## Abstract

Applications of correlated vanadium dioxides VO_{2}(A) and VO_{2}(B) in electrical devices are limited due to the lack of effective methods for tuning their fundamental properties. We find that the resistivity of VO_{2}(A) and VO_{2}(B) is widely tunable by doping them with tungsten ions. When *x* < 0.1 in V_{1−x}W_{x}O_{2}(A), the resistivity decreases drastically by four orders of magnitude with increasing *x*, while that of V_{1−x}W_{x}O_{2}(B) shows the opposite behaviour. Using spectroscopic ellipsometry and X-ray photoemission spectroscopy, we propose that correlation effects are modulated by either chemical-strain-induced redistribution of V−V distances or electron-doping-induced band filling in V_{1−x}W_{x}O_{2}(A), while electron scattering induced by disorder plays a more dominant role in V_{1−x}W_{x}O_{2}(B). The tunable resistivity makes correlated VO_{2}(A) and VO_{2}(B) appealing for next-generation electronic devices.

### Similar content being viewed by others

## Introduction

Since the level of correlation effects between neighbouring electrons in the outermost 3*d* orbitals of vanadium ions differs depending on the crystal structure, the polymorphs of vanadium dioxide (VO_{2}) show a wide range of electrical properties, acting as an insulator in monoclinic VO_{2}(M1) and tetragonal VO_{2}(A), a semiconductor in monoclinic VO_{2}(B), and a conductor in tetragonal VO_{2}(R)^{1,2,3}. Therefore, VO_{2} polymorphs have been extensively studied for a range of interesting applications. VO_{2}(M1) and VO_{2}(R) have attracted wide interest for electronic devices since they show heat-, light-, electric field-, and chemical-induced reversible metal-insulator transitions near room temperature (*T*_{MI} = 340 K in bulk)^{4,5,6,7}. On the other hand, VO_{2}(A) and VO_{2}(B) have been mainly used for energy applications, including redox-flow batteries, ion batteries, solid oxide fuel cells, hydrogen storage devices, and catalysts^{8}. Compared to the many studies on electronic devices using VO_{2}(M1) and VO_{2}(R), however, the scarcity of published methods to tune the electrical properties of correlated VO_{2}(A) and VO_{2}(B) has limited their potential applications in electrical devices.

Several methods have been suggested for tailoring the correlation effects of VO_{2}(M1) since broad tunability of electrical properties is valuable for electrical devices. In addition to metallization induced by pressure application^{9}, hydrogen doping^{10,11}, or ionic liquid gating^{12,13}, cation doping has been widely used as a method with high efficacy. Chromium doping can result in transition of the dimerization of V−V chains in the VO_{2}(M1) phase into partial dimerization in the VO_{2}(M2) phase^{14}. Substituting a small amount of tungsten for vanadium in VO_{2}(M1) causes notable changes in its electrical property^{15}. For convenience, V_{1−x}W_{x}O_{2} will be used herein to denote VO_{2} with tungsten doping of *x* × 100%. As *x* increases, the resistivity of V_{1−x}W_{x}O_{2} films decreases by several orders of magnitude, and *T*_{MI} also rapidly decreases at a rate of *dT*_{MI}/*dx* = 2,100–2,800 K. For 0.08 < *x* < 0.09, the V_{1−x}W_{x}O_{2} epitaxial films have a metallic ground state for a wide temperature range of 50–400 K. When *x* is increased beyond this range, V_{1−x}W_{x}O_{2} reenters its insulating state.

Exploring tuning knobs of the electrical properties of VO_{2}(A) and VO_{2}(B) would not only enable their usage in electrical devices but also promote rich functionalities for energy devices. Motivated by research on correlated VO_{2}(M1), here, we investigate the effects of tungsten doping on the resistivity of correlated VO_{2}(A) and VO_{2}(B). We organize this paper as follows: first, we describe how, as the tungsten concentration increases, chemical strain increases the lattice parameters of both VO_{2}(A) and VO_{2}(B). Next, we find that tungsten doping is effective for tuning the resistivity of correlated VO_{2}(A) and VO_{2}(B) over a broad range. For *x* < 0.1–0.15, VO_{2}(A), an insulator in the pure phase, exhibits a monotonic decrease in resistivity of four orders of magnitude with increasing *x*. VO_{2}(B), a semiconductor in the pure phase, shows a monotonic increase in the resistivity of two orders of magnitude. Finally, to understand these opposite dependences, we explore the systematic evolution of the electronic structures and vanadium oxidation states by performing spectroscopic ellipsometry and X-ray photoemission spectroscopy (XPS), respectively.

