🔬 optics

Probe- and Substrate-Dependent Visibility of Mie Resonances in Silicon Nanospheres

This study demonstrates that the observable Mie resonances of silicon nanospheres are not intrinsic properties but are significantly altered by substrate interactions and excitation methods, necessitating specific guidelines for interpreting and designing substrate-supported dielectric resonators.

Original authors: Yonas Lebsir, Huatian Hu, P. A. D. Gonçalves, Hiroshi Sugimoto, Minoru Fujii, N. Asger Mortensen, Christos Tserkezis, Sergii Morozov

Published 2026-05-13

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Yonas Lebsir, Huatian Hu, P. A. D. Gonçalves, Hiroshi Sugimoto, Minoru Fujii, N. Asger Mortensen, Christos Tserkezis, Sergii Morozov

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 a tiny, perfect glass marble made of silicon. In the world of light, this isn't just a speck of dust; it's a high-performance musical instrument. When light hits it, the marble doesn't just scatter the light; it "sings" specific notes. These notes are called Mie resonances. Some notes are like a deep drumbeat (magnetic modes), and others are like a sharp whistle (electric modes).

Scientists have long known how these marbles sing when they are floating freely in a vacuum. But in the real world, you can't just float a marble in mid-air to study it; you have to put it on a table. The problem is, the "table" (the substrate) changes how the marble sings.

This paper is like a detective story where the authors investigate how different "tables" change the song, and how different "microphones" hear that song differently.

The Three Tables (Substrates)

The researchers placed their silicon marbles on three very different surfaces to see how the environment changed the music:

The Thin Plastic Sheet (Silicon Nitride): This is like placing the marble on a very thin, almost invisible piece of plastic. It barely touches the marble. The marble's song remains mostly the same, just like a singer on a quiet stage.
The Concrete Floor (Bulk Silicon): This is a thick, heavy surface that looks just like the marble itself. When the marble sits here, the floor acts like a giant echo chamber. It grabs the "whistle" notes (electric modes) and makes them much louder, while the "drumbeat" notes (magnetic modes) stay mostly the same.
The Mirror (Gold): This is the most dramatic change. Placing the marble on a gold mirror is like putting a singer in front of a giant, perfect mirror. The mirror creates a "ghost" version of the singer right underneath the floor. The real singer and the ghost start singing together, creating a brand new, hybrid song that neither could sing alone. Some notes become incredibly sharp and clear, while others become a long, blurry hum.

The Two Microphones (Probes)

Here is the twist: The paper shows that how you listen to the marble changes what you hear. They used two different methods to "listen" to the marbles:

The Spotlight (Dark-Field Microscopy): Imagine shining a broad, white spotlight on the marble from the side and listening to the light that bounces off. This is like listening to the marble from a distance in a concert hall. It gives you a good overview of the whole song, but it can't hear the quiet details or the specific notes that are hidden by the "table."
The Electron Pen (Cathodoluminescence): This is a much more invasive tool. Instead of light, they shoot a tiny, focused beam of electrons (like a microscopic pen) directly at the marble. This is like tapping the marble with a needle to make it vibrate. Because the pen is so small, it can tap specific spots on the marble.
- If they tap the edge, they hear the "whistle" notes.
- If they tap the center, they hear the "drumbeat" notes.

The Big Discovery

The main lesson of the paper is that you cannot judge a marble's song just by looking at the marble; you have to know what table it's sitting on and what microphone you are using.

The "Invisible" Note: On the gold mirror, the "whistle" note (electric mode) was so messy and broad that the "Spotlight" (Dark-Field) couldn't see it clearly. It looked like static noise. However, the "Electron Pen" (Cathodoluminescence) tapped the marble in just the right way to make that specific note ring out loud and clear.
The "Ghost" Effect: On the thick silicon floor, the "Electron Pen" at low energy lost its precision. Instead of tapping one spot, the electrons scattered inside the marble like a pinball, tapping everywhere at once. This made it impossible to tell which spot was making which note, blurring the song.
The Mirror's Secret: On the gold mirror, the "Electron Pen" revealed a special, high-quality note that the "Spotlight" completely missed because the mirror blocked the light from reaching the detector in that specific way.

The Takeaway

The authors conclude that if you want to build devices using these tiny silicon marbles (which they call "Mie-tronics"), you can't just rely on one type of measurement. You need to combine the "Spotlight" and the "Electron Pen" to get the full picture.

If you only look at the marble from the outside, you might think it's singing a simple song. But if you tap it with an electron pen, you realize the "table" it's sitting on is actually part of the instrument, changing the music entirely. The song isn't just in the marble; it's in the marble and the table and the way you are listening.

Technical Summary: Probe- and Substrate-Dependent Visibility of Mie Resonances in Silicon Nanospheres

Problem Statement
Silicon (Si) nanospheres serve as high-quality optical resonators and fundamental building blocks for "Mie-tronic" devices, leveraging coexisting electric dipole (ED) and magnetic dipole (MD) modes. While the Mie resonances of isolated spheres are well-described by theory, practical implementations invariably require substrates. These substrates modify the modal landscape by shifting resonances, lifting degeneracies, and reshaping radiation patterns. A critical ambiguity remains in the field: to what extent does the measured optical spectrum reflect the intrinsic resonances of the particle versus the specific excitation mechanism and environmental coupling? Different experimental probes, such as dark-field (DF) microscopy and cathodoluminescence (CL) microscopy, access different physical observables and excitation pathways, potentially leading to conflicting interpretations of the same mode, particularly when electron-beam induced effects (e.g., transition radiation) are involved.

