Spectral and Polarization Multiplexing

• Spectral and polarization multiplexing is a technique that exploits the wavelength and polarization degrees of freedom to encode and transmit optical signals in systems like fiber communications and metasurface devices.
• It enhances channel capacity by enabling simultaneous transmission over multiple spectral bands and polarization states, as demonstrated in NFDM and DSCM architectures.
• Its integration leads to improved imaging, sensing, and holographic applications with advanced noise mitigation and reduced system complexity.

Spectral and polarization multiplexing denotes the joint or separate exploitation of wavelength (spectral) and polarization degrees of freedom to encode, transmit, and demultiplex information-bearing optical signals or to control light–matter interaction phenomena. Both dimensions are individually central in fiber-optic communications, integrated photonics, spectroscopy, and emerging on-chip metasurface devices. Their combination offers enhanced channel density, information capacity, and tunable optical functionalities. Architectures and methodologies realizing such multiplexing range from dual-polarization nonlinear Fourier methods in fiber (Ros et al., 2019, Goossens et al., 2017, Span et al., 2018), digitally subsampled coherent subcarrier multiplexing (Wang et al., 2024), adaptive polarization control for self-homodyne links (Anghan et al., 2018), broadband metasurface holography (Bao et al., 2024, Wei et al., 2024), interferometric modal splitting in chip-based spectroscopy (Johnson et al., 9 Apr 2025), to passive spectral-polarimetric encoders (0903.2735). Distinct physical, signal-theoretic, and material-science principles underlie spectral and polarization multiplexing, yet their integration is driving record-breaking performance in data, imaging, and sensing applications.

1. Physical Principles and Mathematical Formalism

Spectral multiplexing uses either discrete (e.g., WDM, OFDM) or continuous spectral degrees of freedom to encode and separate optical channels. In its signal-theoretic form, time-domain signals are represented as superpositions of narrowband components at different carrier frequencies; in the nonlinear domain, the nonlinear Fourier transform allows modulation over continuous and discrete (solitonic) components of the spectral data (Ros et al., 2019, Goossens et al., 2017).

Polarization multiplexing leverages the vectorial nature of light. In fiber systems, a two-mode representation $q(t,z) = [q_1, q_2]^T$ describes the orthogonal polarization states. The Manakov system,

$$j\partial_z q_1 = \partial_t^2 q_1 + 2(|q_1|^2 + |q_2|^2)q_1, \quad j\partial_z q_2 = \partial_t^2 q_2 + 2(|q_1|^2 + |q_2|^2)q_2,$$

captures polarization-averaged Kerr nonlinearity under fast mixing (Ros et al., 2019, Goossens et al., 2017). In metasurfaces, each nanostructure is characterized by a Jones matrix $J(\lambda, \theta)$, with up to three independent polarization channels per wavelength in a passive, planar, birefringent system (Bao et al., 2024).

Integrated, both spectral and polarization channels can be assigned, balanced, or adaptively controlled by optical, electronic, or computational means.

2. Techniques for Spectral and Polarization Multiplexing

2.1 Optical Fiber Communications

Linear and Nonlinear Domain Transmission: Dual-polarization nonlinear frequency division multiplexing (NFDM) encodes data jointly over the continuous and solitonic spectra in both polarizations, fully harnessing the Manakov-integrable channel. Mapping QAM or QPSK symbols onto both $R(\lambda)$ (continuous) and $b_{1,2}(\lambda_k)$ (discrete) allows up to $4\times$ capacity scaling over single-polarization linear methods, with propagation-induced nonlinearities integrated into the transmission model (Ros et al., 2019, Goossens et al., 2017, Span et al., 2018).

Orthogonal versus Non-Orthogonal Multiplexing: Power-domain multiplexing (NOMA concepts) introduced into dual-polarization OFDM systems overlays multiple QPSK-OFDM signals at different power levels, separated post-detection with successive interference cancellation. This hybrid spectral/polarization/power-division approach relaxes spectral orthogonality for improved aggregate rates (Wu et al., 2017).

