Integrated optics is primarily based on planar designs due to the availability of mature lithographic manufacturing and optical confinement constraints. These 2D designs with finite thickness in the third dimension are often referred to as 2.5D. Full 3D photonic architectures, with refractive-index variations along the three dimensions, hold the promise of higher integration density and novel, to the best of our knowledge, light control capabilities, but require advanced multi-layer stacking techniques with precise alignment and planarization. Nanoscale 3D printing techniques, such as multi-photon lithography, can address these challenges, and are gaining momentum thanks to their cost-effectiveness and rapid prototyping capabilities compared to silicon foundries. Despite this potential, the exploration of freeform polymer optics at the nanoscale remains limited due to challenges associated with low-index materials and a lack of design tools. Here, we address these limitations by applying a multi-layered inverse design approach for polymer-based integrated optics. We systematically compare 3D with 2.5D designs (all simulations are conducted in 3D), for the task of demultiplexing two wavelengths with spectral spacing from 100 nm to 20 nm. Our numerical results show that fully 3D polymer designs consistently outperform their 2.5D counterparts, achieving higher efficiencies at equal footprint. These findings propel the advancement of a next generation of miniaturized 3D devices for polymer-based integrated optics.
Simultaneously achieving giant chirality, high-efficiency nonlinearity, and robust nonreciprocity remains elusive due to the inherent trade-off between the structural asymmetry required for chiroptics and the high quality (Q) factors essential for nonlinearity. Here, we overcome this fundamental limitation in silicon metasurfaces by exploiting symmetry-protected bound states in the continuum (BICs). By implementing a two-stage symmetry-breaking mechanism that combines an in-plane structural rotation with out-of-plane height perturbations, we engineer high-Q (104) resonances dominated by magnetic quadrupole modes. Linearly, the device exhibits robust asymmetric transmission with a forward circular dichroism (CD) of 0.98 and a backward CD of -0.79. Leveraging intense near-field confinement, we realize third- and fifth-harmonic generation efficiencies of 3.6 × 10-3 and 2.9 × 10-9 at a moderate pump intensity of 1 MW/cm2. Crucially, the chiral-nonlinear coupling enables perfect nonlinear nonreciprocity, yielding near-unity nonlinear CD values of ±0.99. This work establishes a compact paradigm for high-performance chiral nonlinear isolators.
Non-modulated pyramid wave-front sensors (PWFS) deliver high sensitivity for extreme adaptive optics (ExAO) yet suffer from severe nonlinearity. While deep learning can alleviate these issues, conventional models rely heavily on large-scale labeled datasets and lack physical interpretability. To overcome these limitations, we propose PINN-Pyr, a physics-informed neural network for self-supervised wavefront reconstruction. By embedding a differentiable forward optical model into a U-Net, the network maps intensity patterns to Zernike coefficients without paired training data. PINN-Pyr significantly extends the linear dynamic range of the PWFS owing to its nonlinear mapping capability. Furthermore, among the conventional matrix-vector-multiplication (MVM), pure data-driven, and PINN-Pyr methods, PINN-Pyr achieves the lowest residual root-mean-square (RMS) error and superior Strehl ratio (SR) stability under strong turbulence and low signal-to-noise ratio (SNR) conditions. This physics-constrained scheme provides a robust, efficient solution for next-generation large-aperture telescopes.
Holographic optical elements (HOEs) manipulate light via refractive index modulation and are widely used in imaging, display, and extended reality systems. However, accurate phase characterization remains challenging due to the coupling of surface figure and internal refractive index effects. In this Letter, we present a phase measurement method based on four-step phase-shifting interferometry that separately measures the total and surface phase distributions of HOEs by leveraging their angular selectivity. The internal phase component, induced by refractive index modulation, is decoded from the total phase to eliminate the surface effect. A phase compensation experiment is conducted to validate the accuracy of the measured phase, demonstrating the effectiveness and practicality of the proposed method for HOEs phase characterization and correction.
