Band Gap Variation of Size- and Shape-Controlled Colloidal CdSe Quantum Rods

合集下载

PVP保护下纳米银颗粒的液相化学还原法制备及表征

PVP保护下纳米银颗粒的液相化学还原法制备及表征王春霞;李英琳;徐磊【摘要】Nano-silver particles were prepared by using polyvinyl pyrolidone as the dispersing agent and ammonium as the reducing agent. The resulting specimens were characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM) and UV-Vis adsorption spectrum. The result revealed that a mixture of cube and hexagonal prisms of nano-silver particles was obtained when the mass ratio of PVP to AgNO3 was 2.2:1 and the aging time was 24 hours.%以PVP为表面活性剂，甲酸铵为还原剂，采用液相还原法制备了纳米银颗粒。

采用X射线衍射(XRD)、透射电子显微镜(TEM)和紫外-可见光吸收光谱(UV-Vis)对所制备样品进行表征。

结果显示：当PVP与AgNO3的质量比为2.2：1，陈化时间24 h，得到立方块和六棱柱的银混合颗粒。

【期刊名称】《贵金属》【年(卷),期】2014(000)004【总页数】5页(P30-34)【关键词】纳米化学;纳米银;化学还原法;PVP;陈化时间【作者】王春霞;李英琳;徐磊【作者单位】天津工业大学纺织学院，天津 300387;天津工业大学纺织学院，天津 300387;天津工业大学材料科学与工程学院，天津 300387【正文语种】中文【中图分类】TG146.3+2纳米银作为贵金属纳米材料的一种，具有比表面积大，表面活性高，导电性优异，催化性良好等优点[1]，在物理、化学、生物等方面具有显著的优势，包括表面增强拉曼散射[2]、导电[3]、催化[4]、传感[5]以及广谱抗菌活性[6]等。

多尺度特征融合的脊柱X线图像分割方法

脊柱侧凸是一种脊柱三维结构的畸形疾病，全球有1%~4%的青少年受到此疾病的影响[1]。

该疾病的诊断主要参考患者的脊柱侧凸角度，目前X线成像方式是诊断脊柱侧凸的首选，在X线图像中分割脊柱是后续测量、配准以及三维重建的基础。

近期出现了不少脊柱X线图像分割方法。

Anitha等人[2-3]提出了使用自定义的滤波器自动提取椎体终板以及自动获取轮廓的形态学算子的方法，但这些方法存在一定的观察者间的误差。

Sardjono等人[4]提出基于带电粒子模型的物理方法来提取脊柱轮廓，实现过程复杂且实用性不高。

叶伟等人[5]提出了一种基于模糊C均值聚类分割算法，该方法过程繁琐且实用性欠佳。

以上方法都只对椎体进行了分割，却无法实现对脊柱的整体轮廓分割。

深度学习在图像分割的领域有很多应用。

Long等人提出了全卷积网络[6]（Full Convolutional Network，FCN），将卷积神经网络的最后一层全连接层替换为卷积层，得到特征图后再经过反卷积来获得像素级的分类结果。

