We report a high-stability ultrafast ultraviolet (UV) laser source at 352 nm by exploring an all-fiber, all-polarization-maintaining (all-PM), Yb-doped femtosecond fiber laser at 1060 nm. The output power, pulse width, and optical spectrum width of the fiber laser are 6 W, 244 fs, and 17.5 nm, respectively. The UV ultrashort pulses at a repetition rate of 28.9 MHz are generated by leveraging single-pass second-harmonic generation in a 1.3-mm-long BiB₃O₆ (BIBO) and sum frequency generation in a 5.1-mm-long BIBO. The maximum UV output power is 596 mW. The root mean square error of the output power of UV pulses is 0.54%. This laser, with promising stability, is expected to be a nice source for frontier applications in the UV wavelength window.

Laser processing with high-power ultrashort pulses, which promises high precision and efficiency, is an emerging new tool for material structuring. High repetition rate ultrafast laser highlighting with a higher degree of freedom in its burst mode is believed to be able to create micro/nanostructures with even more variety, which is promising for electrochemical applications. We employ a homemade high repetition rate ultrafast fiber laser for structuring metal nickel (Ni) and thus preparing electrocatalysts for hydrogen evolution reaction (HER) for the first time, we believe. Different processing parameters are designed to create three groups of samples with different micro/nanostructures. The various micro/nanostructures not only increase the surface area of the Ni electrode but also regulate local electric field and help discharge hydrogen bubbles, which offer more favorable conditions for HER. All groups of the laser-structured Ni exhibit enhanced electrocatalytic activity for HER in the alkaline solution. Electrochemical measurements demonstrate that the overpotential at 10 mA cm ^{- 2} can be decreased as much as 182 mV compared with the overpotential of the untreated Ni (-457 mV versus RHE).

In this work, we present a high-power, high-repetition-rate, all-fiber femtosecond laser system operating at 1.5 $\unicode{x3bc}$ m. This all-fiber laser system can deliver femtosecond pulses at a fundamental repetition rate of 10.6 GHz with an average output power of 106.4 W – the highest average power reported so far from an all-fiber femtosecond laser at 1.5 $\unicode{x3bc}$ m, to the best of our knowledge. By utilizing the soliton-effect-based pulse compression effect with optimized pre-chirping dispersion, the amplified pulses are compressed to 239 fs in an all-fiber configuration. Empowered by such a high-power ultrafast fiber laser system, we further explore the nonlinear interaction among transverse modes LP₀₁, LP₁₁ and LP₂₁ that are expected to potentially exist in fiber laser systems using large-mode-area fibers. The intermodal modulational instability is theoretically investigated and subsequently identified in our experiments. Such a high-power all-fiber ultrafast laser without bulky free-space optics is anticipated to be a promising laser source for applications that specifically require compact and robust operation.

We present a hybrid coupling scheme of a magnetic toroidal and electric dipole metasurface with suppressed radiation loss, which can produce the tunable plasmon-induced transparency (PIT) with an enhanced slow-light effect in the terahertz regime. The terahertz metasurface is constructed by nesting a dual-split ring resonator (DSRR) inside a ring resonator (RR) to exploit the destructive coherence of hybrid electromagnetic mode coupling at the PIT resonance. The polarization-dependence excitation performs the active tunability of a PIT-induced group slowing down by rotating the polarization angle, experimentally achieving a maximum group delay of 3.5 ps. Furthermore, the modified terahertz metasurface with a four-split ring resonator (FSRR) nested in an RR is prepared on photoconductive silicon, demonstrating the pump-controllable group delay effect at the PIT resonance. The large group delay from 2.2 to 0.9 ps is dynamically tunable by adjusting the pump power. The experimental results are in good accord with the theoretical simulations.

酿酒葡萄中的总酚含量是影响葡萄品质的重要指标, 也是影响葡萄酒质量的关键因素。 为了快速准确地检测赤霞珠葡萄的总酚含量, 利用近红外光谱技术结合GA-ELM预测模型对赤霞珠葡萄总酚含量进行预测研究。 试验采用5个收获期(每期采集40串, 每串取10个)的赤霞珠葡萄, 采集200组葡萄的12 500~4 000 cm^-1波段范围内的近红外光谱。 基于福林酚比色法原理对赤霞珠葡萄的总酚含量进行测定, 使用SPXY算法将样品按照3:1比例分为校正集和预测集, 共计150个校正集和50个预测集。 分别采用多元散射(MSC)、 标准正态变换(SNV)、 数据中心化(MC)、 移动窗口平滑(MA)和一阶导数+SG方法对原始光谱进行预处理, 优选出最佳的预处理方法为MSC。 并进一步采用竞争性自适应重加权算法(CARS)、 遗传算法(GA)、 联合区间偏最小二乘算法(si-PLS)和连续投影算法(SPA)分别对光谱波段进行提取, 经对比分析发现CARS提取的69个特征波长数据能有效提高模型的稳定性和预测结果。 在MSC预处理和特征波长提取的基础上, 引入极限学习机(ELM)算法, 建立赤霞珠葡萄总酚含量的预测模型, 在总酚含量预测过程中, 采用遗传算法(GA)对ELM模型进行优化, 并探究了不同的激活函数和隐含层神经元个数对GA-ELM模型预测能力的影响, 确定最优的激活函数为Sigmoidal, 最优的神经元个数为50个。 最后, 将ELM和GA-ELM模型的预测能力进行对比, 结果显示GA-ELM模型的预测能力高于ELM模型的预测能力, 其中MSC+CARS+GA-ELM模型预测能力最好, 校正相关系数(R_c)为0.901 7, 预测相关系数(R_p)为0.901 3, 校正均方根误差(RMSEC)为2.112 4, 预测均方根误差(RMSEP)为1.686 8, 剩余预测偏差(RPD)为2.308 0。 研究结果表明: 利用近红外光谱技术结合变量优选建立的GA-ELM模型可实现对赤霞珠葡萄的总酚含量的预测, 为赤霞珠葡萄品质的检测奠定了理论基础。