Modeling of Graphene Oxide Memory based on Hybridization State Modulation

Abstract

Graphene oxide (GO)-based resistive random access memory (RRAM) is one of the most promising emerging non-volatile memories for flexible electronics because of its simple structure and low fabrication cost. The reported switching mechanism can be classified into either localized or bulk effect. The localized effect is commonly observed in GO-RRAM with active electrode e.g. Cu and Al. The switching mechanism includes metallic conduction filament evolution or modification of interfacial resistance between GO layer and electrodes. On the other hand, the bulk effect involves hybridization state change of the entire GO layer and is typically observed in devices with inert electrode e.g. Pt and Au. Numerous compact models have been proposed for RRAM that are based on localized switching mechanism. However, compact modeling of GO-RRAM based on bulk mechanism is still lacking. Thus, this paper presents a compact model for GO-RRAM that relies on hybridization state modulation. The proposed model associates the resistance switching mechanism with the electricaldriven modification of sp2 clusters density in the switching layer. A SPICE model incorporating the proposed compact model is developed for parameter calibration. The simulated I-V characteristic exhibits strong correlation with the experimental data. Hereby, we regard the proposed compact model as an acceptable and close description of GO-RRAM operation that depends on the bulk switching mechanism.