7 November 2018 in Nature Communications
Neuromorphic networks of artificial neurons and synapses can solve computationally hard problems with energy efficiencies unattainable for von Neumann architectures. For image processing, silicon neuromorphic processors outperform graphic processing units in energy efficiency by a large margin, but deliver much lower chip-scale throughput. The performance-efficiency dilemma for silicon processors may not be overcome by Moore’s law scaling of silicon transistors. Scalable and biomimetic active memristor neurons and passive memristor synapses form a self-sufficient basis for a transistorless neural network. However, previous demonstrations of memristor neurons only showed simple integrate-and-fire behaviors and did not reveal the rich dynamics and computational complexity of biological neurons. Here we report that neurons built with nanoscale vanadium dioxide active memristors possess all three classes of excitability and most of the known biological neuronal dynamics, and are intrinsically stochastic. With the favorable size and power scaling, there is a path toward an all-memristor neuromorphic cortical computer.
1 July 2001 in Review of Scientific Instruments
We present a method for simultaneously measuring the specific heat and thermal conductivity of a rod- or filament-like specimen using a way similar to a four-probe resistance measurement. The specimen in this method needs to be electrically conductive and with a temperature-dependent resistance, for acting both as a heater to create a temperature fluctuation and as a sensor to measure its thermal response. With this method, we have successfully measured the specific heat and thermal conductivity of platinum wire specimens at cryogenic temperatures, and measured those thermal quantities of tiny carbon nanotube bundles some of which are only in mass.
1 April 1999 in Physical Review B
The specific heat and thermal conductivity of millimeter-long aligned carbon multiwall nanotubes (MWNT’s) have been measured. As a rolled-up version of graphene sheets, a MWNT of a few tens nm diameter is found to demonstrate a strikingly linear temperature-dependent specific heat over the entire temperature range measured (10–300 K). The results indicate that interwall coupling in MWNT’s is rather weak compared with its parent form, graphite, so that one can treat a MWNT as a few decoupled two-dimensional single wall tubules. The thermal conductivity is found to be low, indicating the existence of substantial amounts of defects in the MWNT’s prepared by a chemical-vapor-deposition method.
18 June 2013 in IEEE Transactions on Electron Devices
A key requirement for using memristors in circuits is a predictive model for device behavior that can be used in simulations and to guide designs. We analyze one of the most promising materials, tantalum oxide, for high density, low power, and high-speed memory. We perform an ensemble of measurements, including time dynamics across nine decades, to deduce the underlying state equations describing the switching in Pt/TaOx/Ta memristors. A predictive, compact model is found in good agreement with the measured data. The resulting model, compatible with SPICE, is then used to understand trends in terms of switching times and energy consumption, which in turn are important for choosing device operating points and handling interactions with other circuit elements.
12 March 2012 in Applied Physics Letters
Although TaOx memristors have demonstrated encouraging write/erase endurance and nanosecond switching speeds, the linear current-voltage (I-V) characteristic in the low resistance state limits their applications in large passive crossbar arrays. We demonstrate here that a TiO2-x/TaOx oxide heterostructure incorporated into a 50 nm× 50 nm memristor displays a very large nonlinearity such that I(V/2) ≈ I(V)/100 for V ≈ 1 volt, which is caused by current-controlled negative differential resistance in the device.
26 January 2011 in Applied Physics A