Direct Observation of Conducting Filaments in Tungsten Oxide Based

Oct 5, 2016 - Transparent nonvolatile memory has great potential in integrated transparent electronics. ... Citation data is made available by partici...
1 downloads 11 Views 4MB Size
Research Article www.acsami.org

Direct Observation of Conducting Filaments in Tungsten Oxide Based Transparent Resistive Switching Memory Kai Qian,† Guofa Cai,† Viet Cuong Nguyen,† Tupei Chen,‡ and Pooi See Lee*,† †

School of Materials Science and Engineering, and ‡School of Electrical and Electronic Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore, 639798 S Supporting Information *

ABSTRACT: Transparent nonvolatile memory has great potential in integrated transparent electronics. Here, we present highly transparent resistive switching memory using stoichiometric WO3 film produced by cathodic electrodeposition with indium tin oxide electrodes. The memory device demonstrates good optical transmittance, excellent operative uniformity, low operating voltages (+0.25 V/−0.42 V), and long retention time (>104 s). Conductive atomic force microscopy, ex situ transmission electron microscopy, and X-ray photoelectron spectroscopy experiments directly confirm that the resistive switching effects occur due to the electric field-induced formation and annihilation of the tungsten-rich conductive channel between two electrodes. Information on the physical and chemical nature of conductive filaments offers insightful design strategies for resistive switching memories with excellent performances. Moreover, we demonstrate the promising applicability of the cathodic electrodeposition method for future resistive memory devices. KEYWORDS: transparent resistive memory, tungsten oxide, CAFM, XPS, ex situ TEM



hinder multilevel cell storage22 and result in high energy consumption. As one of the most promising switching materials for future memristors, tungsten oxides have been intensively studied due to their potential advantages (e.g., high compatibility with the back-end-line process in the CMOS (complementary metal oxide semiconductors) technology, easy fabrication, and thermal stability).23−28 Although several kinds of mechanisms have been proposed to study the resistive switching behaviors, there is still a lack of direct evidence for the switching mechanism of the tungsten oxide based memory device. Therefore, to achieve high optical transparency, the tradeoffs on excellent electronic performance and low switching voltages are still an impending challenge. In order to achieve these superiorities in a practical device, a thorough understanding of the resistive switching mechanism is indispensable to gain insights into the governing physics and continued optimization of memristor devices, especially on the atomic level. In this work, the highly transparent tungsten oxide (WO3) thin film was prepared by cathodic electrodeposition, which presents several extraordinary features, such as time-effective process, low production cost, and large-area fabrication.29,30 The ITO/WO3/ITO glass resistive memory devices featured

INTRODUCTION With the significant impact of transparent electronics and the recent advances in nanotechnology, the development of transparent circuits has attracted tremendous amount of attention in various areas (e.g., emerging consumer electronics and defense applications).1−3 Moreover, great effort has been made to construct different transparent electronics, such as transistors,4,5 artificial skins,6 sensor implants,7 and solar cells.8 The memory unit, which is an indispensable element in a myriad of electronic products, also requires high transparency for integrated transparent circuit realization. Due to the tradeoff between the efficient charge trapping in the floating gate and high transparency, there is a sacrifice in nonvolatility aspect for the charge-based transparent memory with a transistor structure.9 In this case, the transparent nonvolatile resistive memories, which possess the advantages of both invisible electronics and resistive switching memories (RSM),10−15 have aroused extensive interest.9,16−19 For the realization of a high transparency memory device, wide band gap (band gap > 3.1 eV) resistive switching materials are always necessary.1,9 Therefore, the wide band gap metal oxides, silicon oxides, and nitrides were generally investigated to fabricate transparent memristors, which were sandwiched between transparent conductive electrodes.9,16,17,20,21 However, the fluctuation in the aspect of set/reset voltages and ON/OFF ratio20 and high switching voltages9,16,17,19,21 were still an obstacle for transparent RSM development, which would © XXXX American Chemical Society

Received: July 5, 2016 Accepted: September 28, 2016

A

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. (a) Schematic of the ITO/WO3/ITO device arrays for electrical characterization. (b) The optical image of the ITO/∼50 nm WO3/ITOlayered structure on a glass substrate. Scale bar, 5 mm. (c) Optical transmittance in ITO/WO3/ITO/glass devices with different WO3 thin film thicknesses. The left inset is the cross-sectional SEM image of ∼50 nm WO3 film, corresponding to the curve 1. The right inset is the cross-sectional SEM image of ∼200 nm WO3 film, corresponding to the curve 2. (d) I−V characteristics of the ∼50 nm WO3-based memory cell.