### Chemical tensile strain in VO_{2}(A) and VO_{2}(B) induced by tungsten doping

Using pulsed laser epitaxy, we grew (100)-oriented V_{1−x}W_{x}O_{2}(A) and (001)-oriented V_{1−x}W_{x}O_{2}(B) epitaxial films on (011)SrTiO_{3} and (001)LaAlO_{3}, respectively, for *x* from 0 to 0.25. We provided details of the deposition conditions in our previous reports^{2,3,16,17,18} as well as in the Methods section. Figure 1a,b show the X-ray diffraction (XRD) *θ* − 2*θ* scans of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B), respectively (see Fig. S1 for XRD *θ* − 2*θ* scans in a wider 2*θ* range). We clearly observed (600)VO_{2}(A) diffraction peaks for *x* < 0.1 and (002)VO_{2}(B) diffraction peaks for *x* < 0.15, indicating preservation of the crystal structures for *x* < *x*_{c} ≈ 0.1–0.15. However, for *x* > *x*_{c}, these XRD intensities significantly decreased, and some peaks disappeared, indicating that heavy doping of tungsten could lower the quality of our epitaxial VO_{2}(A) and VO_{2}(B) films [we ruled out possible effects of film thickness on the reduction of the XRD intensities since the X-ray reflectivity results (Fig. S2) revealed an almost invariant thickness of ~100 nm with tungsten doping.]. This observation is different from the seemingly preserved crystallinity up to *x* = 0.33 in V_{1−x}W_{x}O_{2} epitaxial films grown on (001)-oriented TiO_{2}^{15}, which might be due to a strong strain effect between isostructural VO_{2}(R) and TiO_{2}. Hereafter, we will mainly focus on the resistivity of V_{1−x}W_{x}O_{2} (*x* < *x*_{c}) to avoid unwanted effects from film deterioration.

Substitution with tungsten ions, which are larger (0.60 Å)^{19} than vanadium ions (0.58 Å), causes chemical tensile strain in V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) epitaxial films. As shown by the XRD *θ* − 2*θ* scans, the (600)VO_{2}(A) diffraction peak gradually shifts to a lower 2*θ* angle with increasing tungsten concentration (Fig. 1a), indicating an increase in the *a*-axis lattice parameter. We observe a similar behaviour for V_{1−x}W_{x}O_{2}(B) films. The observation of the (002)VO_{2}(B) diffraction peak at a lower 2*θ* angle with tungsten doping (Fig. 1b) indicates an increase in the *c*-axis lattice parameter. As shown in Fig. 1c, the *a*- and *c*-axis lattice parameters of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) increase by 4.8% from 8.52 to 8.91 Å and by 2.4% from 6.17 to 6.32 Å, respectively, for *x* = 0–0.25. This chemical tensile strain is similar to the increase in the *c*-axis lattice parameter by 5.7% from 2.83 to 2.99Å for *x* = 0–0.33 in V_{1−x}W_{x}O_{2} epitaxial films grown on (001) TiO_{2} (for convenience of this calculation, we assumed that V_{1−x}W_{x}O_{2} epitaxial films have a tetragonal structure on (001)TiO_{2,} although R, M1, and the intermediate phases could coexist in one film, as indicated by a broadened *T*_{MI})^{15,20}.

