Turbulence and air bubbles jointly affect optical propagation in near-sea-surface underwater wireless optical communication (UWOC) links, causing attenuation, irradiance fluctuations, and multipath dispersion. We present a semi-empirical model for a horizontal near-surface link with a spatially inhomogeneous bubble distribution associated with anisotropic turbulence. A midpoint-corrected Monte Carlo method (Midpoint MC) evaluates the spatially varying channel, and a reduced-order on-off keying (OOK) model characterizes its dispersion and bit error rate (BER) under weak irradiance fading. Spatially nonuniform bubble distributions broaden the channel impulse response, whereas increasing turbulence anisotropy smooths the modeled bubble-density field and reduces the estimated inter-symbol interference. For the simulated 30 m, 532 nm link at 600 Mbps, increasing the anisotropy factor from 1.0 to 5.0 lowers the estimated statistical-average BER floor from approximately 10 −3 to 10 −5 .
Semiconductor quantum dots embedded in photonic structures are perhaps the most promising sources of photons for on-chip classical and quantum photonic circuits. Recently, a quantum dot positioned in a relatively long, multimode photonic crystal cavity has been observed to emit light to the far-field asymmetrically from different parts of the cavity at different frequency detunings from the first-order cavity mode. In this work, we further investigate this phenomenon and find that the far-field radiation patterns are formed by the interference of light emitted directly by the quantum dot to free space and scattered by the cavity structure, while the field inside the cavity remains symmetric, in contrast to what has been assumed previously. Our further investigations show that, at larger detunings from the cavity resonance, the intracavity field also exhibits strong asymmetries due to the mode interference inside the cavity. This effect can be used to realize directional coupling of the intracavity light to nearby photonic-crystal waveguides. These results open, what we believe to be, new possibilities for controlling the emission of quantum dots in multimode photonic-crystal cavities, especially in the creation of reconfigurable and tunable photonic chips.
Silicon photonic design requires layout generation, simulation, interpretation, and verification across multiple tools, resulting in fragmented workflows that depend on expert operation and limit reproducibility. We present a tool-grounded, rule-guarded large language model (LLM) framework integrating domain knowledge, task parsing, tool invocation, workflow orchestration, and database checking. Starting from a LangChain-style A-flow, we develop an enhanced B-flow through deterministic execution and orchestration. Experiments comparing key system inputs with outputs show that constraining the LLM within a structured workflow improves execution reliability and substantially reduces token consumption. The enhanced B-flow provides a blueprint for model context protocol (MCP) services, enabling cross-client tool interoperability.
We report an experimental implementation of deep reinforcement learning (RL) for optimizing the loading stage of a 6 Li magneto-optical trap (MOT) in a high-dimensional continuous control space. An off-policy actor-critic agent observes in-situ fluorescence images at 111 ms intervals and updates seven experimental parameters over 18 decision steps. Learning is performed directly on the apparatus using a sparse terminal reward R = N × A extracted from an absorption image, where N is the atom number, and A is the peak optical depth. We benchmark Deep Deterministic Policy Gradient (DDPG) and Soft Actor-Critic (SAC) across three independent 300-episode training runs under identical alignment. Both algorithms surpass the human-optimized (HO) baseline, but the central finding concerns reliability rather than peak performance. In deterministic replay validation, the three SAC waveforms show reward coefficients of variation of 9–14%, and the minimum reward of each waveform exceeds the HO mean reward. The DDPG waveforms show up to three times the dispersion, and even the best-performing DDPG run contains a catastrophic near-zero shot. The SAC solutions are nearly time-independent and consistently converge to a blue-detuned repump laser with an elevated field gradient, reminiscent of compressed-MOT operating conditions. Our results show that off-policy actor-critic RL can autonomously optimize high-dimensional laser-cooling sequences in an operating ultracold-atom apparatus, and that SAC has a reproducibility advantage over DDPG in a real noisy environment.
To address spectral distortion from power fluctuations and the high trigger frequency and complex synchronization of acquisition cards in wavelength-swept interrogation of large-scale ultra-weak FBG arrays, we propose an auxiliary spectral–temporal reconstruction method using a broadband scattering enhancement point at the fiber front end. The scattering signal serves as both an optical power reference and a wavelength-division reference, enabling simultaneous segmentation, power calibration, and spectral stitching of multi-wavelength signals within a single scan. This eliminates power-fluctuation-induced spectral distortion and removes reliance on the source’s electrical pulses, reducing the timing frequency and hardware requirements. A 450 Hz system interrogated 1644 gratings with 0.6 m spacing; strain linearity R 2 = 0.997283 and sensitivity of 1.19 pm/με were achieved. Dynamic vibrations at 40.26, 57.63, and 74.93 Hz were accurately demodulated. The method shows great potential for rapid demodulation in large-scale ultra-weak grating sensor networks.
We propose a visible-light transparent ultra-broadband infrared reflectance asymmetric gradient multilayer film (VT-UBIRA film) composed of indium tin oxide (ITO) and silica. The proposed structure exhibits high visible-light transmittance and remarkable infrared reflectance contrast. In the visible spectral range (0.38–0.78 µm), the simulated peak transmittance reaches 0.9 with a band-averaged value of 0.72; the experimental peak transmittance is 0.8 with a band-averaged value of 0.6. In the infrared spectral range (1.1–14 µm), the simulated peak reflectance contrast is 0.88, with a band-averaged value of 0.71; the experimental peak reflectance contrast is 0.69, with a band-averaged value of 0.5. More importantly, the VT-UBIRA film avoids lithography and etching, enabling scalable, large-area transparent infrared devices. Based on the impedance matching principle, the asymmetric infrared reflection of the film is clarified. The epsilon-near-zero characteristic of ITO enables the decoupling of visible and infrared regions. Combined with the asymmetric geometry of the VT-UBIRA film, our design maintains high and symmetric transmittance in the visible band while independently manipulating infrared responses, thereby achieving asymmetric infrared reflection. The proposed structure possesses great application potential in various fields such as smart windows, infrared camouflage, and optical communication, offering new routes for asymmetric-reflectance device development.
