We propose and demonstrate an optical camera communication (OCC)-based novel optical over-the-air computation (OAC) method. In this method, a defocused camera image of an LED array realizes the optical aggregate computation, where the receiver estimates an aggregate function of the emitted intensities directly from the pixel-domain optical observation. Hence, it eliminates the need to decode each LED data stream separately, reducing the processing complexity significantly. The proposed direct OAC method is experimentally demonstrated using a proof-of-concept camera receiver. Across a wide range of transmission powers and aggregation weights, a single linear read-out attains a normalized mean-square error (NMSE) of 10[-5]-10[-4] (coefficient of determination R[2]>0.94), substantially outperforming channel-blind image summation. Because a learned linear read-out makes direct aggregation and separate-then-compute mathematically equivalent, the direct method matches the learned separate-then-compute baseline in accuracy while using a single aggregate-specific model instead of K per-emitter models. This complexity advantage grows in a strongly overlapped reflection channel, where the direct read-out stays accurate while a calibrated zero-forcing per-emitter inversion degrades by up to two orders of magnitude, suggesting a simple optical front end for aggregate computation.

A broadly wavelength-tunable Er-doped fluoride fiber laser with high power flatness was demonstrated in this work based on a 1650 nm pump scheme. Cascaded transitions of ground- and excited-state absorption processes at 1650 nm were used to motivate the population recycling among energy levels and suppress the reabsorption-induced wavelength tunability range shrinkage, as well as a red-shifted and steeper high-gain band. A wavelength-tunable range up to 157 nm from 2702 nm to 2859 nm was achieved at the watt level for the first time, to the best of our knowledge. High power flatness with a power difference less than 5% was realized in a wavelength-tunable range of 117 nm.

The fixed refractive indices of natural materials constrain the versatility of optical metasurfaces. Here, we introduce sub-5-nm-period nanohole arrays in monolayer MoS2, where edge-induced defect states reshape the excitonic optical constants. Combining GW-BSE many-body calculations with full-wave simulations reveals that these modified constants in a reflective metasurface configuration govern Fano-type interference between a grating-coupled hybrid resonance and the cavity background, tuning the response from resonant absorption enhancement to interference-dominated suppression. These findings establish few-nanometer defect engineering as a material-level design parameter for tailoring the optical response of metasurfaces.

On-chip terahertz (THz) photonics holds great promise for compact, high-speed communication systems that could empower artificial intelligence and cloud-based applications. Lithium niobate (LN), an excellent platform simultaneously enabling THz generation and modulation, unfortunately suffers from severe intrinsic material absorption, restricting the propagation length to the millimeter scale. Here, we propose a grating-assisted, heterogeneously integrated LN-on-silicon (LNOS) waveguide that enables low-loss on-chip THz transmission through non-Hermitian mode engineering. By tailoring the significant material loss contrast between the highly dissipative LN and the low-loss Si, we construct a two-level non-Hermitian framework where the THz wave preferentially evolves along the low-loss eigenstate within the silicon core. This approach suppresses the effective propagation loss to 0.05 mm[-1], representing much lower propagation loss compared to an uncoupled LN waveguide. The resulting structure also features a broadband 3 dB operational bandwidth of 0.16 THz, offering a viable route toward long-range, energy-efficient on-chip THz photonics.

In this study, UV phase mask and femtosecond (fs)-laser line-by-line (LbL) techniques are employed to inscribe tilted fiber Bragg gratings (TFBGs) into in-house fabricated all-cyclo olefin polymer (COP) optical fibers. Analysis of cross-sectional refractive index profiles shows that the UV phase mask technique induces a higher refractive index change in the cladding (3.3 × 10[-3]) than in the core (3.6 × 10[-4]), as the laser beam is absorbed by the cladding before reaching the core, despite similar absorption properties of both regions at the inscription laser wavelength. This technique also causes undesired physical deterioration within the irradiated fiber cladding region, which hinders the formation of cladding mode resonance comb. In contrast, TFBGs are successfully inscribed using the fs-laser LbL technique at tilted angles of 2°, 4°, and 6°, where the tightly focused laser beam results in a significantly lower spectral density in the cladding compared to the core, thereby avoiding mechanical degradation of the cladding. The structural changes of the all-COP optical fiber core and cladding are smooth and uniform, with a decrease in refractive index of 6.6 × 10[-3] and 3.2 × 10[-3], respectively. The resonance strength and spectral width of the cladding mode comb are up to 8.2 dB and 126.9 nm, respectively. This study serves as a foundation for TFBG fabrication in all-COP optical fibers, which offer advantages of long-term thermal stability and negligible water uptake over other polymers.

