Ultrafast lasers, owing to their high peak power and ultrashort pulse duration, enable highly precise and localized energy deposition inside transparent glass through nonlinear absorption. This process can induce a variety of micro-modifications, including refractive index changes, nanogratings, and microvoids, and is often accompanied by stress-field modulation and elemental migration in the near-focus region. Such a unique processing mechanism provides an important foundation for three-dimensional micro/nanofabrication inside glass materials. Starting from the interaction mechanism between ultrafast lasers and glass, this paper systematically reviews different types of material modification and their corresponding processing windows, and further summarizes recent progress in applications such as on-chip photonic device fabrication, high-precision cutting, optical waveguide writing, stress-based waveplate fabrication, microchannel processing, and burst-mode ultrafast laser machining. Finally, the paper analyzes the current bottlenecks in processing consistency, mechanistic understanding, and industrial implementation, highlights the importance of inverse engineering for process optimization, and discusses the potential of artificial intelligence technique in complex parameter optimization and intelligent laser processing.

Dispersion scan (D-scan) is an ultrashort laser pulse characterization technique based on dispersion modulation and nonlinear spectral response. Owing to its extremely simple optical configuration and high sensitivity to broadband spectra and phase evolution, it has developed into an important tool in the field of ultrashort pulse characterization. Focusing on the ability of D-scan to meet the demands of real-time operation and robustness, as well as its extension toward extreme parameters such as single-cycle pulses and the deep-ultraviolet region, this paper systematically reviews the key progress of D-scan technology in terms of retrieval algorithm optimization and experimental scheme expansion. First, the evolution of D-scan retrieval algorithms is summarized. This progression traces the shift from early Nelder-Mead and differential evolution algorithms to the current standard generalized pulse retrieval algorithm, and ultimately to deep-learning-based techniques that enable millisecond-level, real-time reconstruction. Particular emphasis is placed on the improvements in computational speed, algorithmic robustness, and noise immunity achieved across these diverse approaches. Regarding experimental techniques, the paper reviews second-harmonic-generation (SHG) D-scans based on second-order nonlinearities and details the technological transition from conventional scanning methods to real-time, single-shot measurements, and highlights recent progress in applying SHG D-scans to vectorial optical field characterization. Subsequently, to break the physical limitations of second-order nonlinearities—specifically concerning multi-octave spectral overlap and phase matching in the DUV region—this review further explores D-scan techniques leveraging third-order nonlinear effects and their derivatives. It elucidates how these methodologies push the application boundaries of D-scan toward the single-cycle limit and into the DUV regime. Finally, current challenges confronting D-scan technology are outlined, including its reliance on external components and its extension to longer wavelengths and longer pulse durations. The paper concludes with an outlook on the future trajectory of D-scan technology within strong-field physics and attosecond science.

To achieve the generation of ultrashort isolated attosecond pulses using few-cycle pulses, it is necessary to study the precise manipulation of the electric field of few-cycle light field through double optical gating technology. In conventional experiments, double optical gating typically regulates multi-cycle pulses, and the analysis does not consider higher-order dispersion during laser propagation in media, second-harmonic conversion efficiency, or the exact waveform of the second-harmonic electric field. However, such approximations are no longer valid for few-cycle pulses. This paper accurately simulates the propagation and second-harmonic generation process of few-cycle pulses in nonlinear crystals based on a coupled-wave equation model, revealing the key influence of dispersion effects and other factors on the gating waveform. The research shows that when the driving light field is a few-cycle laser pulse, the traditional electric field estimation method for double optical gating is no longer applicable. Few-cycle pulse lasers have an ultra-broad spectrum, and the effects such as group velocity mismatch, phase mismatch, and dispersion caused by differences in phase accumulation among different wavelength components become significantly more pronounced compared to long pulses. For a few-cycle pulse, the optimal gating light field can be achieved by adjusting the thickness of the beta-barium borate (BBO) crystal in the double optical gating setup to 126.4 μm. This paper proposes that coordinated adjustment of the waveplate and BBO crystal thickness can finely tune the relative delay between the driving field and the second-harmonic field, thereby optimizing both the gating and driving electric fields, providing effective parameter optimization guidance for the generation of ultrashort isolated attosecond pulses.

