We studied using organic liquids (cyclohexane, n-butanol, and ethylene glycol) to modulate the transport properties at room temperature of an epitaxial VO2 film on a VO2/Al2O3 heterostructure. The resistance of the VO2 film increased when coated with cyclohexane or n-butanol, with maximum changes of 31% and 3.8%, respectively. In contrast, it decreased when coated with ethylene glycol, with a maximum change of −7.7%. In all cases, the resistance recovered to its original value after removing the organic liquid. This organic-liquid-induced reversible resistance switching suggests that VO2 films can be used as organic molecular sensors.
I. INTRODUCTION
The metallization of correlated insulators induced by high electric fields is a forceful method of creating novel electronic phases. However, it is often difficult to achieve such high electric fields by conventional dielectric gates.1,2 Recently, extra-high electric fields are achieved by approaches such as Schottky junctions,3,4 polar adsorbates5,6 or ionic liquids7–12 as dielectric gates in field-effect transistor devices.13,14 For example, Xu et al.15 studied the modulation of grain boundary barriers in ZnMgO/ZnO/sapphire heterostructure by DI water and they found that the electronic conduction property of ZnO layer far beneath the surface could be switched from Ohmic to Schottky junction. Recently, Xie et al.6 reported that the conductivity of an LaAlO3/SrTiO3 interface could be changed in the presence of water because it increased the sheet carrier density of the two-dimensional electron gas. Soon after, Au et al.4 demonstrated that the LaAlO3/SrTiO3 interface could be used as a molecular sensor and explained the sensing mechanism using a Schottky junction model. These results strongly suggest that liquid can dramatically modulate transport properties, which has practical application in molecular sensors.
Similarly, the electronic states of vanadium dioxide (VO2) are very sensitive to external perturbations such as liquids, and these behaviors are being studied extensively.7,8,12,16–18 VO2 is a typical strongly correlated electron system that exhibits a characteristic metal-insulator transition (MIT) at ∼68° C, where its resistance decreases by more than four orders of magnitude.19 Tokura et al. used ionic liquids to reversibly switch epitaxial VO2 films between insulating and metallic states. They found a electrostatic charge induced at the interface of the VO2 film and that the ionic liquid drove all the localized charge carriers into motion from the surface through the whole bulk, generating a three-dimensional metallic ground state.7,8 However, S. P. Parkin et al. proposed an entirely different mechanism to explain the ionic-liquid-controlled MIT in VO2 films.12,16 They believed that the electric-field-induced migration of oxygen vacancies in the VO2 film by ionic liquids was critical for understanding the mechanism of the electrically manipulated MIT. Thus, it remains unclear how ionic liquids influence the transport properties of VO2 films, and the tuning mechanism is not yet fully understood. However, note that the liquids used in these reports were always water or ionic liquids (essentially, a kind of salt). Although organic liquids are extremely important in scientific20–22 and industrial areas, few studies have explored how organic liquids influence switching of transport properties in correlated electron oxide (CEO) films. Furthermore, when the organic liquids are gating on the CEO material, its transport properties might be modulated because of the different polarity of the liquid.15 Motivated by this gap in the literature, here we study organic liquids with three representative relative polarities:23 cyclohexane (weak), n-butanol (moderate), and ethylene glycol (strong).
Motivated by the liquid-induced novel and interesting physical properties mentioned above in the oxide films and heterostructures, we fabricated epitaxial VO2/Al2O3 heterostructures and then placed organic liquids on their surfaces to modulate the conductive properties of the VO2 films at room temperature. Our results may have a promising application in novel VO2-based electronic devices such as organic molecular sensors.
