Although the inference task has been successfully demonstrated using the well-trained analogous conductance states, the neuromorphic hardware system can further be made to be energy efficient by making a device environment, where the weight update can be driven by an identical pulse scheme,[60] as discussed earlier. As identical pulses are successively addressed to the HfO2-based RRAM, the asymmetric response of the conductance due to nonlinear potentiation was observed,[56] which was an exactly opposite property of the PCM. Once abrupt conductance jumped at the initial pulse due to the formation of the filament, no more conductance increase was observed in the potentiation. The conductance was adjusted by the number of negative pulses and the slope of the decrease in conductance was determined by the amplitude and width of the pulse. A microscopic physical description of the RRAM that investigated the link between the filament evolution and the electrical behavior revealed that the formation of a strong filament caused the binary state during the potentiation.[61] In contrast, it was discovered that an alternative scenario, where the radial size of the filament is changed, is preferred to have a linear current response. The first attempt was to engineer the filament dynamics from the next cycle as desired because an abruptly grown filament in the initial state was difficult to control in working principle. Introducing an additional barrier layer of AlOx featuring a slower oxygen mobility compared with that of the HfO2 caused the dissolution process of the filament during the reset to be retarded, as shown in Figure 4.[56] It resulted in an incompletely disconnected filament. In the subsequent set cycle, the weakest constriction part of the filament, where the bilayer was contacted, was to be a plausible switching region by moving back-and-forth in the vacancy while the filament was still connected between two electrodes. Instead of growing in a vertical direction, the lateral expansion of the filament was discovered to be facilitated to update the conductance linearly, depending on the identical pulses. Other methodologies to manage the generation and migration of the oxygen vacancies in the initial stages, prior to these vacancies making the strong filament, have been proposed. By using a thermal barrier of TaOx with low thermal conductivity, the heat that is produced during the device operation can be confined into the HfO2 layer.[62] The heat spreads the distribution of the vacancies extensively while the vacancies were electrically driven to form the filament as is normal. The laterally expanded filament shape seemingly enabled the analog set transition in the IV curve and pulse switching. In addition, to exploit the temperature as another kinetic terminology in ion transport, formation energy of the vacancies was reduced to lower the probability of generating the vacancies using an electric field.[63] It was realized that by incorporating dopants into the HfO2 matrix, bonding strength was reduced. The uniformly distributed dopants facilitated the broad making of multiple filaments, resulting in analogously updated behaviors in both polarities. Even at a high temperature, the multiple states were distinguishable, ensuring the information at the peripheral sensing circuit was accurate. Thus, hardware systems with eight processing blocks comprising 128 × 16 TaOx/HfO2-based 1T–1R analog synapse arrays were successfully integrated to implement a five-layer convolutional neural network to perform MNIST image recognition.[58] The clear distribution of 1024 RRAM devices in 5-bit state within the current range of 0.4–4 μA without any overlap was also achieved by an identical pulse train with a substantially fast speed of 50 ns. Consequently, a high accuracy of more than 96% can be achieved. More importantly, the neuromorphic systems indicated that more than two orders of magnitude resulted in better power efficiency, whereas one order of magnitude resulted in better performance density compared with the CPU-based accelerator.
Similar hardware performance was also verified through mass-produced Ta2O5/TaOx-based 1T–1R synapses.[59] The uniform analog states linearly tuned from 20 to 50 μA with a verification technique that allowed a maximum recognition accuracy of 90% on the MNIST database. The 180 nm Ta2O5/TaOx-based RRAM exhibited a similar number of synapses per unit area when compared with a 65 nm SRAM. However, due to the reduced operational power of the RRAM device, the efficiency in performance and acceleration inference was more than three times, which was sufficiently high to enable a real-time recognition service. Furthermore, due to the local filamentary switching, the RRAM was scaled in a 40 nm test-chip, the efficiency running the neural network workloads can be further improved. In Table 1, reported array-level RRAM-based synapses were compared to identify the normal range of the conductance states and the pulse conditions that were typically required to control the states in most of the HfO2 device stacks.