To enable low-cost and efficient characterization of EUV photoresists, a lithographic evaluation system based on a tabletop high-harmonic generation (HHG) source and a reflective interferometer was developed. High-order harmonics were generated using a 515 nm femtosecond laser to excite argon gas. After focusing by a toroidal mirror and spectral dispersion by a blazed grating, the 11th harmonic (46.8 nm) was selected through a slit as the EUV source. Nonchemically amplified photoresists based on oxime sulfonate-functionalized polystyrene (PSOS) were used as the test material, and interference exposure was performed with a Lloyd’s mirror and a symmetric dual-mirror configuration. To solve the alignment problem between the sample plane and the symmetric dual-mirror interferometer, an optical ranging method based on diffraction-fringe inversion was proposed, enabling precise positioning of the symmetric dual-mirror interferometer. The results show that the Lloyd’s mirror produces clear high-contrast line-space patterns with a period of 125 nm and enables characterization of the tested resist at the 100 nm scale. After precise positioning by diffraction-fringe inversion, the symmetric dual-mirror configuration produces line-space patterns with a period of 60 nm, significantly improving the system resolution. This evaluation system provides a low-cost and scalable experimental platform for rapid screening of photoresist materials, investigation of resolution limits, and development of related lithographic processes.

This article provides an overview of the implementation of self-reference spectral interferometry (SRSI) from two main aspects: the optical path and the reconstruction algorithm. At the optical path level, from the earliest proposed SRSI based on cross-polarization wave generation (XPW-SRSI) to the SRSI based on transient grating effect (TG-SRSI) with a compact total reflection configuration, the sensitivity, wavelength coverage, and compactness of the implementation path have been continuously iteratively upgraded. At the algorithm level, from pulse reconstruction methods targeting near Fourier transform limits to reconstruction algorithms for large chirp pulses with temporal broadening exceeding twice the Fourier transform limit, three evolutionary paths are discussed: spectral stitching schemes, reconstruction schemes incorporating supervised deep learning, and neural networks without training embedded in physical forward models. The latter achieves single-shot accurate reconstruction in large chirp and high noise scenarios without any pre-training dataset. In the face of emerging new beam characterization demands, the temporal measurement of ultrashort laser pulses still require significant attention in the future.

Femtosecond-laser-induced self-organized interference, together with near-field enhancement and incubation effects, enables deep-subwavelength nanoetching. However, the formation of such structures usually relies on random surface scattering centers and multi-pulse feedback, leading to poor uniformity, limited repeatability, and high sensitivity to laser parameters. In this work, a spatiotemporal laser-energy modulation strategy is proposed to improve the writing quality of nanogrooves on 4H-SiC single-crystal surfaces. In the spatial domain, single-slit and double-slit beam shaping are employed to reconstruct the focal-plane energy distribution, suppress lateral energy spreading, and enhance the incubation and self-organized interference effects. In the temporal domain, a GHz burst mode is introduced to regulate the energy-deposition sequence, reduce the instantaneous energy-deposition intensity, and extend the annealing time. Static irradiation and dynamic scanning experiments were conducted to systematically investigate the surface morphology evolution and nanogroove formation mechanism under different modulation strategies. The results show that the elliptical focal spot formed by single-slit shaping enhances the incubation and annealing effects, producing nanogrooves with improved uniformity and edge regularity and a minimum groove width of 62 nm. Double-slit shaping further strengthens the interference-field confinement, reduces the dependence of nanogroove formation on random scattering centers, and decreases the minimum controllable groove width to 34 nm. By introducing GHz burst pulses on the basis of double-slit shaping, stepwise energy deposition through sub-pulses with a 400 ps interval reduces instantaneous strong excitation and suppresses molten redeposition and particle attachment, further decreasing the minimum controllable groove width to 24.5 nm. This spatiotemporal modulation strategy provides an effective technical approach and theoretical basis for high-precision, low-damage, and highly repeatable nanomanufacturing on third-generation semiconductor surfaces.

