Highly Efficient Midinfrared On-Chip Electrical Generation of Graphene Plasmons by Inelastic Electron Tunneling Excitation

Inelastic electron tunneling provides a low-energy pathway for the excitation of surface plasmons and light emission. We theoretically investigate tunnel junctions based on metals and graphene. We show that graphene is potentially a highly efficient material for tunneling excitation of plasmons because of its narrow plasmon linewidths, strong emission, and large tunability in the midinfrared wavelength regime. Compared to gold and silver, the enhancement can be up to 10 times for similar wavelengths and up to 5 orders at their respective plasmon operating wavelengths. Tunneling excitation of graphene plasmons promises an efficient technology for on-chip electrical generation and manipulation of plasmons for graphene-based optoelectronics and nanophotonic integrated circuits.

Figure 1

Schematic of the plasmon excitation from a metal-insulator-metal or -graphene tunnel junction. (a) Metal-insulator-metal tunnel junction. (b) Metal-insulator-graphene tunnel junction.

Figure 2

(a) Mechanism of the inelastic electron tunneling in generating a gap plasmon. Upon applying a bias voltage $V_z$, the energy level of the left electrode is shifted and electrons tunnel inelastically to the right electrode, exciting a gap plasmon with energy $\hslash \omega =e{V}_z$. (b) Model of the 2D FDTD simulation. The gap-plasmon mode is modeled as a line dipole source with a normalized power of 1 W at all wavelengths. The power is coupled out as SPP modes and propagates along the metal or graphene surface.

Figure 3

Comparison of intensity spectrum of the $\mathrm{Al}\text{−}{\mathrm{Al}}_2{\mathrm{O}}_3\text{−}\mathrm{Ag}$ IETP system: (a) simulated and (b) experimental [4].

Figure 4

Comparison of simulated (solid lines) and experimental (dashed lines) intensity spectrum of Au-air-Au IETP system under bias of (a) 2 V [6] and (b) 2.5 V [10].

Figure 5

IETP gap plasmons from tunneling from an aluminum tip to metals (dashed lines) and to graphene of various Fermi levels (solid lines).

Figure 6

(a) Schematic of the FDTD electromagnetic simulation structure. (b),(c),(d) are coupling efficiencies of gap plasmons to gold, silver, and graphene substrates, respectively. (e),(f),(g) are the total efficiencies in generating SPPs in gold, silver, and graphene, respectively.

Figure 7

(a) Frequency linewidths for the gap plasmons for graphene, gold, and silver. Graphene gap plasmons have narrower linewidths compared to metal gap plasmons. (b) and (c) are gap-plasmon power plots for each plasmon vector $q$ and photon energy in eV, for gold and 0.4-eV graphene, respectively. The gap-plasmon power for graphene is more confined along the plasmon dispersion curve compared to gold. These results show that graphene exhibits a much larger $Q$ factor and is tuneable compared to gold and silver materials.