Decoupling degrees of freedom in optical encryption

Information security is a defining challenge of the digital era. Conventional electronic encryption faces an existential threat from quantum computing given that such algorithms can reduce problems that underpin classical cryptography to polynomial complexity¹. Physical-layer optical encryption offers a qualitatively different resistance; by encoding information in the physical properties of light itself, the computational complexity of security measures can, in principle, increase much faster than that of potential attacks^2–6. Landmark demonstrations spanning orbital angular momentum (OAM) holography⁷, stimuli-responsive chiral photonic crystals⁸, cascaded metasurface secret sharing⁹, and reprogrammable meta-holograms¹⁰ have each extended the nominal key space of optical encryption systems.

Nonetheless, a persistent and underappreciated subtlety limits many of these demonstrations. In single-component chiral photonic crystals, the degrees of freedom of light are not independently controllable; rather, they are correlated through a given structural parameter. In a cholesteric liquid crystal (CLC), the helical pitch simultaneously determines the reflection wavelength, the circular polarisation-handedness selectivity, and the Bragg–Berry geometric phase^11,12. Similarly, although polymer-stabilised blue-phase liquid crystals (PS-BPLCs) exhibit a remarkable thermal robustness of their 3D photonic lattice, they operate at a single fixed Bragg wavelength and one circular polarisation handedness¹³. Cascaded liquid crystal (LC) and metasurface devices push the channel count higher but typically leave spin and wavelength entangled through the shared Bragg condition¹⁴. This means that in practice, multi-parameter systems offer a key space that is at best additive and lack the combinatorial growth that genuine independence would provide¹⁵.

Ouyang et al. recently reported a biphasic chiral photonic crystal platform that fundamentally resolves the coupling problem¹⁶. A right-handed PS-BPLC is formed inside a photopatterned cell and treated with solvent to remove unreacted monomers. The polymer network contracts irreversibly, which opens an atmospheric gap that is subsequently refilled with a left-handed CLC (LH-CLC). This results in two co-existing photonic crystal phases within a single cell. Their opposite handedness provides structurally orthogonal spin selectivity, and the PS-RH-BPLC reflects only right-handed circular polarisation (RCP) at a thermally stable wavelength fixed by its locked polymer lattice, while the LH-CLC reflects only left-handed circular polarisation (LCP) at a wavelength that shifts continuously from 650 to 580 nm as the temperature falls from 80 °C to room temperature. Each phase carries its own independently photopatterned alignment distribution, and thus neither phase optically perturbs the other, the schematic of multichannel optical encryption is shown in Fig. 1.

Fig. 1  Schematic of multichannel optical encryption using photopatterned biphasic chiral photonic crystals. Reprinted with permission from Ref. 16. Copyright 2026, Springer Nature.

This orthogonality has been shown to enable dual-channel operation unavailable in prior single-medium devices. In the near-field channel, circular-polarisation interference in a 4f imaging system revealed an image of a “flower” encoded in the LH-CLC and an image of “leaves” in the PS-RH-BPLC. The flower was visible only below the CLC clearing temperature, and its colour drifted thermochromically upon cooling. Meanwhile, the leaves remained spectrally and spatially stable at all temperatures. In the far-field holographic channel, two Gerchberg-Saxton holograms encoding the geographic coordinates of Nanjing were reconstructed only when σ, λ, and T were simultaneously correct. At elevated temperatures, CLC was isotropic, and the longitudinal hologram was simply absent regardless of polarisation or wavelength. Meanwhile, at room temperature, 500 nm LCP revealed the longitude, and 540 nm RCP revealed the latitude. Spin crosstalk was negligible not due to algorithmic filtering but rather because the two layers were physically transparent to each other’s operating conditions.

Research on optical encryption has increasingly pursued multi-parameter platforms, with the implicit assumption that each additional degree of freedom expands the key space proportionally. This work exposes a structural flaw in that logic: in a single-phase chiral photonic crystal, spin, wavelength, and phase modulation all share a common origin and cannot be varied independently. Thus, nominally multi-parameter systems only offer an additive key space at best. By decoupling these axes at the fabrication level rather than through modulation schemes, the biphasic platform reframes the central design question from that of how many parameters can be addressed to that of how many such parameters are structurally orthogonal. Table 1 positions this result against some representative approaches in the field.

Looking forward, the principle of structural orthogonality demonstrated here points to several underexplored directions in optical encryption. In terms of materials, the biphasic architecture is not restricted to the CLC/BPLC pair. Any two photonic crystal phases with opposite handedness and non-overlapping Bragg conditions could be integrated by the same route. Thus, the operating range could be extended into the infrared region, or nonlinear and luminescent channels could be incorporated for additional encryption dimensions. In terms of encryption systems themselves, combining such platforms with on-chip photonic integration could bring multi-key optical security within reach to enable practical deployment at scale for applications in secure communications, hardware authentication, and anti-counterfeiting technologies^22,23.