Figure 2. (a) I−V characteristics of 320 consecutive voltage sweep tests from an ITO/WO3/ITO cell. (b) Statistical distribution/cumulative probability of the switching voltage for the ITO/WO3/ITO device. (c) Enduring test and (d) retention time of the memory device under a readout voltage of +0.1 V for both an ON and an OFF state.



RESULTS AND DISCUSSION ITO was chosen as the both top and bottom electrodes material, which represents the mainstream choice for transparent conductive material. Vertical sandwiched pillar structures ITO/WO3 (∼50 nm thick)/ITO were prepared on a glass substrate (Figure 1a). After annealing at 300 °C in the ambient atmosphere, the transparent ITO/WO3/ITO device was fabricated (Figure 1b), which showed ∼80% transparency in the visible region and the highest transparency of ∼86% at 482 nm (Figure 1c, curve 1). As shown in Figure S1a (Supporting Information), the WO3 switching layer on the ITO substrate

high transmittance (∼80%) over the visible region, low programming voltages (+0.25 V/−0.42 V), long retention time (>2 × 104 s), and highly uniform switching cycling. The electric field-induced formation and annihilation of tungstenrich conductive filaments in the WO3 switching layer are found to account for the stable resistive switching of the transparent WO3-based memristors, based on the elucidation of the resistive switching behavior via conductive atomic force microscopy (CAFM), ex situ transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS). B

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 3. CAFM mapping images (8 μm × 10 μm) of the WO3-based memory at (a) OFF and (b) ON states. The reading voltage was −0.5 V. W 4f XPS spectra of the WO3 film in the (c) initial and (d) ON states.

was flat and uniform with about ∼50 nm thick (Figure 1c, left inset). As shown in Figure 1c (curve 2), compared with the ∼50 nm thick WO3 thin film, the transparency of the ITO/ WO3 (∼200 nm)/ITO device (optical image shown in Figure S1b) decreased a little (∼75% transmittance), suggesting that the optical transparency of the WO3-based memory device can be further enhanced by employing a thinner thickness of WO3 film. To investigate the memory device switching behaviors, direct current (DC) electrical bias was applied on the ITO top electrode (TE), while the bottom electrode (BE) was grounded. In order to avoid the breakdown of the memory device, a compliance current (CC) of 0.1 mA was applied during the electroforming process for repeatable voltage sweeping. Figure S1e shows the representative initial I−V curve for the ITO/WO3/ITO device; the Vforming is about ∼3.1 V where the current increased abruptly. For the highly transparent ITO/WO3/ITO memory device, the memory cell was shifted from the high resistance (OFF) state to the low resistance (ON) state at ∼0.25 V, which is defined as the “SET” process (Figure 1d). Then, a negative voltage switched the memory cell back to the OFF state. During the RESET process, there are two intermediate states at ∼ −0.36 V and ∼ −0.42 V where the resistive switching occurs, respectively. It is noteworthy that the multistep RESET processes which are frequently observed (Figure 1d) might result from the formation and annihilation of multifilaments with different threshold voltages (Vth).31 Notably, the WO3-based memory device fabricated via cathodic electrodeposition has lower set/ reset voltage than most of the reported tungsten oxides memory devices (Table S1, Supporting Information), indicating that the cathodic electrodeposition method can be a promising technique for memristor fabrications. In addition, the ITO/WO3/ITO memory device also showed the same switching characteristics with Figure 1d when reversed polarity voltage was applied (Figure S2 in the Supporting Information) due to its symmetric electrode structure. To survey the uniformity of the switching behaviors, cyclic programming operations of the highly transparent WO3-based

