Resistive Switching Memory of TiO2 Nanowire Networks Grown on Ti Foil by a Single Hydrothermal Method
- 2.7k Downloads
The resistive switching characteristics of TiO2 nanowire networks directly grown on Ti foil by a single-step hydrothermal technique are discussed in this paper. The Ti foil serves as the supply of Ti atoms for growth of the TiO2 nanowires, making the preparation straightforward. It also acts as a bottom electrode for the device. A top Al electrode was fabricated by e-beam evaporation process. The Al/TiO2 nanowire networks/Ti device fabricated in this way displayed a highly repeatable and electroforming-free bipolar resistive behavior with retention for more than 104 s and an OFF/ON ratio of approximately 70. The switching mechanism of this Al/TiO2 nanowire networks/Ti device is suggested to arise from the migration of oxygen vacancies under applied electric field. This provides a facile way to obtain metal oxide nanowire-based ReRAM device in the future.
KeywordsTiO2 nanowire networks Resistive switching memory Ti foil Hydrothermal process Al electrode
TiO2 nanowire networks were grown on Ti foil by a one-step hydrothermal method.
Obtained Al/TiO2 nanowire networks/Ti devices showed forming-free resistive switching behavior. Good retention and endurance performance was achieved for the fabricated devices.
Switching mechanism is due to migration of oxygen vacancies under electric field.
Resistive switching random access memory (ReRAM) utilizing an electric-field-induced resistance switching phenomena has attracted great attention for next-generation nonvolatile memory due to its advantages of simple sandwich structure of metal/insulator/metal, high storage density, and fast operation speed [1, 2]. Among different metal oxide materials that demonstrate potential for ReRAM, including NiO [3, 4], TiO2 [1, 5, 6], ZnO [2, 7, 8], VO2 , Ta2O5 [10, 11], CuO [12, 13], WO3 , etc., TiO2 nanomaterial-based memory has been widely studied due to its ease of fabrication [1, 15] and its ability to demonstrate both unipolar [16, 17] and bipolar [18, 19, 20] resistive switching behavior.
Compared to TiO2 thin films used for ReRAM [6, 19, 20, 21, 22], few studies based on one-dimensional TiO2 nanomaterials for ReRAM have been reported. It was recently shown that a single TiO2 nanowire-based resistive switching device demonstrated multilevel memory behavior [23, 24]. But the fabrication process of Au electrodes bridging a single nanowire required costly and time-consuming electron-beam lithography. Therefore, a facile way to fabricate TiO2 nanowire-based ReRAM is required. Furthermore, TiO2 nanorod [25, 26] and nanotube  arrays grown on fluorine-doped tin oxide (FTO) glass substrate by hydrothermal synthesis were also employed in resistive switching memory devices, however transparent conductive glass was required as a substrate. It was recently reported that TiO2 nanowire networks could be grown directly on Ti foil via a hydrothermal method [28, 29, 30, 31] or oxidation process [32, 33], and the applications of these nanowires in dye-sensitized solar cells [29, 30] and field emission  were investigated. But the suitability of these TiO2 nanowires for ReRAM devices and the corresponding switching mechanism has not been reported yet.
In this paper, TiO2 nanowire networks were directly grown on Ti foil by a hydrothermal method and their resistive switching behavior was investigated. Since the Ti foil serves both as the source of Ti during the synthesis of the TiO2 nanowire, as well as a bottom electrode for the device, preparation of the device is straightforward, cost effective and highly reproducible. Notably, the electrical contact between the nanowires and the bottom metal substrate is ensured. According to the current–voltage (I–V) measurements of the fabricated Al/TiO2 nanowire networks/Ti device, a switching mechanism based on the migration of oxygen vacancies is proposed. The reliability of the fabricated device was examined by studying its retention and endurance performance.
