Self-Assembled VO2 Mesh Film-Based Resistance Switches with High Transparency and Abrupt ON/OFF Ratio.

Vanadium dioxide, a well-known phase transition material with abrupt resistance change during its transition temperature, is herein used to fabricate the transparent mesh film onto a glass slide through self-assembly mesh printing. A record high ON/OFF ratio up to 104 is achieved together with high visible transmittance of 86% compared to the normal glass slide with visible transmittance at 88%. The high transparent properties make the resistive switches applicable for next-generation electronics, such as see-through computing device and beyond. A simple and scalable mesh printing approach-integrated phase change material may provide a promising way to fabricate transparent resistance switches for next-generation electronics.

Introduction

Resistive switching behavior has been reported among conductive organic materials like poly(N-vinylcarbazole),[1] perovskite oxide like SrTiO3,[2] as well as common binary transition metal oxides (TMO), such as ZnO,[3,4] TiO2,[5] Nb2O5,[6] and VO23. Among all these materials, organic resistance switches suffer from high operating voltage, while most TMO switches such as ZnO, TiO2, and Nb2O5 face the problem of low ON/OFF ratio. Vanadium dioxide (VO2) is a compound that undergoes a metal-to-insulator transition (MIT) at a temperature of ∼68 °C,[7] which gives a high resistance change[8] and makes it one of the most potential materials for the resistive switching application.[9] Besides, the near room temperature MIT and the abrupt change of infrared transmission make VO2 a potential candidate for some novel applications, such as thermochromic smart windows,[10−13] camouflage,[14] and thermal actuator.[15,16]

The resistive switches of VO2 have been classified into nonvolatile and volatile types. Nonvolatile resistive switches can be integrated into memristors,[17] while the volatile ones can be assembled as selectors in memristors or transistors or can be coupled with capacitors or resistors into oscillators, which shows potential applications in artificial neural networks and associative neuromorphic computing.[18] However, their resistive switching behavior in the VO2 device has not been fully exploited, as up to 104 resistance contrast has been observed across the MIT.[9] The earlier VO2 resistance switch was constructed on Si/SiO2 substrate through chemical vapor deposition and the MIT occurred at around 2 V, and the ON/OFF resistance ratio was less than 1 order of magnitude.[19] The recent works on VO2 resistance switches, such as VO2 electrodeposition on Pt substrate[20] and VO2 RF sputtering on SiO/Si substrate[21] or c-sapphire,[22,23] improved the ON/OFF ratio within a certain degree. However, several challenges still need to be tackled for the next generation of electronic devices, such as high transparency as a component for transparent circuits, high ON/OFF ratio, and low operation voltage.[24] Moreover, most of the current works on VO2 resistance switches focused on nonvolatile random access memory.[25] With the development of artificial intelligence and the internet of things, volatile threshold switches, which could be applied in artificial neural networks and associative, are also highly in demand.

In this study, we introduce an easy handling and inexpensive method-mesh printing to fabricate VO2 film on the most widely used glass substrate,[26] which gives the largest ON/OFF ratio, high transparency, and low operation voltage in the application of volatile resistive switches compared with the best-reported results.[27]

Results and Discussion

Transparent VO2 Mesh Film Fabrication and Characterization

Figure a shows the photograph of the as-fabricated VO2 mesh film on a glass slide, which displays the high optical transparency of the mesh film. The detailed morphology of the VO2 mesh film was observed through a scanning electron microscope (SEM), as shown in Figure b. The inset images demonstrate the enlarged SEM image of one knot within the VO2 mesh film and the corresponding energy-dispersive X-ray spectroscopy (EDX) mapping of the vanadium element, which exhibits uniform and continuous distribution of VO2 nanoparticles within the VO2 mesh film. The SEM figures of the VO2 mesh film overlapping with the EDX mapping (Figure S1) is followed by the height profile of the VO2 mesh film structure, measured by a ZEISS Smartproof 5 widefield confocal microscope (Figure S2). Figure c is the visible transmittance spectrum range from 400 to 750 nm. From this, one can see that the continuous VO2 mesh film exhibits 86% of visible transmittance at 550 nm, which is comparable with that of a pure glass slide (around 88%) and is much higher than our previous VO2 mesh work.[26]

Figure 1

Characterization of a transparent VO2 mesh film. (a) Photography of a VO2 mesh sample within the blue circle. (b) SEM image of the VO2 mesh film; the inset is the enlarged SEM of one knot and the corresponding EDX result of the vanadium element. (c) Visible transmittance spectrum of the VO2 mesh film and a bare glass slide.

