Characterization of Information-Transmitting Materials Produced in Ionic Liquid-based Neuromorphic Electrochemical Devices for Physical Reservoir Computing.

Device implementation of reservoir computing, which is expected to enable high-performance data processing in simple neural networks at a low computational cost, is an important technology to accelerate the use of artificial intelligence in the real-world edge computing domain. Here, we propose an ionic liquid-based physical reservoir device (IL-PRD), in which copper cations dissolved in an IL induce diverse electrochemical current responses. The origin of the electrochemical current from the IL-PRD was investigated spectroscopically in detail. After operating the device under various operating conditions, X-ray photoelectron spectroscopy of the IL-PRD revealed that electrochemical reactions involving Cu, Cu2O, Cu(OH)2, CuSx, and H2O occur at the Pt electrode/IL interface. These products are considered information transmission materials in IL-PRD similar to neurotransmitters in biological neurons. By introducing the Faradaic current components due to the electrochemical reactions of these materials into the output signal of IL-PRD, we succeeded in improving the time-series data processing performance of the nonlinear autoregressive moving average task. In addition, the information processing efficiency in machine learning to classify electrocardiogram signal waveforms was successfully improved by using the output current from IL-PRD. Optimizing the electrochemical reaction products of IL-PRD is expected to advance data processing technology in society.

Dynamic Nonlinear Behavior of Ionic Liquid-Based Reservoir Computing Devices.

Herein, a physical reservoir device that uses faradaic currents generated by redox reactions of metal ions in ionic liquids was developed. Synthetic time-series data consisting of randomly arranged binary number sequences ("1" and "0") were applied as isosceles-triangular voltage pulses with positive and negative voltage heights, respectively, and the effects of the faradaic current on short-term memory and parity-check task accuracies were verified. The current signal for the first half of the triangular voltage-pulse period, which contained a much higher faradaic current component compared to that of the second half of the triangular voltage-pulse period, enabled higher short-term memory task accuracy. Furthermore, when parity-check tasks were performed using a faradaic current generated by asymmetric triangular voltage-pulse levels of 1 and 0, the parity-check task accuracy was approximately eight times higher than that of the symmetric triangular voltage pulse in terms of the correlation coefficient between the output signal and target data. These results demonstrate the advantage of the faradaic current on both the short-term memory characteristics and nonlinear conversion capabilities and are expected to provide guidance for designing and controlling various physical reservoir devices that utilize electrochemical reactions.

Reservoir computing with dielectric relaxation at an electrode-ionic liquid interface.

A physical reservoir device with tunable transient dynamics is strongly required to process time-series data with various timescales generated in the edge region. In this study, we proposed using the dielectric relaxation at an electrode-ionic liquid (IL) interface as the physical reservoir by making the most of designable physicochemical properties of ILs. The transient dynamics of a Au/IL/Au reservoir device were characterized as a function of the alkyl chain length of cations in the IL (1-alkyl-3-methylimidazolium bis(trifluoromethane sulfonyl)imide). By considering a weighted sum of exponentials expressing a superposition of Debye-type relaxations, the transient dynamics were well reconstructed. Although such complex dynamics governed by multiple relaxation processes were observed, each extracted relaxation time scales with a power law as a function of IL's viscosity determined by the alkyl chain length of cations. This indicates that the relaxation processes are characterized by bulk properties of the ILs that obey the widely received Vogel-Fulcher-Tammann law. We demonstrated that the 4-bit time-series signals were transformed into the 16 classifiable data, and the data transformation, which enables to achieve higher accuracy in an image classification task, can be easily optimized according to the features of the input signals by controlling the IL's viscosity.

Three-Dimensional Nanoconfinement Supports Verwey Transition in Fe3O4 Nanowire at 10 nm Length Scale.