To meet the stringent requirements of attosecond science, precision manufacturing, and ultrafast spectroscopy for high-performance ultrashort pulses, we report the development of a highly stable, self-starting all-polarization-maintaining few-cycle femtosecond fiber laser. An all-polarization-maintaining erbium-doped fiber seed source with repetition frequency of 100.45 MHz is constructed using a single-walled carbon nanotube film as a saturable absorber. By employing an all-polarization-maintaining bidirectional amplification scheme combined with nonlinear spectral broadening and dispersion management, the system achieves a pulse duration of 30.1 femtoseconds after nonlinear broadening in a polarization-maintaining highly nonlinear fiber and subsequent dispersion compensation. This corresponds to a few-cycle pulse train containing approximately 5.8 optical cycles, with an amplified average output power exceeding 200 milliwatts. Stability measurements demonstrate that the seed source maintains a power root-mean-square fluctuation of only 0.4 percent over 10 hours of continuous operation. Furthermore, after the implementation of temperature control and piezo-electric transducer stabilization for repetition rate locking, the Allan deviation reaches 9.98 × 10⁻¹⁴ over 100 seconds. With its compact architecture and exceptional operational stability, this seed source establishes a robust experimental foundation for the subsequent generation of octave-spanning supercontinuum and high-performance frequency comb applications.

To retrieve the pulse information from the dispersion-scan (d-scan) trace, a differential evolution (DE) algorithm is used. A partially coherent pulse train is generated and then test by traditional DE algorithm and its improved version. The errors retrieved using the traditional and improved DE algorithms are 7% and 1%, respectively. The improved algorithm can more accurately retrieve the d-scan trace of partially coherent pulse train.

To address the signal reconstruction problem of high-Q torsion pendulum systems under ground seismic excitation in ordinary laboratory environments—pertinent to ground testing for space gravitational wave detection—a curriculum learning-driven Fourier feature network (FouCLNet) is proposed. Existing hardware isolation strategies impose stringent environmental requirements that are difficult to meet in ordinary ground laboratories. Traditional physics-informed neural networks suffer from spectral bias, gradient conflict, and an inherent tendency for outputs to decay toward zero in high-Q oscillatory systems. Log-spaced Fourier feature mapping is employed with frequency parameters uniformly distributed over the 0.003–0.02 Hz band to match the 0.007 Hz natural frequency of the pendulum. A four-layer fully connected network is constructed, and a hard amplitude constraint loss function is designed to prevent attenuation of micro-radian-scale signals. Through a three-stage curriculum learning strategy, weak physical constraints are progressively introduced with partial differential equation (PDE) residual weights annealed to the order of 1×10⁻⁸. Using the torsional response under seismic noise excitation as the target physical quantity and fourth-order Runge-Kutta integration results as ground-truth labels, we randomly sample 80% of time points for training and 20% for in-distribution validation. Experimental results demonstrate that the proposed method achieves a correlation coefficient of 0.9988 on the training set, peak error below 0.4% full scale, 100% amplitude matching, and power spectral density agreement at the 0.007 Hz resonance peak. The in-distribution discrete validation set achieves a correlation coefficient of 0.9958, and the single-point inference latency is merely 0.524 ms. Ablation studies indicate that removing Fourier features or hard amplitude constraints leads to severe model degradation. The pure data-driven configuration achieves good accuracy on the discrete validation set, yet its correlation coefficient drops sharply to 0.511 in continuous-segment reconstruction, whereas the full method maintains 0.986, indicating that weak physical constraints stabilize continuous temporal consistency. Temporal extrapolation beyond the training domain yields a correlation coefficient of merely 0.019, accompanied by spurious low-frequency drift. Furthermore, the model fails to generalize to continuous unseen periods within the training domain (R ≈ 0.029). Consequently, the proposed method is unsuitable for predictive extrapolation beyond the temporal scope of the training data and should be strictly confined to interpolation scenarios within the established training domain. The method provides a high-fidelity in-distribution interpolation reconstruction reference for torsion pendulum systems in ordinary laboratory environments and offers guidance for developing causal real-time vibration suppression algorithms.

This study addresses the difficulty of interpreting and separating multi-source coupling noise in gravitational reference sensors (GRSs) for spaceborne gravitational-wave detection. A unified acceleration-noise spectrum model is established for Brownian noise, thermal-field coupling, magnetic noise, electrostatic noise, drive-voltage noise, and residual low-frequency noise, with key parameters calibrated against LISA Pathfinder measurements. CNN layers are used to extract local transient features, BiLSTM layers are used to capture long-range temporal dependence, and adaptive spectral subtraction is then applied sequentially by physical noise category. At an input SNR of 10.2 dB, the proposed method achieves a recovery fidelity of 0.9694 and a waveform overlap of 0.9695, outperforming matched filtering, pure CNN, and pure BiLSTM baselines. Across an SNR range from −15 dB to +25 dB, the method shows a slower performance degradation in the negative-SNR regime. Combining physics-guided noise classification with CNN-BiLSTM temporal modeling improves signal recovery under complex GRS noise backgrounds and provides a useful reference for noise budgeting, simulation pipelines, and onboard denoising algorithms in spaceborne gravitational-wave missions.