memory devices were conducted. As shown in Figure 2a, the resistive switching operation can be repeated uniformly, and there was no obvious degradation in the 320 consecutive cycles. Due to the multistep RESET processes, the voltage where the second resistive switching occurs is defined as V reset corresponding to reset 2 in Figure 1d, which induces the sharpest change in the device current during the RESET processes. A narrow distribution of Vset (0.3 ± 0.08 V) and Vreset (−0.39 ± 0.05 V) is shown in Figure 2b, indicating the uniform distribution of the set and reset voltages. In addition, both the ON and the OFF states (read at 0.1 V) were highly stable during the repeated cycles (Figure 2c), indicating the reliable and uniform resistive switching operation. Meanwhile, it is found that the ITO/WO3/ITO memory devices also showed unipolar switching with high RESET current. However, for the unipolar operation mode, the large fluctuations of switching voltages are an obstacle for the memristor applications (Figure S3, Supporting Information). Figure 2d shows the retention capabilities of the ON and OFF states at room temperature, which featured high stability without significant degradation over 2 × 104 s. These memory characteristics confirmed the highly transparent and uniform ITO/WO3/ITO resistive switching memory under bipolar switching mode. The formation and annihilation of the localized conductive filament path in the oxide is widely accepted to be responsible for the bipolar resistive switching in crystalline memory devices, which is related to the redox reaction of defects present in the switching materials.16,32 For this valence-change based switching mechanism, the oxygen anions (O2−) will drift across the switching layer to the electrodes under high electric field, leading to a possible new phase with a different valence state of cations and conducting filament. Then, the resistive memory will be switched between ON and OFF states due to the formation and annihilation of oxygen deficient conducting filaments.33 To investigate the bipolar switching mechanism, C-AFM with a diamond coated tip electrode was employed to observe the ON and OFF states of the WO3/ITO glass structure C

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (Figure 3a,b). Chemical composition of the switching layer has been investigated by XPS analysis of the WO3/ITO BE cells in their respective pristine state and ON state (Figure 3c,d). Compared with the OFF state (Figure 3a), it is concluded that filamentary conduction accounts for the ON state (Figure 3b) in the WO3-based memory device. With electric field increased, the metallic tungsten-rich conducting filaments formation (Figure 3b) due to the migration of oxygen anions can be expected. At ON state, as shown in Figure 3b, current will preferentially flow from the localized conductive filaments. When switched to OFF state, the current distributed uniformly over the entire device area (Figure 3a). The deconvoluted W 4f peaks of the pristine WO3 memory cell is shown in Figure 3c, which were composed of W 4f7/2 (35.8 eV), W 4f5/2 (37.8 eV), and W 5p3/2 (41.3 eV) peaks and well consistent with previous reported results for WO3.29,34 It is concluded that W is at the highest +VI oxidation state in the films. Here, to avoid the valence change of tungsten ions caused by the introduction of argon-ion etching treatment,34 for the ON state of the WO3 memory cell, the XPS measurement was carried out on the surface of WO3/ITO glass instead of depth-profiling analysis. It is anticipated that the chemical composition change of WO3 can be detected due to that the maximum analysis depth of XPS can be up to 10 nm. Before preparing the ON state sample for XPS observation, the 10 μm × 10 μm memory device was switched to the ON state via CAFM through 10 repetitive voltage sweeps. Figure 3d shows the multiple W−O bonding states of the ON state memory cell, i.e., W6+ (35.5 eV), W5+ (34.3 eV), W4+ (33.3 eV), and Wx+ (32.1 eV) oxidation states,34 indicating oxygen vacancies existence due to the oxygen ions migration under an electric field. Therefore, it is reasonably concluded that metallic tungsten-rich conductive filaments formation under an electric field is responsible for bistable resistive switching, which also corresponds to the results of the CAFM image (Figure 3b) in the ON state. In addition, we would like to state briefly that the switching behaviors of the ITO/WO3/ITO device are different from the switching characteristics in the Pt/WO3−x/Pt memristor, where the device showed rectification bipolar RS properties.23 In that case, the rectification property resulted from the Schottky-like barrier at the Pt TE/WO3−x interface. On the other hand, the dependence of the electroforming voltage on the switching layer thickness confirmed the bulk nature of the switching mechanism in the ITO/WO3/ITO device (Figure S4, Supporting Information), which excluded a possible interfacial switching mechanism. To obtain superb performances of the memory devices, a thorough understanding of the switching mechanism is indispensable to effectively control the switching behaviors.10,35 Herein, to further investigate the bipolar switching mechanism, atomic-resolution TEM was conducted to provide information on the physical and chemical changes inside WO3 thin film in their initial resistance, low resistance, and high resistance states, respectively. Here, all the TEM specimens were obtained from the same batch of memory device and prepared via focus ion beam (FIB). As shown in Figure 4a and a1, the fresh memory samples contain lots of small dark spots with the diameter of 1−2 nm which may result from the reduction of WO3 during the FIB process. As shown in Figure 4a and a1, the darkcontrasted dots should be associated with a tungsten-rich composition which is heavier than that of the WO3 matrix.36 Unlike quantum dots in WO3−x thin film,37 there was no lattice in these black dots (Figure 4a1). In addition, for the fresh

