An Efficient Method to Reduce Crosstalk for Multi-layered GNR Interconnects at 32 nm Technology

Abstract

In recent, the coupling capacitance becomes prominent in deep submicron and nanotechnology. It can significantly affect the performance of a digital circuit that causes digital logic fault and degrades the performance of Integrated Circuits (ICs). To address this problem, the research paper introduces two shielding methodologies (active and passive shielding) of a multi-layered graphene nanoribbon (MLGNR). An RLC based multi-conductor transmission (MTL) line model is simplified to an equivalent single conductor (ESC) to address the impact of the crosstalk-induced delay. The model parameters are used with reference to the ITRS 32 nm technology in order to demonstrate the performance for different interconnect spacing and lengths. For better understanding, shielded and unshielded methodologies are demonstrated to provide a comparative analysis and improvement with respect to unshielded interconnect lines. Irrespective of the shielding methodology, it is evident that an active shielding provides an improved crosstalk-induced delay compared to passive shielding for different interconnect spacing. At 32nm, the simulation results demonstrate that an active shielded interconnect outperforms passive and without shielding techniques by on an average 6.31% and 44.51% for the crosstalk-induced delay, respectively for different spacing at global lengths. In addition to that, active shielding provides an improved power delay product (PDP) of approximately 6.75% and 16.83% than passive shielded interconnects for 5nm and 20nm spacing, respectively at 1000µm length. Therefore, in order to accomplish the constraint of next-generation technology, an MLGNR can be adopted as a desirable shielding nanomaterial.