The test mass (TM) in orbit is subject to the charging effect of high-energy particles, which interferes with gravitational wave detection. This paper addresses the charge management requirements of the TianQin Project. First, an electrostatic force model is established to determine the charge threshold of the TM, which is limited to less than 2×10⁻¹³ C. Then, a charge management structure using a UV LED light source is designed, and an engineering prototype is developed accordingly. Finally, a ground testing system based on a torsion pendulum is constructed to evaluate the charge management performance. Experimental results show that at 1 mHz, the resolution of charge measurement is better than 2×10⁻¹⁴ C, and the resolution of charge control is approximately 6×10⁻¹⁴ C, which meet the requirements of space gravitational wave detection. These achievements provide a solid foundation for TianQin project.

Space-based gravitational wave detection uses laser heterodyne interferometry to measure picometer-level displacement fluctuations of test masses separated by millions of kilometers. The interferometric system must achieve picometer-level accuracy in the millihertz frequency band. In the interferometer, test-mass rotation limits system sensitivity through two types of coupling errors: rotation–rotation coupling and rotation–translation coupling. This paper systematically investigates the mechanisms of these two errors and adopts a sequential suppression strategy: rotation–rotation coupling is first suppressed, followed by rotation–translation coupling. A test-mass interferometer is developed based on laser heterodyne interferometry and wavefront sensing, enabling high-sensitivity displacement and angular measurement as well as noise analysis. The coordinate transformation between the steering mirror and the detector is experimentally calibrated. The steering mirror is then rotated to the minimum-coupling angle, aligning the two coordinate systems and suppressing rotation–rotation coupling. An optical model is further established based on geometric relationships, and its parameters are experimentally calibrated. A real-time compensation system is developed to dynamically suppress rotation–translation coupling. After suppression, the rotation–rotation coupling coefficient is approximately 12.5 mrad/rad, and the rotation–translation coupling error is reduced by about 90% in the time domain and by approximately one order of magnitude in the frequency domain. These results provide a theoretical and experimental foundation for multi-degree-of-freedom decoupling and noise suppression in interferometers for space-based gravitational wave detection.

To satisfy the thermal deformation testing requirements of ceramic-based ultra-stable structures for space gravitational wave detection, and address the difficulty of verifying the measurement link prior to component fabrication, a ground-based vacuum interferometric measurement system operating within 1 mHz–0.1 Hz is developed. We use an equivalent 4J32 Invar sample for tests. It matches ceramic parts in geometry, interfaces and optical path. We calibrate thermal expansion, test displacement stability, and decompose low-frequency noise systematically. In the experiment, the tested structure, fiber-optic measuring probe, reflector and temperature sensors as a whole in a vacuum environment, and the natural cooling of the cavity is utilized to realize the overall thermal response calibration of the measurement link. Under mK-level steady-state temperature control, the Welch method is employed to acquire the amplitude spectral density of displacement noise, and a comparative analysis is conducted on the intrinsic noise of the interferometer, thermal equivalent noise and system comprehensive noise. The experimental results show that the relative deviation between the system comprehensive thermal expansion coefficient and the intrinsic value of Invar is about 8.3%, which can effectively characterize the overall thermal response of the measurement link. The amplitude spectral density of displacement noise of the system in the mHz band is approximately 26.6 nm/$ \sqrt{{\mathrm{Hz}}} $, and the consistency of long-period multi-segment steady-state measurement results is high, demonstrating high reliability. The intrinsic noise of the interferometer is about two orders of magnitude lower than the system comprehensive noise, and the thermal equivalent displacement noise converted from temperature fluctuation only accounts for 7.8%~14.1% of the system noise. Both of them are far lower than the comprehensive measurement noise and do not constitute major interference. The proposed system and method complete the full measurement link verification and noise benchmark establishment, and can provide an equivalent verification foundation and reusable technical scheme for subsequent thermal deformation testing of ceramic-based ultra-stable structures.

