Plasmonic hollow nanocavities tune exciton selectivity in monolayer MoS2

Researchers have unveiled a new class of hollow gold nanocavities that can selectively amplify specific excitonic transitions in atomically thin molybdenum disulfide, offering a geometry-based route to control how these remarkable two-dimensional semiconductors absorb and emit light. Writing in the journal Results in Physics, Abdullah Efe Yildiz and Emre Ozan Polat present detailed finite-difference time-domain simulations combined with first-principles excitonic calculations, showing that simply changing the inner radius of a hollow gold nanocylinder can shift which exciton—the A or the B—in a monolayer of MoS2 receives the lion’s share of the optical enhancement. The work addresses one of the most persistent bottlenecks in two-dimensional optoelectronics: although monolayer MoS2 hosts extraordinarily strong light-matter interactions and excitons that survive at room temperature, its atomic thinness means it absorbs and emits only a tiny fraction of the light that strikes it, limiting the performance of photodetectors, light-emitting devices and valleytronic circuits built from it.

The central innovation lies in the topology of the cavity itself. Unlike the metal-mirror gap cavities that have dominated previous high-enhancement demonstrations, the Turkish researchers modeled a vertically oriented hollow gold nanocylinder resting on a dielectric spacer atop a MoS2 monolayer supported by SiO2 on silicon. The hollow architecture is crucial because it supports coupled charge oscillations on both the inner and outer metal surfaces simultaneously, a phenomenon the authors describe through a plasmon hybridization picture analogous to that developed for metallic nanoshells. This dual-interface coupling produces hybrid plasmon modes with deeply subwavelength mode volumes and strong near-field confinement, and it introduces an additional degree of freedom for spectral engineering that solid nanostructures such as nanocubes cannot offer. Critically, the inner radius and wall thickness can be varied while keeping the outer footprint fixed, allowing the resonance wavelength and the spatial distribution of the near field to be tuned without redesigning the entire array.

The team quantified this tunability using a parameter they call the cavity aspect ratio, defined as the cavity height divided by the difference between the outer and inner radii. Increasing the aspect ratio drives a pronounced redshift of the localized surface plasmon resonance, the collective oscillation of conduction electrons that concentrates electromagnetic energy at the nanoscale. From a systematic parametric study, two optimized geometries emerged: a cavity with an inner radius of 30 nanometers and an aspect ratio of 5, whose resonance aligns with the A exciton of MoS2, and a cavity with an inner radius of 20 nanometers and an aspect ratio of 3.33, aligned with the higher-energy B exciton. The distinction between the A and B excitons arises from the direct bandgap at the K and K′ valleys of monolayer MoS2 combined with spin-orbit splitting of the valence band, which produces two spectrally distinct and optically bright excitonic transitions that dominate the material’s optical response.

A dielectric spacer layer—either aluminum oxide or the polymer PMMA—separates the gold from the semiconductor, and its thickness provides a second control knob. As the spacer thickens from 5 to 25 nanometers, the dominant plasmon resonance redshifts monotonically, shifting by as much as 54.4 nanometers for the A-exciton-targeted geometry on Al2O3 and 26.9 nanometers on PMMA. The authors explain this behavior with a sensing-volume model in which the resonance wavelength scales with the effective refractive index sampled by the evanescent near field, with the shift saturating exponentially as the spacer moves the semiconductor beyond the field’s decay length. Al2O3 produces larger shifts than PMMA because its higher refractive index imposes stronger dielectric screening on the plasmonic mode. Notably, the secondary high-energy resonance remains nearly insensitive to spacer thickness, indicating that it involves charge oscillations confined within the cavity walls with little interaction volume extending into the surrounding dielectric. The absence of significant linewidth broadening confirms that these shifts stem from dielectric screening rather than changes in plasmon damping.

To capture the intrinsic excitonic physics of the monolayer accurately, the researchers computed MoS2’s complex dielectric function from first principles using density functional theory within the G0W0-Bethe-Salpeter equation framework, including spin-orbit coupling to correctly resolve the A-B splitting. This wavelength-dependent dielectric function was then imported into the full-wave electromagnetic simulations as a surface conductivity layer, allowing the hybrid system’s response to emerge self-consistently from the interplay between the plasmonic cavity and the excitonic sheet. The optical constants of gold, silicon, silica, alumina and PMMA were drawn from experimentally validated datasets, ensuring that the simulated enhancement factors rest on realistic material parameters rather than idealized dispersion models.

