PdCoNi nanoparticles supported on nitrogen-doped porous carbon nanosheets for room temperature dehydrogenation of formic acid

The development of cost-effective heterogeneous catalysts for the dehydrogenation of formic acid (FA) is the key challenge for the commercialization of FA as a hydrogen-storage medium. Herein, PdCoNi nanoparticles (NPs) with different element ratios supported on N-doped carbon nanosheets (N-CN) were designed, which exhibit excellent catalytic dehydrogenation performance for FA. Compared with PdCoNi NPs loaded on the carbon nanosheets (CN), the introduction of pyrrolic N to CN induces the formation of ultrafine, monodispersed and amorphous Pd_0.6Co_0.2Ni_0.2 NPs with a size of 1.60 nm, which significantly increases the number of active sites and the instant contact between FA and catalysts. The as-prepared Pd_0.6Co_0.2Ni_0.2/N-CN catalyst shows more than 99% conversion and 100% H₂ selectivity at room temperature, with a record-high initial turnover frequency (TOF_initial) of 1249.0 h⁻¹ among non-noble containing Pd-based catalysts, which demonstrates the high potential of Pd_0.6Co_0.2Ni_0.2/N-CN as a practical catalyst for the hydrogen generation from FA.     

Investigation of the role of Nb on Pd−Zr−Zn catalyst in methanol steam reforming for hydrogen production

Methanol steam reforming is regarded as a very promising process to generate H₂ suitable for fuel cells. Typically, the Pd-based catalysts can catalyze efficiently methanol steam reforming for hydrogen production. But their high selectivity to CO, a byproduct of methanol reforming reaction, severely limits their potential application. In this work, a series of Nb-modified Pd−Zr−Zn catalysts with different Nb loadings were prepared to study their catalytic activities with more focus on the role of Nb on Pd−Zr−Zn catalyst for methanol steam reforming. The prepared catalysts were fully analyzed by using various characterization techniques, for example, ICP, BET, SEM, XRD, H₂-TPR, NH₃-TPD, HRTEM, CO chemisorption, XPS, and Raman. The experimental results showed that an increase in Nb loading for the Nb-modified Pd−Zr−Zn catalysts led to a decrease of the methanol conversion and H₂ production rate. This was probably due to the decrease in the amount of oxygen vacancies on the catalyst surface. However, introduction of Nb into Pd−Zr−Zn catalyst increased the acid strength on the catalytic surface. The aldehyde species derived from methanol decomposition were readily transformed to HCOOH, thus yielding high selectivity to CO₂ for the Nb-modified Pd−Zr−Zn catalysts. Significantly, the addition of Nb to Pd−Zr−Zn catalyst facilitated the incorporation of Pd into the ZnO lattices, which led to the formation of Pd−Zn alloy. Consequently, the Nb-modified Pd−Zr−Zn catalysts exhibited significantly lower CO selectivity and production rate than the Pd−Zr−Zn catalyst. From the results, this work offers a new way to the rational design of selective methanol steam reforming catalysts to decrease the formation of byproduct CO.     

Facile synthesis of PdAu/C by cold plasma for efficient dehydrogenation of formic acid

Decomposition of formic acid biomass to generate hydrogen is vital for coping with fossil energy depletion, environmental pollution, and developing clean, efficient, safe, and sustainable modern energy system. In this study, a PdAu/C−C bimetallic catalyst was prepared by the co-impregnation method followed by an atmospheric pressure (AP) cold plasma treatment to synthesize PdAu/C−P catalysts. The resulting PdAu/C−P showed excellent catalytic activity for the formic acid dehydrogenation (FAD) reaction. The total volume of H₂ and CO₂ released from the FAD reaction was about 375 mL after 4 h at 50 °C, and the initial turnover frequency (TOF_initial) was 808.6 h⁻¹. We used X−ray diffractometry (XRD), temperature programmed reduction (TPR) and high-resolution transmission electron microscopy (HRTEM) to show that plasma can effectively promote the redispersion of Pd−Au particles on the surface of the support. The average particle size of PdAu/C−P (3.5 ± 1.5 nm) was less than PdAu/C−C (4.4 ± 1.9 nm) and uniformly distributed. X-ray photoelectron spectroscopy (XPS), TPR, and HRTEM showed that PdAu/C−P has a higher degree of alloying. In addition, the strong electric field in the plasma facilitated more metal sites located on the outer surface of the support in PdAu/C−P, and the atomic ratio of M/C (M = Pd and Au) (0.0134) was much larger than that of PdAu/C−C (0.0060). The apparent activation energy (E_a) of PdAu/C−P for the FAD reaction was only 27.25 kJ mol⁻¹, and it had much higher activity and stability than the commercial Pd/C (Sigma−Aldrich). The total volume of H₂ and CO₂ produced over the PdAu/C−P for three cycles was 1.33, 5.87, and 8.56 times that of commercial Pd/C. Overall, the cold plasma enhanced the degree of alloying, promoted the redispersion of agglomerated particles, and regulated the surface enrichment of the active metal components. This is of great significance for guiding the preparation of high−performance multi-metal catalysts by cold plasma.     

Bimetallic PdAu catalysts for formic acid dehydrogenation

A series of monometallic and bimetallic palladium gold catalyst were prepared and studied for the formic acid dehydrogenation reaction. Different Pd/Au compositions were employed (Pd_xAu_100-x, where x = 25; 50 and 75) and their impact on alloy structure, particle size and dispersion was evaluated. Active phase composition and reaction parameters such as temperature, formic acid concentration or formate/formic acid ratio were adjusted to obtain active and selective catalyst for hydrogen production. An important particle size effect was observed and related to Pd/Au composition for all bimetallic catalysts.     

