Multimode mode-locked fiber lasers provide an effective route for realizing high-energy, high-power fiber oscillators, but the speckled beam profiles limit their broader applications. Utilizing the beam self-cleaning effect induced by high-power pulse propagation, we experimentally obtain nearly single-mode output beam profiles across multiple pulse regimes, including dissipative solitons, soliton molecules, noise-like pulses (NLPs), and Qswitched noise-like pulses (QSNLPs). Furthermore, transitions among these regimes are controlled by adjusting the pump power and polarization state. In the multimode QSNLP regime, a high Q-switched envelope energy of 60.6 mu J is achieved at a pump power of 6.5 W. This regime also supports central-wavelength tunability for singlewavelength operation and spacing tunability for dual-wavelength operation, both related to the multimode interference effect. To the best of our knowledge, this work presents the first experimental demonstration of beam self-cleaning in the soliton molecule and QSNLP regimes, as well as the first report of unique features of multimode QSNLPs. The results contribute to a better understanding of nonlinear dynamics in multimode fiber lasers and facilitate their practical applications.

Interfacial excitons (IXs) in CuPc/CdSe nanowire (NW) heterostructures (HS) show enhanced nonlinear optoelectronic properties through hybridization, but their resonant dynamic behavior is still not well understood. Our combined experimental and theoretical study shows the resonant excitonic nature of these hybrid states: enhancement of second-harmonic generation (SHG) around 725 nm (B-exciton) and 835 nm (A-exciton) demonstrates the coexistence of IXs (760 nm) with efficient charge transfer, increasing the SHG intensity by about more than 14 times in CuPc/CdSe NW interface. First-principles DFT calculations confirm the type-II band alignment and hybridized interfacial states (1.62 eV) resulting from orbital interactions between the 2D-organic (CuPc) and 1D-inorganic (CdSe) components. However, transient reflection (TR) spectroscopy confirms the modified ultrafast charge transfer at the CuPc/CdSe NW interface, driven by type-II band alignment. TR spectroscopy supports intralayer A- and B- excitons resonance, attributed to Forster resonance energy transfer. On the other hand, the IXs peak shows a faster decay than bare CdSe NW, enhancing excitonic resonance signatures due to strong interlayer coupling. These findings advance our understanding of many-body exciton physics in 2D-organic/1D-inorganic hybrid systems, underscoring how nonlinear optics can probe charge-transfer-mediated SHG enhancements- a finding supported by traditional TR spectroscopy that offers valuable insights for designing optoelectronic devices.

Investigating interfacial charge carrier dynamics is important for improving the efficiency of photovoltaic devices with organic and inorganic heterostructures to exceed the Shockley-Queisser limit. Charge transfer dynamics at the organic/inorganic interfaces with two different types of excitons are still unclear. In this work, we reveal the photogenerated charge carrier dynamics at the interface of CuPc/CdSe nanoflakes using steady-state and transient reflection spectroscopy. The CuPc layer deposited on CdSe effectively modifies the charge carrier dynamics, reducing the fast electron lifetime from 10.96 to 3.12 ps. Following photonic interaction with ZB-CdSe, the photogenerated electrons are transferred to CuPc, forming a singlet charge transfer state (1CT). Rapid intersystem crossing converts this into a triplet state (3CT), preventing electrons return to CdSe and enabling efficient exciton dissociation into long-lived polarons in CuPc with longer lifetimes. The results show that an enhanced red-shift around 20 nm, caused by a decrease in the bandgap, ultimately improves the overall charge transfer efficiency, eta similar to 70% at CuPc/ZB-CdSe interfaces, comparable to that achieved in quantum dot systems. Our work demonstrates an effective pathway for improving the photoelectric performance of CuPc/CdSe composites to get electrons out from CdSe, which involves transporting them across CuPc/CdSe interfaces.

