Minimization of crosstalk noise and delay using reduced graphene nano ribbon (GNR) interconnect

Abstract

In this paper, we proposed a side-contact reduced graphene nano-ribbon (SC-RGNR) interconnect model to reduce crosstalk noise and delay for next generation high performance integrated circuit (IC) design. The SC-RGNR interconnect structure is designed with latest 11 nm ITRS (International Technological Road Map for Semiconductor) technology node. A comparative crosstalk noise, delay and power consumption analysis has been performed with side-contact graphene nano-ribbon (SC-GNR) interconnect model. The SC-RGNR and SC-GNR interconnect both model parameters are applied in standard (STD) bus and staggered (STAGG) bus model with different interconnect lengths (10 μm–500 μm) and four standard mean free paths (MFP) (i.e., 300 nm, 419 nm, 1000 nm and 1200 nm) values. Five different switching conditions (i.e., SP1 to SP5) are applied on tri-interconnect driver-load circuit to check the performance of SC-RGNR and SC-GNR interconnect in-terms of crosstalk noise and delay. In our analysis, it is shown that, the higher MFP based SC-RGNR interconnect shows less dynamic delay compared with SC-GNR interconnect for both STD and STAGG interconnect model with worst-case switching condition (SP5) and best-case switching condition (SP2). For STD interconnect model with worst-case switching condition (SP5) the dynamic delay of SC-RGNR interconnect is ∼1.33–2.62 × lesser than SC-GNR interconnect. Similarly, for STAGG model with worst-case switching condition (SP5) the dynamic delay of SC-RGNR interconnect is ∼1.26–2.16 × lesser than SC-GNR interconnect. Although, the STAGG interconnect model performs ∼1.5–2 × faster than STD interconnect model in terms of delay. In our analysis, it is also observed that, the output noise peak is ∼2.91–17.5 × less in SC-RGNR compared with SC-GNR at 500 μm interconnect length and higher MFP. Our analysis also shows that, the switching power consumption in SC-RGNR is ∼1.66–2.2 × less than SC-GNR for both STD and STAGG model for both best and worst-case switching condition (SP2 and SP5).