast, a novel strategy based on interfacial oxygen defect migration induced by the mismatch of defect formation energy of two dissimilar oxides, is much less studied. This "oxygen diode" effect essentially utilizes the oxygen chemical potential gradient from the difference in defect formation energy (E f ) across oxide interfaces. In order to balance the oxygen chemical potential, a unidirectional flow of oxygen defects can be induced during the fabrication process of oxide heterostructures. Recently, this approach has been successfully applied to tune the charge carrier density of LaNiO 3-δ (LNO) by using oxide capping layers with different oxygen vacancy formation energies 26 , and also to transform thin films of La 0.67 Sr 0.33 CoO 3 (LSCO) from its perovskite form to the brownmillerite form La 0.67 Sr 0.33 CoO 2.5 by capping thin film with Gd 27 . Since the underlying mechanism of this effect is not material specific, it should be feasible to apply this approach to a broader spectrum of oxide material systems.
In this work, we use a TiO 2 capping layer to construct "oxygen diodes" of TiO 2 /VO 2 heterostructures, in order to tune the MIT behavior of VO 2 . Interfacial oxygen vacancy migration from rutile TiO 2 to VO 2 is induced by the much higher oxygen vacancy formation energy (E f ) of TiO 2 compared with VO 2, as predicted by our ab initio many body diffusion Monte Carlo (DMC) as well as DMC-benchmarked density function theory (DFT + U) based calculations of E f across this heterostructure. The formation of oxygen vacancies in VO 2 induced by a TiO 2 capping layer was also confirmed by an observed c-axis lattice expansion using X-ray diffraction (XRD), as well as by electron microscopy and X-ray photoelectron spectroscopy characterization, in agreement with our computational predictions. The MIT behavior in VO 2 was drastically altered by the incorporation of oxygen vacancies. While a VO 2 thin film without a TiO 2 capping layer showed a sharp transition in resistivity vs. temperature, the transition was shown to be suppressed in VO 2 with TiO 2 capping layers. Our combined experimental and computational investigation on the TiO 2 /VO 2 system underscores the effects of oxygen defect redistribution in oxide heterostructures on altering electronic structure and demonstrates how to use the 'oxygen diode' effect to control a MIT.

Methods
Thin film deposition. VO 2 and TiO 2 /VO 2 thin films were grown by using pulsed laser epitaxy (PLE) on TiO 2 (001) single crystals (CrysTec, Germany). Growth temperature was fixed at 300 °C, and VO 2 layer was grown in oxygen partial pressure (pO 2 ) of 15 mTorr, while the TiO 2 capping layers were grown in three different pO 2 , i.e. 10 mTorr, 15 mTorr and 20 mTorr. X-ray diffraction (XRD) was performed on the grown thin films by using a four-circle X-ray diffractometer (Panalytical X'Pert MRD). Transport measurement. Electrical contacts to thin film samples were made by using ultrasonic Al wire bonding. A Physical Property Measurement System (PPMS, Quantum Design) was used for measuring resistivity as a function of temperature by performing a warming up and cooling down cycle.

STEM and EELS.
Cross-sectional STEM specimens were prepared to see along the [100]TiO 2 substrate direction using mechanical thinning and precision polishing followed by ion milling. High-angle annular darkfield (HAADF) imaging and STEM-EELS analysis were carried out in Nion UltraSTEM200 operated at 200 keV. For HAADF imaging, inner detector angle of 65 mrad was used, and the convergence semi-angle for the electron probe was set to 30 mrad. X-ray photoelectron spectroscopy (XPS). XPS measurements were performed on a VO 2 thin film sample and a VO 2 thin film with a very thin TiO 2 capping layer (~ 2 nm). The thin TiO 2 capping layer was deposited by using the same condition as the LP sample (i.e., 10 mTorr pO 2 ). The reason for choosing a thin thickness for the TiO 2 capping layer is due to the shallow probing depth of XPS (estimated ~ 5 nm for V 2p spectra). XPS spectra of O 1s and V 2p were collected by using a system equipped with a monochromated Al Kα X-ray source (hν = 1486.7 eV) and a multi-channel detector (Sigma Surface Science, Germany). XPS data collection, as well as XRD and STEM characterizations, were performed at room temperature.