通过对FCN结构改进，Ronneberger等人提出了一种编码-解码的网络结构U-Net[7]解决图像分割问题。

Wu等人提出了BoostNet[8]来对脊柱X线图像进行目标检测以及一个基于多视角的相关网络[9]来完成对脊柱框架的定位。

上述方法并未直接对脊柱图像进行分割，仅提取了关键点的特征并由定位的特征来获取脊柱的整体轮廓。

Fang等人[10]采用FCN对脊柱的CT切片图像进行分割并进行三维重建，但分割精度相对较低。

Horng等人[11]将脊柱X线图像进行切割后使用残差U-Net 来对单个椎骨进行分割，再合成完整的脊柱图像，从而导致分割过程过于繁琐。

Tan等人[12]和Grigorieva等人[13]采用U-Net来对脊柱X线图像进行分割并实现对Cobb角的测量或三维重建，但存在分割精度不高的问题。

以上研究方法虽然在一定程度上完成脊柱分割，但仍存在两个问题：（1）只涉及椎体的定位和计算脊柱侧凸角度，却没有对图像进行完整的脊柱分割。

lecture_16

Lecture 16Biosensors1. What are biosensors?The term is used in the literature in many ways. Some definitions:a) A device used to measure biologically-derived signalsb) A device that “senses” using “biomimetic” (imitative of life) strategiesex.,“artificial nose”c) A device that detects the presence of biomoleculesWe will adopt a recent IUPAC definition:“A self-contained integrated device which [sic] is capable of providing specific quantitative or semi-quantitative analytical information using a biological recognition element which is in direct spatial contact with a transducer element.”2. Uses of biosensorsx Quality assurance in agriculture, food and pharmaceutical industries ex.E. Coli, Salmonellax Monitoring environmental pollutants & biological warfare agents ex.,Bacillus anthracis (anthrax) sporesx Medical diagnosticsex., glucosex Biological assaysex., DNA microarrays3. Classes of biosensorsA) Catalytic biosensors:kinetic devices that measure steady-state concentration of a transducer-detectable species formed/lost due to a biocatalytic reaction Monitored quantities:i) rate of product formationii) disappearance of a reactantiii) inhibition of a reactionBioca ta lysts used:i) enzymesii) microorganismsii) organellesiv) tissue samplesB) Affinity biosensors:devices in which receptor molecules bind analytemolecules “irreversibly”, causing a physicochemical change that is detected by a transducerReceptor molecules:i) antibodiesii) nucleic acidsiii) hormone receptorsBiosensors are most often used to detect molecules of biological origin, based on specific interactions.4. Biosensor ComponentsSemipermeablebiological element(electrochemical)Analyte:chemical/biological targetSemipermeable Membrane (1): allows preferential passage of analyte(limits fouling)Detection Element (Biological): provides specific recognition/detection of analyteSemipermeable Membrane (2): (some designs) preferential passage of by- product of recognition eventElectrolyte:(electrochemical-based) ion conduction medium betweenelectrodesTransducer:converts detection event into a measurable signalA) Detection Elements1) Catalysis Strategies: enzymes most commonex., glucose oxidase, urease (catalyzes urea hydrolysis), alcohol oxidase, etc. Commercial Example: glucose sensor using glucose oxidase (GOD)Glucose + O2 + H2O o Gluconic acid + H2O2GOD3 potential measurement routes: 1. pH change (acid production)2.O2 consumption (fluorophore monitor)3.H2O2 production (electrochemical) Commercially Available Biosensors: glucose, lactate, alcohol, sucrose, galactose, uric acid, alpha amylase, choline, L-lysine—all amperometric based (O2 /H2O2)2) Affinity Binding strategies: antibodies & nucleic acid fragments most commonCommercial Example: DNA chipB) Transducers1) Electrochemical:translate a chemical event to an electrical event bymeasuring current passed (amperometric = most common), potential change between electrodes, etc.Oxidation reaction of the reduced chemical species C red :red ox C C no e Measured current is mass transport limitedAmperometric Devices C redC red *G= 96,487 coulombs(Faraday const.) A = electrode areaG = boundary layer widthlim i i n AJwhere J is the flux:*0red reddC C J DD dxG |*redC iD n A G|Example:Glucose sensor based on oxidation of peroxide(most commercial devices)Gel incorporating glucose oxidaseElectrolyteAu counter electrodeAu working electrodeGlucose + O 2 + H 2O o Gluconic acid + H 2O 2GOD Anodic: H 2O 2o O 2 + 2H + + 2e -current passed thru working electrode(Recall: oxidation occurs at anode; here,O -1o O 0)2) Photochemical:translate chemical event to a photochemical event, measure light intensity and wavelength (O)a) Colorimetric: measure absorption intensityExamplesIndirect: H2O2 + Dye Precursor Colored DyeperoxidaseenzymeDirect: flavin adenine dinucleotide (FAD) bound cofactors (redox sites on GOD) absorption at 377nm & 455nm disappears in presence of glucoseb) FluorescenceExample 1: DNA microarrays– fluorophores selectively bound to detected molecule via avidin-biotin complex; commercialized by Affymetrix (S. Fodor)Example 2: fiber optic sensors: fluorophores incorporated into tip change fluorescence level depending on level of target presentTypically:- Oxygen present at tip quenches fluorescence from trapped fluorophore (ex., tris(4,7-diphenyl-1,10 phenantroline) Ru(II) dichloride = Ru(dpp)32+Cl2)- Action of trapped oxidase (biological element, ex., GOD) depletes O2, causing n fluorophore emissionGlucose + O2 + H2O o Gluconic acid + H2O2GODHow can we account fornatural O2 fluctuations? Multichannel fiber optic: 1. enhancing selectivity and/or2. multianalyte detectionHow can we measure multipleanalytes?Ref: MD. Marazuela et al., “Fiber-optic biosensors-an overview”, Anal. Bioanal. Chem.372, 664 (2002).Example 3: Semiconductor nanoparticles (quantum dots)currently in development, ex., Quantum Dot Corp. (P. Alivasatos)ADVANTAGES:i) QD band gap (and hence emission)antibodyvaries with size multiple analyte capability 2nm CdSe green5nm CdSe redii) sharp, intense emission spectra (higher signal/noise)IntensityOiii) can be used for surface or solution-based approachesRef: A.P. Alivisatos, Science 271, 2013 (1998).c) ReflectanceExample 1: “Nanobarcodes” – reflection from surface of multilayer metallic rods provides optical signature; being developed by Surromed, Inc. (M. Natan) Affinity-binding based0.04 – 15 PmPdAuAgAuPtMade by electrochemical reduction of aseries of metal salts into template pores(dissolve w/HNO3) 20-500 nmReflectance microscopy gives unique signature for each rod ReflectedIntensitylengthADVANTAGES:i) solution based (not limited by surface area)ii) many combinations of lengths/sequences multiple analyte capability Multianalyte transduction uses a single fluorophoreFluorophore – indicatesBarcode—identifiesanalyte Detection limit: 1-10 ng/ml Challenges: will require high-throughput readout mechanism Ref : S.R. Nicewarner-Pena et al., Science294, 137(2001)3) Piezoelectric:translate a mass change from a chemical adsorption event to electrical signalExample:Quartz Crystal Microbalanceattached biomolecules- Crystal vibrates at resonantfrequency parallel to appliedfield:Q = (k/m)1/2typical: 5 MHz;research grade: 100-200MHz applied alternating E-field- A change in quartz mass (due to adsorption) changes Q. Advantage: high sensitivity-- 10’s of nanograms/cm2 Disadvantage:highly sensitive to nonspecific adsorptionRef: C.K. O’Sullivan and GG. Guilbault, Biosensors& Bioelectronics14, 663 (1999).5. Detection Element Immobilization MethodsPhysical entrapment—viscous aqueous soln trapped by membrane permeable to analyteMembranes: cellophane, cellulose acetate, PVA, polyurethaneEntrapment Gels: agarose, gelatin, polyacrylamide, poly(N-methyl pyrrolidone) Microencapsulation:inside liposomes, or absorbed infine carbon particles that are incorporated in a gelor membraneAdsorption:direct adsorption onto membrane or transducer; can also be adsorbed onto pre-adsorbed proteins, e.g., albumin; avidin (via biotin linker)Covalent binding (via –COOH, -NH2, -OH chemistries) or crosslinking (ex., via glutaraldehyde) to transducer or membrane surface6. Ideal Biosensor Characteristics1. Sensitivity: high 'S/ 'c analyte (S = signal)2. Simple calibration (with standards)3. Linear Response: 'S/ 'c analyte constant over large concentration range4. Background Signal: low noise, with ability for correction (ex., 2nd fiber sensor head lacking biological species to measure background O2 changes)5. No hysteresis—signal independent of prior history of measurements6. Selectivity—response only to changes in target analyte concentration7. Long-term Stability—not subject to fouling, poisoning, or oxide formation that interferes with signal; prolonged stability of biological molecule8. Dynamic Response—rapid response to variation in analyte concentration9. Biocompatibility—minimize clotting, platelet interactions, activation of complement when in direct contact with bloodstream7. Future Directions1. Multianalyte capability (proteins, biowarfare agents, pathogens, etc.)Cholera B. B.F.Toxin anthracisMS2SEB F1globigiiRicintularensisSalmonellaNaval Research Lab biowarfare agent multianalyte antibody array x Bacteriax Bacteriophage x Toxic proteinsAfter C.R.Taitt et al, Anal. Chem.74,6114 (2002)2. Integration/Miniaturization (microfluidic “lab on a chip” devices)Motorola Labs prototype microfluidic biochip for full DNA analysis from blood samples (60u 100u 2 mm 3)x cell separation x cell lysisx DNA amplification xDNA detectionR.H. Liu et al, Anal. Chem.76, 1824 (2004)Image removed due to copyright considerations.3. Implantable Devicesex., Medtronic glucose sensor implant in major vein of heart—shear from blood flow inhibits cell attachmentImplantable glucose sensorImage removed due to copyright considerations.Implantable insulin pumpImage removed due to copyright considerations.Ref: R.F. Service, Science297, 962 (2002).4. Living cells/tissues as biological elementBioImage screening platform for protein translocations (e.g.,cytoplasmĺ nucleus) associated with the activation ofsignaling pathways (from )Image removed due to copyright considerations.。

