Atomically thin semiconductors such as tungsten disulfide (WS2) are emerging as key materials for next-generation photonics technologies. Although they are just a single layer of atoms, they can host tightly bound excitons, electron-hole pairs that interact strongly with light. These materials can also generate new colors of light through nonlinear optical effects such as second harmonic generation. These properties hold promise for applications in quantum optics, sensing, and compact on-chip light sources. However, its extreme thinness also creates challenges. With so little material available, optical interactions are limited, and unless the surrounding photonic environment is carefully designed, emission is often weak and frequency conversion is inefficient.
Research published in advanced photonics present a new strategy to overcome this limitation by modifying the space beneath the material rather than the material itself. In this approach, a single layer of WS2 is placed over nanoscale cavities called meavoids carved into a high-index crystal of bismuth telluride (Bi2Te3). These small air gaps greatly increase the emission and nonlinear optical signal. It also allows direct observation of local optical modes, providing new insights into how light behaves on very small scales.
Turn empty space into a resonator of light
Conventional dielectric nanocavities trap light inside solid materials such as silicon. Although effective in many cases, this design moves the strongest optical fields away from surfaces where atomically thin materials reside. Additionally, when a material absorbs light, its efficiency decreases, weakening the resonance and reducing the electric field strength.
Me Void works differently. Rather than confining light within a solid material, it confines light within a subwavelength cavity etched into a material with a very high refractive index. Strong reflections at the air-dielectric interface keep the light circulating within the cavity. As a result, the light field is concentrated in the air region and near the top surface, exactly where the WS2 layer is located.
This “reverse” confinement approach has several advantages. The enhanced field has direct access to the surface material, the resonant wavelength can be tuned by adjusting the cavity shape, and the design remains effective even in materials that strongly absorb light. Although Bi2Te₃ is not ideal for conventional resonators, it performs well in this void-based configuration.
Structure design and construction
The researchers used detailed electromagnetic simulations to design a cavity that supports a dipole resonance consistent with WS2’s main emission feature known as the A exciton. By carefully tuning the radius and depth of each cavity, we were able to control both the resonant wavelength and the vertical position of the optical mode.
Cavities were created using focused ion beam milling in mechanically exfoliated thick Bi2Te3 flakes. They were spaced far enough apart to act as individual resonators rather than interacting with each other. A continuous WS2 monolayer was then transferred over the entire patterned surface, covering the resonant cavities, nonresonant cavities, and flat areas. This design ensured that the differences in optical behavior were due to the shape of the cavity rather than variations in the material itself.
Optical reflectance measurements confirmed that the cavity behaves as expected. As the cavity became larger, the resonance shifted smoothly towards longer wavelengths, while the change in depth changed both the spectral and vertical positions of the optical modes. Importantly, the resonance remains stable even when the geometry is not fully optimized, demonstrating that the design can tolerate manufacturing imperfections.
Increase the amount of light emitted by WS2
To understand how the cavity affects light emission, the research team measured photoluminescence from WS2 under laser excitation while varying the depth of the cavity. When the cavity resonance coincided with the WS2 emission band, the optical output increased by approximately 20 times compared to the cavity with the least resonance.
Further analysis showed that this increase was not due to stronger absorption of the incident light. Simulations showed no significant enhancement at the excitation wavelength, and experiments using different pump wavelengths consistently produced the strongest emission at the same cavity depth. This confirms that the improvement is due to emissions-related effects. The resonant cavity increases the local optical density of the state, helping the emitted light to escape more efficiently.
Because the WS2 layer was continuous throughout the sample, researchers were able to directly compare the emission from different regions under the same conditions. This demonstrated that the enhanced emission was caused by the designed cavity modes rather than differences in the materials themselves.
Visualization of nonlinear optics and optical modes
The research team also investigated nonlinear optical effects by adjusting the shape of the cavity so that the resonance shifted into the near-infrared region. Under these conditions, the second harmonic signal from WS2 was increased by approximately 25 times compared to the non-resonant cavity. The signal peaked when the excitation wavelength coincided with the cavity resonance.
In addition to improved performance, this system allows direct visualization of optical modes. Far-field imaging of the second harmonic signal revealed bright local hotspots above individual cavities. As the excitation wavelength or cavity depth was varied, these hotspots moved in a predictable pattern across the array. This provided a clear real-space view of how the optical field changes within individual resonators without the need for specialized near-field techniques.
A new platform for atomic thin film photonics
Combining tunable optical enhancement and precise spatial control in van der Waals-compatible systems, meavoid heterostructures provide a powerful new platform for working with atomically thin materials. Unlike traditional approaches, this method does not rely on large metasurfaces and remains effective even in materials that strongly absorb light.
This technology has the potential to enable advances in nonlinear light generation, surface-enhanced sensing, and programmable photonic devices based on two-dimensional semiconductors. More broadly, we show that shaping empty spaces can be as important as choosing the right materials when designing nanoscale light-matter interactions.