We report a fully fibered, polarization-maintaining, single-frequency thulium-doped fiber amplifier operating between 1730 and 1762 nm and delivering more than 30 W of output power. Relative intensity noise is preserved after amplification, with a noise floor of -158 dBc/Hz between 100 kHz and 4 MHz. These results demonstrate the suitability of this source for quantum technology, including direct operation within the tuning range at 1762 nm for barium ion excitation or a possible nonlinear frequency conversion toward 813 nm for strontium atom trapping.
A 1D/2D switchable grating based on polymer-stabilized liquid crystal (PSLC) is proposed using a two-step exposure method. The first exposure anchors the tilt angle of liquid crystal (LC) molecules via the polymer network, forming a 1D grating in the absence of an electric field. The second exposure modulates the threshold voltage of LCs through the polymer, and a phase difference is generated between the exposed region and the unexposed region under an external field, thus forming a 2D grating. The device features low driving voltage (1 V), fast response, high diffraction efficiency, no scattering, and low cost. Moreover, the gratings corresponding to the two exposures are independent and non-interfering, and diverse diffraction patterns can be flexibly constructed by adjusting the period and stripe orientation of the two gratings, showing great potential in photonic applications.
We demonstrate the generation of 20.2 W of continuous-wave, low-intensity noise, single-frequency light at 840 nm, obtained via sum-frequency mixing of two high-power infrared fiber laser systems in an MgO-doped periodically poled lithium niobate (ppLN) crystal. Subsequent cavity-enhanced frequency doubling to 420 nm yields an output power of 15.2 W. Respective conversion efficiencies of 48% and 79% are achieved for the sum-frequency and second-harmonic stages, with good beam quality (M2 < 1.3), low intensity noise (<0.05% RMS [1 kHz-10 MHz]), and long-term continuous operation. Relative intensity noise and frequency noise of the generated signals are experimentally characterized.
The strong circular dichroism (CD) of chiral structures is essential in applications such as sensing, polarization manipulation, and imaging. However, to achieve strong CD for practical applications in most chiral metamaterials and metasurfaces, the symmetry of the structure must be deliberately broken, which greatly increases the complexity and process errors of micro-nano fabrication. Here, we achieve the multi-wavelength excellent CD governed by a pair of chiral bound states in the continuum (BICs) and a pair of chiral quasi-BICs in a magneto-optical photonic crystal (MO PhC) slab without structural symmetry breaking. By applying an external magnetic field instead of breaking the structural symmetry, both the chiral BICs and the quasi-BICs exhibit high Q-factors and CD values, with the Q-factors of up to 106 and 103 and the maximum CD values of up to 0.98 and 0.99, respectively. Furthermore, the results show that the Q-factor of the chiral BICs/quasi-BICs perfectly aligns with a linear function relationship with the external magnetic field, thereby enabling the tunable quantitative Q-factor. The performance of the proposed MO PhC at four wavelengths is comparable to that of traditional single-wavelength chiral structures, giving it significant prospects for various applications, including optical security, chiral lasers, and chiral optoelectronic devices.
How nonspherical raindrops scatter light and how their scattering deviates from spheroidal approximations remain poorly quantified owing to complex geometries and extreme size parameters (χ>104). Here, we extend the three-dimensional vectorial complex ray model (VCRM3D) to investigate light scattering by Beard-Chuang (BC) raindrops, the established equilibrium model for realistic falling drops. Validated against Lorenz-Mie theory and the multilevel fast multipole algorithm, our results characterize the light scattering signatures of BC raindrops and quantify their discrepancies from those of spheroids. Unlike the monotonic increase for spheroids, BC backscattering efficiency Qback exhibits non-monotonic behavior governed by surface curvature, peaking at D≈4.2 mm before declining.
