Bipolar resistive switching characteristics in LaTiO₃ nanosheets

Abstract

In this work, we reported a facile approach to fabricate LaTiO₃ (LTO) nanosheets for resistive switching memory applications. Different from other lanthanum titanates synthesized via solvothermal approaches, herein the unique composition ratio of La : Ti : O = 1 : 1 : 3 has been found. The drop-coating method was utilized to deposit LTO films followed by gold top electrode deposition to complete device fabrication. The pristine device was found to exhibit excellent bipolar resistance switching characteristics with resistance ON/OFF ratio of ∼100, high uniformity in switching parameters and stability at elevated temperatures as well. The origin of switching behaviour in these devices on the basis of formation and annihilation of conducting filaments was addressed.

---

Introduction

Recently, perovskite oxide nanomaterials have aroused wide research interest in the applications of nanoelectronics^1–3 owing to their unique physical and chemical properties which are highly shape and size dependent.^1,4 A wide range of nanostructures such as nanodots,⁵ nanorods,⁶ nanowires,⁷ nanocubes,⁸ and nanosheets⁹ have been reported for diverse applications. Different from other types of nanostructures, 2-D nanostructures such as nanosheets express some unique electronic, optical, catalytic and magnetic properties because of their large surface area and quantum confinement effects.^10–17 In particular, the nanosheets architecture could be one of the ideal building blocks for future nanoelectronics, thus offering high capacitor density by high permittivity¹⁸ and molecular thickness for electronic applications.¹⁹

Resistive random access memories (RRAM) is a non-volatile memory type with unique advantages of smaller bit cell size, faster reading and writing rate and lower operating voltages.^20,21 The memory cell of a RRAM is usually built on a configuration of a capacitor with a switching layer sandwiched between two metal electrodes, where the central layer behave like an active layer for a resistive-switching (RS) device. To date, a large variety of solid-state materials ranging from amorphous materials based on chalcogenide glasses, and oxide semiconductors have been investigated for RRAM applications owing to their distinct and specific properties.^4,22,23 Among these materials, the transition-metal oxides with perovskite structure, such as titanates, zirconates or manganites,^3,24 have attracted special attention due to the great variability in their electrical properties. Perovskite-type metal oxide devices generally exhibit bipolar resistance switching, in which the resistance state can be changed by altering the voltage polarity. Two types of polarity behaviour under the same bias voltage exist in the bipolar resistance switching. For the positive bias voltage, one is eight-wise polarity, in which resistance can be changed from high state to low state and the other one is counter eight-wise polarity which alter the state resistance from low to high. Up to now, various models had been described or proposed to explain the switching mechanics in various materials. Among all, the conducting filament model had gain considerable interest being close to reality.^25,26

Lanthanum titanate is an interesting defect perovskite, with transport properties varying from insulating to metallic based on its oxygen stoichiometry. A number of methods were reported to synthesize lanthanum titanate (LTO) nanostructures including hydrothermal,²⁷ sol–gel,²⁸ RF sputtering²⁹ and pulsed laser deposition³⁰ etc. Among all, hydrothermal/solvothermal method had been identified as most facile and efficient approach to synthetize well-defined structure with high crystallization and fine control of composition. The fabrication of ordered LTO 2-dimensional nanostructures such as nanosheets for photocatalytic water splitting²⁷ and photocatalytic properties⁹ had been reported before. To the best of our knowledge, the resistive switching characteristics of LTO nanosheets have never been reported. Herein, we reported the fabrication of LTO nanosheets for their applications as resistive random access memory cells. The LTO nanosheets were found to exhibit bipolar, highly uniform and stable RS characteristics even at elevated temperatures. A conducting filament based model is proposed to explain conduction process in these devices.

