Emulation of Pavlovian conditioning and pattern recognition through fully connected neural networks using Holmium oxide (Ho2O3) based synaptic RRAM device

In this manuscript, we report on the paramagnetic Ho2O3-based synaptic resistive random-access memory device for the implementation of neuronal functionalities such as long-term potentiation, long-term depression and spike timing dependent plasticity respectively. The plasticity of the artificial synapse is also studied by varying pulse amplitude, pulse width, and pulse interval. In addition, we could classify handwritten Modified National Institute of Standards and Technology data set (MNIST) using a fully connected neural network (FCN). The device-based FCN records a high classification accuracy of 93.47% which is comparable to the software-based test accuracy of 97.97%. This indicates the highly optimized behavior of our synaptic device for hardware neuromorphic applications. Successful emulation of Pavlovian classical conditioning for associative learning of the biological brain is achieved. We believe that the present device consists the potential to utilize in neuromorphic applications.

α-Fe2O3-based artificial synaptic RRAM device for pattern recognition using artificial neural networks

We report on theα-Fe₂O₃-based artificial synaptic resistive random access memory device, which is a promising candidate for artificial neural networks (ANN) to recognize the images. The device consists of a structure Ag/α-Fe₂O₃/FTO and exhibits non-volatility with analog resistive switching characteristics. We successfully demonstrated synaptic learning rules such as long-term potentiation, long-term depression, and spike time-dependent plasticity. In addition, we also presented off-chip training to obtain good accuracy by backpropagation algorithm considering the synaptic weights obtained fromα-Fe₂O₃based artificial synaptic device. The proposedα-Fe₂O₃-based device was tested with the FMNIST and MNIST datasets and obtained a high pattern recognition accuracy of 88.06% and 97.6% test accuracy respectively. Such a high pattern recognition accuracy is attributed to the combination of the synaptic device performance as well as the novel weight mapping strategy used in the present work. Therefore, the ideal device characteristics and high ANN performance showed that the fabricated device can be useful for practical ANN implementation.

Spin transfer torque bias (STTB) due to domain wall resistance in an infinitely long ferromagnetic nanowire

The shift of a magnetization loop along the magnetic field axis for a ferromagnetic (FM)/anti-ferromagnetic (AFM) system when it is cooled through Néel temperature of AFM layer is called exchange anisotropy or exchange bias. Here, using micromagnetic simulations we propose that spin transfer torque (STT) mechanism would indeed be helpful in realizing the shift of the magnetization loop along magnetic field axis through domain wall (DW) resistance for an infinitely long FM nanowire without having AFM layer, which we call as spin transfer torque bias (STTB). Essentially, STTB is realized on both positive and negative magnetic field axes by varying the angle between spin polarized current and Zeeman field from 0° to 180° respectively and the origin is attributed to helical motion of the DW. However, we do not see STTB at 90° due to coherent rotation of domain. We also ascertain that STTB is also a function of magnetic anisotropy, current density, polarization strength and non-adiabatic STT term. Variation in STTB for different FM systems such as Fe₂CoSi, Ni₈₀Fe₂₀and Fe is attributed to a change in DW width. We believe that present results would lead to a new dimension in the field of spintronics.

Graphene oxide based synaptic memristor device for neuromorphic computing

Brain-inspired neuromorphic computing which consist neurons and synapses, with an ability to perform complex information processing has unfolded a new paradigm of computing to overcome the von Neumann bottleneck. Electronic synaptic memristor devices which can compete with the biological synapses are indeed significant for neuromorphic computing. In this work, we demonstrate our efforts to develop and realize the graphene oxide (GO) based memristor device as a synaptic device, which mimic as a biological synapse. Indeed, this device exhibits the essential synaptic learning behavior including analog memory characteristics, potentiation and depression. Furthermore, spike-timing-dependent-plasticity learning rule is mimicked by engineering the pre- and post-synaptic spikes. In addition, non-volatile properties such as endurance, retentivity, multilevel switching of the device are explored. These results suggest that Ag/GO/fluorine-doped tin oxide memristor device would indeed be a potential candidate for future neuromorphic computing applications.

