Planar Superchiral Fields in Nanophotonics

• Planar superchiral fields are enhanced electromagnetic near fields in 2D nanophotonic platforms that exceed the chirality of traditional circularly polarized light.
• They are engineered using plasmonic, dielectric, and inverse-designed architectures that optimize local field overlap, phase shift, and modal interference.
• Implementations demonstrate marked improvements in circular dichroism, vibrational chiral sensing, and enantioselective optical forces for advanced spectroscopic and separation applications.

Planar superchiral fields are electromagnetic near fields or surface-bound waves generated by planar nanophotonic structures whose local optical chirality exceeds that of circularly polarized plane waves. In the surveyed literature, they arise in plasmonic films, dielectric metasurfaces, photonic-crystal slabs, and surface-wave platforms, and they are engineered to strengthen enantiospecific light–matter coupling while retaining the fabrication and integration advantages of planar geometries. Reported implementations range from chiral slit-pair lattices in the mid-infrared to achiral dielectric metasurfaces based on merged bound states in the continuum, vector exceptional points, and inverse-designed freeform cavities, as well as surface-wave schemes for enantiomer separation (Mattioli et al., 2019, Barkaoui et al., 2022, Jiang et al., 20 Nov 2025, Pellegrini et al., 2018).

1. Definition, optical-chirality metrics, and scope

The central quantity in this literature is optical chirality density. In the Tang–Cohen framework used for mid-IR plasmonic slit arrays, the dissymmetry factor in the lowest-order dipolar approximation is written as

$$g = \frac{G''}{\alpha''}\frac{2C}{\varepsilon_0 \omega U_e^*},$$

with

$$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$

Within that same source, “superchiral” originally referred to field configurations that give a larger measured $g$ than a circularly polarized plane wave, but in later usage it broadly means fields with enhanced optical chirality $C$ (Mattioli et al., 2019).

Across later metasurface work, the same physical quantity is written in alternative conventions, including

$$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$

and

$$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$

These forms differ in normalization and field variables, but they are used for the same design objective: maximizing a local handed electromagnetic state rather than merely maximizing intensity (Jiang et al., 20 Nov 2025, Romashkina et al., 22 Dec 2025).

A recurring physical criterion is that the electric and magnetic fields must be strongly enhanced, spatially overlapped, directionally aligned or nearly parallel, and phase shifted by about $\pi/2$. The merged-BIC literature states these conditions explicitly as strong spectral overlap, strong spatial overlap, and a phase difference of $\pi/2$, while inverse-designed dielectric metasurfaces formulate the objective directly as the pointwise quantity $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$ at an accessible probe point (Barkaoui et al., 2022, Jiang et al., 20 Nov 2025). This emphasis on local, molecule-accessible field topology distinguishes planar superchirality from far-field circular polarization alone.

2. Principal planar architectures

The literature contains several distinct planar routes to superchirality, differing mainly in whether chirality is encoded in geometry, in modal interference, or in optimization targets.

Platform	Mechanism	Representative outcome
Chiral plasmonic slit-pair square lattice	Coupled slit-pair hotspot plus collective lattice resonance	Enhanced optical chirality and sharp mid-IR CD features (Mattioli et al., 2019)
Real chiral gammadion array	Resonant plasmonic near fields modified by roughness, protrusions, and indentations	Maximum optical chirality up to about 3 times greater than the ideal model (Gilroy et al., 2021)
Achiral rectangular nanoslit in gold	Magnetic-dipole antenna with incident–radiated field interference	Purely superchiral light with tunable sign and no background in transmission (Cui et al., 2023)
1D photonic crystal supporting surface waves	Simultaneous excitation of TE and TM Bloch surface waves	Superchiral surface waves with chiral optical forces 1–2 orders of magnitude larger than plane waves (Pellegrini et al., 2018)
Achiral dielectric metasurface at merged BICs	TE-like and TM-like symmetry-protected BICs tuned to merge	$C/C_{\mathrm{CPL}} > 3000$, and up to about $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$0 near $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$1 (Barkaoui et al., 2022)
Asymmetric photonic-crystal slab at a vector EP	Coalesced TE-like and TM-like modes under two-sided excitation	Uniform single-handed superchiral fields (Wu et al., 2020)
Inverse-designed dielectric metasurfaces and meta-cavities	Direct optimization of $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$2 or $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$3	$$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$4 optimization-objective enhancement, 116-fold peak CD increase, and simulated $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$5 up to about $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$6 (Jiang et al., 20 Nov 2025, Romashkina et al., 22 Dec 2025)

In this context, “planar” denotes 2D patterned slabs, films, apertures, and surface-bound wave platforms rather than volumetric chiral metamaterials. The surveyed systems show that planarity does not imply geometrical achirality, and geometrical achirality does not preclude strong superchirality. Structurally chiral slit pairs and gammadions coexist in the literature with mirror-symmetric nanoslits, freeform achiral metasurfaces, and square-lattice dielectric slabs.