Experimental

All chemicals were purchased from Sigma without further purification. At the first stage, bis (ammonium lactate) titanium dihydroxide (TALH) (Sigma-Aldrich, C₆H₁₈N₂O₈Ti, 50 wt% in H₂O, 0.05mol L⁻¹) was added into a aqueous solution of lanthanum(III) nitrate hexahydrate (La(NO₃)₃·6H₂O, ≥99.0%), La : Ti = 1 : 1. Then 5 mol L⁻¹ aqueous NaOH solution was poured into the mixture with mechanical stir in order to achieve a NaOH concentration of 1 mol L⁻¹. The mixture was transferred to a 50 mL autoclave and the combined solutions of oleic acid (OLA) and tert-butylamine (molar ratio: Ti : OLA : tert = 1 : 8 : 8) were added into the mixture. In the second stage, the sealed autoclave was heated at 220 °C for 24 hours and then cooled to room temperature. After that, the product in the form of suspension solution was collected by centrifugation at 4000 rpm for 4 minutes followed by twice washing with 95% ethanol. The LaTiO₃ nanosheets were dispersed into absolute ethanol and then drop coating method was used to deposit thin films.

Structural analysis of as-synthesized LaTiO₃ films was carried out by using an X-ray diffractometer with Cu Ka radiation (The PANalytical Xpert Multipurpose X-ray Diffraction System). The morphology and microstructures were characterized by transmission electron microscopy (Philips CM200). To measure the electrical properties (resistive switching characteristics) of the prepared films, first Au was deposited on Si wafer to form a conducting bottom electrode followed by deposition of LaTiO₃ nanosheets by drop coating method. The drop-coated method was repeated five times (10 minutes UV treatment between each drop-coating sequence) in order to get a considerable film thickness. The deposited film was then treated with UV radiation for one hour to get rid of all organics and so. Finally, a small area of the electrode (gold, Au) with round patterning and size of about 250 μm diameter was sputtered through a shadow mask to complete the device fabrication process. The whole process is demonstrated in a schematic diagram as shown in Fig. 1.

Fig. 1 (a and b) Schematic diagram for the device preparation process. (c) Setup for I–V measurements.

Current–voltage curves (RS properties) of the films were measured by using an Autolab 302N electrochemical workstation controlled with Nova software.

Result & discussion

Fig. 2 shows the XRD pattern of the as prepared film. All diffracted peaks can be indexed as the cubic phase [space group Pm3̄m (no. 221)] of lanthanum titanium oxide (JCPDS 01-075-0267 and 01-073-0069) with excellent crystallinity. The broadness of these peaks indicates that the nanosheets are composed of small nanoparticles and that the structure of Ti precursor plays a dominant role in the synthesis of lanthanum titanate nanoparticles.

Fig. 2 XRD pattern of as prepared film composed of LaTiO₃ nanosheets.

Fig. 3a and b show the TEM images of the LTO nanosheets at different magnifications. From Fig. 3b, it is quite clear that that the nanosheets are well uniform and highly dense in nature. The size of a single nanosheet was found to be between 100 nm to 200 nm indicated by Fig. 3b. The HRTEM image of a single nanosheet can be depicted from Fig. 3c, the lattice fringes shows the lattice spacing of 0.27 nm, which corresponds to the (110) plane of the LaTiO₃ matched well with the XRD pattern in Fig. 1. Due to their well dispersion in ethanol, it could be imagined that the LaTiO₃ nanosheets can form close packed structure on the gold coated silicon substrate for electrical characterization.

Fig. 3 (a) TEM image of LaTiO₃ nanosheets dispersed in ethanol (b) TEM image of a single LTO nanosheet (c) HRTEM image of a single nanosheet.

The current–voltage (I–V) characteristics of LTO films are illustrated in Fig. 4a. The bipolar current–voltage (I–V) characteristics of the Au/LTO/Au/Si device under voltage sweeping mode, at a speed of 1 V s⁻¹, in the sequence of 0 → 3 → 0 → −3 → 0 V was measured. In the positive voltage sweep region, the device switches its resistance state from high resistance state (HRS/OFF) to low resistance state (LRS/ON) at about +2.2 V with a sharp increment in device current value, which is referred as set process. The device retains its LRS state until the polarity reverse. At about −2.4 V, a subsequent transition from LRS/ON state to HRS/OFF state was observed, which is called as reset process. The sample exhibits a typical bipolar nature of resistive switching. It shows that the positive biasing can be utilized for writing data while the negative biasing for erasing.