Synthesis of highly active and stable Pd/C catalysts toward HCOOH dehydrogenation by a simple NH3 adsorption method

Pd catalysts supported on activated carbon (Pd/C–NH₃) toward HCOOH dehydrogenation were prepared by a simple adsorption method using ammonia (NH₃) and Ar as the working gas. The results show that the TOF_initial of Pd/C–NH₃ was 459.8 h⁻¹ at 50 °C. When the reaction was carried out for 4 h, the HCOOH dehydrogenation ratio over Pd/C–NH₃ was about 81.2%, which was 1.15 and 1.13 times, respectively, as that of the as-prepared Pd/C catalyst without any treatment (Pd/C–As) and the Pd/C catalyst purchased from Sigma-Aldrich (Pd/C-CM). The total amount of H₂ and CO₂ produced by using Pd/C–NH₃ to decompose HCOOH in the third cycle was 99.4% of the gas produced by the first reaction cycle, and 1.80 and 12.60 times, respectively, as that of Pd/C–As and Pd/C-CM. The characterization results indicated that the Pd active species in Pd/C–NH₃ migrated to the outer surface of the carbon support during the reaction, and the pore volume of the carbon support became larger, which were beneficial to the reaction. These factors made Pd/C–NH₃ exhibit excellent HCOOH dehydrogenation activity and stability. NH₃ adsorption is a simple and effective method for preparing high-performance Pd/C HCOOH dehydrogenation catalysts, and has important guiding significance for the preparation of other carbon supported noble metal catalysts.     

可穿戴脑电图设备关键技术及其应用综述

可穿戴脑电图（EEG）设备是一种用于日常实时监测的无线EGG系统,因其便携性、实时性、无创性及低成本等优势迅速发展并得到广泛应用。该系统主要由信号采集模块、信号处理模块、微控制模块、通信模块及电源模块等硬件部分以及移动终端模块和云存储模块等软件部分组成。就可穿戴EEG设备关键技术进行论述。首先,阐述了对EGG信号采集模块的改进,另外对可穿戴EEG设备信号预处理模块、信号的降噪、伪影处理及特征提取技术进行比较;然后,对机器学习、深度学习分类算法的优缺点进行分析,并对穿戴式EEG设备的应用领域进行总结;最后,提出可穿戴EEG设备的关键技术未来的发展趋势。     

基于卷积神经网络的图像分类算法综述

卷积神经网络（CNN）是目前基于深度学习的计算机视觉领域中重要的研究方向之一。它在图像分类和分割、目标检测等的应用中表现出色,其强大的特征学习与特征表达能力越来越受到研究者的推崇。然而,CNN仍存在特征提取不完整、样本训练过拟合等问题。针对这些问题,介绍了CNN的发展、CNN经典的网络模型及其组件,并提供了解决上述问题的方法。通过对CNN模型在图像分类中研究现状的综述,为CNN的进一步发展及研究方向提供了建议。     

pH对离子液体辅助水热制备介孔二氧化钛结构的影响

以钛酸四丁酯为钛源, 离子液体1-丁基-3-甲基咪唑四氟化硼([C₄MIM]BF₄)为辅助剂, 采用水热法制备了锐钛矿相的介孔二氧化钛。用盐酸和氨水调节反应前驱体的pH, 考察了pH在1~11变化范围内对离子液体-水制备二氧化钛体系的影响。采用X射线衍射和N2吸附?脱附对样品的晶相和孔结构进行表征。结果表明: 离子液体抑制了板钛矿二氧化钛的形成, 随着pH的增大样品的晶化程度增强。pH在1~9较大范围变化时, 其孔径分布较窄, 平均孔径随pH增大由6.6 nm减小到5.0 nm, 此时离子液体能较好的将样品粒径控制在9 nm左右。这与透射电镜分析结果相一致。     

离子液体中Co掺杂介孔TiO2可见光催化剂的制备及性能研究

以离子液体([C₄MIM]BF₄)为模板剂, 采用溶胶-凝胶法制备了介孔TiO₂和Co掺杂的介孔TiO₂光催化剂。研究了Co掺杂对样品的比表面积、晶相、元素价态、吸光特性及可见光活性的影响。结果表明: 所制备的Co/TiO₂光催化剂为介孔结构的具有较大比表面积的锐钛矿相纳米颗粒; XPS分析表明: Co以Co²⁺取代Ti⁴⁺进入TiO₂晶格形成杂质能级, 降低了TiO₂带隙能, 有效拓展了TiO₂的光响应范围。以亚甲基蓝水溶液为降解对象, 在可见光 (λ>420 nm)下考察制备样品的光催化活性, 结果表明: Co掺杂的TiO₂具有可见光活性, 且0.3%Co/TiO₂的活性最高。     

基于卷积神经网络的机械故障诊断技术综述

针对传统机械故障诊断方法难以解决人工提取不确定性的问题,提出了大量深度学习的特征提取方法,极大地推动了机械故障诊断的发展。作为深度学习的典型代表,卷积神经网络（CNN）在图像分类、目标检测、图像语义分割等领域都取得了重大的发展,在机械故障诊断领域也有大量文献发表。为了进一步了解利用CNN的方法进行机械故障诊断的问题,首先简单介绍了CNN的相关理论,然后从数据输入类型、迁移学习、预测等方面对CNN在机械故障诊断中的应用进行了归纳总结,最后展望了CNN及其在机械故障诊断应用中的发展方向。