3. Plasmonic realizations

A canonical planar plasmonic implementation is the mid-IR array of chiral slit pairs in a 50 nm gold film on CaF$$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$7. Its elementary building block is a pair of subwavelength slits arranged in a chiral configuration. The fundamental electric dipole resonance yields electric and magnetic fields that superimpose favorably, generating a hotspot of enhanced optical chirality right above the sample surface. Simulations evaluate the optical chirality on a plane 10 nm above the surface and show a localized hotspot in the center of the coupled-slit pair. A lone slit pair already produces a superchiral response, but the localized resonance is relatively broad, with a quality factor of only about 5. Placing the motif in a square lattice of pitch 6 $$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$8m sharpens and amplifies the response by coupling the localized plasmon resonance to diffraction anomalies. The same square symmetry removes the strong uniaxial anisotropy of a single slit pair, thereby reducing extrinsic-chirality artifacts. Diffraction-related features appear around 1666 cm$$C = -\frac{\varepsilon_0 \omega}{2}\, \mathrm{Im}\left(\mathbf{E}\cdot \mathbf{B}\right).$$9 for the first order in air, around 1235 cm$g$0 for the first order in the CaF$g$1 substrate, and around 1160 cm$g$2 for an additional possible order along the lattice diagonal (Mattioli et al., 2019).

A second plasmonic route emerges from realistic rather than idealized geometry. In gammadion arrays, the strongest fields are surface-bound and localized, but the “real” structures reconstructed from AFM differ materially from flat ideal models. The reported real geometry has top-surface roughness $g$3 nm, compared with $g$4 in the idealized top surface and $g$5 nm for the glass substrate. Simulations show additional electric-field hotspots at protrusions on the rough top Au surface, whereas the highest optical chirality values occur near the bottom of the gammadion close to the Au–quartz interface. The maximum optical chirality values for the real structure are up to about 3 times greater than those of the ideal structure, and the AFM-based model gives much better qualitative agreement with the observed CD resonances I–III than the ideal model (Gilroy et al., 2021).

A third plasmonic example is the achiral rectangular nanoslit in a thin gold film. Here the planar aperture behaves as a magnetic dipole antenna, consistent with Babinet’s principle. For a gold film thickness $g$6, slit width $g$7, and slit length $g$8 to $g$9, the structure supports a resonance near $C$0 for $C$1. The superchiral field does not come from the magnetic-dipole near field alone; it comes from interference between the incident wave and the radiated field. Because transmission occurs through an opaque film, the useful region is dominated by the local aperture field rather than by a propagating background. The sign of the chirality is externally tunable: it reverses across the resonance, with extrema reported at $C$2 and $C$3, and it also reverses when the incident linear polarization is rotated from $C$4 to $C$5. In the comparison given there, the nanoslit yields a volumetric chirality density about 9 times larger than that of a gold nanorod for the same interaction volume (Cui et al., 2023).

4. Dielectric, topological, and non-Hermitian routes

Achiral dielectric metasurfaces offer a distinct mechanism: superchirality emerges from high-$C$6 modal hybridization rather than from metallic plasmonic confinement. In a square lattice of air holes in a silicon membrane on a glass substrate, with lattice constant $C$7 nm, hole radius $C$8 nm, and thickness $C$9 nm, TE-like and TM-like symmetry-protected BICs occur near the Brillouin-zone center $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$0. Under quantum spin-Hall-effect-assisted excitation, each BIC near $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$1 already supports a strong superchiral field because the finite slab thickness supplies a small complementary field component to the dominant polarization. Near $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$2, the quality factor follows $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$3, and values $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$4 are reported. For an isolated BIC, the enhancement reaches about $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$5 at $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$6, though this regime is highly angle sensitive. When orthogonally polarized TE-like and TM-like BICs are tuned to merge, the system satisfies the superchirality criteria simultaneously and becomes more robust; numerically, the chirality increases from about $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$7 for isolated quasi-BICs to more than $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$8 at the merging point, and up to about $$C = -\frac{k_0}{2\omega_0}\,\mathrm{Im}(\mathbf{E}\cdot \mathbf{H}^*)$$9 is reported near $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$0 (Barkaoui et al., 2022).