II. EXPERIMENT
Reactive rf-magnetron sputtering was employed to deposit VO2 films on commercial single-crystal (0001) Al2O3 substrates with dimensions of 10×5 ×0.5 mm.24–26 A vanadium metal target (99.99% purity) was sputtered at an rf-power of 60 W, a sputtering pressure of 0.43 Pa, and an Ar:O2 flow ratio of 60:0.5. The substrate temperature was optimized to 350 ° C. The sputtering distance was ∼12 cm and the base pressure was better than 1.4×10−4 Pa. The structure, thickness, and epitaxial quality of the VO2 films were characterized by high-resolution X-ray diffraction (XRD) with Cu Kα1 (λ = 1.5406 Å) radiation (Rigaku SmartLab Film Version). Using indium as contact electrodes to form Ohmic electrical contacts,27–31 current–voltage curves were measured using the two-probe method at room temperature. The resistance of the VO2/Al2O3 heterostructure was measured by Keithley 2400 Source Meter while the liquid was placed on the center of the VO2 film. To analyze the valence states of the vanadium ions in the VO2 film, X-ray photoelectron spectroscopy (XPS) with Al Kα radiation (Thermo ESCALAB 250) was carried out.
III. RESULTS AND DISCUSSION
To evaluate the epitaxial quality of the VO2 film, we collected XRD patterns (step size of 0.01 ˚) of the VO2/Al2O3 heterostructure, as shown in Fig. 1(a). Only the (020) VO2 peak appeared, and no others from vanadium oxides, suggesting the high orientation and good purity of the VO2 film. From this (020) peak, we calculated the out-of-plane lattice constant to be 4.52 Å. This value equals that of bulk VO2,32 indicating the epitaxial strain was fully relaxed. Using the rocking curve of the (020) peak (inset of Fig. 1(a)), we found the full width at half maximum (FWHM) to be only ∼0.07 ˚, narrower than that reported by Fan et al.,33 suggesting the VO2 film had high quality. Another way to demonstrate the high quality of VO2/Al2O3 films is a sharp metal-insulator transition.34 As shown by the small-angle X-ray reflectivity (XRR) curve in Fig. 1(b), the presence of interference peaks confirms the uniformity of the film, and its thickness is measured non-destructively to be ∼92 nm. To confirm the epitaxial relationship between the VO2 film and Al2O3 substrate, the representative XRD φ scans of VO2 (220) and Al2O3 (116) peak are shown in Fig. 1(c). Both of these peaks exhibited six-fold symmetry,33,35,36 which agrees with the crystal symmetry of the VO2 and Al2O3. Furthermore, the azimuth angle φ of the VO2 (220) peak matches well with that of the Al2O3 (116) peaks, which demonstrates the perfect epitaxial relationship between the VO2 film and Al2O3 substrate. To give a more intuitive understanding of the in-plane epitaxial relationship between the film and the substrate, Figure 1(d) shows sketches of the VO2 (220) and Al2O3 (116) crystal facets. We derived their epitaxial relationship to be along the in-plane direction and [010]V O2//[0001]Al2O3 along the out-of-plane direction.33,36
(a) XRD pattern of the VO2 film grown on a c-Al2O3 substrate by rf-magnetron sputtering. Inset: Rocking curve of the VO2 (020) peak. (b) Small-angle XRR oscillation of the VO2 film. (c) In-plane φ scans of the VO2 film and Al2O3 substrate. (d) Sketch of the epitaxial relationship between the VO2 film and Al2O3 substrate.
(a) XRD pattern of the VO2 film grown on a c-Al2O3 substrate by rf-magnetron sputtering. Inset: Rocking curve of the VO2 (020) peak. (b) Small-angle XRR oscillation of the VO2 film. (c) In-plane φ scans of the VO2 film and Al2O3 substrate. (d) Sketch of the epitaxial relationship between the VO2 film and Al2O3 substrate.