硫脲热解制备G-C3N4分解NO如何获得更佳G-C3N4

E ﬃcient and Durable Visible Light Photocatalytic Performance of Porous Carbon Nitride Nanosheets for Air Puri ﬁcationFan Dong,*,†Meiya Ou,†Yanke Jiang,†Sen Guo,‡and Zhongbiao Wu ‡†Chongqing Key Laboratory of Catalysis and Functional Organic Molecules,College of Environmental and Biological Engineering,Chongqing Technology and Business University,Chongqing 400067,China‡Department of Environmental Engineering,Zhejiang University,Hangzhou,Zhejiang 310027,China*Supporting InformationThe development of visible light driven photocatalysts has been the focus of considerable worldwide attention as photocatalysis technology is intensively applied in several important areas,including especially environmental pollution control and solar energy conversion.1−5In general,most of the photocatalysts are metal-containing,such as metal oxide,metal sul ﬁde,tungstates,niobates,tantalates,and vandates.6−8Until recently,a new kind of conjugated polymer semi-conductor (graphitic carbon nitride,g-C 3N 4)has been discovered as a fascinating metal-free organic photocatalyst working under visible light.9−13Graphite-like covalent g-C 3N 4is constructed by poly(heptazine)heterocyclic planes packed closely in a way similar to graphite.9The g-C 3N 4is multifunctional with broad applications (energy conversion and storage,contaminants degradation,carbon dioxide storage and reduction,catalysis,solar cells,and sensing)owing to its high stability,appealing electronic structure,and medium band gap.14,15The g-C 3N 4can be facilely prepared by pyrolysis of nitrogen-rich precursors via polycondensation.9−15The texture,electronic structure,and performance of g-C 3N 4are largely depended on the condensation conditions and the types of precursor.14,15The precursors employed for synthesis of g-C 3N 4include cyanamide,dicyandiamide,trithiocyanuric acid,melamine,triazine,heptazine derivatives,and more recently discovered urea and thiourea.16−23The texture and band structure of g-C 3N 4can also be tuned by templating,doping,heterostrucutre design,and postfunctionalization in order to enhance the reactivity in photocatalysis,selective synthesis,and CO 2reduction.24−29increase of surfaceareas could improve the photocatalytic activity of materials.30,31The former factor is favorable for the reduction of defects and inhibiting charge carriers recombina-tion,while the later one could provide more active sites foradsorption and reaction.30,31However,high crystallinity andlarge surface areas are contradictory in most of the cases.In another word,the synthesis of catalytic materials with high crystallinity can be normally realized at the expense of large surface areas.Thermal treatment is the most common way toenhance crystallinity of the catalytic materials.For example,byincreasing the annealing temperature and prolonging the annealing time during synthesis of TiO2and other inorganic photocatalysts,the crystallinity could be enhanced,which however inevitably resulted in the decrease of surface areas.30,31It is highly desirable that high crystallinity and large surface areas for a catalyst can be achieved simultaneously.In spite of the advances made on the synthesis of g-C3N 4as a photocatalyst for hydrogen evolution and aqueous pollutant degradation,the micro/nanostructuresof g-C 3N 4need to be improved for better photocatalysis.16−23Moreover,the photo-catalytic treatment e ﬃciency of g-C3N 4for gaseous air pollutants has seldombeen reported.Previously,we have synthesized g-C3N 4by pyrolysis of urea and found that the pyrolysis conditions have signi ﬁcant e ﬀects on the microstructure and photocatalytic activity of g-C 3N 4.16,22Received:November 11,2013Revised:January 21,2014Accepted:January 23,2014Published:January 23,2014In the present work,we develop a simple method to engineer the micro/nanostructures of g-C 3N 4from pyrolysis of thiourea and apply the as-prepared g-C 3N 4in visible light photocatalytic air puri ﬁcation.The easily available thiourea is a superior precursor because it is nontoxic,low-cost,and earth-abundant.A layer-by-layer coupled with layer-splitting process was proposed for the gradual reduction of layer thickness and size of g-C 3N 4obtained at elevated temperature and prolonged time.The formation mechanism of g-C 3N 4from thiourea was also revealed.Interestingly and importantly,we ﬁnd that both the crystallinity and the surface areas of g-C 3N 4increase spontaneously with elevated pyrolysis temperature and prolonged pyrolysis time,which is very important to enhance the activity of g-C 3N 4.The morphology and band structure of g-C 3N 4can also be simply engineered by variation of pyrolysis conditions.The optimized g-C 3N 4nanosheets exhibit e ﬃcient and durable visible light photocatalytic performance in NO removal.This unique ﬁnding will shed new light on synthesis and engineering of organic photocatalysts for large-scale environmental applications.2.EXPERIMENTAL SECTION 2.1.Synthesis of g-C 3N 4from Thiourea.All chemicalsused in this study were analytical