Abstract:This paper introduces an innovative Multifunction Integrated Optic Circuit （MIOC） design utilizing thin-film lithium niobate， surpassing traditional bulk waveguide-based MIOCs in terms of size， half-wave voltage requirements， and integration capabilities. By implementing a sub-wavelength grating structure， we achieve a Polarization Extinction Ratio （PER） exceeding 29 dB. Furthermore， our electrode design facilitates a voltage-length product （V_πL） below 2 V·cm， while a double-tapered coupling structure significantly reduces insertion loss. This advancement provides a pivotal direction for the miniaturization and integration of optical gyroscopes， marking a substantial contribution to the field.

Abstract:Aircraft contrail detection remains crucial for maintaining airspace safety and addressing the greenhouse effects caused by the aviation industry. Existing methods for detecting aircraft contrails primarily relied on the radiance or temperature differences between specific channels in multispectral images. But they did not fully exploit the potential of spectral features. The advancement of satellite-borne hyperspectral imaging technology has provided a new data foundation for aircraft contrail detection. However， methods that rely solely on either the spatial or spectral dimension of the image are unlikely to achieve satisfactory results in the task of aircraft contrail detection using satellite-based hyperspectral imagery. Therefore， a detection algorithm for potential aircraft contrails was explored using shortwave infrared hyperspectral images from the GF-5 AHSI. A spatial-spectral feature extraction method was proposed， which utilized the complementary nature of spatial and spectral information in hyperspectral images. The method achieved an accuracy of over 97% and a false alarm rate of less than 2% on GF-5 hyperspectral image data. It not only provides an innovative technical approach for aircraft contrail detection， but also offers valuable insights for future researchers and promotes further development of hyperspectral imaging in practical applications.

Abstract:Metasurfaces are artificial structures that can finely control the characteristics of electromagnetic waves at subwavelength scales， and they are widely used to manipulate the propagation， phase， amplitude， and polarization of light. In this work， a bound state in the continuum （BIC） structure based on a metallic metasurface is proposed. By adjusting the metallic structure using CST and COMSOL software， a significant quasi-BIC peak can be achieved at a frequency of 0.8217 terahertz （THz）. Through multi-level expansion analysis， it is found that the electric dipole （ED） is the main factor contributing to the resonant characteristics of the structure. By leveraging the characteristics of BIC， an imaging system was created and operated. According to the simulation results， the imaging system demonstrated excellent sensitivity and resolution， revealing the great potential of terahertz imaging. This research not only provides new ideas for the creation of BIC structures but also offers an effective reference for the development of high-performance terahertz imaging technology.

Abstract:The Fourier transform spectrometer （FTS） is a precision infrared detection instrument. It adopts Michelson interference splitting， and the moving mirror is one of the core components. The uniformity and stability of the moving mirror’s speed directly affect the quality of the subsequent interferogram， so it is necessary to carry out high-precision motion control of the moving mirror. For some FTS with moving mirror in low-speed motion， the traditional M-method can no longer meet the requirements of speed measurement accuracy. In addition， when the moving mirror moves at a low speed， the speed stability is more easily affected by external mechanical disturbance. Based on the stability requirement of the low-speed moving mirror， this paper studies the motion control of the moving mirror based on the T-method measuring speed. It proposes a high-precision algorithm to obtain the measured and expected value of the velocity. By establishing the mathematical model and dynamic equation of the controlled object， the speed feedforward input is obtained， and then the compound speed controller based on the feedforward control is designed. The control algorithm is implemented by the FPGA hardware platform and applied to the FTS. The experimental results show that the peak-to-peak velocity error is 0.0182， and the root mean square （RMS） velocity error is 0.0027. To test the anti-interference ability of the moving mirror speed control system， the sinusoidal excitation force of 5 mg， 7.5 mg， and 10 mg is applied in the moving mirror motion direction on the FTS platform. Under each given magnitude， the scanning of each frequency point in 2~200 Hz is carried out. The experimental results show that the peak-peak velocity error and the RMS velocity error are proportional to the excitation magnitude. Under the 10 mg excitation， the maximum peak-to-peak velocity error is 0.1405， and the maximum RMS velocity error is 0.0448. After analysis， the speed stability of the moving mirror can still meet the performance requirements of the FTS. This design provides a technical means for realizing the speed control of the moving mirror with low speed and high stability. Also， it makes the FTS have wider applications.