Figure 4. TEM images of transparent WO3-based memory device samples in their initial resistance state (a, a1), ON state (b, b1), and OFF state (c, c1), respectively. The red rectangular area indicated there were no big black dots near the top electrode.

sample without ITO TE (Figure S5a, Supporting Information), there were also lots of black dots inside the WO3 film, indicating that the ITO TE is not involved in the generation of these dark-contrasted grains. For the ON state TEM specimen, the memory device was switched to the ON sate via 10 repetitive voltage sweeps. As shown in Figure 4b and b1, it is noteworthy that there are some bigger black-contrasted grains with the diameter of 4−5 nm across the WO3 thin film, which are different from the fresh sample (Figure 4a and a1). It is reasonably expected that the existence of these bigger black dots is the result of the reduction of WO3 from the high oxidation state to the low state with oxygen ions migration under an electric field. Thereafter, the tungsten-rich WO3−x conductive filaments (CFs) in the switching layer (Figure 4b) will be formed, corresponding to the XPS analysis results (Figure 3d). For the OFF-state TEM specimen, it was switched to the OFF state via 10 repetitive voltage sweeps with negative voltage on the TE. Compared to the ON state memory cell (Figure 4b and b1), it is worth to note that there were no big black-contrasted grains beneath the TE (red rectangular area); however, lots of black dots with the diameter of 4−5 nm still existed near the BE (Figure 4c and c1). To further investigate the switching mechanism, EDS analysis was employed to characterize the chemical composition at different positions in the OFF-state memory cell (positions 1 and 2 in Figure S5b, D

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

demonstrated, which show high transmittance over the visible region, low programming voltages (+0.25 V/−0.42 V), excellent uniformity of operation parameters, and long retention time. With the use of CAFM, XPS, and high spatial resolution TEM analysis, we have directly confirmed that the underlying mechanism results from the oxygen ions migration under an electric field, leading to formation/annihilation of the tungsten-rich conducting filament. This experiment presents fresh insights into the chemical and physical nature of resistive switching behaviors for metal oxide memristors, which may serve as a guide to improve the performance and provide general guidance for novel memory devices designs.

respectively). Although the nature of EDS analysis is semiquantitative, the W atomic concentration near the top electrode (position 1, 40.2 at %) is clearly lower than the one near the bottom electrode (position 2, 64.5 at %, Figure S5c,d, Supporting Information). The decline in tungsten concentration near the top electrode indicates that the oxidation reaction occurred near the TE area, leading to the OFF state with oxidized filaments. Application of reverse bias on TE causes the oxygen ions migration from the anode interface to oxidize the tungsten-rich WO3−x filaments (i.e., oxygen vacancies),33 leading to the rupture of CFs and OFF state (Figure 4c and c1). Hence, the ex situ TEM experiments directly demonstrated that the switching behavior of the WO3based memory device was associated with the conductive filaments formation/annihilation due to the oxygen ions migration under an electric field. In addition, for the WO3 film prepared via cathodic electrodeposition, it is composed of WO3 nanoparticles (Figure 1c, Figure S1a), leading to lots of grain boundaries or other defects in the switching layer. Therefore, oxygen ions may drift and diffuse easily along the grain boundaries or other defective sites to form the large dimension tungsten-rich filaments,38−40 leading to a lower SET/RESET voltage. On the basis of the above-mentioned experimental results, schematic diagrams of the bipolar switching WO3-based memory device are shown in Figure 5 to illustrate the RS



