Control of 2D plasmons in the topological insulator Bi2Se3 with highly crystalline C60 overlayers

Topological Insulators (TIs) present an interesting materials platform for nanoscale, high frequency devices because they support high mobility, low scattering electronic transport within confined surface states. However, a robust methodology to control the properties of surface plasmons in TIs has yet to be developed. We propose that charge transfer between Bi$_2$Se$3$ and crystalline C${60}$ films may provide tunable control of the two-dimensional plasmons in Bi$_2$Se$_3$. We have grown heterostructures of Bi$_2$Se$3$/C${60}$ with exceptional crystallinity. Electron energy loss spectroscopy (EELS) reveals significant hybridisation of $π$ states at the interface, despite the expectation for only weak van der Waals interactions, including quenching of 2D plasmons. Momentum-resolved EELS measurements are used to probe the plasmon dispersion, with Density Functional Theory predictions providing an interpretation of results based on interfacial charge dipoles. Our measurements suggest a robust methodology for tuneable TI interfaces that can be engineered for plasmonic applications in computing, communications and sensing.

💡 Research Summary

The authors present a systematic study of how highly crystalline fullerene (C₆₀) overlayers can be used to tune two‑dimensional (2D) plasmons in the topological insulator Bi₂Se₃. Using a multi‑functional molecular beam epitaxy (MBE) system, they first grow 15 nm (≈15 quintuple layers) of Bi₂Se₃ on c‑plane sapphire, then deposit an ≈80 nm thick C₆₀ film at low temperature via a Knudsen cell. Cross‑sectional lamellae are prepared by focused ion beam (FIB) thinning to ~35 nm and examined with scanning transmission electron microscopy (STEM). The Bi₂Se₃/C₆₀ interface is atomically sharp, and the C₆₀ layer exhibits exceptionally high crystallinity, with only occasional stacking faults spaced ~20 nm apart.

Electron energy‑loss spectroscopy (EELS) mapping across the heterostructure reveals the characteristic π⁻ (6.8 eV), π⁺ (10 eV) and π + σ (≈25 eV) plasmons of bulk C₆₀, as well as the 7 eV (π) and 16.8 eV (π + σ) plasmons of bulk Bi₂Se₃. At the Bi₂Se₃/C₆₀ interface a new loss feature appears at 6.3 eV, positioned midway between the C₆₀ π⁻ and Bi₂Se₃ π modes. The authors label this a “hybrid π‑plasmon” and attribute it to strong interfacial charge transfer: density‑functional theory (DFT) calculations show electron accumulation on the C₆₀ cage and depletion in the topmost Bi₂Se₃ layer, creating a localized dipole and inducing Rashba‑type spin splitting of the surface state.

In contrast, the Bi₂Se₃/Al₂O₃ interface displays a well‑defined 2D π‑plasmon that emerges within ≈2 QL (≈2 nm) of the interface and follows the expected q¹ᐟ² dispersion of a true 2D plasmon. Momentum‑resolved EELS (QEELS) performed along the Γ‑M direction (q = 0–1.43 Å⁻¹) confirms that the Bi₂Se₃/Al₂O₃ π‑plasmon disperses as a 2D mode, while the hybrid π‑plasmon at the Bi₂Se₃/C₆₀ interface shows essentially zero dispersion, indicating a localized dipolar oscillation rather than a propagating surface charge wave.

The authors also compare experimental spectra with analytical models based on bulk dielectric functions. The model reproduces the Al₂O₃‑capped Bi₂Se₃ spectra well, but fails to capture the red‑shifted and intensified hybrid π‑plasmon at the Bi₂Se₃/C₆₀ interface because it does not account for molecular orbital hybridisation with the TI surface π electrons. This mirrors observations in graphene where π‑plasmon energies are highly sensitive to doping and hybridisation.

Key insights from the work include: (1) C₆₀, owing to its high electron affinity, can draw charge from the TI surface, creating an interfacial dipole and Rashba‑type band modifications; (2) this charge transfer quenches the conventional 2D Dirac plasmon of Bi₂Se₃, replacing it with a non‑dispersive hybrid mode; (3) the hybrid mode’s intensity is enhanced by surface‑enhanced EELS (SEELS), reflecting resonant coupling between molecular interband transitions and the TI surface electrons; (4) the approach provides a robust, chemically stable, and potentially reversible method to electrically tune TI plasmonic response, unlike alkali‑metal dopants that are highly reactive.

Overall, the study demonstrates that highly ordered organic overlayers can serve as an effective, tunable gate for topological‑insulator plasmonics. By engineering charge transfer and orbital hybridisation at the interface, one can switch on/off 2D plasmon modes, tailor dispersion, and introduce new hybrid excitations. This opens pathways toward TI‑based plasmonic devices for terahertz/infrared computing, high‑speed communication, and ultrasensitive sensing, where the plasmonic response can be modulated via external bias or by selecting appropriate molecular overlayers.