### Tunable resistivity of VO_{2}(A) and VO_{2}(B) with tungsten doping

With increasing tungsten concentration (*x* < *x*_{c}), the resistivity of VO_{2}(A) decreased, while that of VO_{2}(B) increased. Figure 2a shows the temperature dependence of the resistivity of V_{1−x}W_{x}O_{2}(A) for *x* = 0–0.25. The resistivity of pure VO_{2}(A) was very high, i.e., 4.37 Ω cm at 400 K, and increased with decreasing temperature. Such insulating behaviour is attributed to correlation-induced bandgap opening between unoccupied and occupied *t*_{2g} orbitals^{1}. It should be noted that the resistivity of V_{1−x}W_{x}O_{2}(A) decreased with *x* (<*x*_{c}, solid lines). The V_{1−x}W_{x}O_{2}(A) epitaxial film for *x* = 0.15 showed a smaller resistivity (by three orders of magnitude) of 0.001 Ω cm at 400 K than that of pure VO_{2}(A). This small resistivity indicated that the film was on the verge of metallicity, in terms of the Mott-Ioffe-Regel limit^{21} (i.e., the material is regarded as a metal when the resistivity is smaller than 0.001 Ω cm.). The resistivity of V_{1−x}W_{x}O_{2}(A) increased for *x* > *x*_{c} (dashed lines), probably due to film deterioration. As shown in Fig. 2b, the resistivity of pure VO_{2}(B) also increased with decreasing temperature, indicating insulating behaviour. However, its resistivity was close to the Mott-Ioffe-Regel limit^{21} at 400 K and significantly smaller (by three orders of magnitude) than that of pure VO_{2}(A). Such a low resistivity in pure VO_{2}(B) is ascribed to thermal electron jumping across the very narrow bandgap (<25 meV) near room temperature^{1}. The resistivity of V_{1−x}W_{x}O_{2}(B) increased with increasing *x* by two orders of magnitude. Moreover, the resistivity increased more for *x* > *x*_{c}, as observed in V_{1−x}W_{x}O_{2}(A).

We noted interesting features of the tungsten doping effects on the resistivity of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) (*x* < *x*_{c}). The dependences were opposite: the resistivity of V_{1−x}W_{x}O_{2}(A) decreased, i.e., 4.37 → 0.01 Ω cm at 400 K for *x* = 0 → 0.1, while V_{1−x}W_{x}O_{2}(B) showed increasing resistivity, i.e., 0.003 → 0.01 Ω cm. It is surprising that the resistivities of VO_{2}(A) and VO_{2}(B) changed to such an extent, although we doped a relatively small amount (*x* < 0.1) of tungsten ions. Therefore, our work indicates that tungsten doping is promising for tuning the resistivity of VO_{2}(A) and VO_{2}(B). This extensive tunability is expected to provide many opportunities to realize both electronic and energy devices using correlated VO_{2}(A) and VO_{2}(B).

To obtain more information about these opposing and large dependences, we compared the tungsten-doping dependences of the resistivity in V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) to those in previous reports on V_{1−x}W_{x}O_{2}(M1) and V_{1−x}W_{x}O_{2}(R). It was simple to evaluate the resistivities of our V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) epitaxial films since, across a wide temperature range, they do not show any phase transitions. However, when we plotted the resistivities of V_{1−x}W_{x}O_{2}(M1) and V_{1−x}W_{x}O_{2}(R), we had to pay attention to the different resistivities of the M1 and R phases due to the *T*_{MI} variation induced by the tungsten doping. Figure 2c–f show the 400 K resistivities of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B), the 135 K resistivity of V_{1−x}W_{x}O_{2}(M1)^{15,20,22}, and the 400 K resistivity of V_{1−x}W_{x}O_{2}(R)^{15,20,22}. We note three features of the tungsten doping effect across the polymorphs. (1) For light doping (*x* < *x*_{c}), at which the crystal structures are well preserved (highlighted in yellow), the resistivities of insulating VO_{2}(A) and VO_{2}(M1) decrease by 3–4 orders of magnitude compared to those in the pure phases, while semiconducting VO_{2}(B) and metallic VO_{2}(R) exhibit increases in the resistivity by 1–2 orders of magnitude compared to those in the pure phases. (2) It is quite surprising that the resistivities of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(M1) can be smaller than those of V_{1−x}W_{x}O_{2}(B) and V_{1−x}W_{x}O_{2}(R) when doped with certain amounts of tungsten (e.g., *x* ≈ *x*_{c} in this work). (3) For heavy doping (*x* > *x*_{c}), all phases show increased resistivity. We suggest that the increases seen in V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) can be attributed to deterioration of the films because we observed suppression of diffraction peaks in the *θ*−2*θ* XRD scans (Fig. 1a,b). The similar dependences between VO_{2}(A) and VO_{2}(M1) and between VO_{2}(B) and VO_{2}(R) suggest that the proposed mechanisms underlying the tungsten doping effects in VO_{2}(M1) and VO_{2}(R) may apply to VO_{2}(A) and VO_{2}(B), respectively.