Bipolar resistive switching in HoCrO3 thin films

We report on the resistive switching characteristics of an HoCrO₃ (HCO) based memristor device. The device comprising Ag/HCO/fluorine doped tin oxide shows stable bipolar resistive switching with a good ON/OFF resistance ratio between high resistance state (HRS) and low resistance state (LRS). Furthermore, the device is capable to show excellent endurance and retentivity characteristics over a period of 30 days. The statistical distribution of the switching parameters (voltage and resistance) show a narrow distribution, hinting reliable memory performance and stability of the device. Impedance spectroscopy analysis of the HRS and LRS illustrates a bulk resistance effect, which is due to formation of multiple ionic conductive channels in the film with oxygen vacancies. Indeed, conducting channels formed by oxygen vacancies are further confirmed by calculating the temperature coefficient of resistance through resistance vs temperature measurements. We believe that these results will be helpful in developing future memory devices based on resistive switching.

Detection of bovine serum albumin using hybrid TiO2 + graphene oxide based Bio - resistive random access memory device

Bio - molecules detection and their quantification with a high precision is essential in modern era of medical diagnostics. In this context, the memristor device which can change its resistance state is a promising technique to sense the bio - molecules. In this work, detection of the Bovine Serum Albumin (BSA) protein using resistive switching memristors based on TiO2 and TiO2 + graphene oxide (GO) is explored. The sensitivity of BSA detection is found to be 4 mg/mL. Both the devices show an excellent bipolar resistive switching with an on/off ratio of 73 and 100 respectively, which essentially demonstrates that the device with GO, distinguishes the resistance states with a high precision. The enhanced performance in the GO inserted device (~ 650 cycles) is attributed to the prevention of multi-dimensional and random growth of conductive paths.

Remote control of resistive switching in TiO2 based resistive random access memory device

We report on the magnetic field control of a bipolar resistive switching in Ag/TiO2/FTO based resistive random access memory device through I-V characteristics. Essentially, in the presence of magnetic field and in the low resistance state, an abrupt change in the resistance of the device demands higher voltage, hinting that residual Lorentz force plays a significant role in controlling the resistance state. Endurance characteristics of the device infer that there is no degradation of the device even after repeated cycling, which ensures that the switching of resistance between 'off' and 'on' states is reproducible, reversible and controllable. Magnetic field control of 'on' and 'off' states in endurance characteristics suggest that this device can be controlled in a remote way for multi-bit data storage.

Ferromagnetism in graphene nanoribbons: split versus oxidative unzipped ribbons

Two types of graphene nanoribbons: (a) potassium-split graphene nanoribbons (GNRs), and (b) oxidative unzipped and chemically converted graphene nanoribbons (CCGNRs) were investigated for their magnetic properties using the combination of static magnetization and electron spin resonance measurements. The two types of ribbons possess remarkably different magnetic properties. While a low-temperature ferromagnet-like feature is observed in both types of ribbons, such room-temperature feature persists only in potassium-split ribbons. The GNRs show negative exchange bias, but the CCGNRs exhibit a "positive exchange bias". Electron spin resonance measurements suggest that the carbon-related defects may be responsible for the observed magnetic behavior in both types of ribbons. Furthermore, information on the proton hyperfine coupling strength has been obtained from hyperfine sublevel correlation experiments performed on the GNRs. Electron spin resonance finds no evidence for the presence of potassium (cluster) related signals, pointing to the intrinsic magnetic nature of the ribbons. Our combined experimental results may indicate the coexistence of ferromagnetic clusters with antiferromagnetic regions leading to disordered magnetic phase. We discuss the possible origin of the observed contrast in the magnetic behaviors of the two types of ribbons studied.