Efficient coupling between standard single-mode fibers (SMF-28) and thin-film lithium niobate (TFLN) waveguides remains challenging due to significant mode mismatch between fibers and submicron waveguides. To address this issue, a two-stage mode-field transformation architecture for low-loss fiber-to-chip coupling is experimentally demonstrated. The approach decomposes the overall mode-matching process into two cascaded stages: an on-chip spot-size converter (SSC) with a triple-tip configuration that enhances fabrication tolerance, followed by fiber-based mode expansion to SMF-28 via an intermediate UHNA fiber. The particle swarm optimization (PSO) algorithm is employed to determine the SSC parameters in a multidimensional design space, significantly reducing computational cost compared with conventional parameter sweeping methods. The fabricated devices exhibit a coupling loss below −0.87 dB per facet at 1550 nm. Across the wavelength range of 1520 to 1630 nm, the coupling loss remains below −1.18 dB per facet. The proposed architecture enables low-loss, broadband, and robust fiber-to-chip coupling in TFLN platforms, providing a practical solution under stringent SMF-28 compatibility and fabrication constraints. Moreover, the PSO-assisted strategy offers an efficient pathway for multi-parameter optimization in integrated photonic devices.
As a cutting-edge signal processing technology, the microwave photonic phase shifter (MPPS) exhibits significant advantages, including wide bandwidth, low insertion loss, and high immunity to electromagnetic interference (EMI), making it highly promising for applications in phased-array antennas, radar systems, 6G research, and low-Earth-orbit (LEO) satellite communications. Conventional frequency-doubling phase-shifting schemes typically necessitate high-performance optical filters to extract specific sidebands, which often leads to limited frequency tunability, degraded environmental stability, and increased system complexity. We proposed a filterless MPPS architecture based on parallel dual-parallel polarization modulators (DPPolMs). The scheme utilizes two parallel DPPolMs to generate ±2nd-order double-sideband carrier-suppressed (DSB-CS) signals, which are then coherently superimposed to synthesize a single-sideband carrier-suppressed (SSB-CS) signal. By modulating the DC bias voltage of the phase modulator (PM) in the lower branch, a continuous 360 deg phase shift of the frequency-quadrupled microwave signal is successfully achieved. Simulation results demonstrate that the system achieves an optical sideband suppression ratio (OSSR) of 62.79 dB and an electrical sideband suppression ratio (ESSR) of 61.68 dB without the requirement of additional optical filters, confirming its robust performance and spectral purity.
Resolving how optical turbulence varies along a propagation path remains a key challenge for designers of free-space optical propagation systems. Instruments such as scintillometers and differential image motion monitors are commonly used, but only provide path-integrated turbulence estimates. Point sensors provide localized estimates of turbulence strength and can be used to generate path-resolved profiles when an array of point sensors are distributed along the optical path. However, this approach can be costly and complex to deploy in certain environments. Alternatively, a single point sensor can be mounted on a mobile platform that collects data while traversing the optical path, although this approach presents different logistical inconveniences. With these challenges in mind, there has been interest in using path-integrated measurements to estimate path-resolved turbulence profiles. One technique that uses this concept is a dual-beacon Shack–Hartmann-based turbulence profiler. We employ wave optics simulations to evaluate the performance of this approach. The turbulence distribution and propagation path are fully controllable, allowing us to compare turbulence profiling estimates to the user-prescribed turbulence conditions. We simulate the dual-beacon Shack–Hartmann sensing geometry and use the output to compute the differential tilt variances, from which we estimate the turbulence profile through regularized inversion. We apply truncated singular value decomposition and Tikhonov regularization to stabilize our profile retrievals across a range of turbulence strengths and distributions. Results show that meaningful profiles can be recovered over the central portion of a 1 km path, whereas sensitivity is reduced near the source and receiver due to the nature of the weighting functions that relate turbulence along the path to differential tilt variances. Additional analyses highlight the influence of noise, smoothing, and scintillation, particularly for cases with strong mid-path turbulence. These findings establish wave optics simulations as a powerful computational tool for evaluating profiling algorithms and highlight both the potential and the limits of dual-beacon tilt-based profiling approaches.
