基于人工神经网络技术的投标前期决策

针对投标前期决策中常用决策分析方法未能有效解决影响因素之间、影响因素与决策结果之间的复杂非线性关系的问题,通过调查、比较和分析,结合人工神经网络分析方法的要求,确定了投标前期决策影响因素的指标体系.将人工神经网络技术应用于投标前期决策,应用定量法和模糊综合评价法提出相应指标的量化标准及方法,据此构建了具有较好泛化学习能力及较好学习速度的人工神经网络投标前期决策模型.在MATLAB平台上,开发出人工神经网络投标前期决策模型应用系统.通过工程实例证明,此方法能有效应用于施工企业的前期投标决策.     

Photoinduced Water Oxidation in Chitosan Nanostructures Containing Covalently Linked RuII Chromophores and Encapsulated Iridium Oxide Nanoparticles

The luminophore Ru(bpy)₂(dcbpy)²⁺ (bpy=2,2’-bipyridine; dcbpy=4,4’-dicarboxy-2,2’-bipyridine) is covalently linked to a chitosan polymer; crosslinking by tripolyphosphate produced Ru-decorated chitosan fibers (NS-RuCh), with a 20 : 1 ratio between chitosan repeating units and Ru^II chromophores. The properties of the Ru^II compound are unperturbed by the chitosan structure, with NS-RuCh exhibiting the typical metal-to-ligand charge-transfer (MLCT) absorption and emission bands of Ru^II complexes. When crosslinks are made in the presence of IrO₂ nanoparticles, such species are encapsulated within the nanofibers, thus generating the IrO₂⊂NS-RuCh system, in which both Ru^II photosensitizers and IrO₂ water oxidation catalysts are within the nanofiber structures. NS-RuCh and IrO₂⊂NS-RuCh have been characterized by dynamic light scattering, scanning electronic microscopy, and energy-dispersive X-ray analysis, which indicated a 2 : 1 ratio between Ru^II chromophores and IrO₂ species. Photochemical water oxidation has been investigated by using IrO₂⊂NS-RuCh as the chromophore/catalyst assembly and persulfate anions as the sacrificial species: photochemical water oxidation yields O₂ with a quantum yield (Φ) of 0.21, definitely higher than the Φ obtained with a similar solution containing separated Ru(bpy)₃²⁺ and IrO₂ nanoparticles (0.05) or with respect to that obtained when using NS-RuCh and “free” IrO₂ nanoparticles (0.10). A fast hole-scavenging process (rate constant, 7×10⁴ s⁻¹) involving the oxidized photosensitizer and the IrO₂ catalyst within the IrO₂⊂NS-RuCh system is behind the improved photochemical quantum yield of IrO₂⊂NS-RuCh.     

糖厂澄清工段生产指标建模及操作参数优化

针对亚法糖厂澄清工段清汁色值和清汁残硫量难以在线测量的问题,提出了一种基于人工蜂群优化的在线极限学习机软测量方法。先用核主元分析法确定影响清汁质量的关键参数,建立基于在线极限学习机的软测量模型。同时利用人工蜂群算法对在线极限学习机的隐层参数进行寻优,优化所建模型。最后,使用带约束的粒子群对软测量模型进行优化求解,得到典型工况下的最优操作设定值,为后续工况操作提供参考依据。仿真结果表明,基于人工蜂群优化的在线极限学习机模型能够准确地预测清汁色值和残硫量,同时基于此模型优化的操作参数设定值能够达到期望的指标。     

Aqueous‐Solution Route to Zinc Telluride Films for Application to CO2 Reduction

As a photocathode for CO₂ reduction, zinc‐blende zinc telluride (ZnTe) was directly formed on a Zn/ZnO nanowire substrate by a simple dissolution–recrystallization mechanism without any surfactant. With the most negative conduction‐band edge among p‐type semiconductors, this new photocatalyst showed efficient and stable CO formation in photoelectrochemical CO₂ reduction at ?0.2–?0.7 V versus RHE without a sacrificial reagent.     