Herein, we construct three-dimensional (3D) Fe3O4 epitaxial nanowires at a 10 nm length scale on a 3D MgO nanotemplate using an original nanofabrication technique that mainly comprises nanoimprint lithography and inclined thin-film deposition. Despite the high density of inevitable nanoscale defects, the ultrasmall Fe3O4 nanowires exhibit a prominent Verwey transition at about 112 K with a maximum relative change in resistance of 9.5, which is 6 times larger than that of the thin-film configuration. Numerous measurements on a large number of Fe3O4 nanowires grown concurrently on the same 3D MgO nanotemplate reveal a dramatic difference in their electrical transport property with the presence/absence of the Verwey transition. A comparative study of Fe3O4 wires of increasing volume and a thin film reveals that a profound change in the Verwey transition is observed only for wires with a volume on the order of 10 nm3. Moreover, a significant decrease in the sharpness of the resistance jump and the transition temperature of the Verwey response is noticed with an increasing volume of Fe3O4. This indicates the potency of the 3D nanofabrication technique in controlling nanoscale defects, which is further reconfirmed through magnetoresistance measurement. A feature of the magnetoresistance curve identifies the antiphase boundaries as a major source of defects. The occurrence of the smallest magnetoresistance in the ultrasmall nanowire with the highest Verwey transition temperature and resistance change ratio proves that 3D isotropic spatial confinement into a length scale comparable to the average spacing between two antiphase boundaries enables the favorable control over nanoscale defects. A simple statistical model satisfactorily illustrates the dependence of electrical transport properties on the volume of Fe3O4 from the macroscale down to the nanoscale. Finally, an ultrasmall nanowire with a low defect concentration allows the estimation of the true coherence length of the fundamental quasiparticle, the trimeron, responsible for the Verwey transition.

Investigation of switching mechanism in HfOx-ReRAM under low power and conventional operation modes.

Low-power resistive random access memory (LP-ReRAM) devices have attracted increasing attention owing to their advantages of low operation power. In this study, a vertical-type LP-ReRAM consisting of TiN/Ti/HfO2/TiN structure was fabricated. The switching mechanism for LP-ReRAM was elucidated as the conductive filament mechanism for conventional mode, and an interface-type switching mechanism for low power mode was proposed. The analysis of low frequency noise shows that power spectral density (PSD) is approximately proportional to 1/f for conventional operation mode. Nevertheless, for low power mode, the PSD of low resistance state (LRS) is proportional to 1/f, while that of high resistance state (HRS) is clear proportional to 1/f2. The envelope of multiple Lorentzian spectra of 1/f2 characteristics due to different traps reveals the characteristics of 1/f. For HRS of low power mode, a limited number of traps results in a characteristic of 1/f2. During the set process, the number of oxygen vacancies increases for LRS. Therefore, the PSD value is proportional to 1/f. Owing to the increase in the number of traps when the operation mode changes to conventional mode, the PSD value is proportional to 1/f. To the best of our knowledge, this is the first study that reveals the different noise characteristics in the low power operation mode from that in the conventional operation mode.

Hierarchical three-dimensional layer-by-layer assembly of carbon nanotube wafers for integrated nanoelectronic devices.

We report a general approach to overcome the enormous obstacle of the integration of CNTs into devices by bonding single-walled carbon nanotubes (SWNTs) films to arbitrary substrates and transferring them into densified and lithographically processable "CNT wafers". Our approach allows hierarchical layer-by-layer assembly of SWNTs into organized three-dimensional structures, for example, bidirectional islands, crossbar arrays with and without contacts on Si, and flexible substrates. These organized SWNT structures can be integrated with low-power resistive random-access memory.

Magnetic properties of monodispersed Ni/NiO core-shell nanoparticles.

We have recently developed a method to fabricate monodispersed Ni/NiO core-shell nanoparticles by pulsed laser ablation. In this report, the size-dependent magnetic properties of monodispersed Ni/NiO core-shell nanoparticles were investigated. These nanoparticles were formed in two steps. The first was to fabricate a series of monodispersed Ni nanoparticles of 5 to 20 nm in diameter using a combination of laser ablation and size classification by a low-pressure differential mobility analyzer (DMA). The second step was to oxidize the surfaces of the Ni particles in situ to form core-shell structures. A superconducting quantum interference device (SQUID) magnetometer was used to measure the magnetic properties of nanostructured films prepared by depositing the nanoparticles at room temperature. Ferromagnetism was observed in the magnetic hysteresis loop of the nanostructured films composed of core-shell nanoparticles with core diameters smaller than the superparamagnetic limit, which suggests the spin of Ni core was weakly exchange coupled with antiferromagnetic NiO shell. In contrast, smaller nanoparticles with core diameters of 3.0 nm exhibited superparamagnetism. The drastic change in the hysteresis loops between field-deposited and zero-field-deposited samples was attributable to the strong anisotropy that developed during the magnetic-field-assisted nanostructuring process.