DFT calculations. Density functional theory (DFT) calculations were performed using the Vienna ab initio simulation package (VASP 28 ). Electron exchange correlation was represented by the functional of Perdew, Burke, and Ernzerhof (PBE) of generalized gradient approximation (GGA 29 ). The correlation effects were considered by using the DFT + U method 30 with U values of (U V ,U Ti ) = (4.0,5.0) eV as determined by scanning the twoparameter space using many-body quantum Monte Carlo (QMC) calculations (further details available in SI and "QMC calculation" section below). The ion-electron interaction was described with the projector augmented wave (PAW) method 31 . A cutoff energy of 400 eV was used for the plane-wave basis set. The internal coordinates were relaxed until the forces were lower than 0.01 eV/Å. The calculations were performed with spin polarization, and a ferromagnetic (FM) ordering was applied to the V sites.
DFT calculations were performed on 48-atom and 96-atom TiO 2 /VO 2 supercells. The 48-atom supercell contains four VO 2 layers and four TiO 2 layers stacked along the rutile c axis (a = 6.4966 Å, b = 6.4966 Å, c = 11.5892 Å), while the 96-atom supercell contains eight VO 2 layers and eight TiO 2 layers (c = 23.1784 Å). Monkhorst-Pack k-point meshes of 6 × 6 × 3 and 6 × 6 × 2 were used for the 48-atom and 96-atom supercells, respectively. The O vacancy formation energy E O f was calculated using the equation: is the total energy of the supercell with one oxygen vacancy, E tot perf is the total energy of the pristine supercell, and µ O is the chemical potential of oxygen. We use a definition of µ O common in the literature 32 is the energy of an oxygen atom, and ε coh O2 is the cohesive energy of oxygen molecule. From experimental data, we determined ε coh O2 to be − 5.21 eV 33 .
QMC calculations. Calculations using the Diffusion Monte Carlo (DMC) 34 flavor of QMC were performed on 48 atom rutile TiO 2 /VO 2 interfacial cells containing a neutral oxygen vacancy for all inequivalent sites. The V site magnetic moments were constrained to a ferromagnetic configuration. High quality pseudopotentials 35 were used to represent the V, Ti, and O species. Fixed node errors were minimized by optimizing LDA + U 36 parameters for both the V and Ti species separately. All other parameters of the trial wavefunction were optimized via the linear method 37 . All supercell results were averaged over a 2 × 2 × 2 Gamma centered supercell twist grid. The calculations were performed with the QMCPACK simulation code 38 at the Argonne Leadership Computing Facility and all simulation workflows were driven with the Nexus workflow automation system 39 .