We propose and experimentally demonstrate a dual-pump-power-regulated chaotic laser system for bandwidth broadening and effective time-delay-signature (TDS) suppression. The system consists of two pump lasers, two erbium-doped fibers, a phase-shifted fiber Bragg grating, and a long single-mode fiber, forming a self-feedback double-loop resonant configuration. Theoretical analysis and experimental results reveal the dynamic mechanism underlying chaotic generation and bandwidth broadening. The effects of pump power variation on chaotic bandwidth and TDS are systematically investigated. By jointly regulating the two pump powers, the directly measured electrical bandwidth reaches 35.8 GHz within the ESA measurement range, accompanied by effective suppression of the delayed autocorrelation peaks associated with the feedback cavity. For higher pump-2 powers beyond the ESA measurement range, an estimated radio frequency bandwidth of 67.7 GHz is obtained from optical-spectrum-based beat-frequency analysis. The proposed scheme provides a simple and effective approach for bandwidth broadening and TDS suppression in optical-feedback chaotic laser systems.
We carried out an analysis of a free-space optical communication system enhanced by both the optical reconfigurable intelligent surface (RIS) and the radio frequency (RF) system, employing a hybrid multipulse position modulation-binary phase shift keying subcarrier intensity modulation. To avoid link blockage, data from the source are forwarded to the relay via a free-space optics (FSO) RIS. Data processing in the relay is managed through the decode and forward (DF) protocol. Next, a hybrid RF/FSO link is deployed from the DF relay to the destination. The destination combined the signals from both the RF and FSO links using maximum ratio combining. As the RIS element operates in the optical domain, the source-to-RIS and RIS-to-relay subchannels as well as the relay-to-destination FSO links all followed a Malaga distribution for the atmospheric turbulence. The relay-to-destination RF channel followed a Rayleigh distribution. The end-to-end outage probability and bit error rate of the system were analytically derived and provided in closed form. Next, the system was deeply analyzed in different strengths of atmospheric turbulence and pointing errors.
We demonstrate an antenna-integrated InGaAs/InP UTC-PD heterogeneously transferred onto a high-thermal-conductivity silicon carbide (SiC) substrate to improve heat dissipation at the substrate level. The device integrates a circular patch antenna (CPA) co-designed for the modified dielectric environment introduced by the SiC substrate and supports backside optical injection through a ground-plane aperture. Full-wave electromagnetic simulations predict efficient impedance matching and a realized antenna gain of 6.7 dBi near 300 GHz, with a broadside radiation pattern exhibiting a 3-dB beamwidth of ∼90 deg in both E- and H-planes. Thermal simulations further indicate that replacing the native InP substrate with SiC reduces the effective thermal resistance by ∼4 times, significantly suppressing junction temperature rise under high electrical dissipation. Experimental measurements using an optical heterodyne setup confirm stable operation with a responsivity of 0.15 A/W and bias-dependent saturation photocurrents reaching 17 mA at −2 V. Frequency-dependent characterization shows a peak THz emission near 300 GHz, and measured far-field radiation patterns agree well with the simulated broadside antenna characteristics. These results demonstrate that SiC-based heterogeneous integration effectively mitigates self-heating while preserving antenna radiation performance, enabling thermally robust and high-photocurrent photonic THz emitters.
We present a programmable envelope shaping technique for generating reconfigurable linear frequency-modulated (LFM) signals with tunable pulse compression characteristics. The proposed scheme employs a cascaded dual-stage modulation architecture: first, a parabolic temporal envelope is imposed on the optical carrier through an intensity modulator, followed by chirped spectrum generation via self-phase modulation in a highly nonlinear fiber; second, a programmable intensity modulator reshapes the envelope of the chirped optical signal by designing the electrical driving waveform. This approach allows dynamic configuration of various window functions (e.g., rectangular, Hanning, and Hamming), facilitating an adaptive trade-off between range resolution and sidelobe suppression performance. Numerical simulations demonstrate that the generated LFM signals, with a center frequency of 30 GHz and an instantaneous bandwidth of 8 GHz, exhibit distinct pulse compression performance under different window configurations, achieving peak sidelobe level variations ranging from 13.35 to 72.67 dB, thereby validating the effectiveness of this technique for microwave photonic radar waveform optimization.
This research work presents significant advancements in multispectral lidar technology, specifically focusing on the development of a sophisticated instrument that employs agile wavelength selection utilizing a Geiger-mode avalanche photodiode (GmAPD) camera. A method for obtaining 2D/3D multispectral point clouds is described and demonstrated. Unlike traditional lidar systems that rely on a single, fixed-frequency pulsed laser, this multispectral lidar simultaneously captures distance and spectral information through a single optical path. This capability is essential for achieving high spatial and temporal resolution in terrestrial mapping. System diagnostics and initial 2D data obtained from the developed system using the GmAPD camera will be discussed, highlighting its effectiveness in capturing pulse returns from targets with distinct spectral properties. The integration of this advanced technology promises to enhance the quality of collected data and improve overall mapping accuracy across diverse environments.