Figure a shows the transmission electron microscopy (TEM) image of VO2 nanoparticles as well as the selected area electron diffraction (SAED) results, which illustrates the polycrystalline characteristic of the VO2 nanoparticles. As shown in Figure b, the X-ray powder diffraction (XRD) results of the as-fabricated transparent VO2 mesh film are consistent with the starting VO2 nanoparticles and both match well with the standard card of VO2 (PDF card no. 43-1051), which implies that the main functional material of VO2 remains unchanged during the fabrication process for the transparent VO2 mesh film. In Figure c, Raman peaks at 191 and 223 cm–1 indicate a V–V bond while 619 cm–1 peak matches well with the V–O bond of the VO2 M phase, as well as the less intense peaks at 260, 308, 392, and 509 cm–1 are known to be Raman modes of M phase VO2.[28] Infrared transmittance results of the transparent VO2 mesh film are collected at a wavelength between 1000 and 2500 nm before (20 °C) and after (90 °C) the MIT. Figure d shows up to 2% decrease of transmittance at high temperature, which proves that the MIT of VO2 occurs. However, the transmittance difference is much smaller than the typical VO2 film,[29] which is due to the limited amount of VO2 nanoparticles and the testing area.

Figure 2

(a) TEM and SAED results of VO2 nanoparticles, (b) XRD results of VO2 nanoparticles and the VO2 mesh, (c) Raman result of the VO2 mesh film, and (d) infrared transmittance results of wavelengths between 1000 and 2500 nm at 20 °C (blue) and 90 °C (red).

Affecting Factors of the VO2 Mesh Film Fabrication

During the fabrication of the transparent VO2 mesh film, several factors could be tuned to control the continuous and transparency of the mesh film, such as the concentration of VO2 nanoparticles dispersion and the usage of the surfactant. First, the effect of surfactant is studied. It is found that the usage of a small amount of the surfactant could dramatically decrease the surface tension of the VO2 solution, which will be easier to wet the glass slide through the stainless steel mesh. As shown in Figure i, the surface tension data of BYK 348 with a different concentration in deionized (DI) water are collected. The pure DI water has a surface tension up to 70.74 mN m–1, which was decreased to around 22.31, 21.48, and 20.52 mN m–1 when the concentration of BYK increased from 0.01, 0.05, to 0.5 vol %, respectively. Similar trends were also observed with different concentrations of VO2 nanoparticles (Figure ii). Due to the similar surface tension results with 0.01–0.5 vol % surfactant, the smallest weight percentage of 0.01 vol % BYK was selected to mix with VO2 dispersion to obtain the high-quality VO2 mesh film. The dispersion with the surfactant amount below 0.01 vol % has a similar surface tension as DI water. Furthermore, the effect of different concentrations of the VO2 dispersion is studied. To avoid the precipitation of VO2 nanoparticles, up to 3 wt % VO2 nanoparticles were dispersed into the DI water. Surface tension results of different concentrations of VO2 solution with and without 0.01% BYK surfactant are also shown in Figure ii.

Figure 3

Surface tension vs (i) surfactant concentration and (ii) VO2 nanoparticles concentration. Optical microscopy images for the VO2 mesh with different mixture percentages, inset with an enlarged mesh structure (iii) 1 wt % VO2, 0.01 vol % BYK; (iv) 2 wt % VO2, 0.01 vol % BYK; (v) 3 wt % VO2, 0.01 vol % BYK; and (vi) 3 wt % VO2, no BYK.