Results and discussion
VO 2 thin films and TiO 2 /VO 2 heterostructures were grown on TiO 2 (001) substrates by pulsed laser epitaxy (PLE) in different oxygen partial pressures (pO 2 ). The deposition conditions for the VO 2 layers were fixed to a substrate temperature of 300 °C and 15 mTorr pO 2 , while the growth pO 2 of the TiO 2 capping layers was varied between 10 mTorr, 15 mTorr and 20 mTorr. Both the VO 2 thin films and the TiO 2 capping layers were ~ 15 nm thick. We refer to the three TiO 2 /VO 2 heterostructures grown in different pO 2 as LP, MP and HP (i.e., low, medium and high pressure). The different growth pO 2 of the TiO 2 capping layer effectively changes the concentration of formed oxygen vacancy in TiO 2 during deposition. Figure 1 shows XRD scans on VO 2 and TiO 2 /VO 2 samples. The clearly resolved thickness fringes indicate the high sample quality. We observed a clear trend of c-axis lattice expansion following the direction of decreasing growth pO 2 of TiO 2 capping layer (i.e. HP < MP < LP). This "chemical expansion" 40 indicates an increased oxygen vacancy concentration in LP and MP samples compared with bare VO 2 and HP samples. Therefore, the XRD data is a strong proof of interfacial migration of formed oxygen vacancies in the TiO 2 capping layer to the VO 2 layer underneath. A lower growth pO 2 introduces more oxygen vacancies into VO 2 layer, which results a larger c-axis lattice spacing. We also performed detailed structural and chemical analysis to verify the reduction of VO 2 induced by TiO 2 capping layers, as shown in Fig. 2. Figure 2a shows a Z-contrast (where Z refers to atomic number) high-angle annular dark-field (HAADF) scanning transmission electron microscopy (STEM) image of a TiO 2 /VO 2 heterostructure grown on a TiO 2 substrate. Clearly resolved atom columns and coherent interface again indicate the high sample quality. Figures 2b,c show electron energy loss spectra (EELS) of the Ti L 2,3 -edge of the TiO 2 substrate and capping layer, as well as V L 2,3 -edge of LP and HP samples. The identical line shapes and peak positions in Ti L 2,3 -edge show the nearly indistinguishable chemical states between TiO 2 substrate and capping layer. The well-resolved t 2g and e g peaks indicate that Ti cations in the TiO 2 capping layer have 4 + oxidation state with no appreciable reduction to 3 + , and do not change valence with decreasing deposition pO 2 . In contrast, a clear peak shift towards lower energy was observed in V L 2,3 -edge spectra comparing LP and HP sample, indicating a lower V oxidation state (high oxygen vacancy concentration) in the LP sample. EELS spectroscopic results on TiO 2 capping layer and VO 2 layer evidently reveal that the oxygen vacancies introduced by lowering growth pO 2 of TiO 2 capping layer migrates to the VO 2 layer, resulting in the reduced V oxidation state. We believe the atomic-scale www.nature.com/scientificreports/ oxygen vacancy redistribution in VO 2 could be affected and non-uniform along the thickness direction due to the vicinity effects associated with the surface and interface. Further investigation to determine both the quantity and spatial nonuniformity of oxygen vacancies would provide useful information to accurately understand the ion transport in VO 2 /TiO 2 heterostructures. The unidirectional oxygen vacancy migration from TiO 2 to VO 2 was also probed by using X-ray photoelectron spectroscopy (XPS), shown in Fig. 2d,e. Due to the shallow probing depth of XPS (~ 5 nm), we decreased the thickness of the TiO 2 capping layer down to ~ 2 nm, while interestingly the oxygen diode effect was still apparent. VO 2 is known to have overoxidized surface layers, therefore a V 5+ oxidation state was present in the V 2p spectra 41,42 . Due to the existence of V 5+ oxidation state, we used the average V oxidation state, represented by the intensity ratio of V 4+ and V 5+ (i.e., I(V 4+ )/I(V 5+ )), as a measure of oxygen vacancy concentration in the VO 2 layer. As shown in Fig. 2d, the TiO 2 capping lowered the average oxidation state of V, as indicated by a higher relative intensity of V 4+ peak (higher I(V 4+ )/I(V 5+ )). On the other hand, the Ti 2p spectrum collected on TiO 2 capping layer (Fig. 2e) can be fitted with only one single component of Ti 4+ , which is consistent with EELS results showing the near absence of oxygen deficiency in TiO 2 capping layer. We would like to point out that, since the quantitative analysis of oxygen vacancies based on the local EELS data and surface sensitive XPS is challenging, we remain our conclusion be only qualitative.
