Effect of Edge Roughness on Static Characteristics of Graphene Nanoribbon Field Effect Transistor

Yaser Banadaki, Ashok Srivastava
2016 Electronics  
In this paper, we present a physics-based analytical model of GNR FET, which allows for the evaluation of GNR FET performance including the effects of line-edge roughness as its practical specific non-ideality. The line-edge roughness is modeled in edge-enhanced band-to-band-tunneling and localization regimes, and then verified for various roughness amplitudes. Corresponding to these two regimes, the off-current is initially increased, then decreased; while, on the other hand, the on-current is
more » ... continuously decreased by increasing the roughness amplitude. However, the device-level analysis requires extensive computational time; therefore, the same statistical approach cannot be used for circuit-level simulations. The ideal smooth-edge GNR FETs give an estimation of the upper bound performance, however, the line-edge roughness needs to be considered for practical GNR FETs which deteriorate their performance. A semi-analytical model for GNR FET with perfectly smooth edges was developed in [10], which involved numerical integrations; thereby, it cannot be used for circuit simulation. In [11], a circuit model was implemented based on lookup table techniques to use the results of device-level quantum transport simulations for circuit simulations. However, with a single change in a design parameter, the intensive device-level simulations need to be repeated to rebuild the model accordingly, which makes it inappropriate for evaluating the optimized design parameters of GNR FET circuits. A SPICE-compatible model of GNR FET including the edge roughness is presented in [12] . In this model, the effect of rough edges on the increasing leakage current of GNR (N,0) is considered by effective bandgap due to the bandgap of GNR (N-1,0) , while the real GNRs with rough edges are composed of all neighboring GNRs. Also, the effect of rough edges on decreasing on-current was modeled by a fitting equation regardless of the physical scattering mechanisms in a GNR channel. In addition, this model cannot capture the effect of large line-edge roughness on localization of carriers, which tends to reduce both off-and on-currents of GNR FETs. It has been shown both experimentally [13] and theoretically [14] that strong localization can appear in single layer GNRs for high line-edge roughness. In this work, we develop a physics-based analytical model for circuit simulation of GNR FETs. The band-to-band-tunneling (BTBT) from drain to channel regions can be important for small bandgap GNRs, which has been modeled by a current source in parallel with another current source for the thermionic current. The line-edge roughness in GNRs is modeled using an exponential autocorrelation function. The model incorporates the effect of edge states on the initial increase of BTBT and high edge scattering of carriers in a localization regime. The device-level simulation is performed to evaluate the static performance of GNR FETs in edge-enhanced BTBT and localization regimes. The results of our analytical model are verified by numerical results from accurate quantum transport simulations based on non-equilibrium Green function (NEGF) formalism. The organization of this paper is as follows: In Section 2, we describe the structure of GNR FETs in all-graphene architecture; Section 3 presents the model equations and equivalent circuit model of GNR FET for the circuit simulation in SPICE. Model validation is described in Sections 4 and 5; finally, the last section draws summarizing conclusions. GNR FET Structure
doi:10.3390/electronics5010011 fatcat:lpriyf7f5zhdtgeo4jjewtdb4a