Figure iii,iv,v shows the corresponding VO2 mesh films on a glass substrate with different concentrations of VO2 nanoparticles and the same amount of wetting dispersant (0.01 vol % BYK). With the lower loading amount of VO2, large numbers of voids due to lack of nanoparticles can be observed when averaged distances of broken lines are 32.5 μm (Figure iii) and 15.1 μm (Figure iv) approximately. With the increment of VO2 loading, such voids diminished (Figure iii). It is illustrated that enough nanoparticles need to be provided to form a complete mesh film. On the other hand, at the same concentration of VO2 nanoparticles (3 wt %), certain defects of blank and extra lines can always be observed in the sample without using a surfactant (Figure vi), which is due to the relatively higher surface tension. Therefore, 3 wt % of VO2 concentration with 0.01 vol % BYK surfactant gives the best quality of the VO2 mesh film (Figure v).

Transparent VO2 Mesh Resistance Switches’ Performance

Figure a shows the schematic of the transparent VO2 mesh resistance switch (fabrication details are shown in the Section ). The Au/Ti electrodes are aligned with the VO2 mesh line for the resistance switching (RS) measurement. For comparison, the Au/Ti electrode deposited on a bare glass substrate with the same pattern is shown in Figure b. To study the RS behavior of VO2, we tested the I–V curves under the voltage sweep, as shown in Figure c. The device with an initial high-resistance state (HRS) changes to a low-resistance state (LRS) when applying the threshold voltage (Vth). This model correlated the coupling effect mechanism of electrical and joule heating when applying voltage on the two-terminal RS device. It is worth noting that the Vth for this VO2 mesh RS is around 1.5 V, which is lower than the previously reported value of around 6 V at room temperature.[30] The switching time for the VO2 mesh switches is less than 50 μs (Figure S3) due to the machine testing limitation, which is much less than those of some reported VO2 optical switches (around micrometers range).[8,31] However, none of the reported VO2 switches are transparent. As illustrated in Figure d, a durability test up to 50 cycles was applied onto the VO2 mesh RS. Within the testing period, the ratio of high- and low-resistance state was roughly maintained at around 104. However, the switching ratio for VO2 mesh switches starts to drop after about 40 cycles, and the XRD examination (Figure S4) shows that the peak intensity of VO2 is diminished after 50 cycles. This could be due to the long-standing durability issue of the VO2, which currently involves intensive research and is out of the scope of this current research.[32,33] More future work could be conducted by the encapsulation of this RS to solve this issue. The visible transmission spectrum was measured by switching ON the threshold voltage, which showed that the VO2 resistance switches remain highly transparent during the switching process (Figure S5).

Figure 4

Characterization of the VO2 mesh resistance switch. Optical images of (a) Au/Ti electrode deposited on the VO2 mesh, (b) bare glass substrate, (c) I–V characteristic of the VO2 mesh resistance switch and the bare glass slide, and (d) retention result of the VO2 resistance switch.

As mentioned in the Section , RS can be divided into two modes, nonvolatile memory switching and volatile threshold switching.[34,35] Here, the I–V characteristics clearly demonstrate that the RS behavior is independent of the voltage polarity and that there is only one stable resistance state at zero voltage, which is corresponding to a typical unipolar volatile threshold switching characteristic. The ON/OFF ratio for HRS and LRS in the transparent VO2 resistance is more than 104, which is the highest result among the transparent resistance switching devices in recent years, as summarized in Figure . Resistance switches made from functional materials including ZnO,[36] ZnSrO3,[37,38] and SrTiO3[39] could exhibit around 80–90% visible transmission. However, the ON/OFF ratio of those resistance switches are no more than 100. For the volatile threshold resistive switches applied in selectors, artificial neuron networks, and associative computing, higher ON/OFF ratio is required. The VO2-resistance change could reach 104 due to its own intrinsic phase transition property. Above all, the VO2 mesh shows high potential for the transparent resistive switching application, due to its high ON/OFF ratio and low-cost fabrication.

Figure 5

Comparison between the result of our work and recently reported resistance switches based on poly-4-vinylphenol (PVP),[40] ZnSnO3,[37,38] WO,[41] SnO2/ZnO,[36] SrTiO3,[39] TiO2,[42] BTCO,[43] and Fe/SnO2.[27]

Conclusions

In this work, we reported a simple and low-cost method to fabricate a transparent VO2 resistance switch. Vanadium dioxide nanoparticles (VO2 NPs) with an average diameter less than 100 nm were screen printed onto a common glass slide by a stainless steel mesh to fabricate the transparent resistance switch device through photolithography and e-beam evaporation process. A volatile resistance switch behavior, as well as the highest ON/OFF ratio up to 10 000 with a low operation voltage, was achieved in the VO2 mesh film on the normal glass slide with the high visible transmission of around 86%. The facile fabrication method and novel micropatterned structures of the transparent mesh VO2 resistance switching devices give a new promising direction for next-generation electronic devices.