The canalization effect of phonon polaritons (PhPs) shows highly directional and diffraction-less propagation characteristics in van der Waals (vdW) materials, offering new opportunities to mold the light flow at nanoscale for near-field energy, information, and thermal management. Previously, canalized PhPs have only been experimentally realized in the hexagonal boron nitride metasurface, heterostructures of twisted alpha-phase molybdenum trioxide (alpha-MoO3) crystal flakes, or the hybridized system. However, these systems typically have complex structures, and require strict operational conditions, such as fine structural parameters, a specific photonic magic angle, or a doping level of graphene, for realizing polariton canalization with a modest performance. Here, we demonstrate the high-quality PhPs canalization in a single natural -alpha-MoO3 crystal flake. The canalized PhPs exhibit the highly directional and diffraction-free propagation features associated with lateral confinement ratios up to lambda(0)/80 (where lambda(0) is the free-space wavelength of the incident laser). We believe this work is important to effectively manipulate PhPs in natural vdW materials, with potential applications in nanoimaging, directional energy transfer, and enhanced nonlinearity at the deep subwavelength scale. (c) 2025 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

Difference frequency generation (DFG), underlying on second‐order nonlinear optical effect, is a parametric down‐conversion process that is widely used to generate and amplify broadband signals. It leads to the wide applications in tunable laser systems and infrared sources, as well as quantum light sources. However, it is still challenging to realize the widely tunable, ultra‐compact, and efficient DFG light sources in nanoscales. Here, a boosting DFG from AA‐stacked WS 2 coupled with a tunable dual‐resonant nanocavity, which is composed of Au nanodisk arrays and TiO 2 /Au film, is experimentally demonstrated. In this nanocavity, the Fabry‐Perot mode and localized surface plasmon resonance mode can be independently tuned to match well with the pump (515 nm) and the signal (750–900 nm) wavelength in DFG, respectively. Over two order‐of‐magnitude enhancement of DFG is successfully achieved ranging from 1200 to 1650 nm (e.g., 812 at 1445 nm). This work proposes a dual‐resonance hybrid nanocavity and shines the way to the DFG enhancement in transitional metal dichalcogenides, expanding the prospects in subwavelength coherent light sources and other optoelectronic devices.

The spatiotemporal mode-locked (STML) fiber laser provides a new way to construct a higher-power, higher-energy mode-locked fiber laser. In this work, a noise-like pulse STML fiber laser with an average power of 4.1 W and an optical-to-optical efficiency of up to 41% is reported. At the same time, a good beam quality ($Mx2=1.35,My2=1.41$) of the output beam is also obtained benefiting from the beam self-cleaning. In addition, the 3 dB spectral bandwidth of the output pulse reaches 54 nm due to the strong nonlinear spectral broadening and the weak spectral filtering in this multimode laser cavity. Our research contributes to a better understanding of the behavior of high-power STML lasers and further studies of noise-like pulse STML fiber lasers for practical applications.

Ultrashort laser pulses combined with on-chip nanostructures have led to a new class of ultrafast electronic devices, which show great potential in time-domain metrology and information processing. Among these nanostructures, nanogaps stand out as highly efficient and compact photoemission emitters, attracting significant attention. Here, we demonstrate on-chip nanogaps exhibiting highly nonlinear photoemission behavior driven by femtosecond laser pulses. The photocurrent follows nearly a seventh power-law dependence on laser power and reaches magnitudes exceeding 10 nA, which has never been achieved in prior research. We attribute this unique photoemission behavior to the laser-induced reduction of work function and the excitation of valence band electrons of emitters, as described in the full F–N tunneling model proposed by Murphy and Good [Phys. Rev. 102(6), 1464–1473 (1956)]. These findings, showcasing highly nonlinear, nA-level photocurrents in on-chip devices, represent a substantial advancement in ultrafast electronics and pave the way for further innovations in the field.