A Doping Technique that Suppresses Undesirable H2 Evolution Derived from Overall Water Splitting in the Highly Selective Photocatalytic Conversion of CO2 in and by Water

Photocatalytic conversion of CO₂ to reduction products, such as CO, HCOOH, HCHO, CH₃OH, and CH₄, is one of the most attractive propositions for producing green energy by artificial photosynthesis. Herein, we found that Ga₂O₃ photocatalysts exhibit high conversion of CO₂. Doping of Zn species into Ga₂O₃ suppresses the H₂ evolution derived from overall water splitting and, consequently, Zn‐doped, Ag‐modified Ga₂O₃ exhibits higher selectivity toward CO evolution than bare, Ag‐modified Ga₂O₃. We observed stoichiometric amounts of evolved O₂ together with CO. Mass spectrometry clarified that the carbon source of the evolved CO is not the residual carbon species on the photocatalyst surface, but the CO₂ introduced in the gas phase. Doping of the photocatalyst with Zn is expected to ease the adsorption of CO₂ on the catalyst surface.     

Voltage‐Driven Reversible Insertion into and Leaving from a Lipid Bilayer: Tuning Transmembrane Transport of Artificial Channels

Three new artificial transmembrane channel molecules have been designed and synthesized by attaching positively charged Arg‐incorporated tripeptide chains to pillar[5]arene. Fluorescent and patch‐clamp experiments revealed that voltage can drive the molecules to insert into and leave from a lipid bilayer and thus switch on and off the transport of K⁺ ions. One of the molecules was found to display antimicrobial activity toward Bacillus subtilis with half maximal inhibitory concentration (IC₅₀) of 10 μM which is comparable to that of natural channel‐forming peptide alamethicin.     

羟基磷灰石/聚乙烯醇/壳聚糖水凝胶在模拟体液中的力学性能

从仿生学角度出发,将自制的人工角膜支架材料羟基磷灰石/聚乙烯醇/壳聚糖(n-HA/PVA/CS)浸泡在模拟体液中,对材料的含水率及力学性能进行了测试,并利用扫描电镜、X射线衍射仪、电感耦合等离子体原子发射光谱仪及热重分析仪研究了材料在模拟体液中的形貌、晶体结构、元素组成及热稳定性.结果表明,在模拟体液中,n-HA/PVA/CS复合水凝胶的含水率为80%~86%,具有较高的拉伸强度,能承受正常眼压,且热稳定性较好.在浸泡后期,n-HA/CS/PVA复合材料对Ca2+的吸附和释放达到动态平衡;而其表面含有微量的纳米羟基磷灰石沉积,有利于纤维细胞的长入.     

Discrete Hf18 Metal-oxo Cluster as a Heterogeneous Nanozyme for Site-Specific Proteolysis

The selective hydrolysis of proteins by non-enzymatic catalysis is difficult to achieve, yet it is crucial for applications in biotechnology and proteomics. Herein, we report that discrete hafnium metal-oxo cluster [Hf₁₈O₁₀(OH)₂₆(SO₄)₁₃⋅(H₂O)₃₃] ( Hf₁₈ ), which is centred by the same hexamer motif found in many MOFs, acts as a heterogeneous catalyst for the efficient hydrolysis of horse heart myoglobin (HHM) in low buffer concentrations. Among 154 amino acids present in the sequence of HHM, strictly selective cleavage at only 6 solvent accessible aspartate residues was observed. Mechanistic experiments suggest that the hydrolytic activity is likely derived from the actuation of Hf^IV Lewis acidic sites and the Brønsted acidic surface of Hf₁₈ . X-ray scattering and ESI-MS revealed that Hf₁₈ is completely insoluble in these conditions, confirming the HHM hydrolysis is caused by a heterogeneous reaction of the solid Hf₁₈ cluster, and not from smaller, soluble Hf species that could leach into solution.     

Accelerated Nuclear Magnetic Resonance Spectroscopy with Deep Learning

Nuclear magnetic resonance (NMR) spectroscopy serves as an indispensable tool in chemistry and biology but often suffers from long experimental times. We present a proof‐of‐concept of the application of deep learning and neural networks for high‐quality, reliable, and very fast NMR spectra reconstruction from limited experimental data. We show that the neural network training can be achieved using solely synthetic NMR signals, which lifts the prohibiting demand for a large volume of realistic training data usually required for a deep learning approach.