Graphene Nanoribbons Based 5-Bit Digital-to-Analog Converter

摘要: McCulloch-Pitts neuron structures are comprised of a number synaptic inputs and decision element, called soma. In this paper, we propose 5-bit Graphene Nanoribbon (GNR)-based DAC to fulfill the role summation element featuring programmable input weights. The proposed GNR-based relies on: (i) GNR unit current cells (ii) logic thermometric decoding block. Our implementation is based on mapping structure's conductance using Matlab performing required SPICE analysis Verilog-A model. cell geometry bias conditions were chosen cell's map from which derived its $I_{ON}$ / $I_{OFF}$ ratio, as well transfer output characteristics, resembling classical MOSFET counterpart. By utilizing devices instead FinFET counterparts, reduction active area by up factor three can be achieved. Furthermore, achieved while maintaining comparable INL DNL performance that variant, i.e., [−0.196, 0.088] LSB [−0.809, 0.364] for operating at supply voltage only 0.2 V.