A related but non-Hermitian route uses a radiation vector exceptional point in an asymmetric Si$$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$1N$$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$2 photonic-crystal slab on silica. The reported EP design uses lattice period $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$3 nm, hole diameter $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$4 nm, slab thickness $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$5 nm, slab refractive index $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$6, substrate index $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$7, and background refractive index tuned near $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$8. Breaking the up–down symmetry couples TE-like and TM-like modes; tuning the coupling strength and radiation losses brings the eigenvalues and eigenvectors to coalescence. A single beam does not efficiently excite the desired superchiral eigenstate, so the scheme uses two counter-propagating beams from opposite directions. One excites the TE-like component more strongly, the other the TM-like component, and their superposition synthesizes a mixed state with strong $$C=\frac{\omega \varepsilon_0}{2}\,\mathrm{Im}\!\left(\tilde{\mathbf E}\cdot \tilde{\mathbf B}^{*}\right).$$9 overlap and an appropriate phase relation. Experimentally, the EP appears as the merging of two transmission dips into one when the surrounding medium is changed from air or water to a 75% sucrose solution, and the near-field maps show a uniform single-handed chirality field (Wu et al., 2020).

Inverse design reframes planar superchirality as a direct optimization problem. In the freeform achiral dielectric metasurface, the objective is the local near-field quantity

$\pi/2$0

evaluated at a probe point inside an intentionally placed air gap. The structure is constrained to be geometrically achiral by mirror symmetry, suppressing structural chiroptical background. The workflow uses the Meep FDTD adjoint solver together with PyTorch/Adam gradient descent, and the optimized silicon meta-atom lies on a square lattice with period $\pi/2$1 nm and silicon thickness $\pi/2$2 nm. The resulting Mie-type electric-dipole and magnetic-dipole resonances are spectrally overlapped with an approximately $\pi/2$3 phase delay. Reported outcomes include more than $\pi/2$4 enhancement of the optimization objective, more than $\pi/2$5 enhancement in the local differential chiral density $\pi/2$6 for the analyte-loaded structure, a simulated 116-fold increase in peak CD relative to a bare substrate, and more than $\pi/2$7 measured CD enhancement for $\pi/2$8-(+)-1,2-propanediol (Jiang et al., 20 Nov 2025).

A second inverse-designed dielectric platform is a planar meta-cavity patterned in a 200 nm silicon film on sapphire, with a $\pi/2$9-period unit cell and a subwavelength hotspot near the center. It is designed under normal-incidence, linearly polarized illumination. The optimization is two-step: first, a fully connected neural network generates pixelated candidate topologies optimized with fully differentiable RCWA and stochastic gradient descent; second, the preoptimized geometry is refined by adjoint FDTD with fabrication-aware filtering, including a circular spatial blurring filter of radius 40 nm and a Heaviside filter. Only about 200 iterations are required in the second stage. The reported normalized chirality reaches about $\pi/2$0 in idealized simulation, while the fabricated prototype shows an enhancement of about $\pi/2$1. The measured transmission CD reaches about 20% near resonance, the measured quality factor is about 83, and the simulated value is about 70. The third Stokes parameter

$\pi/2$2

can jump to nearly 1 under H-polarized excitation at resonance, indicating that the transmitted light becomes nearly circularly polarized (Romashkina et al., 22 Dec 2025).

5. Verification, artifacts, and active control

Because planar systems are vulnerable to anisotropy-driven artifacts, the verification of genuine chirality is a major part of this field. In the mid-IR slit-pair lattice, a single slit pair is strongly uniaxially anisotropic, so the square array was adopted not only to support lattice resonances but also to remove the dominant linear anisotropy. The resulting CD spectra were checked versus azimuthal angle and after flipping the sample, and opposite-handed arrays yielded mirror-image CD spectra. Flipping the sample reversed the CD, consistent with planar chiral optical behavior (Mattioli et al., 2019).

A related strategy is to suppress background by enforcing structural achirality. The freeform inverse-designed dielectric metasurface imposes mirror symmetry so that any measured CD primarily comes from the chiral analyte rather than from the metasurface itself. The achiral nanoslit pushes this logic further in transmission: the paper argues that the useful region is free of any background contribution from the excitation beam because the gold film is opaque and the transmitted region is dominated by the local aperture field (Jiang et al., 20 Nov 2025, Cui et al., 2023). The comparison between ideal and AFM-reconstructed gammadions also corrects a common misconception: localized $\pi/2$3 hotspots and localized $\pi/2$4-hotspots are not necessarily co-located. In that study, rough top-surface protrusions create electric-field hotspots, but the largest optical chirality occurs near the Au–quartz interface (Gilroy et al., 2021). The inverse-designed meta-cavity adds another cautionary point: the near-field optical chirality peak and the far-field CD peak are close in wavelength but not identical (Romashkina et al., 22 Dec 2025).