This work demonstrates an all-diamond solar-blind ultraviolet photodetector employing a heavily boron-doped diamond (BDD) thin film as a transparent electrode. This monolithic structure overcomes the optical blocking inherent to traditional metal electrodes while maintaining excellent interfacial characteristics. Under 222 nm illumination and a 10 V bias, the device achieves a responsivity of 49.8 mA/W, approximately one order of magnitude higher than that of a Ti-electrode counterpart. The UV-to-visible rejection ratio reaches 4.54 × 104, accompanied by a high photocurrent-to-dark current ratio of 441.8. Dynamic response measurements reveal fast rise and decay times of 0.13 s and 0.03 s, respectively, significantly outperforming the metal-electrode device. These results highlight the significant advantages of the all-diamond BDD electrode in responsivity, speed, and spectral selectivity, providing a promising approach for high-performance solar-blind UV detection.
A high-efficiency ultraviolet (UV) optical wireless communication link is proposed and demonstrated between unmanned aerial vehicles (UAVs) employing a novel azimuthally omnidirectional optical antenna based on an annular compound parabolic concentrator (ACPC). The receiving antenna offers a 114° field of view (FOV), covering polar angles from 10° to 124°, while achieving a maximum omnidirectional gain of 21.6 dB at a distance of 9 m. With a data transmission rate of 9.6 kbps, an outdoor transmission of live images between two in-flight UAVs successfully validates the link performance.
Conventional inertia-free multiphoton microscopy (MPM) suffers from depth-dependent degradation of lateral and axial resolution during axial scanning due to defocus-induced aberrations. We introduce truncated non-diffracting beams (tNDBs), generated by segmenting a Bessel beam into multiple axial sections with equalized focal lengths, enabling depth-invariant 3D focusing in laser-scanning MPM. The tNDB maintains constant lateral (~540 nm) and axial (~6 µm) resolutions across a ~29-μm depth range. The method is validated using bead phantoms and mouse brain tissue, consistently revealing sharper structures across depth. tNDBs provide an inertia-free and aberration-robust solution for fast volumetric MPM, enabling constant-resolution 3D imaging without modifying the detection geometry.
The anisotropic electronic properties of SnSe provide distinctive opportunities for studying polarization-dependent carrier dynamics. However, resolving such directional effects remains challenging for conventional spectroscopy techniques. Here, time-resolved high-harmonic spectroscopy (TR-HHS) is utilized to capture the anisotropic ultrafast processes in SnSe thin film. The TR-HHS signal evolves over the hundred-femtosecond to few-picosecond regime and demonstrates strong crystal orientation dependence. These results are attributed to anisotropic electronic relaxation encompassing electron thermalization and phonon-mediated hot-electron cooling. Our findings demonstrate that TR-HHS can serve as a powerful all-optical tool for probing symmetry-governed electron dynamics in anisotropic materials and provide new strategies for developing polarization-tunable optoelectronic devices.
A silicon single-photon avalanche diode (SPAD) fabricated in a 40 nm CMOS image sensor (CIS) technology is reported. An N-well/deep P-well (DPW) structure on a CIS epitaxial wafer is employed, where the N-well and lightly doped epitaxial layer form a virtual guard ring that suppresses edge breakdown. The device exhibits a breakdown voltage of 22.7 V, a peak photon detection probability (PDP) of 89.4% at 660 nm and 3 V excess bias voltage, and a median dark count rate (DCR) of 0.6 cps/µm². A timing jitter of 144 ps full width at half maximum and an afterpulsing probability below 0.3% are achieved. Temperature-dependent measurements show a pronounced PDP redshift at 60 °C, accompanied by enhanced peak and long-wavelength PDP. The SPAD is suitable for high-performance low-light visible imaging applications.
Precise manipulation of on-chip optical modes using subwavelength structures is critical for miniaturizing devices and scaling high-speed optical interconnects, especially in Mode Division Multiplexing (MDM) systems. Leveraging advances in nanofabrication, devices engineered with subwavelength features enable versatile control over on-chip optical fields. Here, we propose a compact mode sorter based on a subwavelength waveguide grating, designed using the eigenmode expansion method. We analyze the optical field distribution at the focal plane of the on-chip lens for various incident modes. Implemented on a silicon-on-insulator (SOI) platform, the device achieves efficient separation of TE0, TE1, and TE3 modes, within a footprint of only 8.3 µm. When assembled into a complete (de)multiplexer, the device yields a 0.79 dB, 0.97 dB, and 0.91 dB insertion loss for the TE0, TE1, and TE3 channels, and 114 nm (1493-1607 nm) bandwidth for all channels with insertion loss is under 1.5 dB while crosstalk below 15 dB. Our simulation of fabrication tolerances demonstrates that this approach offers superior robustness compared to traditional devices. This work presents a generalizable design framework for metamaterial-based on-chip lenses using eigenmode analysis, paving the way for compact mode splitting and reconstruction applications.