METHODS

Device fabrication. All chemicals were used as received and analytical-reagent grade. Sodium tungstate dehydrate (Na2WO4·H2O), hydrogen peroxide (H2O2, 30%), and perchloric acid (HClO4, 70%) were purchased from Sigma-Aldrich. All solutions were prepared using deionized water (Milli-Q, Millipore Corp.). The WO3 thin film was prepared via cathodic electrodeposition using Na2WO4 aqueous deposition solution (concentration: 12.5 mM) with H2O2, where the concentration ratio of H2O2 and sodium tungstate is 3.29,30 Perchloric acid was added to the mixed solution to tune the pH value to 1.2. A Ag/AgCl electrode and a 2 × 2 cm2 platinum foil were used as reference electrode and counter electrode, respectively. A conventional three-electrode system (Solartron 1470 E electrochemical workshop) was employed to prepare the WO3 film with a potential of −0.7 V (vs Ag/AgCl) for 60 s at room temperature. Finally, the sample was thoroughly cleaned with DI water and dried in air. Then, we can obtain the transparent WO3 thin film. Thereafter, the ITO top electrode (TE) was prepared via electron beam evaporation (Edwards Auto 306 Turbo) with a shadow mask. Because the as-deposited ITO TE is not transparent (Figure S1c, Supporting Information), then the device was annealed at 300 °C in the ambient atmosphere to get the transparent ITO top electrode. After annealing at 300 °C in the ambient atmosphere, the amorphous WO3 thin film transformed into polycrystalline film (Figure S1d, Supporting Information). Characterization and Device Measurement. The structure and morphological characterizations of WO3 film were conducted using Xray thin film diffraction (XRD, Shimadzu XRD-600, Cu Kα radiation) at a scan rate of 2° min−1, and field emission scanning electron microscopy (FESEM, JEOL, JSM-7600F). The transmission spectra of the device were carried out in the 300−900 nm visible region with a UV−vis−NIR spectrometer (Cary 5000 and DRA 2500, Agilent Technologies). Energy-dispersive spectrometer (EDS) and TEM analyses were conducted with an FEI Tecnai F20 electron microscope and Philips CM 200 (200 kV) and EDAX, respectively. The crosssectional TEM samples of memory cells were cut using a focused ion beam (FIB, FEI Helios 600i) with a Ga ion beam at 30 kV beam energy, followed by thinning (to ∼70 nm) and cleaning at 8 kV beam energy. A Pt (∼100 nm thick) layer was deposited onto the memory device to avoid ions radiation damage. To survey the WO3 film composition and oxidative states in the initial and ON states, the XPS measurements were performed (PHI Quantera II). Monochromatic Al Ka X-ray (hν = 1486.6 eV) was conducted with a large area lens mode and photoelectron takeoff angle of 90° versus surface plane. The analysis depth can be up to 2−10 nm, and the smallest spot size is about 10 μm in diameter. A conducting atomic force microscopy (NTMDT) with a diamond coated tip was used for CAFM tests via contact mode. In order to obtain the ON-state memory cell, voltage sweeping was applied at the CAFM tip top electrode, while the bottom electrode was grounded. First, a relatively high electroforming voltage of 15 V was applied on the CAFM tip. After the electroforming process, 10 V SET voltage was applied at the CAFM tip to set the memory cell to the ON state. In order to switch the memory cell to the OFF state, −10 V RESET voltage was applied at the CAFM tip again. A Keithley 4200 semiconductor analyzer was employed to obtain the current−

Figure 5. Schematic illustrations of the switching mechanism of the ITO/WO3/ITO device. Application of a forward/reverse bias causes the oxygen ions migration to the anode/cathode, opposite to the oxygen vacancies direction. This leads to the formation/annihilation of tungsten-rich filament in the memory cell.

mechanism. First, during the electroforming process, the oxygen ions migrated to the ITO TE/WO3 interface under high electric fields, where these ions were discharged as neutral oxygen atom, leading to the soft dielectric breakdown (i). Thus, the anode/WO3 interface serves as an “oxygen reservoir”.41 Then, oxygen ions migration causes the WO3−x conductive filament growth from the ITO BE to ITO TE, switching the device to the ON state (ii). Consequently, the current flows through the WO3−x filament in the WO3 thin film. In the RESET operation, oxygen ions at the anode interface will migrate back to recombine with the oxygen vacancies in the WO3 switching layer with the aid of the reverse electric field, leading to the filament rupture and OFF state (iii).



CONCLUSION In this work, we have produced highly transparent WO3 film by cathodic electrodeposition, an easy, time-effective, and largearea fabrication technique under an ambient atmosphere. Fully transparent ITO/WO3/ITO resistive memory devices were E

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces voltage (I−V) characteristics at a sweeping rate of 0.01 V/s in an ambient atmosphere and dark room.