### Electronic structures of tungsten-doped VO_{2}(A) and VO_{2}(B)

To better understand the opposing dependences on tungsten doping, we investigated the electronic structures of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) epitaxial films for various *x* values. Figure 3a shows the optical conductivity *σ*_{1}(*ω*) of V_{1−x}W_{x}O_{2}(A) (*x* = 0, 0.05, 0.1) as a function of photon energy. Pure VO_{2}(A) (first row) exhibited opening of a correlation-induced bandgap^{1}. For *x* = 0.05 (second row), a spectral weight distinctly appeared near 0.8 eV. For *x* = 0.1 (third row), the bandgap might be narrower than 25 meV at room temperature (Figs. S3 and S4), so the electrons in the occupied *t*_{2g} orbital could thermally jump into the unoccupied *t*_{2g} orbitals even at room temperature, consistent with the low resistivity shown in Fig. 2a. To examine the variation in optical spectra with *x* in more detail, we evaluated *σ*_{1}(*ω*) using Lorentz oscillators, \({{\rm{\sigma }}}_{1}(\omega )=\frac{{e}^{2}}{{m}^{\ast }}\frac{{N}_{D}{\gamma }_{D}}{{\omega }^{2}+{\gamma }_{D}^{2}}+\frac{{e}^{2}}{{m}^{\ast }}\sum _{j}\frac{{N}_{j}{\gamma }_{j}{\omega }^{2}}{{({\omega }_{j}^{2}-{\omega }^{2})}^{2}+{\gamma }_{j}^{2}{\omega }^{2}}\), where *m**, *γ*_{j}, and *ω*_{j} are the effective mass, damping coefficient, and angular frequency of the *j*^{th} resonance line, respectively^{23}. The first and second terms in *σ*_{1}(*ω*) represent the metallic Drude response and interband transitions, respectively. The *β*-peak represents an interband transition from occupied *t*_{2g} to unoccupied *t*_{2g} orbitals and shifts to a lower photon energy with increasing *x*. Additionally, a new peak (asterisk) appeared near 0.8 eV, representing the creation of an in-gap state between the occupied *t*_{2g} and unoccupied *t*_{2g} orbitals. The spectroscopic findings are consistent with the decrease in resistivity in V_{1−x}W_{x}O_{2}(A). It should be noted that the spectroscopic results for V_{1−x}W_{x}O_{2}(A) are similar to observations for V_{1−x}W_{x}O_{2}(M1)^{24}. Figure 3b shows proposed changes in the electronic structures of VO_{2}(A) with tungsten doping, i.e., a shift of the unoccupied *t*_{2g} orbital towards the Fermi level and the appearance of a new in-gap state just above the Fermi level.

Different from V_{1−x}W_{x}O_{2}(A), the electronic structure of V_{1−x}W_{x}O_{2}(B) did not show any obvious changes. Figure 3c shows the *σ*_{1}(*ω*) of V_{1−x}W_{x}O_{2}(B) (*x* = 0, 0.05, 0.1) as a function of photon energy. Pure VO_{2}(B) in the first row exhibited a non-negligible spectral weight near zero photon energy^{1}, consistent with its low resistivity near room temperature shown in Fig. 2b. With increasing tungsten concentration, the *β-*peak moved very slightly towards higher photon energy, and the Drude response (yellow line) was suppressed. Although such a blueshift is very weak, it is somewhat consistent with the more resistive V_{1−x}W_{x}O_{2}(B) with increasing *x*. Figure 3d shows a very weak change in the electronic structure of VO_{2}(B) with tungsten doping. Therefore, we suggest that the resistivity increase in V_{1−x}W_{x}O_{2}(B) is attributable to mechanisms other than any simple change in the electronic structure.

### Mechanisms underlying the tunable resistivity of tungsten-doped VO_{2}(A) and VO_{2}(B)

Taking our experimental results together, we found that the resistivities of VO_{2}(A) and VO_{2}(M1) and those of VO_{2}(B) and VO_{2}(R) have similar dependences on tungsten doping. Numerous studies have attributed the stabilization of metallic V_{1−x}W_{x}O_{2}(M1) to correlation variations, with structural distortion of V−V dimers^{24,25,26} and band filling by electron doping^{27}, as we will explain in detail in the following paragraphs. In this stage, we aimed to understand the behaviours of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) (*x* < *x*_{c}) by adapting and modifying the mechanisms in V_{1−x}W_{x}O_{2}(M1) and V_{1−x}W_{x}O_{2}(R).