To study the organic-liquid-modulated transport properties of the VO2 film, we measured the I–V curves of the VO2/Al2O3 heterostructure in a two-probe configuration at room temperature. As shown in Fig. 2(a), two parallel 5 × 1.5 mm indium metals were placed as contact electrodes along the direction. Using a microsyringe, we dropped 5 μL of the organic liquid on the center of the VO2 surface. The organic liquid sat on the center of the VO2 film and did not contact the two electrodes, preventing any possible interaction between liquid and electrodes. Also, all three organic liquids we used had conductivities at least four orders of magnitude less than that of the VO2 thin film,34 which means we can ignore the influence of the liquid on the resistance measurements. Fig. 2(b) shows the I–V curves of the device before and after placing the organic liquids in air at room temperature. The environmental humidity effect on resistance switching could be ignored from the humidity test.34 In all cases, with and without any liquid, the I–V curves exhibited Ohmic behavior. When cyclohexane was placed, the slope of the I–V curve decreased significantly, suggesting the conductivity decreased. The slope also decreased in the presence of n-butanol, but less so. In contrast, when ethylene glycol was placed, the I–V curve became much steeper, suggesting the conductivity increased. These results demonstrate that, on a VO2 film, cyclohexane and n-butanol can increase resistance while ethylene glycol can decrease resistance.
(a) The device configuration. (b) Typical I–V curves of the VO2 film with and without organic liquids.
(a) The device configuration. (b) Typical I–V curves of the VO2 film with and without organic liquids.
To further confirm these resistance-switching behaviors modulated by organic liquids, we show in Figs. 3(a) and 3(b) how the resistance changed over time at measurement voltages of +15 V and −10 V. Without any liquid, the resistance remained at its initial value. Then, at 250 s, the resistance increased as cyclohexane was dropped on the VO2 film. Then the resistance gradually decreased because the cyclohexane evaporated in air. We define the resistance modulation as
where R(PL) and R(0) are the resistances of the VO2 film with and without organic liquids, respectively. By this definition, the maximum resistance modulation of the VO2 film was 31% with cyclohexane gating. At 500 s, the resistance returned to its original state because the cyclohexane was almost completely evaporated. At 750 s, n-butanol was added, and the resistance increased. However, this resistance increase of 3.8% is small compared to that of cyclohexane. The resistance of the VO2 film slowly recovered to its original value because n-butanol is not very volatile. After the n-butanol evaporated, ethylene glycol was placed on the VO2 film, which decreased the resistance with a resistance modulation of about −7.7%. At 1500 s, we flushed dry air over the sample to quickly recover its resistance. To strengthen these results, we repeated the experiment at an applied voltage of −10 V; these results are shown in Fig. 3(b).
Resistance change over time, placing various organic liquids placed on the VO2 surface, under bias voltages of (a) +15 V (a) and (b) −10 V. Also shown are resistance changes over time at a bias voltage of +15 V with only (c) cyclohexane (d) and ethylene glycol.
Resistance change over time, placing various organic liquids placed on the VO2 surface, under bias voltages of (a) +15 V (a) and (b) −10 V. Also shown are resistance changes over time at a bias voltage of +15 V with only (c) cyclohexane (d) and ethylene glycol.
The behaviors and resistance modulations at −10 V and +15 V were quite similar. From this, we conclude that the resistance of the VO2 film can be modulated by organic liquids and that, after removing the organic liquid, the resistance returns essentially to its original state.
To confirm that our results are repeatable, we recorded the resistance of the VO2 film while repeatedly dropping one type of liquid on the VO2 surface. Fig. 3(c) shows these results for cyclohexane, and Fig. 3(d) shows them for ethylene glycol. Between these experiments we removed the cyclohexane using dry air. Both of these experiments exhibited repeatable resistance switching, and these results agree well with the results of Fig. 3(a) and 3(b).
To exclude the influence of thermal effects induced by Joule heating when the droplets were placed on the VO2 film, we recorded the resistance as a function of applied voltage. Fig. 4 depicts the resistance–voltage loops of the three organic liquids. All these loops exhibited thermal hysteresis, which implies that the thermal effect should slightly change the resistance of VO2 film, as labeled in Fig. 4. However, the resistance change caused by the thermal effect is smaller than that induced by the liquids, which lets us distinguish the changes induced by the liquid effect and by the thermal effect. Thus, we can ignore the thermal effect in the relationship between cycling applied voltage and resistance switching.
Resistance as a function of bias voltage. The dotted lines with arrows label the resistance changes induced by thermal effects and the liquids.