To address the issue that traditional static noise superposition methods in space-based gravitational wave detection neglect the dynamic coupling between multi-physics fields and the control system, making it difficult to meet the requirements of high-fidelity mission simulations, a full-chain dynamic simulation system for space-based gravitational wave detection is proposed and designed, which establishes the closed-loop feedback capability between the multi-physics fields and the control system. The system adopts a dynamic closed-loop architecture, comprising a spacecraft multi-physics field simulation module, a full-chain noise simulation module, a drag-free control simulation module, and a data processing and analysis module. In this architecture, the physical field states are computed based on the spacecraft states after control, and noises are generated through full-chain noise models mapped from the physical field states. The system achieves physical source tracing and dynamic simulation of full-chain noises. Simulation experiments reveal that self-gravity acceleration disturbances induced by the motion of the Movable Optical Sub-Assembly (MOSA) and fuel consumption can reach 10⁻¹³ m·s⁻²· Hz^−1/2 level in the observation frequency band（0.1 mHz-1 Hz）, necessitating compensation or subtraction through high-precision in-orbit measurement and calibration methods. The constructed system is capable of capturing dynamic coupling effects neglected by static models, thereby providing a high-fidelity simulation platform for mission design, noise source tracing and sensitivity evaluation in space-based gravitational wave detection.

The magnetic field near the test masses (TM) in space-based gravitational-wave detection spacecraft cannot be measured in situ, and the accuracy of magnetic field reconstruction is strongly affected by the arrangement of magnetic sensors. To address these problems, we investigate a magnetic sensor configuration optimization method under constrained installation conditions, aiming to improve the magnetic field reconstruction accuracy at the test mass locations. The magnetic sensor placement problem was formulated as a discrete combinatorial optimization problem. An improved IVY algorithm-based magnetic sensor configuration optimization method, termed MSC-IVYA, was proposed. The method integrates feasible installation region discretization, default-configuration-based population initialization, dynamic neighborhood updating, and a cumulative fitness function designed for multiple random magnetic source models, thereby enabling efficient search under installation constraints. Simulation evaluations were conducted on two representative space-based gravitational-wave detectors, LISA Pathfinder and Taiji-2, using three magnetic field reconstruction methods: inverse distance weighting (IDW), Taylor expansion (TE), and multipole expansion (ME). For LISA Pathfinder, under the default configuration, the average relative errors of TM1 were 593.74%, 508.04%, and 516.50% using IDW, ME, and TE, respectively. After optimization with MSC-IVYA, these errors were reduced to 390.39%, 357.55%, and 363.89%, respectively. In the Taiji-2 case, MSC-IVYA also achieved consistent improvement. For TM1, the reconstruction errors using IDW and ME decreased from 72.14% and 77.27% to 32.55% and 47.25%, respectively. For TM2, the errors using ME and TE decreased from 97.17% and 112.14% to 74.27% and 80.76%, respectively. Magnetic sensor configuration is an important design variable affecting the magnetic field reconstruction performance at the test mass locations. The proposed MSC-IVYA method can consistently improve magnetic field reconstruction accuracy under different mission conditions. It is particularly suitable for engineering scenarios with a limited number of magnetic sensors and constrained installation regions, and provides methodological support for the design of magnetic diagnostic systems in space-based gravitational-wave detection spacecraft.

This paper studies the phase meter applied to space laser interferometry. The phase-locked loop will suffer from lock loss in actual operation. Researchers commonly adopt the FFT frequency measurement method to re-acquire the signal at the present stage. This method has obvious technical defects. Its frequency measurement accuracy is low at the order of 100 Hz, and the relocking time is long about 7 ms. This paper proposes an automatic relocking technique deployed in collaboration with FFT. This technique adopts a lock-loss detection strategy that combines instantaneous frequency values and frequency change rates. It selects two data sources to judge lock loss, including the original data of the loop filter and the down-sampled data of CIC. It clears the integration error through the reset operation after lock loss occurs, and it receives the predicted value output by the frequency prediction algorithm. The frequency prediction algorithm uses the waveform generation algorithm for periodic signals. It uses the second-order polynomial prediction algorithm for aperiodic signals. It also combines interpolation technology to generate the corresponding frequency predicted value. The automatic relocking technique and FFT are deployed in parallel, and they form a clear functional division. This technique performs frequency prediction based on the inherent regularity of the signal. It deals with lock-loss scenarios of all regular signals regardless of the lock-loss duration. It also realizes fast relocking of short-time irregular signals within 1 s. FFT is responsible for signal re-acquisition in irregular signal scenarios and long-time complex lock-loss scenarios. The two methods form a working mode with complementary advantages. Experimental verification results show that the algorithm proposed in this study has an average relocking time of 32 μs and a maximum relocking time of 60 μs in the scenario of regular signal lock loss. The performance is improved by two orders of magnitude compared with the FFT method. The relocking speed has no correlation with the lock-loss duration. It can still maintain the relocking speed at the order of tens of microseconds when the lock-loss duration reaches 10 s. The frequency estimation error is stably controlled below 10 Hz in the signal-to-noise ratio range from −10 dB to 10 dB. The system can still achieve stable locking even when the signal-to-noise ratio is as low as −10 dB. This architecture deployed in collaboration with FFT retains the wide-band acquisition capability of FFT, at the same time it significantly improves the fast relocking capability in regular signal scenarios. It provides high-precision, fast-response and high-stability phase measurement technical support for space gravitational wave detection missions.