3 Materials and Methods
The synthesis process of TiO2 nanowire networks on Ti foil is referred to [28, 29]. Briefly, a piece of Ti foil with a dimension of 1.5 × 3.0 cm × 0.127 mm (Sigma Aldrich) was ultrasonically cleaned in acetone, isopropanol and Milli-Q water for 10 min in sequence and then placed against the wall of a 125 mL Teflon-lined stainless steel antoclave filled with 40 mL of 1 M NaOH aqueous solution. Then, the sealed autoclave was put into an oven at a temperature of 220 °C for 20 h. Next the Ti foil covered with nanowires was taken out of the autoclave and immersed in 50 mL of 0.6 M HCl solution for 1 h to exchange Na+ with H+. Finally, the foil was annealed inside a furnace at 500 °C for 3 h in air to transform the H2Ti2O5·H2O nanowires to anatase nanowires. The color of the foil turned white after the calcination process.
During device fabrication, the top electrode was prepared by depositing an Al layer with a thickness of 150 nm through a shadow mask having circular holes (1 mm in diameter) using e-beam evaporation process (Intelvac e-beam evaporation system). The pressure was <4 × 10−6 Torr, and the deposition rate was 1 Å s−1. Electrical measurements were performed using a Keithley 2602A source-meter at ambient conditions. The bias voltage was applied to the top Al electrode, and the Ti foil was grounded during electrical measurement.
For the characterization of the TiO2 nanowires, a field-emission scanning electron microscope (FESEM, LEO-1550) was used to check the surface morphology. Transmission electron microscopy (TEM, JEOL 2010F) was used to examine the structure and crystalline defects of TiO2 nanowires. X-ray diffraction analysis (XRD, PANalytical X’pert PRO MRD) and Raman analysis (Reinshaw micro-Raman spectrometer) were used to identify the crystal structure and phase, respectively. Furthermore, X-ray photoelectron spectroscopy measurement (XPS, Thermo VG Scientific ESCLab 250) was carried out to examine the surface chemical states of the nanowires.
4 Results and Discussion
4.1 Characterization of TiO2 Nanowire Networks
A room-temperature Raman spectrum of TiO2 nanowires in Fig. 1c shows peaks at 141, 194, 395, 512, and 634 cm−1. These peaks are characteristic of the anatase phase. The peaks at 141, 194, and 634 cm−1 are assigned to the E g modes, while the other two peaks at 512 and 395 cm−1 are assigned to the B 1g modes in TiO2 . The XRD characterization results in Fig. 1d further confirm the phase of the TiO2 nanowires, as the peaks of (101), (112), and (200) planes of anatase in agreement with the standard spectrum (JCPDS No. 21-1272). It should be noted that one of the anatase peaks at 38.57° overlapped with the peaks of the Ti foil (JCPDS No. 44-1294).
Moreover, the surface chemical states of the TiO2 nanowires were analyzed by XPS. Figure 1e shows peaks at binding energies of 459.4 and 465.1 eV, which can be assigned to Ti 2p3/2 and 2p1/2, respectively. These are typical XPS spectra of Ti4+ in TiO2. The signal from Ti3+ is too small to be detected. Furthermore, two Gaussian peaks are observed in the fit to the O 1s spectrum (Fig. 1f). The binding energy at 529.66 eV is assigned to the O2– bond in TiO2 while the binding energy at 531.33 eV can be attributed to oxygen vacancies in TiO2 . XPS scans show that the synthesized TiO2 nanowires contain locally distributed oxygen vacancies, in agreement with the high-resolution TEM (HRTEM) result in Fig. 1b.
4.2 Electrical Performance Evaluation
4.2.1 Resistive Switching Characteristics
4.2.2 Switching Mechanism Analysis
In the ReSET process, the oxygen vacancies in the TiO2 matrix are repelled towards the bottom electrode, leading to the recovery of a higher concentration of oxygen vacancies near the Ti bottom layer, and widening of the Al–Ti–O layer as oxygen vacancies drift away from this layer . The presence of a potential barrier in the Al/Al–Ti–O interfacial layer would suppress electron tunneling through the interface and inhibit the formation of conducting channels. Conversely, as oxygen vacancies migrate to the top Al electrode under an applied negative bias, the interfacial layer begins to thin, increasing the probability of electron tunneling and enhancing the formation of conductive channels. As a result, the final current flowing at +10 V is much less than that at −10 V. This is characteristic of asymmetrical or self-rectifying resistive switching behavior, as seen in Figs. 2 and 3. To verify the source of the self-rectifying performance, an identical device was measured without the top electrode, that is, the probe tip (which is made of tungsten) is directly in contact with the top surface of the TiO2 nanowire layer. We found the bipolar resistive switching performance existed for the device as well, but no self-rectifying feature was observed.