Suspended monolayer graphene exhibits intrinsically low ultraviolet (UV) absorption, which limits its effectiveness in UV photodetector applications. To overcome this limitation, we propose a plasmonic structure comprising an array of square silver plates positioned atop graphene, which is supported by a dielectric spacer on a silver substrate. This configuration demonstrates polarization-insensitive performance and maintains strong absorption over a wide range of incident angles. By systematically optimizing the geometric parameters and selecting appropriate metal materials, broadband and high-efficiency UV absorption in monolayer graphene is achieved, with a peak absorptance of up to 75% and the full width at half maximum (FWHM) of ∼140 nm. The underlying physical mechanisms are revealed through detailed analysis of the electric field distributions. We offer a promising strategy for the development of high-performance graphene-based UV photodetectors.
Comb- and multi-wavelength-source-based intensity-modulation/direct-detection (IM/DD) photonic convolution often adopts a flatten-then-shape workflow: one spectral shaper equalizes the wavelength-channel powers and another programs per-line powers to realize convolution kernels, with multiarm readout commonly used for signed weights. This programming path introduces shaping loss, calibration overhead, and link-budget penalties that grow with the number of wavelength channels. We present a waveshaper-free kernel-programming workflow based on calibration and selection of nonflat multiwavelength spectral states obtained from an OptiSystem four-wave-mixing simulation. A spectral-state library comprising 40, 145 simulated spectral states generated under admissible pump-switch configurations is constructed using 18 equally spaced continuous-wave pumps with 50 GHz spacing, an auxiliary detuned pump at 191.2 THz, and a 0.15 km highly nonlinear fiber, and target kernels are assigned to contiguous spectral windows via normalized shape matching under a nonnegative power-mapping constraint, i.e., single-ended optical powers directly encode only nonnegative weights, enabling single-ended detection. A physically constrained discrete search (pump sparsity and window feasibility), accelerated by Bayesian optimization, selects the best spectral state and window for a given kernel. In a single-layer replacement study using the LeNet convolutional neural network architecture on the Modified National Institute of Standards and Technology (MNIST) handwritten-digit dataset (the first 5 & times;5 convolution), the mapped kernels achieve competitive accuracy while eliminating per-line spectral control. Link-budget analysis indicates that, for a calibrated 5 & times;5 kernel, removing two-stage shaping can reduce the required kernel-writing optical gain/source power by similar to 11.5 dB. Practical considerations, including source requirements, finite-library coverage limits, spectral-state qualification, and robustness to drift/noise, are discussed.
Designing and testing hyperspectral imaging (HSI) systems with commercial-off-the-shelf (COTS) components, e.g., for satellites, is traditionally time-consuming because precise housing for the optical components must be manufactured for each configuration. When changes are required, costs and development timelines increase. We present a low-cost, open-source, modular HSI test bench for faster prototyping with COTS optics. The setup is 3D-printable via fused filament fabrication (FFF) or stereolithography (SLA), and the design is plug-and-play with parametric design tools, including OpenSCAD hosted on MakerWorld or other standard CAD tools. The test bench has been compared with a traditionally machined monolithic design, the near-infrared HSI V7, developed for the HYPerspectral Smallsat for Ocean observation (HYPSO) satellite mission. The printed parts are adjustable, can be printed quickly, and are easy to assemble. The parametric test bench exhibits an average full width at half maximum (FWHM) difference of 0.01 nm relative to the CNC-machined test bench without a diffuser. With a diffuser, the average FWHM difference is 0.94 nm. The results support the use of the parametric test bench for rapid, early-stage spectral screening based on field-imaging capability, center-line spectral response, spectral shift, preliminary spectral-smile comparison, and FWHM. The design allows the lenses to be removed independently without disassembling the setup, which saves time during testing. This enables rapid, low-cost production of custom optical test benches with interchangeable lens configurations and adjustable focus settings. The modular setup will be used to test different optical configurations for future hyperspectral imaging satellites.