Pilot-Tone Polarization Demultiplexing in Subcarrier Systems: Coherent digital subcarrier multiplexing (DSCM) can employ single-pilot-tone–based adaptive demultiplexing, where a pilot on one polarization enables ultra-fast $2\times2$ Jones-matrix estimation and demultiplexing for all subcarriers. A modified Godard phase detector calibrates receiver-side XY-skew to $<0.3$ ps, and a single-pilot method reduces digital equalizer complexity by $\sim4\times$ over traditional MIMO (Wang et al., 2024).

2.2 Metasurfaces and On-Chip Photonics

Metasurface Holography and Channel Scaling: Planar metasurfaces with birefringent pixels support three independent polarization channels (via Jones matrix decomposition) per wavelength; by encoding information in multiple wavelengths, record-breaking capacities of $3\times N$ independent holographic images have been demonstrated, with gradient-based DNN inverse design for per-pixel parameter optimization (Bao et al., 2024). Full-parameter multiplexing across polarization, spectral, and angle channels has yielded $>150$ independent channels in a single device (Wei et al., 2024).

Spin-Orbit and Pancharatnam–Berry Phase Multiplexing: Chiral meta-atoms (e.g., helices) on waveguides exploit polarization-dependent geometric phases (PB phases), enabling multiplexing and active swapping of Fano, Lorentzian, EIT, and antiresonant transmission line shapes by incident polarization control. This line-shape–momentum locking enables polarization-based multifunctional switching and sensing (Cheng et al., 2023).

Interferometric Modal Splitting in Integrated Spectrometers: In thermally driven silicon-photonic Fourier-transform spectrometers, guided TE/TM modes exhibit distinct temperature-derivative dispersions, which are exploited for broadband, high–polarization-extinction spectral separation on-chip, with >20 dB PER over 1480–1630 nm and without moving parts or polarization-sensitive detectors (Johnson et al., 9 Apr 2025).

Spectral Modulation for Polarimetry: Passive optical setups (achromatic QWR, multiple-order retarder, polarizer) encode linear polarization parameters into a single sinusoidally modulated spectrum, allowing simultaneous extraction of $P_L$ (degree of linear polarization) and $\theta$ (azimuth) via curve fitting, avoiding spatial or temporal multiplexing (0903.2735).

3. System Architectures, Algorithms, and Experimental Realizations

3.1 Fiber and Free-space Implementations

• Adaptive polarization control in self-homodyne, polarization-multiplexed-carrier links uses optical-power minimization in one PBS port with electronic gradient-descent on PC actuators, yielding >15 dB carrier-to-data power difference and robust separation after kilometers of SSMF (Anghan et al., 2018).
• Dual-comb, polarization-multiplexed fiber or solid-state lasers employ birefringent cavities to generate two combs (orthogonal SOPs) with finely tunable, stable repetition-rate differences, supporting high-SNR, phase-coherent dual-comb spectroscopy and polarimetry (Cuevas et al., 2022, Kowalczyk et al., 2020).

3.2 Integrated Photonics and Metasurfaces

• Gradient-based optimization, often incorporating trained DNN surrogates for electromagnetic response, is essential for high-dimensional metasurface channel allocation. Multidimensional inverse design enables channel isolation, crosstalk suppression, and diffraction efficiency control within fabrication-constrained parameter spaces (Bao et al., 2024, Wei et al., 2024).
• On-chip IMS spectrometers exploit thermally induced OPL shifts, with post-FT frequency domain windowing to extract polarization-resolved spectra from one photodiode, and NUDFT-based phase corrections for broad instantaneous bandwidth (Johnson et al., 9 Apr 2025).