The simulations reveal pronounced absorption enhancement in the MoS2 plane when the plasmon resonance overlaps an excitonic transition, with the magnitude and spectral alignment depending strongly on both cavity geometry and spacer material. Importantly, the exciton energies themselves remain fixed throughout—the enhancement shifts the amplitude of absorption at the excitonic wavelengths without shifting their positions—confirming that the system operates in the weak-coupling regime, where selective enhancement is governed by spectral alignment and modification of the local density of optical states rather than by the formation of hybrid plasmon-exciton states. No Rabi splitting or excitonic energy shifts appear in the spectra, distinguishing this geometry-controlled selectivity from the strong-coupling phenomena explored in other nanocavity platforms.

To quantify the practical consequences, the team computed the charge generation rate—the rate at which absorbed photons create electron-hole pairs—at the MoS2 layer. At the tightest spacer separation of 2 nanometers, the A-exciton-targeted cavity boosts charge generation by a factor of 4.34 with an Al2O3 spacer and 4.10 with PMMA, while the B-exciton-targeted geometry achieves factors of 3.94 and 3.67 respectively. The selectivity is unambiguous: when the resonance is tuned to the A exciton, the enhancement near the B exciton is comparatively weak, and vice versa. As the spacer thickens, the enhancement decays steadily toward unity as the monolayer exits the intense evanescent near field, falling to or below the bare-substrate limit beyond roughly 20 to 25 nanometers.

Enhanced excitation alone, however, does not guarantee brighter emission, because a plasmonic environment simultaneously opens non-radiative energy-transfer channels into the metal. The authors therefore modeled the exciton as an in-plane oscillating dipole and separated the radiative and non-radiative decay rates using far-field and near-field power monitors. For a material with intrinsically low photoluminescence quantum yield—around one-thousandth to one ten-thousandth—the photoluminescence enhancement approximates the product of the excitation enhancement and the normalized radiative decay rate. Under optimized conditions the radiative decay rate is enhanced more than forty-fold while non-radiative transfer to the metal is suppressed at larger separations, yielding estimated local photoluminescence enhancement factors of 143.85 for the A exciton and 87.27 for the B exciton—values the researchers note substantially exceed those reported for mirror-free plasmonic systems such as gold nanorods and nanospheres, which typically deliver factors of three to forty-five.

Perhaps the most striking result is the demonstrated ability to reshape the emission spectrum rather than merely amplify it. Introducing a metric called the normalized excitonic peak ratio, which compares the A-to-B peak intensity balance in the enhanced spectrum against that of bare MoS2, the team found values as high as 2.4 for the A-exciton-targeted geometry on a 5-nanometer Al2O3 spacer—meaning the A exciton becomes more than twice as dominant as it would be naturally. The maximum internal quantum efficiency of the hybrid structure reached 0.73 at 620 nanometers for the A-exciton geometry, and 0.54 at 612 nanometers for the B-exciton case, both achieved at the largest spacer thickness where non-radiative quenching is weakest. These figures demonstrate that plasmonic nanocavities can spectrally redistribute excitonic emission, a capability with direct implications for ultrathin light sources, on-chip nanophotonic emitters and valleytronic devices where control over which excitonic channel dominates is essential.

The design strategy also generalizes beyond molybdenum disulfide. The authors point out that hollow nanocylinders with aspect ratios of 1.66 and 4 could target the B exciton of MoSe2 and the A exciton of WS2 respectively, extending the platform across the broader family of transition metal dichalcogenides. Because the dielectric spacer prevents direct metal-semiconductor contact, the architecture avoids Fermi-level pinning, interfacial charge transfer and material-mismatch perturbations that complicate other plasmonic approaches, providing a cleaner electromagnetic handle on the excitonic response. While the results remain computational, they map a practical and fabricable pathway—hollow nanocavities of this type can be produced by template-assisted growth or galvanic replacement—toward spectrally programmable emission and photocurrent control in atomically thin semiconductors, a capability that could accelerate the development of integrated optoelectronic platforms where every layer counts and every photon matters.