grade and were used without further puri ﬁcation.In a typical synthesis,10g of thiourea powder was put into an alumina crucible with a cover.The crucible was heated to 550°C at a heating rate of 15°C/min in a tube furnace in air and maintained for 120min.The released air products during thermal treatment were absorbed by dilute NaOH solution of 0.05M.The resulted ﬁnal yellow powder was ground and collected for use without further treatment.In order to investigate the e ﬀects of pyrolysis temperature,g-C 3N 4was synthesized at 500,525,550,575,and 600°C for 120min,respectively.The resulted samples were labeled as CN-500°C,CN-525°C,CN-550°C,CN-575°C,and CN-600°C.In order to investigate the e ﬀects of pyrolysis time,g-C 3N 4was synthesized at 550°C for 0,15,30,60,120,and 240min,respectively.The resulted samples were labeled as CN-0min,CN-15min,CN-30min,CN-60min,CN-120min,and CN-240min.Note that the pyrolysis time does not include the time the furnace spent to raise the temperature to 550°C.2.2.Characterization Methods.The crystal phase was analyzed by X-ray di ﬀraction with Cu K αradiation (XRD:model D/max RA,Japan).The scan rate was 0.02deg/s.The accelerating voltage and the emission current were 40kV and 40mA,respectively.FT-IR spectra were recorded on a Nicolet Nexus spectrometer on samples embedded in KBr pellets.To perform the thermogravimetric-di ﬀerential scanning calorim-etry analysis (TG-DSC:NETZSCH STA 409PC/PG,German),20mg of dry sample was sealed in an Al 2O3cruciblewith a lid and scanned at a rate of 20°C/min.A scanningelectron microscope (SEM,JEOL model JSM-6490,Japan)was used to characterize the morphology of the samples.The morphology and structure were examined by transmissionelectron microscopy (TEM:JEM-2010,Japan).The UV −visdi ﬀuse re ﬂection spectra were obtained for the dry-pressed disksamples using a Scan UV −vis spectrophotometer (UV −vis DRS:UV-2450,Shimadzu,Japan)equipped with an integrating sphere assembly,using BaSO4as re ﬂectance sample.Nitrogenadsorption −desorption isotherms were obtained on a nitrogen adsorption apparatus (ASAP 2020,USA)with all samplesdegassed at 150°C prior to measurements.2.3.Visible Light Photocatalytic Performance for NO Puri ﬁcation.The photocatalytic activity was investigated by removal of NO at ppb levels in a continuous ﬂow reactor as shown in Figure 1(Figure S1shows the photo of the reactorsystem).The volume of the rectangular reactor,made ofstainless steel and covered with Saint-Glass,was 4.5L (30cm ×15cm ×10cm).A 150W commercial tungsten halogen lampwas vertically placed outside the reactor.A UV cuto ﬀﬁlter (420nm)was adopted to remove UV light in the light beam.Photocatalyst (0.2g)was coated onto a dish with a diameter of 12.0cm.The coated dish was then pretreated at 70°C to remove water in the suspension.The catalyst adhesion on the dish was ﬁrm enough to avoid the erosion (or removal)of the catalyst during air ﬂowing.The NO gas was acquired from a compressed gas cylinder at a concentration of 100ppm of NO (N2balance,BOC gas).The initial concentration of NO was diluted to about 600ppbby the air stream.The desired relative humidity (RH)level of the NO ﬂow was controlled at 50%by passing the zero air streams through a humidi ﬁcation chamber.The gas streamswere premixed completely by a gas blender,and the ﬂow rate was controlled at 2.4L/min by a mass ﬂow controller.After the adsorption −desorption equilibrium was achieved,the lamp was turned on.The concentration of NO was continuously measured by a chemiluminescence NO analyzer (ThermoEnvironmental Instruments Inc.,42i-TL),which monitors NO,NO 2,and NO x (NO x represents NO +NO2)with a samplingrate of 1.0L/min.The removal ratio (η)of NO was calculatedas η(%)=(1−C /C0)×100%,where C and C0areconcentrations of NO in the outlet streamand the feeding stream,respectively.Figure 1.Schematic ﬂow diagram of the reactor system.3.RESULTS AND DISCUSSION 3.1.Phase Structure and Transformation.Figure 2a shows the XRD patterns of the prepared g-C 3N 4samplestreated under di ﬀerent temperatures in the range of 500−600°C.All of the g-C 3N 4samples in Figure 2a have similardi ﬀraction patterns,suggesting that all samples are similar in crystal structure.A typical (002)peak around 27.5°is observed,which indicates the graphite-like stacking of the conjugated aromatic units of CN with an interlayer distance of 0.33nm.9A typical (100)di ﬀraction peak around 13.0°corresponding to a distance of 0.68nm could be assigned to the in-plane repeated units.9Further observation on an enlarged view of (002)peak in Figure 2b shows that the di ﬀraction angle 2θof (002)peak increases from 27.31°for the CN-500°C sample to 27.73°for the CN-600°C sample when the pyrolysis temperature increases from 500to 600°C.This result implies that g-C 3N 4becomes more compact when thiourea is treated at ahigher pyrolysis temperature.Figure 2c shows the XRD patterns of the prepared g-C 3N4treated at 550°C for di ﬀerenttimes in the range of 0−240min.The two peaks at around 27.5°and 13°can be observed for all