(10) Yang, J. J. S.; Strukov, D. B.; Stewart, D. R. Memristive Devices for Computing. Nat. Nanotechnol. 2013, 8, 13−24. (11) Waser, R.; Dittmann, R.; Staikov, G.; Szot, K. Redox-Based Resistive Switching Memories - Nanoionic Mechanisms, Prospects, and Challenges. Adv. Mater. 2009, 21, 2632−2663. (12) Qian, K.; Nguyen, V. C.; Chen, T.; Lee, P. S. Amorphous-SiBased Resistive Switching Memories with Highly Reduced Electroforming Voltage and Enlarged Memory Window. Adv. Electron. Mater. 2016, 2, 1500370. (13) Qian, K.; Tay, R. Y.; Nguyen, V. C.; Wang, J.; Cai, G.; Chen, T.; Teo, E. H. T.; Lee, P. S. Hexagonal Boron Nitride Thin Film for Flexible Resistive Memory Applications. Adv. Funct. Mater. 2016, 26, 2176−2184. (14) Chen, G.; Song, C.; Chen, C.; Gao, S.; Zeng, F.; Pan, F. Resistive Switching and Magnetic Modulation in Cobalt-Doped ZnO. Adv. Mater. 2012, 24, 3515−3520. (15) Kudo, M.; Arita, M.; Ohno, Y.; Takahashi, Y. Filament Formation and Erasure in Molybdenum Oxide during Resistive Switching Cycles. Appl. Phys. Lett. 2014, 105, 173504. (16) Shang, J.; Liu, G.; Yang, H. L.; Zhu, X. J.; Chen, X. X.; Tan, H. W.; Hu, B. L.; Pan, L.; Xue, W. H.; Li, R. W. Thermally Stable Transparent Resistive Random Access Memory based on All-Oxide Heterostructures. Adv. Funct. Mater. 2014, 24, 2171−2179. (17) Kim, H. D.; An, H. M.; Seo, Y.; Kim, T. G. Transparent Resistive Switching Memory Using ITO/AlN/ITO Capacitors. IEEE Electron Device Lett. 2011, 32, 1125−1127. (18) Seo, J. W.; Park, J. W.; Lim, K. S.; Yang, J. H.; Kang, S. J. Transparent Resistive Random Access Memory and its Characteristics for Nonvolatile Resistive Switching. Appl. Phys. Lett. 2008, 93, 223505. (19) Yang, P. K.; Ho, C. H.; Lien, D. H.; Duran Retamal, J. R.; Kang, C. F.; Chen, K. M.; Huang, T. H.; Yu, Y. C.; Wu, C. I.; He, J. H. A Fully Transparent Resistive Memory for Harsh Environments. Sci. Rep. 2015, 5, 15087. (20) Liu, K. C.; Tzeng, W. H.; Chang, K. M.; Chan, Y. C.; Kuo, C. C. Bipolar Resistive Switching Effect in Gd2O3 Films for Transparent Memory Application. Microelectron. Eng. 2011, 88, 1586−1589. (21) Zheng, K.; Sun, X. W.; Zhao, J. L.; Wang, Y.; Yu, H. Y.; Demir, H. V.; Teo, K. L. An Indium-Free Transparent Resistive Switching Random Access Memory. IEEE Electron Device Lett. 2011, 32, 797− 799. (22) You, B. K.; Park, W. I.; Kim, J. M.; Park, K. I.; Seo, H. K.; Lee, J. Y.; Jung, Y. S.; Lee, K. J. Reliable Control of Filament Formation in Resistive Memories by Self-Assembled Nanoinsulators Derived from a Block Copolymer. ACS Nano 2014, 8, 9492−9502. (23) Yang, R.; Terabe, K.; Liu, G. Q.; Tsuruoka, T.; Hasegawa, T.; Gimzewski, J. K.; Aono, M. On-Demand Nanodevice with Electrical and Neuromorphic Multifunction Realized by Local Ion Migration. ACS Nano 2012, 6, 9515−9521. (24) Lee, S.; Lee, J.; Park, J.; Choi, Y.; Yong, K. Resistive Switching WOx-Au Core-Shell Nanowires with Unexpected Nonwetting Stability Even when Submerged Under Water. Adv. Mater. 2012, 24, 2418− 2423. (25) Hong, D. S.; Chen, Y. S.; Li, Y.; Yang, H. W.; Wei, L. L.; Shen, B. G.; Sun, J. R. Evolution of Conduction Channel and its Effect on Resistance Switching for Au-WO3‑x-Au Devices. Sci. Rep. 2014, 4, 4508. (26) Liang, L.; Li, K.; Xiao, C.; Fan, S. J.; Liu, J.; Zhang, W. S.; Xu, W. H.; Tong, W.; Liao, J. Y.; Zhou, Y. Y.; Ye, B. J.; Xie, Y. Vacancy Associates-Rich Ultrathin Nanosheets for High Performance and Flexible Nonvolatile Memory Device. J. Am. Chem. Soc. 2015, 137, 3102−3108. (27) Chien, W. C.; Chen, Y. C.; Lai, E. K.; Yao, Y. D.; Lin, P.; Horng, S. F.; Gong, J.; Chou, T. H.; Lin, H. M.; Chang, M. N.; Shih, Y. H.; Hsieh, K. Y.; Liu, R.; Lu, C.-Y. Unipolar Switching Behaviors of RTO WOX RRAM. IEEE Electron Device Lett. 2010, 31, 126−128. (28) Tsai, T. L.; Lin, Y. H.; Tseng, T. Y. Resistive Switching Characteristics of WO3/ZrO2 Structure With Forming-Free, SelfCompliance, and Submicroampere Current Operation. IEEE Electron Device Lett. 2015, 36, 675−677.