V_{1−x}W_{x}O_{2}(A) shows tunable properties due mainly to correlation effects being modulated by chemical-strain-induced redistribution of V−V distances. Goodenough noted that vanadium oxides are metallic when the distance between vanadium ions is less than 2.94 Å^{28}. VO_{2}(M1) is insulating because correlated electrons are localized in V−V dimers, where the asymmetric distances of V−V atoms inside and between dimers are 2.65 and 3.12 Å, respectively. It is widely accepted that the transition from insulating VO_{2}(M1) to metallic VO_{2}(R) is accompanied by a symmetric redistribution of V−V chains, with an even distance of 2.88 Å. When vanadium ions are replaced with tungsten, the X-ray absorption fine structure indicates that the local structure around each tungsten atom is intrinsically symmetric, with a tetragonal-like structure. Therefore, the nearby V−V dimers in a VO_{2}(M1) lattice are rearranged to form rutile-like VO_{2} nuclei^{25,26}. In VO_{2}(A), vanadium ions along the *c*-axis have alternating distances of 3.25, 3.11, and 2.77 Å at room temperature (<162 °C)^{29}. Although the XRD results imply that V_{1−x}W_{x}O_{2}(A) mostly has a tetragonal structure similar to pure VO_{2}(A), expansion of the local structure due to tungsten could symmetrize the V−V chains. This rearrangement would weaken the correlation effect in V_{1−x}W_{x}O_{2}(A), leading to lower resistivity.

As an alternative mechanism for V_{1−x}W_{x}O_{2}(A), we also considered electron doping since V^{4+} ions neighbouring the site of W^{6+} dopants change to V^{3+} ions to maintain charge neutrality^{30}. This band filling drastically decreases the Coulomb repulsion energy and accordingly weakens the electron correlation^{27}, metallizing V_{1−x}W_{x}O_{2}(M1). We also found a significant evolution of V^{3+} oxidation states in V_{1−x}W_{x}O_{2}(A) with tungsten doping (*x* = 0, 0.05, 0.1). Figure 4 shows XPS V 2*p*_{3/2}, V 2*p*_{1/2}, and O 1 *s* spectra in the binding energy range of 505–535 eV. We fitted the XPS spectra of pure VO_{2}(A) (Fig. 4a) with V^{4+}2*p*_{3/2} (red pattern) at 515.84 ± 0.2 eV and V^{4+}2*p*_{1/2} (orange pattern) at V^{4+}2*p*_{3/2} + 7.33 eV^{31}, indicating that the oxidation state of our non-doped epitaxial films was firmly V^{4+}. Interestingly, V^{3+} peaks [V^{3+}2*p*_{3/2} (blue pattern) at 515.29 ± 0.2 eV and V^{3+}2*p*_{1/2} (purple pattern) at V^{3+}2*p*_{3/2} + 7.33 eV^{31}] were additionally required to resolve the XPS spectra of V_{1−x}W_{x}O_{2}(A) (*x* = 0.05, 0.1). It should be noted that the V^{3+} peaks were stronger with increasing *x*. This similar observation between V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(M1) indicates that electron-doping-induced band filling also plays an important role in tungsten-doped metallization^{30}.

Although we also found significant evolution of the V^{3+} XPS peaks in V_{1−x}W_{x}O_{2}(B) (Fig. 4b), it is quite interesting to note that VO_{2}(B) became more insulating with increasing *x*. Therefore, we hypothesize that another mechanism, different from those for V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(M1), plays an important role in the tunable properties of V_{1−x}W_{x}O_{2}(B) (*x* < *x*_{c}). It should be noted that the correlation effects in VO_{2}(B) and VO_{2}(R) are weaker than those in VO_{2}(A) and VO_{2}(M1), considering their lower resistivities and narrower bandgaps^{1,2}. Therefore, we suggest that the increase in resistivity in V_{1−x}W_{x}O_{2}(B) originates from disorder-induced electron scattering. Since more dopants will scatter more electrons, the resistivity will increase with heavier doping^{32,33}.