the as-prepared samples in Figure 2c.From Figure 2d,we can see that the di ﬀraction angle 2θof (002)peak increases from 27.24°for CN-0min to 27.66°for CN-240min when the pyrolysis time increases from 0to 240min (Figure 2d).This result suggests that the interlayer distance of g-C 3N 4decreases with prolongedpyrolysis time,which is similar to the e ﬀects of pyrolysis temperature on the crystal structure.Figure 2also illustrates that the di ﬀraction peak intensity become stronger when the pyrolysis temperature is increased and pyrolysis time is prolonged.This fact implies that the crystallinity of g-C 3N 4is improvedwith the elevated pyrolysis temperature and prolonged pyrolysis time.In order to understand the phase transformation during pyrolysis of thiourea,TG-DSC was carried out.The range of temperature is from room temperature to 800°C at a heatingrate of 20°C/min.An alumina crucible with a cover was used during thermal analysis to simulate the actual thermal environment of thiourea pyrolysis.The DSC and TG thermograms for thiourea (Figure 3)clearly show that several phase transformations can be observed in the semiclosed system.An endothermic peak at 190°C is the melting point of thiourea.The strongest endothermic peak appears in the temperature range 210−295°C,and the weight of the sample decreased rapidly by 70.5%.The peak at 236°C (overlapped bythe strong peak)indicates the reaction of thiourea into cyanamide.Cyanamide is a common precursor to synthesize g-C 3N 4.The sharp peak at 266°C implies that the thermal condensation of cyanamide into melamine occurred in this temperature range.The weak endothermic peak at 312°C corresponds to the further condensation processwhereFigure 2.XRD pattern of g-C 3N 4treated under di ﬀerent temperatures (a)and enlarged view of (002)peak (b),XRD pattern of g-C 3N 4treated fordi ﬀerent times (c)and enlarged view of (002)peak (d).melamine is transformed to melem.The weight loss in this temperature range is about 10.4%.The further weight loss (about 6.4%)with an endothermic peak at 422°C can be ascribed to the phase formation from melem to graphitic carbon nitride.Finally,the endothermic peak at 707°C with a weight loss of 12.7%can be attributed to the sublimation of carbon nitride.The TG-DSC results imply the mechanistic transformation process of carbon nitride from pyrolysis of thiourea.93.2.Chemical Composition.The FT-IR spectra of all the samples are shown in Figure 4.We can observe the absorption band at 801cm −1corresponding to a breathing mode of triazine,the absorption bands in the range of 1200−1600cm −1attributing to stretching mode of C −N heterocycles,and the broad bands in the range of the 3000−3700cm −1region attributing to the adsorbed H 2O molecules and N −H vibration.22For the samples treated under lower temperature and for shorter time,the incomplete condensation of thiourea results in the weak vibration of the C 6N 7units.This poor condensation can be improved by increasing the pyrolysis temperature and time to promote the formation process of g-C 3N 4.3.3.Morphology and Nanostructure Formation Mechanism.The typical SEM images of the as-prepared samples are illustrated in Figure 5.Figure 5a shows that the CN-500°C sample is composed of thick layers attached with some agglomerated particles.Figure 5b demonstrates that the CN-550°C sample is mainly composed of interconnected thin layers with some pores that may result from the gas bubbles during pyrolysis of thiourea.In the case of the CN-600°C sample,as shown in Figure 5c,a large number of small thin layers with abundant pores can be observed.The gas bubbles play a key role in the formation of porous structure.Figure 5d demonstrates that the CN-0min sample is composed of large irregular particles with some layered structure.For the CN-30min sample as shown in Figure 5e,thick plates with some particles can be observed.Increasing the pyrolysis time to 240min,the thickness of the sample is signi ﬁcantly reduced,and porous structure is generated at the same time (Figure 5f).By summarizing the above observations,we can conclude thatelevating the pyrolysis temperature and prolonging the pyrolysis time could make the resulted g-C 3N 4samples possess small size,thin layers,and porous structure.This is a facile way to tune the microstructures of g-C3N 4.The EDX elemental mapping of the typical CN-120min sample (Figure 5g)is shown in Figures 5h,5i,and 5j.It can be seen that the C 3N 4sample prepared from thiourea was composed of C,N,and O elements,indicating S was released from the pyrolysis.The microstructure was further investigated by TEM.Figure 6a shows that the CN-500°C sample has a bulk structure composing of large particles with a layer structure.When the pyrolysis temperature was increased to 550°C,the resulted g-C 3N 4sample was of a sheetlike structure with reduced thickness (Figure 6b).When the pyrolysis temperature was further raised up to 600°C,the resulted g-C 3N4sample was composed of a thinner sheetlike porous structure due to the successful introduction