Silicon-based heterojunction phototransistors (HPTs) demonstrate advantages of high internal gain and compatibility with CMOS technology. A novel, to the best of our knowledge, nanohole-assisted SiGe/Si HPT was designed and fabricated to achieve high absorption and alleviate the contradiction between responsivity and speed of detectors. A 2D nanohole array with a 1000 nm period to support a lateral transmission mode was embedded into the absorption region through the optical window located on the base mesa. The diameter and depth of the nanohole were optimized to reach a higher optical absorption efficiency of 0.80. The nanohole-assisted SiGe/Si HPT was fabricated using a BiCMOS-compatible mesa process. Under 850 nm incident light, the novel device with a window radius of 20 µm exhibited a responsivity of 13.02 A/W and a specific detectivity of 2.56 × 10[12] cm·Hz[0.5]/W, representing an enhancement of 525.9% and 498.1% compared with the standard SiGe/Si HPT with the same structure but without a nanohole array. The nanohole-assisted Si-HPT demonstrates highly promising potential for applications in low-light detection.

Free-space optical transmission at 0.532, 1.55, 4, and 10 µm through real controlled fog with simultaneously measured droplet size distributions representative of continental and maritime fogs is reported. Measured attenuation is compared with predictions from Mie theory and two visibility-based models, the latter of which fail to predict transmission performance beyond the near-infrared. Although the 4 µm wavelength shows no meaningful advantage over 1.55 µm, long-wave infrared transmission at 10 µm is shown to provide a gain in transmission exceeding 20 dB in continental fog conditions.

Spatial optical TRNGs require entropy media with random texture generation, passive room-temperature readout, and global refreshability. Here, we demonstrate a spatially indexed TRNG based on thermally reconfigurable ferroelectric-nematic textures. Thermal reconfiguration creates random POM-readable textures that are experimentally verified to exhibit short-term temporal stability at room temperature and can be repeatedly read from predefined coordinates. Same-index scans are stable, whereas different coordinates and regenerated textures show low Pearson correlation coefficients and Hamming distances near 0.5. The median-thresholded raw stream gives a NIST SP 800-90B non-IID min-entropy rate of 0.0766 bits per raw bit, enabling SHA-256 extraction that passes NIST SP 800-22. A 6 × 6 cm[2] cell gives an upper min-entropy budget of ~0.11 Gbit. These results establish ferroelectric-nematic textures as addressable and globally thermally refreshable entropy media.

This paper proposes a new, to the best of our knowledge, method for measuring the focal length of lenses based on lateral shearing interference. In this method, two shearing interferograms are recorded at the initial CCD position and after a known axial displacement of the charge-coupled device camera (CCD). By using the resulting variation in normal fringe spacing of the shearing interferograms, the wavefront curvature radius at the initial CCD position can be determined without directly measuring the distance between the lateral shearing interferometer and the CCD. Subsequently, the wavefront curvature radius is combined with the vertical beam diameter ratio measured without and with the test lens, establishing a geometric relationship between the wavefront curvature radius and the lens focal length to calculate the focal length. Experimental results show that, for a biconvex lens with a focal length of 175.5784 mm and a biconcave lens with a focal length of -50.9785 mm, the relative measurement errors are 0.1018% and 0.3736%, respectively. The proposed method does not require complex optical path adjustment or an additional reference beam, providing a practical approach for rapid and stable focal length measurement.

Computer-generated holography enables precise modulation of optical fields, while generating high-quality quantized holograms remains challenging due to the non-differentiable quantization operation. In this Letter, we propose a gradient-based optimization algorithm with a statistical constraint strategy for quantized hologram generation. By incorporating the characteristics of the phase modulator, a statistical constraint is introduced during the optimization process to encourage phase values to cluster around target quantization levels. This strategy improves the optimization stability and bridges continuous optimization with discrete modulation. Numerical simulations and optical experiments demonstrate that the proposed method consistently improves reconstruction quality across a wide range of quantization bit depths.