## Conclusion

Tungsten doping provided an effective way to tune the resistivity of correlated VO_{2}(A) and VO_{2}(B). XRD revealed that the crystal structures of V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) expanded and were well preserved for *x* < 0.1–0.15. At this low doping concentration, the resistivity of V_{1−x}W_{x}O_{2}(A) decreased, similar to that of V_{1−x}W_{x}O_{2}(M1), with increasing tungsten concentration; in contrast, the resistivity of V_{1−x}W_{x}O_{2}(B) increased, similar to that of V_{1−x}W_{x}O_{2}(R). Spectroscopic ellipsometry revealed that tungsten doping resulted in a redshift of the unoccupied *t*_{2g} orbital, the creation of an in-gap state in V_{1−x}W_{x}O_{2}(A), and a slight blueshift of unoccupied *t*_{2g} orbitals in V_{1−x}W_{x}O_{2}(B). Both vanadate films showed evolution of the V^{3+} oxidation states based on the XPS study. Referring to the mechanisms in correlated V_{1−x}W_{x}O_{2}(M1), we proposed that V_{1−x}W_{x}O_{2}(A) and V_{1−x}W_{x}O_{2}(B) showed opposite dependences due to either chemical-strain-induced redistribution of V*−*V distances or electron-doping-induced band filling and disorder-induced electron scattering, respectively. We leave further consideration of the mechanisms for future studies.

The extreme tunability of correlated VO_{2}(A) and VO_{2}(B) enables their use in next-generation electronic devices, as well as energy devices. As we mentioned in the Introduction section, VO_{2}(M1) has shown a reversible resistivity change due to intercalation of hydrogen^{10,11} and ionic liquid gating^{12,13}. The similar doping dependence between VO_{2}(A) and VO_{2}(M1) suggests that these dynamic control methods would enable application of correlated VO_{2}(A) [also VO_{2}(B)] in memories, transistors, and gas sensors, as has been extensively studied for VO_{2}(M1).

## Methods

### Epitaxial film growth of tungsten-doped VO_{2}(A) and VO_{2}(B)

We recently optimized the growth conditions for epitaxial films of VO_{2}(A) and VO_{2}(B) on either perovskite oxides or Y-stabilized ZrO_{2}^{2,3,16,17,18}. Using pulsed laser epitaxy, we deposited VO_{2}(A) and VO_{2}(B) epitaxial films on (011)-oriented SrTiO_{3} and (001)-oriented LaAlO_{3} substrates, respectively. We ablated a tungsten-doped V_{2}O_{5} target with a KrF (248 nm wavelength) pulsed laser at a rate of 10 Hz and an intensity of 1 J cm^{−2}. For the targets, we mixed WO_{3} and V_{2}O_{5} powders in the desired molar ratio and sintered pellets at 650 °C for 12 hours in air. We used this low sintering temperature due to the low melting point (690 °C) of V_{2}O_{5}. We fixed the substrate temperature at 420 °C since VO_{2}(A) and VO_{2}(B) are thermodynamically unstable above 430 °C and transition into VO_{2}(R) above 470 °C^{2,29}. We used a flow of oxygen gas with a partial pressure, \({P}_{{{\rm{O}}}_{2}}\), of 8 mTorr for VO_{2}(A) and 15 mTorr for VO_{2}(B) since VO_{2} stably forms in only a narrow range of 5 mTorr <\({P}_{{{\rm{O}}}_{2}}\) < 30 mTorr^{16,34}.

### Measurement of electrical and optical properties

To investigate the electrical transport properties, we used a physical property measurement system (Quantum Design Inc.). We used the four-point probe method, which is the most common method for measuring the resistivity^{35}. We deposited evenly spaced Pt contacts on the middle of the film surface. We applied a small constant current through the outer two contacts and measured the voltage between the inner two contacts. We swept the temperature in the range of 10–400 K. We measured the reflectance, *R*(*ω*), spectra in a photon energy range of 0.1–1 eV via a Fourier transform-type infrared spectrometer (model VERTEX 70 v; Bruker). We employed an *in situ* gold overcoating technique to obtain an accurate absolute value of *R*(*ω*). We obtained the optical conductivity of the VO_{2} film via a two-layer model fit of the measured *R*(*ω*) with Drude-Lorentz oscillators^{23,36}. We used spectroscopic ellipsometers (models V-VASE and M-2000; J. A. Woollam Co.) to obtain the complex dielectric constants, \(\epsilon (\omega )={\epsilon }_{1}(\omega )+i{\epsilon }_{2}(\omega )\), in the energy region between 1 and 5 eV. The optical conductivity, *σ*_{1}(*ω*), in this energy range can be calculated by \({\sigma }_{1}(\omega )={\epsilon }_{0}\omega {\epsilon }_{2}(\omega )\)^{23}, where \({\epsilon }_{0}\) is the vacuum permittivity.

