Breathing solitons, i.e., dynamic dissipative solitons with oscillating pulse shape and energy caused by different mechanisms of spatiotemporal instabilities, have received considerable interest from the aspects of nonlinear science and potential applications. However, by far, the study of breathing solitons is still limited within the time scale of hundreds of cavity round trips, which ignores the slow dynamics. To fill this lacuna, we theoretically investigate a new type of vector dissipative soliton breathing regime and experimentally demonstrate this concept using mode-locked fiber lasers, which arise from the desynchronization of orthogonal states of polarization (SOPs) in the form of complex oscillations of the phase difference between the states. The dynamic evolution of polarization states of the vector breathings solitons takes the form of a trajectory connecting two quasi-equilibrium orthogonal SOPs on the surface of the Poincaré sphere. The dwelling time near each state is on the scale of a tenth of a thousand cavity round trip times that equals the breathing period, which is up to 2 orders of magnitude longer than that for common breathers. The obtained results can reveal concepts in nonlinear science and may unlock approaches to the flexible manipulation of laser waveforms toward various applications in spectroscopy and metrology.

We propose and experimentally demonstrate a high quality (Q)-factor all-silicon bound state in the continuum (BIC) metasurface with an imperforated air-hole array. The metasurface supports two polarization-insensitive BICs originated from guided mode resonances (GMRs) in the frequency range of 0.4 to 0.6 THz, and the measured Q-factors of the two GMRs are as high as 334 and 152, respectively. In addition, the influence of the thickness of the silicon substrate on the two resonances is analyzed in detail. The proposed all-silicon THz metasurface has great potential in the design and application of high-Q metasurfaces.

随着人工智能技术的不断发展，激光工业制造的智能化已成为重要趋势。机器学习作为人工智能的主要技术之一，已经在相关领域得到了广泛应用，并促进该领域实现巨大突破，推动下一代激光微纳加工技术的发展。为此，本文综述了机器学习在激光微纳加工领域各工艺过程中的重要应用，包括激光微纳加工参数优化与工艺窗口预测、加工过程的实时监测与控制、加工结果的预测以及辅助物理机制的研究，总结并展望了当前已有的机器学习与激光微纳加工交叉方向的改进方案，以及未来可能出现的机器学习与激光微纳加工交叉技术的进一步应用。

对于高灵敏原子磁力计极弱磁测量，激光温度的精确稳定控制是一项必不可少的工作。激光温度不稳定会导致激光波长波动和漂移，从而降低原子磁力计的灵敏度。为了降低激光器温度波动对原子磁力计的影响，本文设计并实现了一个基于ADN8834温度控制芯片的高精度DBR激光器自动温度控制系统。首先，基于ADN8834和高精度模/数转换芯片LTC2377设计了温度反馈电路，成功采集到了与温度对应的模拟电压信号并将其转换为数字信号送入FPGA。然后，在FPGA中实现了增量式数字PID算法，自动计算温度控制信号。最后，设计了数/模转换电路将该温度控制信号转换为模拟信号传递给ADN8834，ADN8834输出加热或冷却信号来控制半导体热电制冷器，从而实现闭环温度自动控制。实验结果表明，当目标温度分别设定在20，25，30 ℃时，该温度自动控制系统的温度稳定性均在±0.005 ℃，测试DBR激光器输出波长稳定性范围为±2 pm。该激光器自动温度控制系统温度稳定性高，且操作方便，设计灵活，基本满足原子磁力计系统对激光温度控制器的要求。

In this work, the surface morphology and internal defect evolution process of GaAs substrates implanted with light ions of different fluence combinations are studied. The influence of H and He ions implantation on the atomic mechanism of the blister phenomenon observed after annealing is investigated. Raman spectroscopy is used to measure the surface stress change of different samples before and after implantation and annealing. Optical microscopy and atomic force microscopy are used to characterize the morphology changes of the GaAs surface under different annealing conditions. The evolution of bubbles and defects in GaAs crystals is revealed by transmission electron microscopy. Through this study, it is hoped that ion implantation fluence, surface exfoliation efficiency and exfoliation cost can be optimized. At the same time, it also lays a foundation for the heterointegration of GaAs film on Si.

研究一种基于富铟团簇（IRC）效应的新型高应变InGaAs/GaAs自适应阱簇复合（WCC）量子结构，该结构具有比较灵活的能带调控能力，可产生一种特殊的偏振双峰光谱。为探究该WCC结构的偏振双峰辐射机制及能带特征，利用IRC效应生长获得了自适应WCC结构，并测量了该新型结构的偏振光致发光（PL）光谱，在横向电场（TE）和横向磁场（TM）模式下，PL光谱皆显示出特殊的双峰结构。这主要是铟原子的自适应迁移导致有源层同时存在铟组分正常和铟组分降低的In_xGa_1-xAs材料所致。根据半导体激光器的跃迁矩阵元理论和PL光谱的双峰能量间距，阐明不同偏振模式下的光谱双峰与多组分In_xGa_1-xAs材料的对应辐射机制，同时确定该WCC结构的混合能带特征。该研究内容对IRC效应和新型WCC发光结构的发展和应用具有重要研究意义。