🔬 mesoscale physics

Optimizing strong light-matter coupling of plasmonic lattices and monolayer semiconductors

This paper demonstrates that embedding gold nanodisk arrays into van der Waals heterostructures with optimized layer ordering and encapsulation mitigates strain and contamination-induced degradation, thereby enabling the realization of high-quality, homogeneous large-area polariton lattices for applications ranging from polarization control to topological polaritonics.

Original authors: Lukas Krelle, Lukas Husel, Kenji Watanabe, Takashi Taniguchi, Ismail Bilgin, Alexander Högele, Farsane Tabataba-Vakili

Published 2026-05-15

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Lukas Krelle, Lukas Husel, Kenji Watanabe, Takashi Taniguchi, Ismail Bilgin, Alexander Högele, Farsane Tabataba-Vakili

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to build a super-fast, high-speed dance floor where two very different partners can hold hands and move as one. One partner is a light particle (a photon), and the other is a matter particle (an exciton, which is an electron and a hole stuck together in a special material). When they hold hands tightly, they become a new creature called a polariton. Scientists love these polaritons because they can do amazing things, like creating lasers or behaving like a single giant quantum wave.

The problem is that the "matter" partner (the exciton) is extremely sensitive. It's like a dancer who gets dizzy if the floor is bumpy, dusty, or if they are stretched out of shape. If the floor isn't perfect, the dance falls apart, and the connection between light and matter gets weak.

The Setup: A Golden Dance Floor

In this paper, the researchers built a special dance floor using gold nanodisks (tiny, flat gold coins) arranged in a grid. These gold coins act like mirrors that trap light very tightly, creating a strong "near-field" where the dance happens.

They wanted to put their sensitive exciton dancers (made from a material called MoSe2, a single layer of atoms) right next to these gold coins. However, to keep the dancers healthy and happy, they needed to wrap them in a protective, perfectly flat blanket made of a material called hBN (hexagonal boron nitride).

The Experiment: Who Goes on Top?

The researchers had a clever way to build this sandwich, but they wanted to test one specific thing: Does the order of the layers matter?

They built two identical sandwiches, but with the layers flipped:

Sample A (The "Top-Down" Sandwich): The gold coin grid was embedded in the hBN blanket, and this whole "gold-in-blanket" unit was placed on top of the MoSe2 dancer.
Sample B (The "Bottom-Up" Sandwich): The gold coin grid was embedded in the hBN blanket, but this unit was placed underneath the MoSe2 dancer.

What Happened?

The results were like a tale of two very different dance floors:

1. Sample A (Gold on Top): The Smooth Sailing
In this version, the gold coins were sitting on a flat, clean surface. The MoSe2 dancer could rest on a perfectly smooth, flat hBN surface without being stretched or scratched.

The Result: The dancers were happy. They moved smoothly, and the connection between the light and the matter was very strong. The "dance" (the polariton) was clear, sharp, and energetic.

2. Sample B (Gold on Bottom): The Bumpy Ride
In this version, the MoSe2 dancer had to lie directly on top of the gold coins. Since the gold coins were slightly raised, the thin MoSe2 layer had to stretch over them like a sheet of plastic over a bumpy rock. Also, the process of making the gold coins created some "dust" and chemical changes on the surface.

The Result: The dancers were stressed. The stretching (strain) and the surface dirt made them wobble. The connection between light and matter was weaker, and the "dance" was blurry and less energetic.

The Big Discovery

The researchers measured how tightly the light and matter held hands (the "coupling strength"). They found that Sample A had a 25% stronger connection than Sample B.

Why? Because in Sample A, the environment was clean and strain-free. The "dancer" kept its full energy and didn't get tired out by a bumpy floor. In Sample B, the physical stress and surface contamination made the dancer weaker, so the light couldn't grab onto it as tightly.

Why This Matters (According to the Paper)

The paper concludes that if you want to build these advanced light-matter devices, how you stack the layers is crucial. Putting the plasmonic (gold) structure on top of the sensitive material, rather than underneath it, preserves the quality of the material.

This method gives scientists a reliable way to build large, uniform "dance floors" for polaritons. This could help in creating better tools for controlling light polarization (how light waves wiggle) and topological polaritonics (a fancy way of describing light waves that flow in specific, protected paths, like a train on a dedicated track).

In short: To get the best performance out of these tiny light-matter hybrids, you have to treat the sensitive material gently. Don't make it climb over bumps; give it a flat, clean surface to stand on.