### Characterization of structural properties and oxidation states

We investigated the crystal structures via a four-circle high-resolution X-ray diffractometer (model Empyrean; PANalytical) using Cu radiation with a wavelength of 1.5406 Å. Using the fringe patterns obtained in X-ray reflectivity measurements, we confirmed that the films had an ~100-nm thickness (Fig. S2). To determine the oxidation states of vanadium, we carried out XPS (model ESCALAB 250Xi; Thermo Scientific) using a monochromatic Al source with a photon energy of 1486.6 eV under an environmental pressure of 10^{−8} Torr. To remove contamination, the film surface was sputtered with argon ions for 10 seconds^{16}. We used the O 1 *s* peak at 530.0 eV as the energy reference. We supplied electrons using an electron gun to avoid any charging effect.

## References

Lee, S.

*et al*. Electronic structure and insulating gap in epitaxial VO_{2}polymorphs.*APL Mater.***3**, 126109 (2015).Lee, S., Ivanov, I. N., Keum, J. K. & Lee, H. N. Epitaxial stabilization and phase instability of VO

_{2}polymorphs.*Sci. Rep.***6**, 19621 (2016).Choi, S., Chang, S.-J., Oh, J., Jang, J. H. & Lee, S. Electrical and optical properties of VO

_{2}polymorphic films grown epitaxially on Y-stabilized ZrO_{2}.*Adv. Electron. Mater.***4**, 1700620 (2018).Driscoll, T.

*et al*. Memory metamaterials.*Science***325**, 1518 (2009).Chang, S. H.

*et al*. Oxide double-layer nanocrossbar for ultrahigh-density bipolar resistive memory.*Adv. Mater.***23**, 4063 (2011).Yang, Z., Ko, C. & Ramanathan, S. Oxide electronics utilizing ultrafast metal-insulator transitions.

*Annu. Rev. Mater.***41**, 337 (2011).Lee, S.

*et al*. Anomalously low electronic thermal conductivity in metallic vanadium dioxide.*Science***355**, 371 (2017).Lee, S.

*et al*. Persistent electrochemical performance in epitaxial VO_{2}(B).*Nano Lett.***17**, 2229 (2017).Arcangeletti, E.

*et al*. Evidence of a pressure-induced metallization process in monoclinic VO_{2}.*Phys. Rev. Lett.***98**, 196406 (2007).Wei, J., Ji, H., Guo, W., Nevidomskyy, A. H. & Natelson, D. Hydrogen stabilization of metallic vanadium dioxide in single-crystal nanobeams.

*Nat. Nanotechnol.***7**, 357 (2012).Yoon, H.

*et al*. Reversible phase modulation and hydrogen storage in multivalent VO_{2}epitaxial thin films.*Nat. Mater.***15**, 1113 (2016).Nakano, M.

*et al*. Collective bulk carrier delocalization driven by electrostatic surface charge accumulation.*Nature***487**, 459 (2012).Jeong, J.

*et al*. Suppression of metal-insulator transition in VO_{2}by electric field–induced oxygen vacancy formation.*Science***339**, 1402 (2013).Zylbersztejn, A. & Mott, N. F. Metal-insulator transition in vanadium dioxide.

*Phys. Rev. B***11**, 4383 (1975).Shibuya, K., Kawasaki, M. & Tokura, Y. Metal-insulator transition in epitaxial V

_{1−x}W_{x}O_{2}(0 ≤*x*≤ 0.33) thin films.*Appl. Phys. Lett.***96**, 022102 (2010).Choi, S.

*et al*. Sharp contrast in the electrical and optical properties of vanadium Wadsley (V_{m}O_{2m}+ 1, m > 1) epitaxial films selectively stabilized on (111)-oriented Y-stabilized ZrO_{2}.*Phys. Rev. Mater.***3**, 063401 (2019).Choi, S., Oh, J., Lee, J.-H., Jang, J. H. & Lee, S. Degradation mechanism of vanadium oxide films when grown on Y‐stabilized ZrO2 above 500 °C.

*Adv. Eng. Meter.***21**, 1900918 (2019).Choi, S.

*et al*. High infrared transparency up to 8-μm-wavelength in correlated vanadium Wadsley conductors.*APL Mater.***8**, 041111 (2020).Shannon, R. D. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides.

*Acta. Cryst.***A32**, 751 (1976).Okuyama, D.

*et al*. X-ray study of metal-insulator transitions induced by W doping and photoirradiation in VO_{2}films.*Phys. Rev. B***91**, 064101 (2015).Ioffe, A. F. & Regel, A. R. Non-crystalline, amorphous and liquid electronic semiconductors.