Focusing on inter-spacecraft laser link acquisition for space-borne gravitational-wave (GW) detection spacecraft, this paper investigates high-precision beam pointing control based on multi-source information fusion. Considering the structural and payload characteristics of the spacecraft, a complete system state-space model incorporating the motion of movable optical components is established. Using error quaternions, the nonlinear measurement equations described by non-independent variables are transformed into linear equations expressed in terms of independent variables. Measurements from inertial sensors are further fused for filtering and estimation, thereby improving the attitude determination accuracy of the spacecraft platform. In addition, to address the point-ahead requirement caused by the long laser propagation time over ultra-long inter-spacecraft distances and the high relative motion between spacecraft, a time-varying analytical expression for the point-ahead angle (PAA) is derived. The variation of the PAA over an orbital period is analyzed through simulation, providing a theoretical basis for servo compensation of the point-ahead angle during inter-spacecraft link acquisition and maintenance. Closed-loop simulations of a three-spacecraft system demonstrate the advantages of the proposed high-precision pointing estimation algorithm in terms of stability and estimation accuracy. Combined with a robust disturbance-rejection controller, the proposed method achieves high-precision beam pointing and provides essential technical support for the implementation of space-borne gravitational-wave detection missions.

To improve the multi-channel optical axis consistency of photoelectric pods and solve the problem of insufficient target center detection accuracy under complex imaging conditions in optical axis calibration, we propose two high-precision target center detection methods. For pulsed laser images affected by flicker and burn-mark interference, a staged weighted centroid method is proposed. This method performs coarse localization using large laser spots and then refines the center position using burn marks. For infrared cross targets with blurred edges and noise interference, an intersection optimization strategy based on Hough line screening and directional constraints is proposed. This strategy improves center localization accuracy under weak-edge conditions. Experiments were carried out using multiple laser image sequences and infrared cross-target images. The results show that the root mean square error of the staged weighted centroid method is 0.237 pixels, which is lower than those of the least-squares circle fitting method and the conventional centroid method. For 15 infrared cross-target images, the proposed Hough line screening method achieves an average error of 0.499 pixels and a maximum error of 0.974 pixels. These results demonstrate that the proposed methods can achieve sub-pixel center detection and show good stability and strong potential for engineering applications.

To address the high sensitivity of thermally induced diopter and the limited stable operating range of near-collimated propagation thin-disk multi-pass amplifiers under high-power and high-energy conditions, this work investigates the suppression of the thermal lensing effect based on pump light intensity distribution control. First, the relationship between thin-disk diopter variation and the pump light intensity distribution is analyzed based on experimental measurements of the thin-disk diopter. On this basis, an M-shaped pumping is adopted to replace the conventional super-Gaussian pumping. A theoretical model is established to comparatively analyze the thin-disk temperature distribution and diopter variation under both pumping techniques within a pump power density range of 0−8.13 kW/cm². The simulation results show that when the super-Gaussian order of the central depression region of the M-shaped pump is 8, the diopter variation of the thin-disk is minimized, with values of 0.00283 m⁻¹ and −0.00455 m⁻¹ in the horizontal and vertical directions, respectively. Compared with the traditional pump with a super-Gaussian order of 10, the diopter variations in the two directions are reduced by 0.05171 m⁻¹ and 0.06355 m⁻¹, corresponding to reductions of 94.7% and 93.3%, respectively. The M-shaped pumping can significantly reduce the thermally induced diopter variation of the thin-disk, which can provides more favorable conditions for mode matching over the full pump power density range and substantially mitigates the risk of optical damage caused by pump power fluctuations.