Fig. 4 (a) I–V measurements on the semi-log scale (b) endurance performance of the device for 2 × 10³ switching cycles. (c) Data retention performance of the device for 10⁵ s. (d) Statistical distribution of set and reset voltages for 2000 consecutive DC voltage sweeping cycles.

In order to evaluate the potential for the application in non-volatile memory devices, a cycling switching test (endurance) were implemented. The electrical pulse-induced resistance change effect in the memory device was shown in Fig. 4b and memory cell exhibits a minimal degradation during 2 × 10³ switching cycles. This shows that the device decent reliability under stress conditions and therefore indicates the potential of LTO as highly promising material for RRAM application.^31,32 The memory window margin of the as-fabricated device is found to be more than 10², which is comparable to many perovskite materials fabricated by thin-film technology.³³ Subsequently, the data retention test was carried out in both the OFF and ON states for 10⁵ s at room temperature and at 85 °C (as shown in Fig. 4c), showing that the device switching capability is quite stable even at higher temperature.

Fig. 4d shows the distribution of set and reset voltages for the LaTiO₃ nanosheets-based RS device, while the set of data was dragged by implementing 2000 consecutive DC cycles on the sample. The distribution of device set voltage (∼1.75 V) and (∼1.54 V) for reset voltage reflects device decent uniformity and stability for 2000 sweeping cycles.

The typical programming and erasing characteristics of LaTiO₃ nanosheets were also carried out, which are shown in Fig. 5. The pulse characteristics of the devices were investigated to define the program and erase (P/E) conditions, in which the typical programming and erasing characteristics under ac pulse biases were measured over pulse widths ranging from 10 ns to 100 ms and with pulse heights of +3 V and −3 V. A frequency generator and an extra source meter were used in this setup to employ voltage ac pulses and to read device response current against applied voltage pulse.

Fig. 5 Programming and erasing characteristics of LaTiO₃ nanosheets based prototypes.

During the programing process (Fig. 5a), the current was measured in order to record the actual switching process time. It was found that the device was switched from HRS/OFF state to LRS/ON state within 10⁻¹ s at ∼3 V/2 × 10⁻⁴ S. In the erasing process (Fig. 5b), the switching characteristics were similar to the programing process. The device transit its resistance state from LRS to HRS at ∼−3 V/1 × 10⁻⁵ S indicates excellent device response time. These characteristics showed excellent device response time for writing and erasing data.

To explain resistance switching characteristics in these devices, a model comprised of oxygen vacancies and their distribution within the device is proposed.^1,34 It is well-known that due to lower oxygen vacancy formation energy on/near surface, their density would be higher at or near surface³⁵ of nanosheets. Also, due to very thin nanosheets structures, five times repeated drop-coating may introduce few tenths of nanosheets layer in the device. So, there is a greater chance of having a considerable amount of oxygen vacancies in the overall device structure as shown in Fig. 6a. At start, when there was no potential, the oxygen vacancies (V_O) were randomly distributed throughout the nanostructure. With the application of positive voltage sweep, the oxygen ions start migration towards anode and as a result, oxygen vacancies rearrange themselves and proceed towards cathode thus forming a chain of vacancies/conducting paths throughout the nanostructure. The oxygen ions will finally reach to the top electrode or the Au/LaTiO₃ interfaces, but could not penetrate into Au gate electrode and remains trapped at the interface. As the density of oxygen vacancies increase to a crucial value, conducting channels/filaments between two electrodes will be formed as a soft breakdown, allowing electrons to overcome the potential barrier, thereby transiting device resistance from high resistance state (HRS) to low resistance state (LRS) as shown in Fig. 6b.

Fig. 6 Schematics of the proposed model demonstrating the conduction process (a) without applied potential, (b) with positive potential, (c) with negative potential.

