Band Gaps and Quasiparticle Energies of Graphene Nanoribbons

Yang Li, Cheol Hwan Park, Young‐Woo Son, Marvin L. Cohen, Steven G. LouieBulletin 2007

Abstract: We present calculations of the quasiparticle energies and band gaps of graphene nanoribbons (GNRs) carried out using a first-principles many-electron Green's function approach within the GW approximation. Because of the quasi-one-dimension nature of a GNR, electron-electron interaction effects due to the enhanced screened Coulomb interaction and confinement geometry greatly influence the quasiparticle band gap. Compared with previous tight-binding and density functional theory studies, our calculated quasiparticle band gaps show significant self-energy corrections for both armchair and zigzag GNRs, in the range of 0.5-3.0 eV for ribbons of width 2.4-0.4 nm. The quasiparticle band gaps found here suggest that use of GNRs for electronic device components in ambient conditions may be viable.

Introduction

While graphene is celebrated for its incredible conductivity, its lack of a natural band gap—the "off" switch required for transistors—limits its use in digital electronics. Graphene nanoribbons (GNRs), which are thin strips of graphene, solve this by using quantum confinement to create a band gap.

However, standard computational tools like Density Functional Theory (DFT) often fail to accurately predict these gaps. In low-dimensional systems like GNRs, electrons have limited space to move, which enhances their interactions. Standard DFT tends to underestimate these "self-energy" effects, leading to predicted band gaps that are much smaller than what would be observed experimentally. This study provides the first accurate, first-principles calculation of these gaps using the GW approximation to determine if GNRs are truly suitable for electronic applications.

Methods

The authors employ a many-body perturbation theory approach known as the GW approximation. This method calculates the Green’s function ($G$) and the screened Coulomb interaction ($W$). Unlike standard DFT, which treats electrons as moving in an average potential, GW accounts for the complex "quasiparticle" nature of electrons—how an electron moves while being influenced by the surrounding electron cloud.

The computational workflow began with a standard Local Density Approximation (LDA) calculation to establish the initial electronic structure. To simulate an isolated ribbon within a periodic supercell, a mathematical truncation of the Coulomb interaction was used to prevent artificial interactions between periodic images. Furthermore, the edges of the ribbons were passivated with hydrogen atoms to eliminate unphysical electronic states from dangling bonds.

Structural Models and Classification

This theoretical study evaluated structural models for two primary types of GNRs: Armchair GNRs (AGNRs) and Zigzag GNRs (ZGNRs).

Atomic structures of (a) armchair and (b) zigzag graphene nanoribbons (GNRs) with width w.
  • Armchair GNRs (AGNRs): Modeled in widths ranging from 0.4 nm to 2.4 nm. These fall into three distinct families based on their width ($n = 3p, 3p+1, \text{or } 3p+2$ dimer lines).
  • Zigzag GNRs (ZGNRs): Specifically focused on 8-ZGNR and 12-ZGNR structures to study edge-state effects and spin polarization.

Key Insights

In one-dimensional systems, "screening" (the ability of other electrons to dampen the electric field of a single electron) is much weaker than in 3D materials. This leads to a massive increase in the self-energy correction, making the real band gap much larger than the DFT-predicted one.

While GW corrections shift the absolute values of the gaps upward, the relative hierarchy between the three AGNR families remains consistent with LDA predictions. Additionally, the researchers found that the "electronic" width of a ribbon is slightly larger than its physical carbon-to-carbon width because electron wavefunctions "leak" slightly beyond the edge atoms.

Results and Discussion

The GW calculations revealed that self-energy corrections add between 0.5 eV and 3.0 eV to the band gaps predicted by LDA. For ribbons with widths between 0.4 and 2.4 nm, the quasiparticle band gaps were found to be in the range of 0.5 to 3.0 eV.

Calculated band gaps of armchair GNRs as a function of width, comparing LDA (open symbols) and GW (filled symbols) results for the three families.

For context, Silicon has a gap of ~1.1 eV; the GNR gaps identified here are large enough to ensure stable device operation at room temperature without excessive thermal leakage. Even for zigzag ribbons, which simpler theories predicted to be metallic, the inclusion of spin polarization and GW corrections resulted in significant gaps (enlarged by 0.8 to 1.5 eV over previous estimates). Furthermore, the GW bands are generally "steeper" than LDA bands, implying a smaller effective mass for charge carriers and suggesting superior carrier mobility.

Limitations and Future Work

These calculations were performed for GNRs in a vacuum. In a practical device, the ribbon would sit on a substrate (like $SiO_2$), which would provide additional screening and likely reduce the band gap. Additionally, direct experimental comparison remains a challenge as current fabrication techniques typically produce wider ribbons (15–90 nm) than the sub-3nm models studied here. Future research must also consider the impact of different edge chemistries, as real-world processing may introduce functional groups other than hydrogen.

Summary updated

BibTeX
@article{li2007band,
  author = {Yang Li and Cheol Hwan Park and Young‐Woo Son and Marvin L. Cohen and Steven G. Louie},
  title = {{Band Gaps and Quasiparticle Energies of Graphene Nanoribbons}},
  journal = {Bulletin},
  year = {2007},
}

Loading related content...

Figure