We hypothesize that the increased oxygen deficiency is due to the oxygen diode effect, and results in strongly affecting the MIT of VO 2 , as shown by transport (resistivity ~ temperature plot) data in Fig. 3. We observed a clear MIT in the TiO 2 /VO 2 HP sample with a sharp change in the resistivity at T c of ~ 300 K. This is consistent with the XRD data showing that there is no difference between the c-axis lattice spacing of HP and bare VO 2 , indicating a nearly identical oxygen stoichiometry between these two samples. Contrarily, the sharp MIT was gradually suppressed with decreasing deposition pO 2 . While the MP sample showed a much lower ON/OFF ratio (R OFF /R ON , defined as the ratio of resistance measured during warming up and cooling down exepriments at T c , see Fig. 3b) compared with the HP sample, the resistivity hysteresis was completely suppressed in the LP sample. The observed suppression of the MIT is consistent with a previous report on oxygen deficient VO 2-δ thin films 17 . We also note that our finding is consistent with a recent repot on VO 2 /TiO 2 heterostructures 43 . By utilizing the proposed "oxygen diode effect" we could tune the MIT and thereby change the ON/OFF ratio by three orders of magnitudes. This therefore provides a novel route of controlling oxide electronic properties.
We further explored the mechanism of the unidirectional oxygen vacancy migration in TiO 2 /VO 2 by performing ab initio many body diffusion Monte Carlo (DMC) as well as DMC-benchmarked density function theory (DFT + U) calculations. Because of two different 3d metal atoms, we used a hybrid U value of (U V ,U Ti ) = (4.0,5.0)  www.nature.com/scientificreports/ eV as benchmarked by surveying the two-dimensional parameter space using total-energies from many-body quantum Monte Carlo (QMC) calculations (further details about the calculations as well as benchmarking can be found in the SI). A 96-atom supercell consisting of TiO 2 /VO 2 bilayers with an atomically 'sharp' interface was used in the DFT calculations to evaluate the (neutral) oxygen vacancy formation energy (E f ) in each oxygen atomic layer, as shown in Fig. 4. A gradient of E f across the TiO 2 /VO 2 interface was observed, with E f decreasing from the TiO 2 layer to the VO 2 layer. E f at oxygen layers away from the interfaces approaches the bulk E f value of TiO 2 (~ 5.8 eV) and VO 2 (~ 4.6 eV). In spite of quantitative differences (Fig. S2, discussion in SI), that can be rationalized by comparing the underlying electronic-structure changes (Fig S3), the considerably lower (~ 1 eV) formation energy of oxygen vacancies in VO 2 compared to TiO 2 is further supported by benchmark DMC calculations performed in a smaller 48-atom supercell. These DMC calculations provide further support for a solid theoretical justification for the hypothesis that the TiO 2 /VO 2 heterostructure behaves as an "oxide diode", similar to the perovskite heterostructures 26 . Allowing effects of cation intermixing to model more realistic interface geometries, at the otherwise 'sharp' interface model, does not change this conclusion (Fig. S4). This implies that there is a strong chemical driving force for oxygen vacancy migration from TiO 2 layer to VO 2 layer, which is responsible for tuning the oxygen non-stoichiometry, and thereby achieving the suppression of the MIT, that we achieved by using TiO 2 capping layers grown at different pO 2 conditions. Even though we did not model the dynamic process of oxygen migration from TiO 2 to VO 2 , we would like to point out that the growth pO 2 of the TiO 2 capping layer is the only parameter we altered. Since oxygen vacancies can be readily introduced into grown TiO 2 by lowering growth pO 2 (See Figure S5 in SI), the only feasible mechanism that responsible for the observed suppressed MIT is the oxygen vacancy mirgration we discussed above.

Conclusions
In summary, by utilizing the "oxygen diode effect" in TiO 2 /VO 2 heterostructure we have demonstrated tunability of the MIT and thereby changes in the ON/OFF resistivity ratio by three orders of magnitudes. This, therefore provides a novel route of controlling oxide electronic properties. Electronic-structure calculations as well as multiple characterization methods including XRD, STEM-EELS and XPS were used to prove the unidirectional oxygen vacancy migration across the interface from TiO 2 to VO 2 . Our results highlight the importance of oxygen defect redistribution by design in oxide heterostructures on determining physical properties and functionalities.

Data availability
The data that support the findings of this study are available in the supplementary material and from the corresponding author upon reasonable request.