During a negative voltage sweep, the local Joule heating may induce the local thermal excitation³⁶ to melt those filaments and simultaneously the oxygen ions trapped at the Au/LaTiO₃ interfaces would move towards the bottom electrode through the electrical field, causing the elimination of the oxygen vacancies which ultimately reset the device from LRS to HRS as shown in Fig. 6c. Also V_O–V_O interaction could be important by creation/breaking of V_O–V_O interactions and thus forming/rupturing of the conductive filaments, which leads to novel switching properties in these devices. Migration of oxygen vacancies can be activated by appropriate Joule heating (electric field), and it is energetically favourable to form conductive filaments composed of oxygen vacancies due to the V_O–V_O interaction, which results in the non-volatile switching from OFF to ON state. For the reset process, oxygen vacancies reorder themselves by breaking the V_O–V_O interactions and thus rupturing the conductive filaments, which are responsible for the switching process from OFF to ON state.³⁷

To further strengthen this argument of critical role of oxygen vacancies in the conduction process of these devices, the concentration of oxygen vacancies in the as-prepared films was reduced through thermal annealing at 300 °C for overnight in ambient atmosphere. In order to investigate the role of thermal treatment on crystal structure, the X-ray diffraction studies were conducted on post-annealed samples. No considerable effect of heat treatment was observed in the crystal structure of the LTO samples as shown in ESI Fig. S1.† The current–voltage characteristics were also tested for these thermally annealed samples. The treated samples didn't show any distinct resistive switching behaviour with any significant observed ON/OFF ratio (∼4) as shown in ESI Fig. S2 and S3.† It further authenticates the critical role of oxygen in the resistive switching behaviours of these devices.

Conclusion

In summary, LaTiO₃ nanosheets were fabricated through a facile solvothermal approach which exhibit bipolar resistance switching characteristics. The pristine device was found to exhibit good endurance of 2 × 10³ switching cycles to maintain the switching ON/OFF ratio of 10². Also, a decent trend in data retention performance was demonstrated at both room and high temperatures (85 °C) with minimal degradation in device resistance ratio for more than 10⁵ S. The prototype also expressed efficient data writing and erasing response time at about 2 × 10⁻⁴ S and 1 × 10⁻⁵ S. Finally, a model based on oxygen vacancies migration and annihilation under electric field was proposed to explain the conduction process in these devices. The present work demonstrates that LTO nanosheets could be a potential material for next generation non-volatile memory applications.

Acknowledgements

The authors would like to acknowledge the financial support from the Australian Research Council Project of DP110102391. Financial supports of the Chancellor’s Postdoctoral Research Fellowship Program from University of Technology, Sydney are greatly appreciated.

