Ion Beam-Mediated Defect Engineering in TiO_x Thin Films for Controlled Resistive Switching Property and Application

Abstract

Defect density and its migration under the influence of an applied electric field plays a crucial role for memristors toward designing a fundamental element of neuromorphic computing, e.g., an artificial synapse. Therefore, to have tunable performance, even at the nanoscale, it is essential to have better control over the defect density, albeit achieving it with direct growth methods remains a challenge. Here, we demonstrate an effective and robust approach to tune the defect density and consequently resistive switching behavior in TiO x layers using low energy argon ion implantation as a tool. Interestingly, by selecting appropriate implantation parameters an adjustable nanoscale resistive switching property, in terms of increased memory window and lower operative voltage, is achieved. These findings are attributed to ion-beam induced controlled defect engineering and are well supported by our experimental results and the predictions of Monte Carlo simulation results. Further, learning ability with an input electric stimuli strongly depends on the implantation fluence, revealing a striking similarity with biological synapses. The present study provides a one-step processing to tune the defect density and in turn resistive switching behavior for designing advanced memory devices and neuromorphic computing.

Tuning resistive switching properties of WO3−x-memristors by oxygen vacancy engineering for neuromorphic and memory storage applications

“…Hwang et al reported significant improvement of conduction modulation linearity and pattern recognition efficiency upon increasing V O s concentration in Ta 2 O 5 thin films [20]. In case of TiO x -based memristors increase of V O s by means of ion beam irradiation, as well as incorporation of a V O -rich TiO y layer have been found to improve their neuromorphic characteristics [21,22].…”

Site‐Specific Emulation of Neuronal Synaptic Behavior in Au Nanoparticle‐Decorated Self‐Organized TiOx Surface

“…[30] In general, various intrinsic and extrinsic methods can be employed to gain control over the defect-dependent switching properties. [3,[30][31][32][33][34][35] In the former case, the concentration of cationic/anionic vacancies is modified during the film growth process, [34] whereas, in the latter case, impurity doping of the switching medium is considered either during growth or post-growth processing. [30] For instance, Ag-cluster-doped TiO 2based two-terminal memristor exhibits an improved artificial synaptic performance.…”

“…[26] In addition, recently Choi et al have reported better switching performances in Si-Ge-based epitaxial memristors via defect engineering by means of structural modifications. [3] On another note, ion implantation is also considered to be an extremely useful technique not only to modulate post-growth defect density within the switching medium [30] but also to give rise to self-organized pattern formation on its surface, depending on the ion-energy, -fluence, and incident angle. [30,32] As a result, anisotropic, electrical, [36] magnetic, [37,38] and optical properties [39,40] are realized.…”

“…[3] On another note, ion implantation is also considered to be an extremely useful technique not only to modulate post-growth defect density within the switching medium [30] but also to give rise to self-organized pattern formation on its surface, depending on the ion-energy, -fluence, and incident angle. [30,32] As a result, anisotropic, electrical, [36] magnetic, [37,38] and optical properties [39,40] are realized. In addition, patterned surfaces are promising for cold cathode electron emission, [41,42] biology [43] as well as nano bio-electronic applications.…”

“…[3,[30][31][32][33][34][35] In the former case, the concentration of cationic/anionic vacancies is modified during the film growth process, [34] whereas, in the latter case, impurity doping of the switching medium is considered either during growth or post-growth processing. [30] For instance, Ag-cluster-doped TiO 2based two-terminal memristor exhibits an improved artificial synaptic performance. [26] In addition, recently Choi et al have reported better switching performances in Si-Ge-based epitaxial memristors via defect engineering by means of structural modifications.…”