Experimental Section

Materials

Vanadium dioxide nanoparticles (VO2, weight-average molecular weight (Mw) = 82.94), purchased from Wenzhou Jingcheng Chemicals and DisperBYK 348 without further purification. Deionized water (18.2 MΩ) was used throughout the experiments.

Fabrication of the Transparent VO2 Mesh Film

The schematic procedures of fabricating the transparent VO2 self-assembly mesh film onto a glass slide are shown in Scheme . The glass slide with a thickness of about 1 mm is as purchased from Sail Brand 7107 Microscope Slides with further washing by DI water and ethanol. The screen-printing mesh (Ponger 2000, Israel) was mounted directly above the clean glass slide. By using a 100 μL pipette, a certain amount of VO2 dispersion was dropped onto the glass slide substrate through the mesh. The VO2 dispersion was made through 10 min sonication of a certain weight percentage of VO2 nanoparticles, DI water, and surfactant BYK 348. A sonicator (Elmasonic S60H, Elma Ultrasonic) with the ultrasonic frequency of 37 kHz with 550 W power consumption was used. When the nanoparticles dispersion was dropped onto the top of the mesh, upon contact, the liquid immediately wetted the mesh walls and space between the mesh and glass slide, and then, with the evaporation of the DI water, the VO2 nanoparticles self-assembled toward the edge of the mesh walls due to the caterpillar force. Upon removing the mesh, the VO2 mesh film was heated at 60 °C for 2 h in vacuum for water evaporation. At last, the final transparent VO2 mesh film was obtained.

Scheme 1

Schematic Diagram for the Fabrication of Self-Assembly Transparent VO2 Mesh Film

Fabrication of VO2 Resistance Switches

As-fabricated transparent VO2 mesh film on the glass substrate was spin-coated (Headway Research Spin Coater) with a layer of photoresist at 3000 rpm for 30 s (AZ5214E, photoresist image reversal, MicroChemicals GmbH) and then prebaked at 105 °C. The alignment was adjusted by microscopy to make sure that the VO2 nanoparticles branches were located between the source and the drain. The electrode patterns were subsequently transferred from the photolithography template to the VO2 mesh film by exposing to UV light for 4 s at ∼44 mW cm–2 (SUSS MicroTec, MJB4) and developed for 30 s (AZ Developer: DI H2O = 1:1, AZ Electronic Materials GmbH). Then, Ti/Au (30/150 nm) films were deposited by electron beam evaporation (Edwards Auto 306). The microelectrodes were finally formed by a lift-off process.

Materials and Device Characterization

Structural analysis of as-purchased and as-fabricated VO2 nanoparticles and the VO2 mesh film were determined by Shimazu X-ray diffractometer with Cu Kα radiation (λ = 0.1541 nm). The morphology of VO2 nanoparticles, as well as the SAED (selected area electron diffraction), were characterized using a transmission electron microscope (JEM-2010, JEOL, Japan) with an accelerating voltage of 200 kV. The microstructures of the VO2 mesh film were observed by the optical and the field emission scanning electron microscopes (JEM-7600F, JEOL, Japan). The element analysis mapping of the VO2 mesh was also characterized by energy-dispersive X-ray spectroscopy (EDX) attached with JEM-7600F. The visible transmittance of transparent resistance switches and the infrared transmittance for VO2 (M) confirmation were carried out using Agilent Cary 5000 UV–vis–NIR spectrometer. A heating stage (PE120, Linkam, U.K.) was employed to control the temperature of the samples mounted on the spectrophotometer. The performance of the VO2 mesh resistance switch was determined under room temperature and atmospheric pressure. The I–V characteristics of the device were measured by Agilent B1500A digital source meter connected with triaxial cables to the probe station.