Resistance as a function of bias voltage. The dotted lines with arrows label the resistance changes induced by thermal effects and the liquids.
According to the results of Stuart S. P. Parkin,12,16 the changes in the oxidation state of V ions observed by XPS strongly indicated the formation of oxygen vacancies in the thin film with consequent migration of oxygen from the oxide film into the ionic liquid. Similarly, to explain the mechanism behind resistance switching induced by organic polar liquids, we propose the model shown in Figs. 5(a) and 5(b): When cyclohexane coats the film, the V5+ ions capture local electrons and reduce into V4+. This behavior decreases the electron density in the VO2 film, increasing its resistance. This electrochemical process can be described as , as shown in Fig. 5(a).37,38 In contrast, when ethylene glycol coats the film, the V4+ ions lose electrons and oxidize into V5+ ions. This behavior increases the electron density, reducing its resistance, described as .37,39 Because of the weaker reducibility of n-butanol than cyclohexane,40–43 n-butanol modulates the resistance less than cyclohexane does.
Proposed mechanism for resistance switching induced by (a) cyclohexane and (b) ethylene glycol. XPS spectra of V 2p lines with Lorentzian–Gaussian fits for the VO2 films (c) without ionic liquids, (d) with cyclohexane, and (e) with ethylene glycol.
Proposed mechanism for resistance switching induced by (a) cyclohexane and (b) ethylene glycol. XPS spectra of V 2p lines with Lorentzian–Gaussian fits for the VO2 films (c) without ionic liquids, (d) with cyclohexane, and (e) with ethylene glycol.
Note that the surface electronic structure of the VO2 film should be modulated even by removing a polar liquid, as reported by Jeong et al.12 Because of this expectation, we used X-ray photoelectron spectroscopy (XPS) to investigate the valence states of vanadium ions in the VO2 film.38,44–46 Figs. 5(c)–5(e) show the Lorentzian–Gaussian fits of V 2p peaks for the VO2 film without polar liquids, with cyclohexane, and with ethylene glycol, respectively. This VO2 film was stressed by voltage before XPS measurements. We found 1/2 and 3/2 spin-orbit doublet components of the V 2p photoelectrons at 523.6 eV and 515.7 eV, respectively.44,46 Additionally, we found a peak at 517.1 eV corresponding to V5+ ions. Integrating the area of the V5+ peaks from 527 eV to 511 eV, we roughly estimated the V5+ ion concentration to be 37.62% without polar liquids (Fig. 5(c)), 28.15% with cyclohexane (Fig. 5(d)), and 44.17% with ethylene glycol (Fig. 5(e)). These results reveal that cyclohexane decreases the V5+ ion concentration, while ethylene glycol increases it. They also suggest that the VO2 film should be reduced by cyclohexane, increasing the film resistance, and oxidized by ethylene glycol, which decreases the film resistance. Unfortunately, XPS is a surface-sensitive tool that can only detect the surface valence of the VO2 film down to several nanometers, but the VO2 resistance is a bulk behavior. This limitation means these XPS results are only corroborative evidence for our proposed mechanism, so future studies should assess the valence states of the vanadium ions in the bulk of the VO2 film.
IV. CONCLUSION
Using reactive rf-magnetron sputtering, we fabricated high-quality VO2/Al2O3 epitaxial heterostructures. The VO2 film exhibited resistance switching when the VO2 surface was coated with cyclohexane, n-butanol, or ethylene glycol. The resistance modulations in response to these organic liquids were 31%, 3.8%, and −7.7%, respectively. Using XPS, we investigated the mechanism behind the differing resistance changes caused by modulation of the VO2 electronic structure. Our results suggest that this simple two-terminal VO2-based device could be used as an organic molecular sensor.
ACKNOWLEDGMENTS
This work was funded by the National Basic Research Program of China (2012CB922004 and 2010CB934501) and the Natural Science Foundation of China. Yuanjun Yang also acknowledges the partial support from the Fundamental Research Funds for the Central Universities (WK2310000043), the Natural Science Foundation of Anhui Province, and the China Postdoctoral Science Foundation.