The I–V characteristic curve for the negative voltage in Fig. 6b also demonstrates SCLC-like behavior but the fitted slope values in different regions are generally larger than those under positive voltage sweeping. This could be due to the concentration gradient of oxygen vacancies that exists in the pristine state (high concentration of vacancies at the bottom Ti/TiO2 interface and low concentration underneath the Al–Ti–O layer), which would lead to diffusion of these oxygen vacancies. The diffusion combined with the drift of the oxygen vacancies under an applied negative bias could lead to accelerated migration of vacancies, resulting in higher slopes when transitioning from the HRS to LRS, as compared to the transition from LRS to HRS under positive sweeping. Furthermore, dissociation of the insulating Al–Ti–O layer due to migration of oxygen vacancies under negative bias also decreases the overall resistance of the device, which would contribute to higher values of the slopes. This SCLC-like behavior for both positive and negative sweeping voltages can also be found with increasing sweeping cycles of the device.
The above analysis indicates that Al/TiO2 nanowire networks/Ti device as fabricated exhibits a similar I–V response and switching mechanism as that seen in devices using a uniform TiO2 layer coated with an Al electrode. Such devices were fabricated by time-consuming and costly reactive sputtering  or plasma-enhanced atomic layer deposition [1, 45, 46, 48, 49] processes. Therefore, our results indicate that the TiO2 nanowire networks grown on Ti foil by a single-step hydrothermal process have potential in the application of ReRAM devices.
4.2.3 Endurance and Retention Study
In summary, electroforming-free bipolar resistive switching behavior was successfully demonstrated in TiO2 nanowire networks directly grown on Ti foil by a one-step hydrothermal process. The prepared Al/TiO2 nanowire networks/Ti device exhibited reproducible and stable electrical performance with a high OFF/ON ratio that persisted for up to 104 s. We found that the interaction of Ti foil with the TiO2 nanowires during the synthesis process results in the generation of large density of oxygen vacancies at the Ti/TiO2 interface, which is likely responsible for the forming-free resistive switching behavior. The switching mechanism of the device is proposed to be the migration of oxygen vacancies under electric field. These results provide an easy way to prepare nanowire-based ReRAM devices with good electrical performance.
This work was supported by the Natural Sciences and Engineering Research Council (NSERC) of Canada. The financial support of the State Scholarship Fund of China (No. 201506160061) is greatly acknowledged. M. Xiao would like to thank Carmen Andrei from the Canadian Center for Electron Microscopy, McMaster University, for help with TEM.
- 10.S. Gao, F. Zeng, M. Wang, G. Wang, C. Song, F. Pan, Tuning the switching behavior of binary oxide-based resistive memory devices by inserting an ultra-thin chemically active metal nanolayer: a case study on the Ta2O5–Ta system. Phys. Chem. Chem. Phys. 17(19), 12849–12856 (2015). doi: 10.1039/C5CP01235J CrossRefGoogle Scholar
- 15.K. Szot, M. Rogala, W. Speier, Z. Klusek, A. Besmehn, R. Waser, TiO2-a prototypical memristive material. Nanotechnology 22(25), 254001(2011). doi: 10.1088/0957-4484/22/25/254001
- 37.P.C. Wang, P.G. Li, Y.S. Zhi, D.Y. Guo, A.Q. Pan, J.M. Zhan, H. Liu, J.Q. Shen, W.H. Tang, Bias tuning charge-releasing leading to negative differential resistance in amorphous gallium oxide/Nb:SrTiO3 heterostructure. Appl. Phys. Lett. 107(26), 262110 (2015). doi: 10.1063/1.4939437 CrossRefGoogle Scholar
- 40.A. Barman, C.P. Saini, P.K. Sarkar, A. Roy, B. Satpati, D. Kanjilal, S.K. Ghosh, S. Dhar, A. Kanjilal, Probing electron density across Ar+ irradiation-induced self-organized TiO2−x nanochannels for memory application. Appl. Phys. Lett. 108(24), 244104 (2016). doi: 10.1063/1.4954166 CrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided 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.