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b08154. XRD, SEM, and optical images for WO3 film; I−V curve during the initial electroforming process for ITO/WO3/ ITO device; performance comparison in terms of set/ reset voltages; I−V curves for ITO/WO3/ITO device under unipolar switching mode; electroforming voltage as a function of WO3 thickness; and cross-sectional TEM images and XRD analysis for WO3-based memory device (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work is supported by the NTU-A*Star Silicon Technologies Centre of Excellence under the program grant no. 112 3510 0003 and the National Research Foundation Competitive Research Programme NRF-CRP13-2014-02. K.Q. acknowledges the scholarship awarded by Nanyang Technological University, Singapore.



REFERENCES

(1) Thomas, G. Materials Science - Invisible Circuits. Nature 1997, 389, 907−908. (2) Wager, J. F. Transparent Electronics. Science 2003, 300, 1245− 1246. (3) Wang, L.; Yoon, M. H.; Lu, G.; Yang, Y.; Facchetti, A.; Marks, T. J. High-Performance Transparent Inorganic-organic Hybrid Thin-Film N-type Transistors. Nat. Mater. 2006, 5, 893−900. (4) Nomura, K.; Ohta, H.; Takagi, A.; Kamiya, T.; Hirano, M.; Hosono, H. Room-Temperature Fabrication of Transparent Flexible Thin-Film Transistors Using Amorphous Oxide Semiconductors. Nature 2004, 432, 488−492. (5) Ju, S. Y.; Facchetti, A.; Xuan, Y.; Liu, J.; Ishikawa, F.; Ye, P. D.; Zhou, C. W.; Marks, T. J.; Janes, D. B. Fabrication of Fully Transparent Nanowire Transistors for Transparent and Flexible Electronics. Nat. Nanotechnol. 2007, 2, 378−384. (6) Lipomi, D. J.; Vosgueritchian, M.; Tee, B. C. K.; Hellstrom, S. L.; Lee, J. A.; Fox, C. H.; Bao, Z. N. Skin-like Pressure and Strain Sensors Based on Transparent Elastic Films of Carbon Nanotubes. Nat. Nanotechnol. 2011, 6, 788−792. (7) Fan, F. R.; Lin, L.; Zhu, G.; Wu, W. Z.; Zhang, R.; Wang, Z. L. Transparent Triboelectric Nanogenerators and Self-Powered Pressure Sensors Based on Micropatterned Plastic Films. Nano Lett. 2012, 12, 3109−3114. (8) Yoon, J.; Baca, A. J.; Park, S. I.; Elvikis, P.; Geddes, J. B.; Li, L. F.; Kim, R. H.; Xiao, J. L.; Wang, S. D.; Kim, T. H.; Motala, M. J.; Ahn, B. Y.; Duoss, E. B.; Lewis, J. A.; Nuzzo, R. G.; Ferreira, P. M.; Huang, Y. G.; Rockett, A.; Rogers, J. A. Ultrathin Silicon Solar Microcells for Semitransparent, Mechanically Flexible and Microconcentrator Module Designs. Nat. Mater. 2008, 7, 907−915. (9) Yao, J.; Lin, J.; Dai, Y. H.; Ruan, G. D.; Yan, Z.; Li, L.; Zhong, L.; Natelson, D.; Tour, J. M. Highly Transparent Nonvolatile Resistive Memory Devices from Silicon Oxide and Graphene. Nat. Commun. 2012, 3, 1101. F

