Using memristive properties common for the titanium dioxide thin film devices, we designed a simple write algorithm to tune device conductance at a specific bias point to 1% relative accuracy (which is roughly equivalent to 7-bit precision) within its dynamic range even in the presence of large variations in switching behavior. The high precision state is nonvolatile and the results are likely to be sustained for nanoscale memristive devices because of the inherent filamentary nature of the resistive switching. The proposed functionality of memristive devices is especially attractive for analog computing with low precision data. As one representative example we demonstrate hybrid circuitry consisting of CMOS summing amplifier and two memristive devices to perform analog multiply and accumulate computation, which is a typical bottleneck operation in information processing. A natural way to tackle variations in switching behavior is to utilize active feedback scheme, e.g. applying iterative write and read (test) pulses to converge to certain desired conductive state of the device. Such scheme has been successfully applied to phase change memories to achieve multilevel memory operation [Pap11,Bed09]. A similar idea to use closedloop circuitry has been proposed and theoretically simulated using Spice model for TiO 2 devices [Yi11]. In this paper, we experimentally demonstrate a simple feedback algorithm which takes into account specific memristive behavior of titanium dioxide devices to tune resistance state of the device within 1% relative accuracy within all the dynamic range. We then use our algorithm to demonstrate one of the most important operations in information processing -analog multiply and accumulate (MAC).
KeywordsThe Pt/TiO 2 (30nm)/Pt devices have been implemented in "bone-structure" geometry with A more accurate control of the device is possible using sequence of relatively large amplitude write pulses followed by smaller non-disturbing read pulses [Pic09]. Note that for all experiments described below we use voltage-controlled pulses for SET switching also. The main reason is that current-controlled switching (or alternatively the utilization of a compliance transistor) is not compatible for large scale crossbar circuits, even though it could be more natural for SET switching in the context of single devices because it allows avoiding overshooting and overheating. In particular, the measurement is composed of two different sequences of pulses: (i) the read pulses of -200 mV and 1 ms width are used to probe the state of the device which is represented by the resistance or by current measured at -200 mV) and (ii) the write pulses which are used to change the state of the memristive device, whether by changing the pulse width and/or the pulse amplitude. The two pulses (read and write) are alternated at a frequency of 0.5 Hz to prevent the accumulative Joule heating effect from single write pulses. highlights that the switching dynamics is exponential with voltage for SET switching and roughly f...
In a neuromorphic computing system, the complex CMOS neuron circuits have been the bottleneck for efficient implementation of weighted sum operation. The phenomenon of metal-insulator-transition (MIT) in strongly correlated oxides, such as NbO2, has shown the oscillation behavior in recent experiments. In this work, we propose using a MIT device to function as a compact oscillation neuron, achieving the same functionality as the CMOS neuron but occupying a much smaller area. Pt/NbOx/Pt devices are fabricated, exhibiting the threshold switching I-V hysteresis. When the NbOx device is connected with an external resistor (i.e., the synapse), the neuron membrane voltage starts a self-oscillation. We experimentally demonstrate that the oscillation frequency is proportional to the conductance of the synapse, showing its feasibility for integrating the weighted sum current. The switching speed measurement indicates that the oscillation frequency could achieve >33 MHz if parasitic capacitance can be eliminated.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.