Perfect vector vortex beam (PVVB) is a specific type of vector vortex beam. Its transverse beam size remains constant regardless of changes in polarization order and topological charge. This unique feature holds significant potential for applications in particle manipulation, optical communication, and laser processing. We propose a innovative method to generate PVVBs with arbitrary distributions on the hybrid-order Poincare sphere (HyPS). The method employs a cascaded system consisting of a spatial light modulator and a q-plate. We constructed a theoretical model of the system using Jones matrices. Based on this model, we derived the analytical relationship between the system parameters (fast axis angles of wave plates and q-plate order) and the coordinates (longitude and latitude) on the HyPS. The effectiveness of the proposed method was successfully verified through experiments. Furthermore, we utilized the cascaded system to generate double-ring PVVBs. Our research confirms that their overall radial dimensions are also independent of the polarization order and topological charge. This method provides a flexible and efficient scheme for PVVB generation and offers important potential value for related fields.
High-performance, miniaturized magneto-optical (MO) devices play a critical role in near-infrared (NIR) opto-electronic systems for energy-related and precision optical applications. However, conventional MO materials generally exhibit weak Faraday rotation and insufficient transmittance in the NIR band. We proposed a chiral magneto-plasmonic heterostructure based on bismuth-substituted yttrium iron garnet, in which a gold nanoparticle array with symmetry-breaking characteristics was fabricated on the surface. The results indicate that localized surface plasmon resonances can be effectively excited, and a synergistic enhancement of Faraday rotation and optical transmission in the NIR band can be achieved. We systematically investigate the effects of particle type and structural parameters on device performance and analyzes the underlying physical mechanisms. The results show that the proposed chiral heterostructure enables a more favorable balance between Faraday rotation and transmittance in the NIR band, enhancing the efficiency of MO devices and expanding their application range. Its overall performance metrics exceed those of conventional periodic structures. We provide a feasible structural approach for achieving high-performance MO responses in the NIR band and offers a theoretical and design basis for the development of compact non-reciprocal MO devices.
Optical frequency comb (OFC) generation using gain-switched (GS) semiconductor lasers combined with optical injection locking (OIL) provides a compact multiwavelength source for photonic systems. In GS-OIL operation, however, OFC flatness is strongly governed by modulation-driven carrier depletion and recovery dynamics in the injection-locked GS laser, which shape the spectral envelope through waveform-dependent time-domain gating. This paper presents a design-oriented numerical study of modulation waveform engineering in GS-OIL lasers, formulating the modulation waveform as a continuous temporal design space and establishing a direct linkage between the depletion-recovery window in the time domain and achievable flatness regions in the frequency domain. Using rate-equation time-domain simulations under fixed OIL conditions, Gaussian and square-wave modulations are employed as representative cases within this design space, and a modulation design map is constructed parameterized by the peak total drive current and an effective duration (duty-cycle-equivalent temporal support). The map identifies practical operating regions that maximize the number of consecutive OFC lines within a 3-dB power window while retaining the spectral-stability benefits of OIL. The resulting framework provides generalized, design-map-based guidelines that extend beyond specific waveform shapes for flat-comb optimization.
The massive expansion of next-generation networks and the Internet of Things has resulted in rising demands for high-data-rate wireless communication, which conventional radio frequency systems are unable to sufficiently accommodate. Free-space optical (FSO) communication has grown as a viable solution to address last-mile connection requirements. We propose a dual-polarized spatial multiplexed FSO system that employs four Hermite-Gaussian modes per polarization, forming 2 & times;4=8 distinct data streams at a single frequency. Each stream is modulated using a duo-binary scheme to enhance error performance. System performance is analyzed under various weather conditions and turbulence levels. The proposed design achieves the best performance under clear weather at a range of 1040 m with log (BER)=-5.75 and the lowest performance under heavy fog at 330 m with log (BER)=-5.46.
Accurate measurement of internal noncircular features remains challenging due to restricted accessibility and complex surface characteristics. In three-lobe bearings, the internal raceway exhibits periodic high-low points along the circumferential direction, and the relative displacement between these characteristic points provides an effective indicator of geometric variation. We present a fully noncontact measurement method based on an optical autofocus imaging principle for high-low point characterization of three-lobe bearing raceways. The method employs a compact camera-based imaging configuration and controlled axial focus scanning to exploit the relationship between image sharpness and axial position. Rather than reconstructing the complete internal profile, a limited number of representative high and low points are measured at selected circumferential angles by identifying best-focus positions to characterize the geometric variation features of the three-lobe raceway. A dedicated optical inspection system integrating a camera, a precision linear displacement stage, and a rotary stage is developed to accommodate the confined internal space of bearing raceways. Experimental measurements performed on a physical three-lobe bearing ring reveal clear circumferential high-low point variations consistent with the expected three-lobe characteristics, demonstrating the applicability of the proposed optical imaging approach for noncontact internal feature inspection.