To overcome the limitations of current technologies, including the slow switching speed, heat accumulation, and high power consumption of thermo-optic waveguide switches, the high transmission loss of electro-optic waveguide switches, and the complex feedback control, difficulty in large-scale array integration, and narrow operating bandwidth commonly found in traditional photonic integrated circuit (PIC) waveguide switches, to meet the future demands of on-chip all-optical switching technologies for optical switches with fast response, low power consumption, broad bandwidth, low cost, and large-scale array fabrication capabilities, this study investigates micro-electro-mechanical systems (MEMS) driven silicon-based waveguide optical switch devices. By integrating silicon photonic waveguide technology with MEMS technology, we designed and fabricated an electrostatically driven MEMS 2×2 silicon waveguide optical switch (silicon photonic MEMS switch) capable of routing optical signals. The monolithic integration of silicon photonic waveguides and MEMS microactuators on a silicon wafer was achieved using electron-beam lithography combined with complementary metal-oxide-semiconductor (CMOS) processes. With a footprint of 192 μm × 192 μm, the device successfully demonstrated 2×2 optical switching functionality. The switching response times were measured at 20 μs and 15 μs, with optical signal rise and fall times of 15 μs and 10 μs, respectively. At a wavelength of 1550 nm, the device achieved an extinction ratio (ER) of 35 dB and an insertion loss (IL) of −0.8 dB. Over the 1500–1600 nm wavelength band, the ER remained above 20 dB, and the holding power consumption in the ON state was less than 0.5 μW. Experimental results demonstrate that this silicon photonic MEMS optical switch features fast response, low modulation power consumption, and excellent broadband performance. Furthermore, it can be fabricated into large-scale arrays with simple control mechanisms using existing process platforms, indicating significant application potential in future on-chip all-optical switching networks.

To meet the demand for multifunctional polarization manipulation and dynamic tunability of terahertz devices, a dual-functional switchable chiral metasurface based on graphene carrier modulation is proposed. By changing the polarization state of the incident wave and continuously tuning the Fermi level of graphene through an external gate voltage, the structure can exhibit circular dichroism (CD) and linear dichroism (LD) responses, enabling switching between different polarization-selective absorption characteristics. Simulation results show that when the graphene Fermi level is 1 eV, the metasurface exhibits pronounced selective absorption for left- and right-handed circularly polarized waves at 2.65 THz, with |A_CD| reaching approximately 0.89, and maintaining |A_CD| above 0.6 within the frequency range of 1.97–3.44 THz. When the Fermi level decreases to 0.2 eV, a significant linear dichroism response appears at 1.91 THz, with |A_LD| reaching approximately of 0.75. Analysis of the electric field and surface current distributions reveals that the difference in the strength of electric resonances excited under different polarization states is the primary mechanism responsible for polarization-selective absorption. In addition, the proposed structure demonstrates good robustness against variations in the incident angle and structural parameters, indicating its potential applications in circular/linear dichroism detection, polarization control, and terahertz photonic devices.

A light field camera incorporates a micro-lens array (MLA) in front of the CCD sensor to simultaneously capture both the spatial and angular information of the incident light rays, thereby enabling digital refocusing after image acquisition and generating sharply focused images at arbitrary depths in the object space. To achieve multi-plane velocity field measurements in three-dimensional (3D) flow fields, a multi-plane particle image velocimetry (PIV) method based on the backward ray tracing-based digital refocusing is proposed in this paper. First, the F-number matching relationship of the light field camera is re-derived to facilitate the implementation of the ray-tracing process. The vignetted and non-vignetted regions within the sub-images of the light field image are then calculated. Subsequently, backward ray tracing is employed to project the principal rays from vignetted and non-vignetted regions through the MLA and the main lens into the 3D physical space, thereby generating refocused particle images at different depth planes. Subsequently, the Expectation-Maximization (EM) algorithm is further applied to deblur the refocused particle images. Finally, the velocity field distributions at multiple depth planes are calculated using a two-dimensional (2D) cross-correlation algorithm. To validate the effectiveness of the proposed method, a submerged water jet experimental system is constructed and corresponding experiments are conducted. The experimental results demonstrate that the proposed multi-plane PIV method based on the digital refocusing technique can achieve measurements of multi-plane velocity fields in 3D flow fields.

Film thickness is a critical parameter that determines optical performance, and the accuracy of its monitoring system directly affects spectral characteristics. To mitigate the significant thickness control errors in conventional direct monitoring systems—caused by light source divergence and weak detector response signals—this paper proposes an externalized optical configuration. In this design, both the optical transmitter and receiver are placed outside the vacuum chamber, thereby avoiding interference from chamber’s vibration, temperature variations, and assembly inconsistencies. Additionally, an optical signal modulation scheme based on fiber coupling and collimation-focusing is introduced. By adopting an external integrated light source combined with multimode optical fibers and a composite optical path, and by optimizing component parameters through optical simulation to improve spot quality and energy density, the stability of both optical and electrical signals is enhanced. After optimization, irradiance at the fiber receiving end increased by 222.7%, signal strength by 156.6%, and the signal-to-noise ratio by 70.38%. The system’s performance was validated by preparing a narrowband filter film with a center wavelength of 2400 nm and a bandwidth of 40 nm, achieving a wavelength deviation within 1 nm over three repeated tests while consistently maintaining the 40 nm bandwidth. These results confirm that the system enables high-precision and stable film thickness monitoring even in spectral bands with weak detector response.

Supercontinuum (SC) generated from femtosecond laser filamentation has found extensive applications due to its broadband spectral properties. In this study, we present a novel method to simultaneously improve the spectral coverage and power density of SC. This is realized by combining two-color femtosecond laser injection and multi-filament array arrangement in fused silica. With this method, high spectral power SC is obtained. The spectral power density is above 0.1 mW/nm over a broad wavelength range from approximately 380 nm to 950 nm. We find that both the spectral range and power density of SC are affected by the input power and the intensity ratio between the fundamental and second-harmonic laser pulses. In addition, the spectral fluctuation of the generated SC is measured to be less than 4% within 6 minutes. These results offer a feasible and effective way to enhance the spectral power and coverage of SC sources. They are of great importance for promoting the practical applications of SC.

The circular swallowtail beams have recently exhibited better autofocusing ability and more tunability compared with low-order Airy or Pearcey catastrophe beams. Our attention is paid to exploring the metasurface generation and dynamics propagation of directional circular swallowtail (DCS) beams carrying power-exponent-phase vortices based on all-dielectric metasurfaces using finite-difference time-domain (FDTD) method, where the directional phase related to launch angles in x- and y- directions is considered. The combined influence of directional and power-exponent phases on dynamics propagation and orbital angular momentum (OAM) of the proposed beams is explored. It is found that their autofocusing positions can be freely adjusted along pre-designed trajectories owing to different launch angles. And rotation behavior and Archimede spiral structure originated from power-exponent phase can be also found during propagation. More importantly, the directional phase associated with launch angles can be regarded as the superposition of spiral spectrum, which can further extend OAM modes to wider multimode states compared with the non-directional cases. The OAM reduction in multiple modes with directional cases is smaller than that in non-directional cases during propagation, which indicates that multimode OAM spectra of our proposed beams provide potential for reducing OAM power attenuation in free-space propagation because the power decay is jointly undertaken by multiple modes rather than a single mode. This work may provide inspiration for guiding or trapping microparticles in three-dimension space as requirement, and for OAM-based optical communication and imaging by the modulation of multi-degrees of freedom associated with directional and power-exponent phases.

Owing to the limited interfacial contact area, insufficient carrier transport efficiency, and the adverse effects of intrinsic defects in ZnO on photoresponse, further improvement in the performance of conventional ZnO/GaN heterojunction ultraviolet photodetectors remains restricted. To address these issues, Ga+Al co-doped ZnO nanorod arrays with different Ga doping concentrations were grown on porous p-GaN/Al₂O₃ substrates by a low-temperature hydrothermal method, and the corresponding Ga+Al co-doped ZnO nanorod/porous GaN heterojunctions were fabricated. The porous p-GaN structure was used to enhance heterojunction interfacial contact and light absorption, while Ga+Al co-doping was employed to regulate the defect states and carrier transport properties of ZnO nanorods, thereby improving the self-powered ultraviolet photodetection performance of the devices. The findings of the study demonstrated that under 365 nm ultraviolet illumination at 0 V bias, the device with a Ga doping concentration of 3% exhibited the best UV photodetection performance, with a light/dark current ratio of 8250, a responsivity (R) of 0.158 A/W, a specific detectivity (D*) of 3.15×10¹² Jones, and an external quantum efficiency (EQE) of 53.5%. This study provides useful theoretical insight and experimental support for the development of next-generation high-performance ZnO/GaN-based heterojunction ultraviolet photodetectors.