Proceedings Article | 15 October 2012
KEYWORDS: Quantum dots, Stochastic processes, Interference (communication), Electrons, Thermal effects, Capacitance, Tolerancing, Monte Carlo methods, Resistance, Device simulation
This paper reports a study of stochastic resonance in a huge quantum dot network for single-electron (SE) circuits. Such
circuits, which are controlled by the Coulomb blockade, are one type of next-generation information-processing device.
However, they are very sensitive to noises such as thermal noise and device mismatch noise. Thus, we introduce the
stochastic resonance phenomenon into the circuit to improve its noise tolerance. Stochastic resonance is a phenomenon
that was discovered in the brains of living things in noisy environments and was modeled for neural networks. When the
phenomenon occurs, its harnessing of noise energy makes weak signals become clear. In current research, SE devices
that operate with stochastic resonance have been reported. However, signals were attenuated in particularly noisy
environments. In contrast, it was reported that a huge molecular network amplified weak signals by harnessing noise
energy. The report said the current-voltage characteristics of the molecular network described the Coulomb blockade
under a noisy environment. Thus, a huge quantum dot network that is partly similar to a molecular network is expected
to amplify the weak signal harnessing noise, when the current-voltage characteristics of the network show the Coulomb
blockade. To confirm this, in this study we use the Monte Carlo method to simulate the noisy-environment operation of a
quantum dot network comprising quantum dots and tunneling junctions. We observe the current-voltage characteristics
of the network, when changing the network size (5×5, 10×10, and 100×100) and the noise intensity (0 K, 2 K, 5 K, and
10 K for operating temperature, and 0%, 5%, 10%, and 30% for device mismatch). As a result, we are able to observe the
Coulomb blockade under the appropriate noise strength, which in this study is 5 K or less with thermal noise, and 30%
with device mismatch. From the results, we conclude the network operates correctly under appropriate noise strength.
Moreover, the noise energy amplifies the network current, indicating that SE circuits can function as signal-amplifying
devices.