of mesopores of several tens of nanometers in the CN-600°C sample (Figure 6c).Further observation in Figure 6a and Figure 6c implies that the average size of the sheets are decreased with increasing temperature probably because the large layers are split into smaller ones under higher temperature.For the g-C3N 4samples prepared at 550°C for di ﬀerent times,Figure 6d shows that the CN-0min sample consists of large particles with irregular shape.A thick and smoothsheetlike structure is clearly observed for the CN-60min in Figure 6e.The morphology of the CN-240min samplewas Figure 3.TG-DSC thermograms for heatingthiourea.Figure 4.FT-IR spectra of g-C 3N 4treated under di ﬀerent temperatures (a)and g-C 3N 4treated for di ﬀerent times (b).quite di ﬀerent,and many thin ﬂat sheets and some mesopores can be clearly seen in Figure 6f.This typical sheetlike morphology imparts CN-240min with a large speci ﬁc surface bining the SEM and TEM results,we can ﬁnd that the thickness and the size of the g-C 3N 4sheetlike nanostructureswere reduced simultaneously when the pyrolysis temperature was increased and the pyrolysis time was prolonged.Such variation in structure would lead to the formation of g-C 3N4with high surface areas and large pore volumes,which is bene ﬁcial for enhancing the photocatalytic activity.The mass of g-C3N 4products obtained under di ﬀerenttemperatures and times with the same amount of thiourea wasmeasured.The weight of g-C 3N 4products was decreasedwithFigure 5.SEM images of CN-500°C (a),CN-550°C or CN-120min (b),CN-600°C (c),CN-0min (d),CN-30min (e),CN-240min (f),SEM image (g),and EDX elemental maping of C,N,and O (h,i,j)in image (g).elevated pyrolysis temperature and prolonged pyrolysis time,resulting from gradual decomposition of solid g-C 3N4due tothermal oxidation in air.The conjugated layered g-C 3N 4isconstructed by the hydrogen bonding between aromatic CN units.The energy of the hydrogen bond is weak and can be destroyed by thermal oxidation.As a result,the layer of CN units would be gradually oxidized and removed in a layer-by-layer way during thermal treatment.16Subsequently,the thickness of g-C 3N 4samples would be decreased with elevated pyrolysis temperature and prolonged pyrolysis time (Figures 5and 6).Meanwhile,large g-C 3N 4layers were split into smaller layers to reduce surface energy (Figures 5and 6).16On this basis,a layer-by-layer coupled with layer-splitting process can be proposed for the explanation of reduction of layer thickness and size of g-C 3N 4samples obtained at elevated temperatureand prolonged time.3.4.Texture Property.The nitrogen adsorption −desorp-tion isotherms and Barrett −Joyner −Halenda (BJH)pore-size distribution of selected samples are displayed in Figure 7.Figures 7a and 7b show that the CN-500°C sample exhibits nonporous structure.When the pyrolysis temperature exceeds 550°C,signi ﬁcant enlargement of surface areas and thegeneration of nanopores (mesopores)can be observed (Figure7b and Table 1).The CN-600°C sample is type IV (Brunauer,Deming,Deming,and Teller,BDDT classi ﬁcation)with a hysteresis loop at high relative pressure between 0.5and 1.0,suggesting the presence of mesopores (2−50nm)and macropores (>50nm).32There are type H3hysteresis loops at 0.45<P /P0<1.00in the isotherms of the optimized samples(CN-600°C and CN-240min),which are often observed onthe aggregates of platelike particles giving rise to slit-shaped pores which agrees well with the nanosheet-like morphology (Figures 5c and 5f).32It can be seen from Figure 7a and Table 1that increasing the condensation temperature from 500to 600°C causes a great enhancement of surface area and pore volume from 5m 2/g and 0.029cm 3/g for the CN-500°C sample to 36m 2/g and 0.25cm 3/g for the CN-600°C sample.The creation of a porous structure can also be observed directly from SEM images (Figures 5a-5c).The e ﬀects of pyrolysis time on the texture property of the as-prepared g-C 3N 4samplesareFigure 6.TEM images for CN-500°C (a),CN-550°C or CN-120min (b),CN-600°C (c),CN-0min (d),CN-60min (e),and CN-240min (f).similar.With increasing thermal treating time,the hysteresis loops shift to the region of lower relative pressure,and the areas of the hysteresis loops gradually become large.Prolonging the pyrolysis time from 0to 240min leads to signi ﬁcant enlargement of surface area from 6m 2/g for the CN-0min sample to 71m 2/g for the CN-240min sample,together with pore volume from 0.036to 0.35cm 3/g (Figures 8a and 8b and Table 1).The change of peak pore size with pyrolysis time also con ﬁrms the introduction of mesopores in the CN-240min sample treated for a longer time (Table 1).The high surface area and large pore volume of CN-600°C and CN-240min samples can be attributed to the reduced layer thickness and size.This interesting result is consistent with SEM and TEM observations (Figures 5and 6).The crystallinity and the surface areas of g-C 3N 4organic photocatalyst can be enhanced with elevated pyrolysis temper-ature and prolonged pyrolysis time (Figures 2and 