Iron-bearing minerals have been reported to be sinks for arsenic, and oxygen vacancies in alpha-Fe2O3 can reduce the bioavailability of arsenic in soil and sediments and further contribute to arsenic immobilization. The mechanism of oxygen vacancies to promote gaseous arsenic immobilization by gamma-Fe2O3 during coal combustion remains to be comprehensively investigated. The oxygen vacancy-enriched gamma-Fe2O3 nanorod (Ovan gamma-Fe2O3) was synthesized by regulating the synthesis conditions through NaBH4 reduction reaction and low-temperature calcination of lepidocrocite. The synthesized Ovan gamma-Fe2O3 was characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM), and it showed excellent gaseous As2O3 immobilization capacity (9.14 mg/g at 600 degrees C and 12.76 mg/g at 900 degrees C), much higher than the commercial gamma-Fe2O3 under the stimulated flue gas (SFG) conditions. This sorbent also endured long-term gaseous As2O3 adsorption without penetration at 600 degrees C for 2 h (47.70 mg/g). Electron spin resonance (ESR) spectra proved that there were more oxygen vacancies (OVs) on Ovan gamma-Fe2O3 than on the commercial gamma-Fe2O3, which provided more arsenic immobilization sites, and it generated more extra OVs under 600 degrees C SFG. X-ray photoelectron spectroscopy (XPS) demonstrated that the immobilized arsenic compounds on the surface of the gamma-Fe2O3 were transformed from As(III) to As(V), indicating that the oxygen vacancies promoted the oxidation of As2O3 to As2O5 with lower toxicity. The Ovan gamma-Fe2O3 performed a strong thermal stability, long time endurance, and excellent immobilization capacity, making it a promising candidate for industrial use in high temperature flue gas.

Hyperbolic phonon polaritons (HPhPs) in van der Waals layered polar crystals exhibit extreme light confinement capability, providing an unprecedented research opportunity to manipulate nanoscale mid-infrared photons. Precise tuning of HPhPs propagation characteristics and near-field energy routing are crucial for applying polaritonic devices. Here, we demonstrate a widely tunable waveguide mode of HPhPs in an α-MoO₃ flake/gold slit composite structure. By varying the width of the gold slit from 3 µm to 220 nm, the compression ratio of HPhPs wavelength relative to the incident light wavelength can be adjusted from 35% to 8.6%, respectively. This is attributed to that the introduction of the gold slit can effectively excite and guide HPhPs within α-MoO₃, forming a confined waveguide mode. Notably, the excitation efficiency of HPhPs in the nanoscale ultra-narrow waveguide can be enhanced by integrating an extended port at the waveguide port. In addition, the routing of polaritons in a Y-shaped waveguide is realized by modulating the frequency of the incident light. This work presents a promising platform for manipulating deep subwavelength polaritons in planar photonic devices for infrared applications.

Dynamically encircling exceptional points (EPs) in the parameter space of a non-Hermitian system has drawn widespread attention for realizing asymmetric mode switching and related applications. While previous works have been restricted to two-level systems, the switching between edge states in a multistate system that contains more complex dynamics remains to be explored. Herein, asymmetric switching of edge modes in multistate systems is demonstrated by encircling multiple EPs. A multistate non-Hermitian system containing multiple EPs is constructed based on a one-dimensional lattice, which can support four edge-localized modes. Dynamically encircling the EPs in parameter space leads to a specific edge mode surviving to the end, determined by the encircling direction. Correspondingly, optical waveguide arrays are designed to investigate mode evolution through refractive index modulation. Simulated results show that clockwise (counterclockwise) encirclement leads to an output mode localized at the left (right) side of the waveguide array with a phase difference of 0 (& pi;), regardless of the input modes; this excellently illustrates the asymmetric switching of edge modes in a multistate system. Moreover, such an effect can be extended to the multistate system with an arbitrary number of EPs. This work enables investigation of interesting effects in non-Hermitian physics by engineering the EPs and topological properties, especially for complex multistate systems, which will be useful for developing functional nanophotonic devices.