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (29) Cai, G. F.; Zhou, D.; Xiong, Q. Q.; Zhang, J. H.; Wang, X. L.; Gu, C. D.; Tu, J. P. Efficient Electrochromic Materials Based on TiO2@WO3 Core/Shell Nanorod Arrays. Sol. Energy Mater. Sol. Cells 2013, 117, 231−238. (30) Pauporte, T. A Simplified Method for WO3 Electrodeposition. J. Electrochem. Soc. 2002, 149, C539−C545. (31) Yang, Y. C.; Pan, F.; Liu, Q.; Liu, M.; Zeng, F. Fully RoomTemperature-Fabricated Nonvolatile Resistive Memory for Ultrafast and High-Density Memory Application. Nano Lett. 2009, 9, 1636− 1643. (32) Sawa, A. Resistive Switching in Transition Metal Oxides. Mater. Today 2008, 11, 28−36. (33) Wong, H. S. P.; Lee, H. Y.; Yu, S. M.; Chen, Y. S.; Wu, Y.; Chen, P. S.; Lee, B.; Chen, F. T.; Tsai, M. J. Metal-Oxide RRAM. Proc. IEEE 2012, 100, 1951−1970. (34) Xie, F. Y.; Gong, L.; Liu, X.; Tao, Y. T.; Zhang, W. H.; Chen, S. H.; Meng, H.; Chen, J. XPS Studies on Surface Reduction of Tungsten Oxide Nanowire Film by Ar+ Bombardment. J. Electron Spectrosc. Relat. Phenom. 2012, 185, 112−118. (35) Park, G. S.; Kim, Y. B.; Park, S. Y.; Li, X. S.; Heo, S.; Lee, M. J.; Chang, M.; Kwon, J. H.; Kim, M.; Chung, U. I.; Dittmann, R.; Waser, R.; Kim, K. In Situ Observation of Filamentary Conducting Channels in an Asymmetric Ta2O5-x/TaO2-x Bilayer Structure. Nat. Commun. 2013, 4, 2382. (36) Wu, W. T.; Wu, J. J.; Chen, J. S. Resistive Switching Behavior and Multiple Transmittance States in Solution-Processed Tungsten Oxide. ACS Appl. Mater. Interfaces 2011, 3, 2616−2621. (37) Cong, S.; Tian, Y. Y.; Li, Q. W.; Zhao, Z. G.; Geng, F. X. SingleCrystalline Tungsten Oxide Quantum Dots for Fast Pseudocapacitor and Electrochromic Applications. Adv. Mater. 2014, 26, 4260−4267. (38) Lee, M. J.; Han, S.; Jeon, S. H.; Park, B. H.; Kang, B. S.; Ahn, S. E.; Kim, K. H.; Lee, C. B.; Kim, C. J.; Yoo, I. K.; Seo, D. H.; Li, X. S.; Park, J. B.; Lee, J. H.; Park, Y. Electrical Manipulation of Nanofilaments in Transition-Metal Oxides for Resistance-Based Memory. Nano Lett. 2009, 9, 1476−1481. (39) Son, J. Y.; Shin, Y. H. Direct Observation of Conducting Filaments on Resistive Switching of NiO Thin Films. Appl. Phys. Lett. 2008, 92, 222106. (40) Shi, Y. Y.; Ji, Y. F.; Hui, F.; Nafria, M.; Porti, M.; Bersuker, G.; Lanza, M. In Situ Demonstration of the Link Between Mechanical Strength and Resistive Switching in Resistive Random-Access Memories. Adv. Electron Mater. 2015, 1, 1400058. (41) Fujimoto, M.; Koyama, H.; Konagai, M.; Hosoi, Y.; Ishihara, K.; Ohnishi, S.; Awaya, N. TiO2 Anatase Nanolayer on TiN Thin Film Exhibiting High-Speed Bipolar Resistive Switching. Appl. Phys. Lett. 2006, 89, 223509.

G

DOI: 10.1021/acsami.6b08154 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX