TL;DR
This paper reveals that the bandgap of graphene nanoribbons (GNRs) can be significantly reduced (by up to 1 eV) when they are placed on a substrate, even if the interaction is weak. This effect, driven by dielectric screening from the substrate, has crucial implications for designing GNR-based electronic devices, suggesting a counter-intuitive approach for interconnects versus transistors.
What Problem Does This Paper Address?
Graphene nanoribbons (GNRs) are promising materials for next-generation electronics due to their tunable electronic properties, particularly their bandgap. The bandgap in GNRs is influenced by two main factors: the inherent electron-electron interactions amplified by quantum confinement, and the dielectric screening from their surrounding environment. While theoretical studies have extensively explored the first factor, the impact of dielectric screening from substrates on GNR bandgaps has been largely overlooked. This is a critical gap because most experimental measurements and potential applications involve GNRs supported on various substrates. Understanding this substrate effect is crucial for accurately predicting and controlling GNR electronic behavior.
Methods
The researchers employed the GW approximation, a sophisticated many-body perturbation theory method, to calculate the quasiparticle bandgaps of GNRs. This approach goes beyond standard Density Functional Theory (DFT) by explicitly accounting for electron-electron interactions and their screening effects, providing a more accurate description of electronic bandgaps. They also used the local density approximation (LDA) for initial electronic structure calculations and the Trouiller-Martins norm-conserving pseudopotentials and Teter-Pade parameterization for the exchange-correlation functional. For geometry optimization and to incorporate van der Waals (vdW) interactions, they utilized the Vienna Ab Initio Simulation Package (VASP) with PAW pseudopotentials and the PBE exchange-correlation functional, along with the DFT-D2 method. The screening was calculated using the plasmon-pole model. To ensure convergence, two monolayers of the substrate were included in the simulation.
Datasets
Not discussed in the paper. The study relies on theoretical calculations rather than experimental datasets.
Key Insights and Ideas
- Substrate-induced bandgap suppression is significant: Even weak van der Waals interactions with a substrate can drastically reduce the GNR bandgap, by as much as 1 eV. This is a substantial change, comparable to the bandgap itself in some cases.
- Dielectric screening is the dominant factor: The reduction in bandgap is primarily attributed to the dielectric screening provided by the substrate, which modifies the self-energy of electrons in the GNR. This screening effect is more pronounced than the inherent electron-electron interactions within the GNRs themselves when on a substrate.
- Edge geometry matters: The degree of bandgap reduction is strongly dependent on the edge geometry of the GNRs. This is due to how electronic confinement varies with edge structure, influencing the localization and delocalization of electron wavefunctions.
- Substrate dielectric constant is key: GNRs deposited on substrates with higher dielectric constants (like graphite) exhibit larger bandgap reductions compared to those on lower dielectric constant substrates (like hexagonal boron nitride, hBN).
- Counter-intuitive design implications: For GNRs used as interconnects, high-κ (high dielectric constant) materials are beneficial to reduce bandgaps and improve performance. Conversely, for GNRs used as channel materials in transistors, low-κ dielectrics are preferred to maintain larger bandgaps. This is the opposite of current trends in silicon-based electronics.
Key Results
- A free-standing armchair GNR (AGNR) with a width of 0.68 nm has a bandgap of 1.72 eV. When deposited on a weakly interacting hBN substrate, its bandgap is reduced to 0.93 eV, a reduction of 0.79 eV.
- For AGNR-6 (width 0.8 nm), the GW bandgap of the free-standing structure is 2.89 eV, while on hBN it is 2.27 eV, a reduction of 0.62 eV.
- AGNR-6 deposited on graphite shows a bandgap reduction of 1.2 eV compared to its free-standing state, significantly larger than the reduction on hBN.
- Zigzag GNRs (ZGNRs) exhibit a bandgap reduction of approximately 0.62 eV on hBN, which is largely independent of their width due to edge-localized wavefunctions.
Limitations and Future Work
- Weak interaction assumption: The study focuses on weakly interacting substrates. The behavior on strongly interacting substrates might differ.
- Specific GNR types: While the study covers AGNRs and ZGNRs, the findings might need to be extended to other GNR configurations.
- Experimental validation: The paper mentions recent experimental studies supporting their predictions, but further experimental validation across a wider range of substrates and GNR types would be beneficial.
The authors suggest that these findings can be exploited to design new electronic materials beyond carbon nanostructures, where long-range polarization effects can be used to tune electronic and magnetic properties of low-dimensional systems. They also highlight the potential for creating novel electronic devices by strategically choosing substrate materials to control GNR bandgaps for specific applications.
Summary updated