为了提高说话人识别系统的性能,提出基于改进语谱图的深度学习说话人识别算法.语谱图当中包含了语音的内容、情绪、语种以及说话人身份等多种信息,在以往的说话人识别算法中,往往没有考虑到说话人身份特性,采用直接提取语音中的语谱图作为网络输入,而说话人识别系统中需要提取语谱图中表征身份的信息,因此需要在原始语谱图的基础上进行改进.在语谱图中,基音频率以及共振峰等信息最能表现说话人的身份特征,从而提出根据语音信号中每一帧的基音频率进行自适应梳状滤波,得到改进后的语谱图,再通过卷积神经网络提取说话人特征,从而达到提升识别准确率的效果.网络模型采用MobileNetv2神经网络,该网络模型具有模型参数少、收敛速度快、识别速度快等优点,有利于实际应用.在对照实验结果中,该方法相对于原始语谱图的准确率分别提高了2.3%、5.2%、3%.

To solve the problem of the low identification rate of language identification in a noisy environment, a language identification method based on the Gammatone-scale power-normalized coefficients spectrograms is proposed, which is obtained by extracting coefficients as features based on the suppression of noise in power and the auditory features of the Gammatone filter-banks. The coefficients are then transformed into images as spectrograms. Then the dark channel prior algorithm and automatic color scale algorithm are applied to enhance and denoise the images. Finally, the residual neural network is used for training and identification. Experiment results show that the identification rate of the proposed method is improved by 39.1%, 12.3%, 19.0%, 5.5%, 28.2% and 28.5% relative to the linear gray-scale spectrograms under the conditions of the signal-to-noise ratio is 0 dB and noise sources are white noise, volvo noise, pink noise, high frequency channel noise, babble noise and factory floor noise respectively. The identification rate under other signal-to-noise ratios is also improved.

为了解决在音频中添加水印信息后如何保持音频质量以及嵌入的水印信息在遭受攻击后的安全问题,提出了一种基于平稳小波变换(SWT)和离散余弦变换(DCT)的音频水印算法.首先,利用Lorenz混沌系统生成密钥对原始水印信息进行加密.然后对原始音频分帧,通过音频帧的能量特征和过零率特征确定水印嵌入帧,对水印嵌入帧进行三级平稳小波变换后,将其三级近似分量平均分成两个一维矩阵,分别进行离散余弦变换,同时计算这两个一维矩阵幅度绝对值的平均值,通过不同的水印信息修改离散余弦变换系数嵌入水印.通过实验选取3种不同类型音乐(classical、hip-hop、rock)测试该算法,结果表明,3种音乐的音频的信噪比分别为25.4517、22.2963、25.2431,高于国际标准;在经过各种攻击后其误码率均在0.02以下;相关系数都在0.98以上.

To solve the issue of low accuracy of language identification in a noisy environment, a language identification method is proposed by combining Mel-scale frequency cepstral coefficients and Gammatone frequency cepstral coefficients. First, the Mel-scale frequency cepstral coefficients and Gammatone frequency cepstral coefficients of speech are extracted, and the feature dimensions are screened based on the language contribution. Then, the feature is mapped in the spatial coordinate system composed of the Mel domain-Gammatone domain to obtain the Mel Gammatone cepstral coefficients（MGCC）. Finally, the fusion feature is input into the deep bottleneck network. The experimental results show that the identification accuracy and speed of the proposed method are much higher than those of the single acoustic feature and other features. The accuracy can reach 99.38% in the clean corpus, and can still reach more than 89% under the-5 dB environment, which fully proves the effectiveness and robustness of the proposed method.

For the problems of unreasonable layout location and low charging utilization rate of electric vehicle（EV）charging stations, this paper proposed a spatial-temporal demand model（STDM） for location optimization of charging stations under spatial-temporal demand. By mining the spatial-temporal distribution characteristics of EV travel data and combining the travel and charging behaviors of EVs, a charging demand prediction model is to obtain the spatialtemporal demand distribution in the region. The method based on spatial-temporal statistics obtained the demand hotspot areas, and the charging coverage rate as the model evaluation parameter considered the service coverage of charging stations. A comprehensive model included the perspectives of users, operators and society with EV distance cost to the station, charging station construction and operation cost and carbon emission cost. Finally, simulation results based on actual data show the feasibility and validity of the model. The results show that the model has set up 12charging stations in the region to reduce the comprehensive cost, while determining the layout location and the number of charging piles. The optimization model results effectively reduce the EV arrival distance and improve the arrival coverage compared with other methods.

