由于道路高排放源所产生的污染气体对环境危害巨大，因此实现对高排放源的准确识别具有重要意义。而传统的基于限值划分的识别方法及新兴的人工智能识别方法在模型选择、评价指标、识别性能等方面都存在一定的改进空间，因此针对以上问题，提出一种基于混核极限学习机的道路高排放源识别方法。该方法使用道路遥感监测设备获取的移动源遥测数据，在核极限学习机的基础上融合不同核函数，可提升模型鲁棒性及道路高排放源识别性能。针对合肥市蜀山区真实道路遥测数据上的分析结果表明，该方法相比于其他方法具有较高的F1分数以及较低的漏报率、虚警率，证实了该方法在高排放源识别中的有效性。因此，该方法有助于对交通路网中高排放车辆进行高效识别，为进一步提升城市空气质量提供支撑。

In this paper, a patch-antenna-array enhanced quantum cascade detector with freely switchable operating modes among mid-wave, long-wave, and dual-color was proposed and discussed. The dual-color absorption occurs in a single active region through an optimized coupled miniband diagonal-transition subbands arrangement, and a successful separation of the operation regimes was realized by two nested antenna arrays with different patch sizes up to room temperature. At 77 K, the 5.7-μm channel achieved a peak responsivity of 34.6 mA/W and exhibited a detectivity of 2.0×1010 Jones, while the 10.0-μm channel achieved a peak responsivity of 87.5 mA/W, giving a detectivity of 5.0×1010 Jones. Under a polarization modulation of the incident light, the minimum cross talk of the mid-wave and the long-wave operating modes was 1:22.5 and 1:7.6, respectively. This demonstration opens a new prospect for multicolor infrared imaging chip integration technology.

Sharing the advantages of high optical power, high efficiency and design flexibility in a compact size, quantum cascade lasers (QCLs) are excellent mid-to-far infrared laser sources for gas sensing, infrared spectroscopic, medical diagnosis, and defense applications. Metalorganic chemical vapor deposition (MOCVD) is an important technology for growing high quality semiconductor materials, and has achieved great success in the semiconductor industry due to its advantages of high efficiency, short maintenance cycles, and high stability and repeatability. The utilization of MOCVD for the growth of QCL materials holds a significant meaning for promoting the large batch production and industrial application of QCL devices. This review summarizes the recent progress of QCLs grown by MOCVD. Material quality and the structure design together determine the device performance. Research progress on the performance improvement of MOCVD-grown QCLs based on the optimization of material quality and active region structure are mainly reviewed.

全无机卤化铅铯钙钛矿纳米晶CsPbX₃（X=Cl，Br，I）因具有独特的光电特性，近年来在固体照明及显示、太阳能电池、阻变存储器等领域成为研究的热门之选。Mn²⁺是一种比Pb²⁺半径小的过渡金属离子，利用Mn²⁺掺杂CsPbCl₃钙钛矿纳米晶能够实现波长位于600 nm左右可见光区域的橘黄色发射，且能够部分替代钙钛矿纳米晶中的Pb²⁺，降低钙钛矿纳米晶的毒性。然而，Mn²⁺掺杂的卤化物钙钛矿纳米晶仍易受环境中水分子等的侵蚀而导致其荧光特性严重退化。本文采用一种利用四甲氧基硅烷（Tetramethoxysilane，TMOS）和聚甲基丙烯酸甲酯（Polymethyl methacrylate，PMMA）对Mn²⁺掺杂CsPbCl₃钙钛矿纳米晶进行双壳层的包覆方法，并分析了双壳层包覆法对钙钛矿纳米晶稳定性提升的机理。此外，还对比了双壳层包覆的Mn²⁺掺杂CsPbCl₃钙钛矿纳米晶在甲苯和二氯甲烷（Dichloromethane，DCM）溶剂下的发光特性，它们呈现的橘黄色荧光发射强度并未明显下降，且荧光量子产率（Photoluminescence quantum yield，PLQY）能够保持在25%左右。并以此为基础制备了相应的荧光纳米晶粉末，将其运用于潜指纹辨识上，可长期有效地检测潜指纹信息。

元素掺杂对调节碳点多色发光有着至关重要的作用。然而，目前碳点的可调荧光发射在固态下难以实现，这是因为会发生严重的聚集诱导猝灭（AIQ）现象以及存在制备工艺繁琐等问题。本文报道了一种以间苯三酚为碳源、硼酸为硼元素掺杂剂、尿素为氮元素掺杂剂，采用固相法，微波一步直接制备的氮硼共掺杂固态荧光碳点。随着氮和硼元素含量的变化，所得固态碳点的发光颜色经历黄色、绿色到蓝色的变化。表征分析发现氮和硼元素的掺杂在碳点表面形成了不同的结构和表面官能团，随着氮和硼元素掺杂含量的提高，其中石墨氮、N—C以及B—O/B—N 基团的协同作用导致了碳点发光颜色蓝移，且荧光发光效率增强。此外，鉴于这些固体碳点材料呈现出优异的多色发光性能，选择发光性能最佳的黄色、绿色和蓝色固态荧光碳点作为固态荧光粉末，制备得到了白光发光二极管（WLED）器件。这些器件均具有优良的色度指标，暖白光区发光和个别器件节能高效的工作特性表明这些发光材料在照明领域具有潜在应用前景。