Technical Summary: Optimizing Strong Light-Matter Coupling of Plasmonic Lattices and Monolayer Semiconductors

Problem Statement
Exciton-polaritons, formed by the strong coupling of matter excitations and cavity photons, offer a platform for phenomena such as polariton lasing, topological polaritons, and bosonic condensation. Monolayer transition metal dichalcogenides (TMDs) are ideal matter constituents due to their large oscillator strengths and binding energies, enabling operation from cryogenic to room temperatures. However, their two-dimensional nature makes them highly sensitive to environmental strain and dielectric disorder, necessitating encapsulation in atomically flat hexagonal boron nitride (hBN) to maintain optical quality.

Integrating these encapsulated TMDs with plasmonic surface lattice resonances (SLRs)—which offer strong field confinement and high quality factors—presents a significant fabrication challenge. Conventional approaches, such as placing the 2D material on top of a nanostructure or patterning the structure directly on the material, often induce strain or fail to bring the emitter sufficiently close to the near-field without compromising the hBN encapsulation. Even a thin hBN layer between the emitter and resonator can detrimentally reduce coupling strength. While recent methods have integrated plasmonic cavities into hBN layers, the specific impact of layer ordering and fabrication-induced surface conditions on exciton quality and coupling strength requires further optimization.

Methodology
The authors employed a fabrication method previously developed to embed gold nanodisk arrays directly into hBN layers, integrating them into van der Waals heterostructures. This process involves:

Mechanical exfoliation of hBN flakes onto a SiO2/Si substrate.
Electron-beam lithography (EBL) using PMMA resist to define a square nanodisk array.
Inductively-coupled plasma reactive ion etching (ICP-RIE) using Ar and SF6 to etch holes through the hBN.
Electron-beam evaporation of a gold film matching the hBN thickness.
Lift-off in acetone/isopropanol and residue removal via O2 plasma.

Two distinct sample geometries were fabricated using large-area, CVD-grown MoSe2 monolayers:

Sample A: The hBN-embedded plasmonic lattice was placed at the top of the heterostack.
Sample B: The hBN-embedded plasmonic lattice was placed at the bottom of the heterostack.

Both samples were fully encapsulated with additional hBN layers. The study utilized cryogenic (4 K) differential reflectance (DR) and photoluminescence (PL) spectroscopy, including momentum-resolved measurements, to characterize exciton quality, homogeneity, and light-matter coupling. A coupled oscillator model was applied to extract coupling strengths ( $g$ ) and Rabi splittings.

Key Results

Exciton Quality and Homogeneity: Sample A exhibited significantly superior exciton quality compared to Sample B. Sample A showed narrow exciton linewidths (average 3.0 meV, down to 2.4 meV) and high spatial homogeneity ( $\sigma \approx 1.0$ meV). In contrast, Sample B displayed broadened linewidths (average 9.6 meV) and reduced spatial homogeneity ( $\sigma \approx 4.0$ meV).
Emission Characteristics: Sample A displayed emission from both neutral excitons and trions with narrow lines. Sample B showed largely suppressed neutral exciton emission, dominated by strong trion emission, indicating charge doping likely induced by O2 plasma exposure during the fabrication of the bottom layer.
Oscillator Strength: Integration of the DR spectra revealed that Sample B suffered a $\sim$ 25% reduction in oscillator strength compared to Sample A, attributed to strain and surface contamination in the bottom-layer configuration.
Light-Matter Coupling: Both samples demonstrated strong coupling, evidenced by avoided crossings in momentum-resolved dispersions. However, Sample A achieved a higher coupling strength ( $g = 23.5$ meV) compared to Sample B ( $g = 19.5$ meV). The Rabi splitting was 40.1 meV for Sample A and 35.5 meV for Sample B.
SLR Properties: The linewidth and quality factors of the Surface Lattice Resonances (SLRs) themselves were found to be largely independent of the layer order, indicating that the differences in coupling strength were driven by the exciton properties rather than the plasmonic mode quality.

Significance and Claims
The paper claims that the layer order in van der Waals heterostructures containing embedded plasmonic lattices critically affects the optical quality of the encapsulated monolayer. Specifically, placing the plasmonic lattice at the top of the stack (Sample A) preserves a clean, strain-free dielectric environment for the TMD monolayer, whereas placing it at the bottom (Sample B) exposes the interface to plasma-induced defects and strain, degrading exciton quality.

The authors conclude that their fabrication approach, which embeds the plasmonic cavity within the hBN layer, enables the creation of homogeneous, large-area polariton lattices. By optimizing the layer order to minimize interfacial irregularities, the study demonstrates a 25% increase in coupling strength. This establishes a versatile platform for integrating sensitive van der Waals materials into nanophotonic architectures, facilitating applications in polarization control and topological polaritonics.

The Setup: A Golden Dance Floor

The Experiment: Who Goes on Top?

What Happened?

The Big Discovery

Why This Matters (According to the Paper)

More like this