语音信号回声隐写后其倒谱系数会在回声延迟出产生峰值,传统回声隐写分析主要采用倒谱系数的统计特征作为隐写检测特征,然而在低回声幅度时隐写信号倒谱系数的峰值并不明显,基于统计特征的方法检测性能并不理想.本文将倒谱分析与图像识别技术结合,提出了一种基于倒谱图像的语音回声隐写分析方法,对语音信号分帧加窗后进行倒谱计算,然后以时间为横轴,倒谱序列点为纵轴,倒谱系数幅值为灰度级生成倒谱图像,将生成的倒谱图像作为隐写检测的输入,采用残差神经网络作为分类器进行回声隐写分析.实验结果表明,在3种经典回声隐写算法上低回声幅度时检测准确率分别达到98.2%、98.6%和96.1%,本文方法在低回声幅度时检测准确率相较传统回声隐写分析方法有较大提升,解决了传统回声隐写分析方法在低回声幅度检测效果不佳的问题.

In this work, we reported a systemic study on the enhanced efficiency of launching hyperbolic phonon polaritons (PhPs) in stacked α-phase molybdenum trioxide (α-MoO3) flakes. By using the infrared photo-induced force microscopy (PiFM), real-space near-field images (PiFM images) of mechanically exfoliated α-MoO3 thin flakes were recorded within three different Reststrahlen bands (RBs). As referred with PiFM fringes of the single flake, PiFM fringes of the stacked α-MoO3 sample within the RB 2 and RB 3 are greatly improved with the enhancement factor (EF) up to 170%. By performing numerical simulations, it reveals that the general improvement in near-field PiFM fringes arises from the existence of a nanoscale thin dielectric spacer in the middle part between two stacked α-MoO3 flakes. The nanogap acts as a nanoresonator for prompting the near-field coupling of hyperbolic PhPs supported by each flake in the stacked sample, contributing to the increase of polaritonic fields, and verifying the experimental observations Our findings could offer fundamental physical investigations into the effective excitation of PhPs and will be helpful for developing functional nanophotonic devices and circuits.

准确的风电功率预测可以合理安排风电场的发电计划和提高电网稳定性.针对单一预测模型预测精度低的问题,提出一种基于MIC-VMD-GWO-LSTM的短期风电功率预测模型.首先使用最大互信息系数法(MIC)对高维特征的风电数据集进行特征提取,以降低数据复杂度;然后采用变分模态分解(VMD)技术将风电功率序列分解为不同频率的模态,以减少功率数据的波动性;接着对每个模态建立GWO-LSTM预测模型,并利用灰狼优化(GWO)算法LSTM模型的相关参数进行优化;最后将每个模态的预测结果求和重构,得到最终的预测结果.仿真结果表明,相对于单一的BP和LSTM预测模型,基于MIC-VMD-GWO-LSTM的组合预测模型的MAPE分别降低了43.16%和31.14%,可有效提高预测精度,证明了该方法在风电功率预测运用中的有效性和可行性.

The generation characteristics of nonlinear optical signals and their multi-dimensional modulation at micro-nano scale have become a prominent research area in nanophotonics, and also the key to developing various novel nonlinear photonics devices. In recent years, the demand for higher nonlinear conversion efficiency and device integration has led to the rapid progress of hybrid nonlinear metasurfaces composed of nanostructures and nonlinear materials. As a joint platform of stable wavefront modulation, nonlinear metasurface and efficient frequency conversion, hybrid nonlinear metasurfaces offer a splendid opportunity for developing the next-generation of multipurpose flat-optics devices. This article provides a comprehensive review of recent advances in hybrid nonlinear metasurfaces for light-field modulation. The advantages of hybrid systems are discussed from the perspectives of multifunctional light-field modulation, valleytronic modulation, and quantum technologies. Finally, the remaining challenges of hybrid metasurfaces are summarized and future developments are also prospected.