Active control has been demonstrated in a magneto-plasmonic metasurface that combines a planar, centrosymmetric hexagonal Au nanohole array with oblique circularly polarized illumination and a magneto-optical oxide stack. The geometry uses period 540 nm and hole radius 185 nm, with a layer sequence of Au 51 nm / Ce:YIG 47 nm / YIG 49 nm / SiO$\pi/2$5 5 nm / TiN 129 nm on silica. Extrinsic chirality arises from the oblique wavevector, while the Ce:YIG layer provides a magnetization-dependent dielectric tensor. The field is concentrated at the Au/Ce:YIG interface by a cavity-like resonance, so magnetic tuning affects both far-field CD and near-field optical chirality. At 950 nm under 45° incidence, the measured CD switches from $\pi/2$6 to $\pi/2$7 under out-of-plane fields up to $\pi/2$8 kOe. The local optical chirality reaches up to 10, the difference between RCP and LCP optical chirality ranges from $\pi/2$9 to 20, and the magnetic-field-induced local optical chirality modulation is about $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$0, corresponding to a $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$1 modulation. A 2 mm $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$2 2 mm “UESTC” patterned metasurface shows clear sign reversal in chiral imaging when the field direction is switched (Qin et al., 2019).

6. Chiral spectroscopy, sensing, and enantioselective manipulation

A central application is vibrational circular dichroism in the mid-infrared. Molecular vibrational resonances lie roughly in the 900–1800 cm$\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$3 range, and the slit-pair platform was designed so that the slit resonance can be tuned across this window by changing slit length while the lattice resonance sharpens both the near field and the CD spectrum. This places molecules near hotspots of enhanced local optical chirality and makes the platform a testbed for coupling plasmonic resonances, lattice resonances, and vibrational resonances of chiral molecules (Mattioli et al., 2019). The inverse-designed achiral dielectric metasurface extends this sensing agenda to quantitative analysis: by varying the Pasteur parameter $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$4, the extracted peak CD scales approximately linearly with $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$5, enabling concentration estimation, and the peak CD also varies linearly with enantiomeric excess in mixtures of $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$6- and $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$7-1,2-propanediol (Jiang et al., 20 Nov 2025).

Planar superchiral fields are also used as force-generating states rather than only as spectroscopic enhancers. In a 1D photonic crystal, simultaneous excitation of TE and TM Bloch surface waves produces a superchiral surface wave extended over a large area. The force analysis emphasizes terms proportional to the chiral polarizability $\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$8, and uses the field helicity

$\mathrm{Im}(\mathbf{E}\cdot\mathbf{H}^*)$9

For the same incident power, the reported chiral optical forces are 1–2 orders of magnitude larger than those produced by circularly polarized plane waves, and the advantage over prism-coupled evanescent waves can reach up to 5 orders of magnitude for smaller particles. For a 5 nm particle placed 30 nm above the surface in a 200 $C/C_{\mathrm{CPL}} > 3000$0m channel, the model predicts 97.4% of the enantiomer displaced to one side after 90 s and 99.3% separation in the stationary limit (Pellegrini et al., 2018).

Other branches of the literature target stronger readout signatures and lower chiral background. In merged-BIC dielectric slabs, introducing an active chiral medium produces a symmetric chiral mode splitting of the transmission dip on the order of $C/C_{\mathrm{CPL}} > 3000$1 THz and a circular-dichroism enhancement more than three orders of magnitude larger in one spectral region (Barkaoui et al., 2022). The achiral nanoslit work frames its magnetic-dipole aperture as a route toward improved signal-to-noise ratios in CD and potentially single-molecule sensitivity, particularly if detection is based on chiral luminescence rather than absorption (Cui et al., 2023). The mid-IR slit-pair study notes a possible future variant based on alternating left- and right-handed slit pairs, intended to preserve strong local chiral near fields while reducing the chiral background in the CD spectrum (Mattioli et al., 2019). Taken together, these results suggest that the field is increasingly organized around three linked objectives: accessible hotspots, minimal structural or propagating background, and direct control over handedness, spectral sharpness, and analyte placement.