We investigate coherent perfect absorption (CPA) and electromagnetically induced transparency (EIT) in a cavity QED system that includes a degenerate optical parametric amplifier (OPA). We derive the condition for CPA and identify two distinct regimes controlled by the control laser. When the control laser is resonant and induces EIT, DQI suppresses CPA, thereby making the system a near-perfect transmitter. When the control laser is detuned, DQI again suppresses CPA, but now the system acts as a near-perfect reflector. In the EIT regime, increasing OPA strength enhances both transmitted and intracavity intensities. At the optimal detuning of the reflector regime, the input light is reflected, and the intracavity intensity drops to zero.

Distributed acoustic sensing (DAS) based on scattering-enhanced fiber (SEF) eliminates interference fading, yet its reliance on discrete differentiation limits it to a quasi-distributed framework, leading to cross-unit signal distortion. We propose a fading-free, fully distributed SEF-DAS scheme that uses a frequency-domain phase-shift transformation. A targeted frequency-spectrum transformation reconstructs spatially complementary optical fields individually for each orthogonal polarization state. A rotational vector sum is then applied to aggregate and suppress interference and polarization fading, filling discrete sensing chasms. Experimentally, continuous, blind-spot-free acoustic field restoration with 5-m spatial resolution is achieved over a 26.4-km range, delivering a strain sensitivity of 2.21 nε/√Hz at the link terminal at 1 mHz.

In this work, we demonstrate a high-range-resolution LiDAR detection scheme based on mode noise characteristics. This approach achieves millimeter-level point-cloud imaging at low sampling rates below 100 MHz, effectively reducing the dependence of high-resolution ranging on high sampling rate digitization and broad tuning bandwidths, thus significantly lowering system costs. The proposed scheme eliminates the need for external modulation and has minimal requirements for the laser linewidth. By employing an unbalanced interferometer, the system exploits intrinsic laser frequency fluctuations, converting them into broadband and sufficiently intense mode noise suitable for ranging. Through differential processing, the two time-delayed mode noise signals undergo coherent cancellation, which yields pronounced frequency dips at specific locations within the power spectrum. Theoretical analysis reveals a direct relationship between the center frequencies of these dips and the relative time delay. By utilizing nonlinear curve fitting to extract the dip profiles, we achieve high-precision center frequency extraction, enabling direct distance retrieval in the frequency domain. Experimentally, the system achieves millimeter-level ranging performance at an exceptionally low sampling rate of only 61.5 MHz. Specifically, the ranging accuracy was measured to be 1.3 mm at a detection distance of 7 m and 3.07 mm at 20 m while successfully resolving adjacent targets separated by a small gap of 4 mm.

Distributed acousto-optic sensing (DAS) enables continuous, high-resolution acoustic monitoring over kilometer-scale distances using standard optical fibers. However, this sensing technique faces limits concerning loss, such as those inherent to passive optical networks (PON). The optical synchronization of remote DAS interrogators with centralized narrow-linewidth light sources that have been off-loaded from the field-distributed sensing units enables to probe the fiber plant of PONs from the branch-side of their power-splitting signal distribution elements. Towards that, acousto-optic acquisition is demonstrated for PON reaches of 14.1 km and large 1 × 128 split ratios.

A high-order spatiotemporal optical vortex (STOV) singularity generally splits into multiple first-order ones upon propagation. Here, we demonstrate the arrest of this splitting via electromagnetically induced transparency. This is achieved by establishing the condition for diffraction-dispersion balance; solving this condition yields the specific system parameters that arrest the splitting. These solutions, located on the red-detuned side of absorption peaks, correspond to anomalous group-velocity dispersion and normal second-order loss dispersion. With the splitting arrested, we further characterize the spatiotemporal drift dynamics of the STOV singularity by deriving its drift trajectory and velocity. This drift, driven by first-order loss dispersion, exhibits a chiral asymmetry in direction dictated by the sign of the spatiotemporal topological charge.