Methodology
The authors investigate individual colloidal crystalline Si nanospheres (diameter $D \approx 160$ nm) placed on three distinct substrates to systematically analyze substrate-probe interactions:

Silicon Nitride ( $\text{Si}_3\text{N}_4$ ): A thin (5 nm) membrane acting as a weakly perturbing, quasi-free-standing environment.
Bulk Silicon (Si): A high-index dielectric substrate common in experimental setups.
Gold (Au): A metallic mirror substrate inducing strong image-dipole coupling and hybridization.

The study employs a dual-probe approach on the same particles:

Dark-Field (DF) Spectroscopy: Probes the far-field scattering cross-section using broadband illumination. Polarization-resolved measurements (s- and p-polarized) are used to selectively excite specific dipolar components.
Cathodoluminescence (CL) Spectroscopy: Uses a focused electron beam (at 5 kV and 30 kV) to excite the particle, mapping the radiative local density of optical states (LDOS) with nanometer-scale spatial resolution.

These experimental results are compared against analytical Mie theory (for isolated spheres) and numerical simulations using the boundary-element method (BEM) for scattering and the finite-element method (FEM) for CL. Monte Carlo simulations are utilized to model electron beam broadening at low energies.

Key Results

Isolated Sphere Reference: For a free-standing 160 nm Si sphere, the dominant resonance is the MD ( $\lambda \approx 642$ nm), followed by the ED ( $\lambda \approx 518$ nm). Near-field analysis reveals that the MD electric field is concentrated internally, whereas the ED field extends significantly beyond the particle boundary. Consequently, the ED is predicted to be more sensitive to substrate perturbations.
Weakly Perturbing Substrate ( $\text{Si}_3\text{N}_4$ ):
- DF: The spectrum remains close to the isolated-sphere limit, with the MD dominating. The thin membrane introduces minimal polarization anisotropy.
- CL: At 30 kV, position-resolved maps show expected mode selectivity (edge excitation favors ED, center favors MD). At 5 kV, spatial selectivity is lost due to strong multiple scattering and beam broadening within the sphere, resulting in an averaged response. A feature at $\lambda \approx 400$ nm observed in experiments but absent in straight-line current models is attributed to transition radiation.
Reflecting Dielectric Substrate (Bulk Si):
- DF: The ED/MD intensity ratio increases substantially compared to the membrane case, indicating preferential enhancement of the ED scattering efficiency due to coupling with the high-index substrate. The ED is strongly polarization-sensitive.
- CL: While spatial selectivity is degraded at 5 kV, the ED signal persists more robustly than on the membrane. The authors suggest the substrate acts as an active participant, providing an additional coupling pathway (via interface-mediated fields) that sustains ED excitation even when beam collimation is lost.
Reflecting Metallic Mirror (Au):
- DF: Strong symmetry breaking occurs. The MD perpendicular to the substrate ( $\text{MD}_\perp$ ) becomes a narrow, high-Q resonance due to destructive interference with the image dipole. The ED perpendicular to the substrate ( $\text{ED}_\perp$ ) forms a broad, gap-confined hybrid mode. In unpolarized DF, the broad $\text{ED}_\perp$ continuum overwhelms the narrow MD peak.
- CL: CL reveals modes hidden in DF. Specifically, the in-plane electric dipole ( $\text{ED}_\parallel$ ) near 580 nm, which requires a normal electric field component not efficiently provided in the DF geometry, becomes clearly visible in CL. CL also resolves the hybrid magnetic and electric modes, demonstrating that the mirror creates qualitatively new particle-on-mirror states.

Significance and Claims
The paper claims that the observable modal spectrum of a dielectric nanoresonator is not solely an intrinsic property of the particle but is the result of an interplay between substrate-induced mode modification and probe-specific excitation pathways.

Probe Complementarity: Neither DF nor CL alone provides a complete picture. DF serves as a far-field reference but can miss modes requiring specific field orientations (e.g., $\text{ED}_\parallel$ on Au). CL offers localized excitation and can reveal resonances that are weak or hidden in far-field scattering, but it is sensitive to electron trajectory and energy-dependent scattering effects.
Substrate Role: Substrates do not merely perturb modes passively; they can actively participate in the excitation process (e.g., sustaining ED excitation on bulk Si at low beam energies) and create new hybrid states (on Au).
Methodological Guidance: Accurate interpretation of Mie resonances in realistic, substrate-supported geometries requires the combined use of multiple experimental techniques alongside electrodynamic modeling that accounts for both the substrate environment and the specific excitation mechanism (including electron beam broadening at low energies).

The authors conclude that these findings provide practical guidelines for designing and interpreting substrate-supported dielectric resonators for Mie-tronic applications, emphasizing that the "visibility" of a resonance is contingent upon the measurement context.

The Three Tables (Substrates)

The Two Microphones (Probes)

The Big Discovery

The Takeaway

More like this