Abstract:Thermo-optic modulators are key components of optical communication systems， and their performance directly affects system efficiency. With the development of silicon optothermonic technology， silicon thermo-optic modulators have been widely used in optothermonic chips. Conventional silicon optical modulators are large in size and have high losses. In recent years， researchers have proposed to use the slow light effect of photonic crystals to reduce the footprint of modulators. Related studies have shown that these devices have advantages， such as small size and low driving voltage. However， the optical transmittance of thermo-optic modulators based on photonic crystals is still affected by defects caused by fabrication errors. Valley photonic crystal optical waveguides can achieve scattering-immune high-efficiency unidirectional transmission， providing a new venue for realizing high-performance photonic devices. In this paper， a new silicon thermo-optic modulator based on a valley photonic crystal Mach-Zehnder interferometer （MZI） is designed. The electrical heating mechanism is introduced on one of the waveguides of the MZI. The thermo-optic effect modulates the refractive index to achieve precise phase modulation of the transmitted light. The thermo-optic modulator device has a small footprint of only 9.26 μm × 7.99 μm， which can achieve a high forward transmittance of 0.91， an insertion loss of 0.41 dB， and a modulation contrast of 11.75 dB. It can also be experimentally fabricated using complementary metal oxide semiconductor （CMOS） technology， so it will have broad application prospects. This modulation principle can be widely used in designing different thermo-optic modulation devices.

Abstract:This paper investigates the impact of extrinsic resistance on the noise performance of deep submicron MOSFETs （metal-oxide-semiconductor field-effect-transistor） using the noise correlation matrix method. Analytical closed-form expressions for calculating the four noise parameters are derived based on the small-signal and noise-equivalent circuit models. The results show strong agreement between simulated and experimental data for MOSFETs with a gate length of 40 nm and dimensions of 4×5 μm （number of gate fingers × unit gate width.

Abstract:This study systematically investigated the influence of deposition rate on the structure， broadband optical properties （1.0-13.0 μm）， and stress characteristics of Germanium （Ge） films. Additionally， a method for enhancing the performance of infrared filters based on rate-modulated deposition of Ge films was proposed. The optical absorption of Ge films in the short-wave infrared （SWIR） and long-wave infrared （LWIR） bands can be effectively reduced by modulating the deposition rate. As the deposition rate increases， the Ge films maintain an amorphous structure. The optical constants of the films in the 1.0-2.5 μm and 2.5-13.0 μm bands were precisely determined using the Cody-Lorentz model and the classical Lorentz oscillator model， respectively. Notably， higher deposition rates result in a gradual increase in the refractive index. The extinction coefficient increases with the deposition rate in the SWIR region， attributed to the widening of the Urbach tail， while it decreases in the LWIR region due to the reduced absorption caused by the Ge-O stretching mode. Additionally， the films exhibit a tensile stress that decreases with increasing deposition rate. Finally， the effectiveness of the proposed fabrication method for an infrared filter with Ge films deposited at an optimized rate was demonstrated through practical examples. This work provides theoretical and technical support for the application of Ge films in high-performance infrared filters.

Abstract:Remote sensing multimodal large language models （MLLMs）， which integrate rich visual-linguistic modal information， have shown great potential in areas such as remote sensing image analysis and interpretation. However， existing knowledge distillation methods primarily focus on the compression of unimodal large language models， neglecting the alignment of features across modalities， thus hindering the performance of large language models in cross-modal tasks. To address this issue， a lightweighting method for remote sensing MLLMs based on knowledge distillation is proposed. This method achieves effective alignment of multimodal information by aligning the outputs across modalities at the feature level. By introducing the reverse Kullback-Leibler divergence as the loss function and combining optimization strategies such as teacher mixed sampling and single-step decomposition， the generalization and stability of the student model are further enhanced. Experimental results demonstrate that the proposed method achieves higher accuracy and efficiency in four downstream tasks of remote sensing image scene classification， visual question answering， visual localization， and image description， significantly reducing the number of model parameters and the demand for computational resources， thereby providing a new solution for the efficient application of MLLMs in the field of remote sensing.

Abstract:Infrared polarization image fusion can fully utilize the polarization information of the scene， compensate for the disadvantage of infrared intensity images in describing high-frequency information such as scene contour edges and texture details， and has unique advantages in target detection and recognition， background noise suppression， and counter camouflage. The article summarized the research progress of infrared polarization image fusion technology from two aspects： single algorithm image fusion and multi-algorithm combination image fusion. It analyzed the design ideas of typical algorithms and summarized the advantages and disadvantages of each algorithm. Based on the current trend where single algorithm serves as the mainstream and multi-algorithm combination as the development trend for infrared polarization image fusion， this paper anticipates its potential future development direction.

Abstract:The InGaAsP material with an energy bandgap of 1.05 eV was grown on InP substrate by all-solid-state Molecular Beam Epitaxy （MBE） technique. The material had no mismatch dislocations between the substrate and the epitaxial layer， and also exhibited high interface quality and luminescence quality. Based on InGaAsP material， single-junction InGaAsP solar cells were grown on InP substrates， and GaInP/GaAs dual-junction solar cells were grown on GaAs substrates. These two separate cells were then bonded together using the wafer bonding technology to fabricate a GaInP/GaAs/InGaAsP triple-junction solar cell. Under the AM1.5G solar simulator， the conversion efficiency of the GaInP/GaAs/InGaAsP wafer-bonded solar cell was 30.6%， achieving an efficiency of 34% under concentration. The results indicate that MBE can produce high-quality InGaAsP material， and that room-temperature wafer bonding technology holds great potential for the fabrication of multi-junction solar cells.