Polarization‑Controlled Quasi‑Far‑Field Superfocusing with Nanoring‑Based Plasmonic Lenses

Background

Super‑resolution imaging, particle acceleration, quantum information processing, and polarization‑dependent optical storage all benefit from subwavelength plasmonic structures that manipulate surface‑plasmon polaritons (SPPs). The plasmonic lens (PL), first proposed by Pendry in 2000, achieves perfect imaging by coupling evanescent waves, but the resulting focus is confined to the near field, limiting its practical use in conventional microscopes. Recent advances have extended PLs to nanostructures that can focus in both the plane and the far field, yet achieving a sub‑diffraction focal spot has remained elusive until phase‑controlled nanoslit‑based PLs employed metal‑insulator‑metal (MIM) waveguides to modulate the phase at the sub‑wavelength scale. However, extending this design to two dimensions with rotational symmetry has proven difficult: a linearly polarized incident beam fails to produce a circular focus, and the focal length deviates significantly from theoretical predictions. NRPLs, with their inherent rotational symmetry, require a radial electric field component for efficient SPP excitation, making radially polarized light a natural match. While prior studies have demonstrated circular foci using donut apertures or Fresnel zone plates, the far‑field superfocusing capability of NRPLs remains unverified, and control over focal length and energy distribution has been limited.

In this work, we design a high‑NA NRPL that exploits the SOP of incident light to achieve sub‑diffraction focusing in the quasi‑far field. Using finite‑difference time‑domain (FDTD) simulations, we examine the effects of linear, circular, azimuthal, and radial polarizations on the focal shape, size, and position. Our analysis reveals that the longitudinal field component dominates the intensity distribution and that careful phase engineering can compensate for focal shift and focal‑length deviations.

Methods

NRPLs are designed via the wavefront reconstruction theory. The required relative phase delay between the innermost nanoring (radius r1) and the i‑th ring is given by

−Δφ(ri) = 2π (√(f₀²+ri²)−√(f₀²+r1²))/λ₀ + 2πn,

where f0 is the target focal length and n is an integer. The nanorings are approximated as MIM waveguides; the propagation constant β satisfies

tanh( w √(β²−k₀² ε_d)/2 ) = −ε_d √(β²−k₀² ε_m)/(ε_m √(β²−k₀² ε_d)),

where k0 is the free‑space wavenumber and εd = 1 for air. A 400‑nm gold film (ε_m = −12.8915 + 1.2044 i at 650 nm) provides the metal layer. To suppress coupling between adjacent rings, the wall spacing is set to 100 nm, well above twice the skin depth (δm ≈ 28 nm). The lens consists of 32 concentric rings with widths ranging from 10 to 100 nm, an inner radius of 800 nm, and a target focal length of 1300 nm (2 λ₀). The total phase span of 10π yields a numerical aperture (NA) of 0.96 and a Rayleigh limit of 0.61 λ₀/NA ≈ 413 nm.

Simulations were performed using a 12‑layer perfectly matched layer (PML) boundary. The mesh was 10 nm in the output region and refined to 5 nm around the focal plane. Incident fields were defined via Jones matrices for linear, circular, azimuthal, and radial polarizations.

Results

I Linear Polarization

With a linearly polarized beam, the lens produces two symmetric foci separated by 400 nm, each with a FWHM of 210 nm (≈0.32 λ₀). The calculated focal length of 1215 nm differs from the design by 6.5 %. The field distribution reveals an elliptical focus (220 nm × 457 nm) dominated by the longitudinal component, which carries ~80 % of the total energy.

Intensity distribution for linear polarization. Two foci 400 nm apart, FWHM ≈0.32 λ₀, DOF ≈1.68 λ₀.

II Circular Polarization

Illumination with circularly polarized light yields a donut‑shaped longitudinal focus (FWHM 210 nm, radius 400 nm) and a circular transverse focus (FWHM 295 nm). The focal length is 1185 nm (≈8.9 % error). The longitudinal component dominates the energy (~81 %).

Donut focus in |E_z|² and circular focus in |E_r|².

III Azimuthal Polarization

Azimuthally polarized light fails to excite SPPs because the local TE component is orthogonal to the metal–dielectric interface, resulting in negligible transmission and no discernible focus.

Intensity distribution for azimuthal polarization; negligible focus.

IV Radial Polarization

Radial polarization, matching the TM SPP excitation, produces a circular focus with a FWHM of 276 nm (≈0.42 λ₀) in |E|² and a similar radius in the longitudinal field. The total intensity is five times higher than the linear case, and the longitudinal component carries ~82 % of the energy.

Radial‑polarized focus: |E|² and |E_z|² distributions.

Discussions

I Superfocusing Capability of NRPLs

Across all SOPs, the NRPL consistently achieves sub‑diffraction foci, with sizes surpassing the Rayleigh limit (413 nm). The intensity profiles resemble Bessel functions, confirming the formation of non‑diffracting beams in the quasi‑far field. Calculated propagation distances (≈357 nm for the dielectric skin depth) support the negligible decay of SPPs beyond the lens surface.

II Shape of Focus

Polarization tuning yields elliptical, circular, and donut foci. The longitudinal field component, which accounts for ~80 % of the energy, dictates the focal shape, while the transverse component contributes to the complementary pattern.

Normalized intensity patterns of |E|² and |H|² for each polarization.

III Modulation of Focal Length

Simulations confirm that increasing the total phase difference from 2π to 16π (NA 0.75–0.96) aligns the longitudinal focus with the design position, while the transverse focus remains stable. The focal length discrepancy between longitudinal and transverse fields is ~200 nm but can be mitigated by optimizing the phase span.

IV Focusing Performance in the Non‑Coaxial Situation

A 3 µm offset of the radial beam from the lens axis reduces the maximum intensity by over 85 % and elongates the focal spot. These results underscore the importance of precise alignment in experimental implementations.

Conclusions

We present a high‑NA NRPL that harnesses SOP‑dependent SPP excitation to achieve quasi‑far‑field superfocusing. By engineering the phase profile of concentric nanorings, we realize elliptical, circular, and donut foci with FWHMs below 0.5 λ₀. Our work demonstrates that polarization control, combined with careful phase and structural optimization, can overcome focal shift and energy‑distribution challenges inherent to plasmonic lenses. These findings open pathways for applications in super‑resolution imaging, particle acceleration, quantum photonics, and high‑density optical storage.

Abbreviations

DOF:

Depth of focus

FDTD:

Finite‑difference time‑domain

FWHM:

Full‑width at half maximum

MIM:

Metal‑insulator‑metal

NRPL:

Nanoring‑based plasmonic lens

PML:

Perfectly matched layer

SOP:

State of polarization

SPPs:

Surface plasmon polaritons

TE:

Transverse electric

TM:

Transverse magnetic