References

1. Z. Yan, Y. Guo, G. Zhang and J. M. Liu, Adv. Mater., 2011, 23, 1351–1355.
2. K. Szot, R. Dittmann, W. Speier and R. Waser, Phys. Status Solidi RRL, 2007, 1, R86–R88.
3. D. Chu, X. Lin, A. Younis, C. M. Li, F. Dang and S. Li, J. Solid State Chem., 2013 DOI:10.1016/j.jssc.2013.10.049.
4. A. Sawa, Mater. Today, 2008, 11, 28–36.
5. J. Li, M. Luo, W. Weng, K. Cheng, P. Du, G. Shen and G. Han, Appl. Surf. Sci., 2009, 256, 342–345.
6. J. J. Urban, W. S. Yun, Q. Gu and H. Park, J. Am. Chem. Soc., 2002, 124, 1186–1187.
7. J. E. Spanier, A. M. Kolpak, J. J. Urban, I. Grinberg, L. Ouyang, W. S. Yun, A. M. Rappe and H. Park, Nano Lett., 2006, 6, 735–739.
8. F. Dang, K.-i. Mimura, K. Kato, H. Imai, S. Wada, H. Haneda and M. Kuwabara, CrystEngComm, 2011, 13, 3878.
9. K. Li, Y. Wang, H. Wang, M. K. Zhu and H. Yan, Nanotechnology, 2006, 17, 4863–4867.
10. L. Zhang, D. Chen and X. Jiao, J. Phys. Chem. B, 2006, 110, 2668–2673.
11. S. J. Chen, Y. C. Liu, C. L. Shao, R. Mu, Y. M. Lu, J. Y. Zhang, D. Z. Shen and X. W. Fan, Adv. Mater., 2005, 17, 586–590.
12. N. Sakai, Y. Ebina, K. Takada and T. Sasaki, J. Phys. Chem. B, 2005, 109, 9651–9655.
13. X.-C. Jiang, L.-D. Sun and C.-H. Yan, J. Phys. Chem. B, 2004, 108, 3387–3390.
14. A. Takagaki, M. Sugisawa, D. Lu, J. N. Kondo, M. Hara, K. Domen and S. Hayashi, J. Am. Chem. Soc., 2003, 125, 5479–5485.
15. J. Zhu, L. Cao, Y. Wu, Y. Gong, Z. Liu, H. E. Hoster, Y. Zhang, S. Zhang, S. Yang, Q. Yan, P. M. Ajayan and R. Vajtai, Nano Lett., 2013, 13, 5408–5413.
16. W. Zhang, M. Han, Z. Jiang, Y. Song, Z. Xie, Z. Xu and L. Zheng, ChemPhysChem, 2007, 8, 2091–2095.
17. M. Osada and T. Sasaki, J. Mater. Chem., 2009, 19, 2503–2511.
18. L. Wang, L.-Y. Ding, S.-T. Zhang, Y.-F. Chen and Z.-G. Liu, Solid State Commun., 2009, 149, 2061–2064.
19. M. Osada and T. Sasaki, Adv. Mater., 2012, 24, 210–228.
20. R. Waser, R. Dittmann, G. Staikov and K. Szot, Adv. Mater., 2009, 21, 2632–2663.
21. A. Younis, D. Chu, C. M. Li, T. Das, S. Sehar, M. Manefield and S. Li, Langmuir, 2014, 30, 1183–1189.
22. K. Szot, W. Speier, G. Bihlmayer and R. Waser, Nat. Mater., 2006, 5, 312–320.
23. T. Yamamoto, R. Yasuhara, I. Ohkubo, H. Kumigashira and M. Oshima, J. Appl. Phys., 2011, 110, 053707.
24. C. C. Lin, J. S. Yu, C. Y. Lin, C. H. Lin and T. Y. Tseng, Thin Solid Films, 2007, 516, 402–406.
25. S. B. Long, L. Perniola, C. Cagli, J. Buckley, X. J. Lian, E. Miranda, F. Pan, M. Liu and J. Sune, Sci. Rep., 2013, 3, 2929.
26. S. Long, X. Lian, C. Cagli, X. Cartoixa, R. Rurali, E. Miranda, D. Jiménez, L. Perniola, M. Liu and J. Suñé, Appl. Phys. Lett., 2013, 102, 183505.
27. H. Song, T. Peng, P. Cai, H. Yi and C. Yan, Catal. Lett., 2007, 113, 54–58.
28. M. M. Sandstrom and P. Fuierer, J. Mater. Res., 2003, 18, 357–362.
29. M. Morales, P. Laffez, D. Chateigner and I. Vickridge, Thin Solid Films, 2002, 418, 119–128.
30. S.-i. Furusawa, H. Tabuchi, T. Sugiyama, S. Tao and J. T. S. Irvine, Solid State Ionics, 2005, 176, 553–558.
31. S. B. Long, X. J. Lian, T. C. Ye, C. Cagli, L. Perniola, E. Miranda, M. Liu and J. Sune, IEEE Electron Device Lett., 2013, 34, 623–625.
32. S. B. Long, X. J. Lian, C. Cagli, L. Perniola, E. Miranda, M. Liu and J. Sune, IEEE Electron Device Lett., 2013, 34, 999–1001.
33. R. Waser and M. Aono, Nat. Mater., 2007, 6, 833–840.
34. Y. B. Nian, J. Strozier, N. J. Wu, X. Chen and A. Ignatiev, Phys. Rev. Lett., 2007, 98, 146403.
35. G. Pacchioni, ChemPhysChem, 2003, 4, 1041–1047.
36. S. H. Chang, J. S. Lee, S. C. Chae, S. B. Lee, C. Liu, B. Kahng, D. W. Kim and T. W. Noh, Phys. Rev. Lett., 2009, 102, 026801.
37. D. S. Jeong, R. Thomas, R. Katiyar, J. Scott, H. Kohlstedt, A. Petraru and C. S. Hwang, Rep. Prog. Phys., 2012, 75, 076502.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra01626b

‡ Authors have equal contribution.

---