4. Performance Metrics, Channel Capacity, and Limitations

Multiplexing Platform	Max Polarization Channels	Spectral Channels	Experimental Capacity	Crosstalk/Efficiency
Fiber NFDM (Ros et al., 2019)	2	Continuous+Discrete	BER<3.8×10⁻³ up to 3200 km	4× single-pol (in principle)
Metasurface (Bao et al., 2024)	3 (indep), 6 (correlated)	5 (demonstrated)	15 holograms (3 pol × 5 λ)	Crosstalk < few %, η ≈25%
Full-param. MS (Wei et al., 2024)	2 (circ. pol.), extendable	3 (demo), extend	150 ch (2 pol × 3 λ × 25 θ)	Crosstalk <5%, η up to 0.89
DSCM (Wang et al., 2024)	2	M subcarriers	SPT up to 10 Mrad/s tracking	4× EQ complexity reduction
IMS (Johnson et al., 9 Apr 2025)	2	1480–1630 nm	≥20 dB PER, Δλ ≈ 1–2.5 nm	PER limited by design/det range

Spectral and polarization multiplexing enable capacity scaling in proportion to the product of independent polarization and (dedicated or computed) spectral degrees of freedom. However, modal crosstalk, thermal/structural nonidealities, and practical DSP or design constraints impose efficiency and orthogonality limitations (Bao et al., 2024, Wu et al., 2017, Johnson et al., 9 Apr 2025). In metasurfaces, three independent linear polarization channels per λ are an upper bound for single-layer designs (Bao et al., 2024); correlated (non-orthogonal) channels can be addressed at reduced isolation or with algorithmic compensation (Bao et al., 2024). Fiber nonlinearities, polarization-mode dispersion (PMD), and noise-induced cross-correlation of nonlinear spectral amplitudes necessitate algorithmic innovations such as common–differential precoding (Span et al., 2018), and pilot-based calibrations in subcarrier-multiplexed systems (Wang et al., 2024).

5. Applications and Representative Use Cases

• Coherent optical communications: Joint spectral and polarization multiplexing in both linear (CO-OFDM, DSCM) and nonlinear (NFDM) domains yields net bit-rate enhancements, longer transmission reaches, and improved robustness against channel impairments (Ros et al., 2019, Goossens et al., 2017, Wang et al., 2024).
• Metasurface-enabled holography and image processing: Generation of multi-wavelength, multi-polarization holograms with dense channel packing and computationally efficient design for display, multiplexed sensing, and integrated photonic neural networks (Bao et al., 2024, Wei et al., 2024).
• Dual-comb spectroscopy/polarimetry: Intracavity polarization multiplexing in fiber and solid-state lasers overcomes the need for phase-locked sources and enables real-time Mueller-matrix imaging with high SNR and sub-Hz linewidths (Kowalczyk et al., 2020, Cuevas et al., 2022).
• Chip-based spectroscopy: IMS leverages modal dispersions for ultrabroadband, dual-polarization detection in integrated, compact, broadband, and high-PER spectrometers (Johnson et al., 9 Apr 2025).
• Programmable spectral–polarization line-shape control: Spin–orbit–PB-phase engineered systems support polarization-controlled switching, spectral filtering, and dispersive modulation with line-shape–momentum locking (Cheng et al., 2023).

6. Outlook and Future Research Directions

Emergent metasurface optimization frameworks are converging on the theoretical limits of independent channel count for static devices; further scaling may require dynamic/tunable meta-atoms, multi-layer designs, or harnessing higher-order (elliptical, OAM) polarization states (Bao et al., 2024, Wei et al., 2024). In fiber systems, extending the nonlinear Fourier framework to multimode and space-division multiplexing is an active area, promising further increases in capacity and nonlinearity mitigation (Goossens et al., 2017, Ros et al., 2019). IMS-based designs point toward generalizable, stimulus-agnostic multi-mode detection paradigms suitable for a range of classical and quantum photonics tasks (Johnson et al., 9 Apr 2025). Fundamental trade-offs among channel independence, loss, complexity, and fabrication tolerance will shape the practical adoption and limits of spectral and polarization multiplexing in all platforms.