Extrinsic Limitations of Stealthy Hyperuniform Optical Metasurfaces

Hyperuniform metasurfaces promise an unusual form of wave control: the suppression of elastic scattering over extended angular ranges without periodic order. Here, we present a comprehensive experimental and theoretical study of 2D stealthy hyperuniform metasurfaces operating at optical frequencies. In agreement with theoretical expectations, we observe a pronounced reduction of elastic scattering around the specular direction in metasurfaces fabricated by electron-beam lithography. However, the measured suppression is substantially weaker than that predicted by structure-factor calculations based on ideal stealthy hyperuniform point-pattern generators. We identify and quantitatively analyze the physical origins of this discrepancy, and establish realistic performance bounds. By isolating the dominant limiting mechanisms, our results provide practical design guidelines for the implementation of stealthy hyperuniform metasurfaces in functional photonic devices.

💡 Research Summary

This paper presents a comprehensive experimental and theoretical investigation of two‑dimensional stealthy hyperuniform (SHU) metasurfaces operating at optical frequencies. SHU point patterns are defined by a structure factor S(q) that vanishes within a finite disc of radius k_c in reciprocal space, which translates into a “quenching zone” where elastic scattering is strongly suppressed for incident in‑plane wavevectors k_∥ < k_c. The authors generate SHU patterns using a potential‑minimization algorithm that controls two key parameters: the degree of stealthiness χ and the point density ρ. By varying χ (0.1–0.5) and ρ (≈10⁸–10⁹ mm⁻²) they create point sets of up to 20 000 particles per unit cell.

The patterns are transferred into a 145‑nm silicon layer on glass by electron‑beam lithography, producing square metasurfaces of 300 µm × 300 µm with identical meta‑atoms (nanodisks or squares). Optical characterization is performed with a supercontinuum laser and a gonio‑spectrometer, yielding the bidirectional reflectance distribution function (BRDF) as a function of polar angle. The measured transition angle θ_c between the low‑scattering quenching zone and the diffuse background follows the analytically derived relation

θ_c ≈ (λ · √(ρ χ))/(2π),

confirming that the angular width of the quenching zone can be engineered by the product ρχ. However, the absolute quenching efficiency (the reduction of scattered intensity inside the zone) is far lower than the theoretical prediction based solely on S(q). Experimentally observed suppression factors are on the order of 10–100, whereas ideal SHU patterns would yield factors exceeding 10⁵.

The authors identify three extrinsic mechanisms that degrade performance: (1) Polydispersity – variations in meta‑atom size or shape cause each element to have a different form factor, breaking the assumption of identical scatterers; (2) Multiple scattering – even identical elements experience different local excitation fields due to inter‑element scattering, invalidating the independent‑scattering approximation, especially at higher densities; (3) Finite‑size and coherence effects – the fabricated metasurfaces have limited lateral extent and are illuminated with a source of finite spatial coherence (tens of micrometers). Consequently, a continuum of wavevectors between the quantized reciprocal‑lattice points becomes accessible, and these unconstrained q‑vectors possess much larger S(q) values, dramatically reducing the averaged quenching efficiency.

Numerical simulations illustrate that increasing the number of points N in a single unit cell reduces the contribution of unconstrained q‑vectors, but practical algorithmic and fabrication limits prevent indefinite scaling. An alternative strategy—tiling a smaller SHU unit cell M × M times—introduces a sinc‑type interference factor that suppresses the continuum contribution. The tiled structure factor follows

S_tiled(q) = S₀(q) ·