*Prog. Semicond.***4**, 237 (1960).Takami, H., Kanki, T., Ueda, S., Kobayashi, K. & Tanaka, H. Filling-controlled Mott transition in W-doped VO

_{2}.*Phys. Rev. B***85**, 205111 (2012).Fox, M.

*Optical properties of solids*. (Oxford University Press, Oxford, United Kingdom, 2010).Lee, J. S., Shibuya, K., Kawasaki, M. & Tokura, Y. Optical investigation of metal-insulator transitions in V

_{1−x}W_{x}O_{2}(0 ≤ x ≤ 0.33).*Phys. Rev. B***85**, 155110 (2012).Booth, J. M. & Casey, P. S. Anisotropic structure deformation in the VO

_{2}metal-insulator transition.*Phys. Rev. Lett.***103**, 086402 (2009).Tan, X.

*et al*. Unraveling metal-insulator transition mechanism of VO_{2}triggered by tungsten doping.*Sci. Rep.***2**, 466 (2012).Sakai, E.

*et al*. Competition between instabilities of Peierls transition and Mott transition in W-doped VO_{2}thin films.*Phys. Rev. B***84**, 195132 (2011).Goodenough, J. B. The two components of the crystallographic transition in VO

_{2}.*J. Solid State Chem.***3**, 490 (1971).Popuri, S. R.

*et al*. VO_{2}(A): Reinvestigation of crystal structure, phase transition and crystal growth mechanisms.*J. Solid State Chem.***213**, 79 (2014).Takami, H., Kanki, T., Ueda, S., Kobayashi, K. & Tanaka, H. Electronic structure of W-doped VO

_{2}thin films with giant metal–insulator transition investigated by hard X-ray core-level photoemission spectroscopy.*Appl. Phys. Express***3**, 063201 (2010).Silversmit, G., Depla, D., Poelman, H., Marin, G. B. & De Gryse, R. Determination of the V 2p XPS binding energies for different vanadium oxidation states (V5

^{+}to V0^{+}).*J. Electron Spectrosc. Relat. Phenom.***135**, 167 (2004).Luo, X.

*et al*. High carrier mobility in transparent Ba_{1−x}La_{x}SnO_{3}crystals with a wide band gap.*Appl. Phys. Lett.***100**, 172112 (2012).Prakash, A.

*et al*. Wide bandgap BaSnO_{3}films with room temperature conductivity exceeding 10^{4}S cm^{−1}.*Nat. Commun.***8**, 15167 (2017).Lee, S., Meyer, T. L., Park, S., Egami, T. & Lee, H. N. Growth control of the oxidation state in vanadium oxide thin films.

*Appl. Phys. Lett.***105**, 223515 (2014).Sze, S. M. & Ng, K. K.

*Physics of semiconductor devices*. (Wiley, New York, United States, 2007).Kuzmenko, A. B. Kramers–Kronig constrained variational analysis of optical spectra.

*Rev. Sci. Instrum.***76**, 083108 (2005).

## Acknowledgements

This work was supported by the Defense Acquisition Program Administration and Agency for Defense Development of Korea (Project No.: 911223001). The work at Hanyang University was supported by the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT, and Future Planning (2019R1A2C1084237). We used the spectroscopic ellipsometer at the Institute for Basic Science, Centre for Correlated Electron Systems, Seoul National University, Korea.

## Author information

### Authors and Affiliations

### Contributions

S.C. conceived and performed the experiments under the supervision of S.L. S.C. and G.A. carried out spectroscopic ellipsometry and used a Fourier transform-type infrared spectrometer under the supervision of S.L. and S.J.M. S.C. and S.L. wrote the manuscript, and the other authors reviewed it.

### Corresponding author

## Ethics declarations

### Competing interests

The authors declare no competing interests.

## Additional information

**Publisher’s note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Supplementary information

## Rights and permissions

**Open Access** This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

## About this article

### Cite this article

Choi, S., Ahn, G., Moon, S.J. *et al.* Tunable resistivity of correlated VO_{2}(A) and VO_{2}(B) via tungsten doping.
*Sci Rep* **10**, 9721 (2020). https://doi.org/10.1038/s41598-020-66439-2

Received:

Accepted:

Published:

DOI: https://doi.org/10.1038/s41598-020-66439-2

## Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.