Free-space and on-chip photonic systems are key components in optical communication networks. While free-space beams allow for the flexible generation and manipulation of spatial modes, integrated waveguides provide compact and stable platforms for on-chip signal processing. Bridging these two domains is essential for scalable multi-mode communication networks. Here, we present an efficient, broadband interface capable of converting multiple higher-order free-space Laguerre-Gauss (LG) modes into corresponding waveguide modes using the multi-plane light conversion (MPLC) scheme. We experimentally demonstrate low-crosstalk mode conversion between various sets of three LG modes and the first three TE modes of a multimode silicon waveguide across the telecom C-band. The system operates passively without active switching and can be adapted to different spatial mode sets. This platform provides a pathway to increased data capacities and may enable more compact and efficient multi-mode optical communication and on-chip processing schemes.
Optically active networks show feature-rich emission that depends on the fine details of their geometry and find diverse applications in random lasers, sensing devices, and photonics processors. In these and other systems, a thorough and predictive characterization of how the network geometry correlates with the resulting emission spectrum would be highly important; however, such an outright description is still lacking. In this work, we take a step toward filling this gap by using the well-known Steady-state Ab initio Laser Theory equations to carry out an extensive set of statistical analyses and establish connections between the random network geometry and their ultimate emission spectrum. Our results show that edge crowding (abundance of short edges in the network) is key to tuning the uniformity of the modal intensity distribution of the emission spectrum. A statistical framework for the comprehensive understanding of the network statistical properties is highly significant to establish precise design rules for network-based photonic devices and intelligent systems.
We study photonic meta-atoms, a unique class of composite solitary wave supported in nonlinear waveguides. We establish an analogy to one-dimensional soft-core atoms, allowing us to describe the complex dynamics via concepts from atomic physics. Higher-order dispersive effects cause specific spectral resonances characteristic for the eigenspectrum of a meta-atom. We demonstrate that subtle changes in this level of spectrum cause frequency shifts of the resonances. These shifts consist of isotopic and isomeric contributions that can be distinguished in terms of a simple model. We further demonstrate a generic mechanism that causes a Zeeman-like splitting of resonance lines.
Improving the wall-plug efficiency (WPE) of AlGaN-based deep ultraviolet light-emitting diodes (DUV LEDs) requires the optimization of both the p-type ohmic contact and the reflectivity, which remains a significant challenge. In this study, we employ a simple and effective Au-sacrificial-layer process, achieving the lowest specific contact resistivity of 6.1 × 10-6 Ω·cm² for the Ni/Rh/Ni/Au contact on p-GaN-6 times reduction compared to the reference device. This approach enhances the p-type ohmic contact via Ga vacancies generated by the annealed Au layer, while the wet-etching reduces optical absorption to preserve the high reflectivity of the p-contact. The DUV LEDs fabricated with the 10-nm Au-sacrificial-layer process exhibit 0.69 V reduction in operating voltage and 15.54% enhancement in WPE at an injection current of 100 mA, compared to the reference device. This research provides a cost-effective strategy to improve the WPE of mass-produced AlGaN-based DUV LEDs.
We report on a high-average-power, high-energy cryogenic Ho:YLF picosecond chirped-pulse amplification system operating at 1 kHz repetition rate. High gain characteristics and excellent thermal management were well achieved by integrating cryogenic cooling with in-band pumping. A maximum compressed average power of 205 W at 1 kHz repetition rate with a pulse duration of 1.7 ps was delivered by the system. A high extraction efficiency of 41.2% and near-diffraction-limited beam quality (M2 = 1.14) were demonstrated.