8).This thermal behavior of g-C 3N 4is contrary to most porous inorganic photocatalysts,which typically undergo structure deformation/pore collapse with decreased surface area upon increasing the heating temperature in order to improve the crystallinity,as it is known that the creation of porous structures with high surface area in g-C3N 4relied largely ontemplates (for example SiO 2,zeolite,and Triton X-100)followed by etching of the templates.33−36Such a process is relatively tedious and thus prevents the large scale applications.This drawback can be overcome by our remarkable observation in this research.The porous nanostructure of g-C 3N 4can be self-generated by a facilely optimized thermal treatment.Porous g-C3N 4with high surface area has been readily synthesized by a template-free method though treating thiourea at higher temperature for a longer time.The creation of porous nanostructure could facilitate catalytic sorption and promote the localization of light-induced electrons in theconjugatedFigure 7.N 2adsorption −desorption isotherms of CN-500°C,CN-550°C,and CN-600°C (a)and the corresponding pore-size distribution curves (b),N 2adsorption −desorption isotherms of CN-0min,CN-30min,CN-60min,CN-120min,and CN-240min (c)and corresponding pore-size distribution curves (d).Table 1.S BET ,Pore Volume,Peak Pore Size,and NO Removal Ratio for Selected g-C 3N 4Samples a sample name S BET (m 2/g)total pore volume (cm 3/g)peak pore size (nm)η(NO)(%)CN-500°C 50.029nonporous 10.2CN-550°C 270.142 2.6/4.122.0CN-600°C 360.25 2.6/3.8/32.632.7CN-0min 60.036 3.87.7CN-30min 100.060 3.814.1CN-60min 120.073 3.817.6CN-120min 270.142 2.6/4.122.0CN-240min 710.35 2.8/3.8/31.132.3C-doped TiO 2451230.25 3.521.8BiOI 5060.027 3.7/18.314.9a The data for C-doped TiO 2and BiOI were collected from references.systems,which are bene ﬁcial for photocatalysis by carbon nitride.113.5.Variation of Band Gap.The relationship between optical property and pyrolysis conditions is investigated by UV −vis DRS,as shown in Figure 9.An absorption edge located in a visible light region is observed for all the samples,which originates from band gap transitions from valence band to conduction band.The absorption edges of g-C 3N 4samples change with the variation of pyrolysis temperature and time.The band gap energy can be estimated from the intercept of thetangents to the plots of (αh ν)1/2vs photon energy,as shown in Figures 9b and 9d.Figures 9a and 9b indicate that when the temperature increases from 400to 550°C,slight reduction band gap energy from 2.49to 2.42eV can be detected.This bathochromic shift in band gap is ascribed to the enhanced structural connections with enhanced van der WaalsinteractionFigure 8.The correlation between S BET and the pyrolysis temperature and time for selected samples (a)and the correlation between pore volume and the pyrolysis temperature and time for selected samples(b).Figure 9.UV −vis DRS (a,c)and plots of (αh ν)1/2vs photon energy (b,d)of g-C 3N 4samples treated under di ﬀerent temperatures and treated for di ﬀerenttimes.between the tri-s-triazine cores as higher pyrolysis temperature results in a higher degree of polymerization and a denser packing of the tri-s-triazine units (Figure 2).37This,in turn,leads to a stronger overlapping of molecular orbitals of the aromatic sheet stacks.Further increasing the temperature from 550to 600°C leads to the hypsochromic shift of the absorption edges from 2.42eV for CN-550°C to 2.57eV for CN-600°C due to the quantum con ﬁnement e ﬀects induced by nanozised particles as high temperatures could signi ﬁcantly reduce the size of g-C 3N 4through layer-by-layer oxidation coupled with layer splitting (Figures 5and 6).38Figures 9c and 9d imply that prolonging the pyrolysis time from 0to 240min causes the band gap energy of g-C 3N 4samples to increase from 2.37to 2.90eV obviously.The relationship between band gap energy of g-C 3N 4and pyrolysis conditions can be found in Figure 9.Recently,Wang et al.developed a novel comonomer strategy to tentatively modify the texture and band structure of g-C 3N 4by chemical incorporation of monomer building blocks into the conjugated polymeric network of g-C 3N 4.39In this research,we can ﬁnd a simple approach to control the microstructure and band gap of g-C 3N 4by tuning the pyrolysis temperature and time,being a potentially valuable way to alter the physical and chemical properties of polymeric semiconductors.3.6.Visible Light Photocatalytic Activity and Stability for NO Removal.3.6.1.Photocatalytic Removal of NO and Monitoring of Reaction Intermediates.The as-prepared g-C 3N 4samples were applied for gaseous NO degradation under visible light irradiation in a continuous reactor in order to demonstrate their potential ability for air puri ﬁcation.Figures 10a and 10b show the variation of NO concentration (C /C 0%)with irradiation time over g-C 3N 4samples treated under di ﬀerent temperatures.Here,C 0is the initial concentration of NO,and C is the concentration of NO after photocatalytic reaction at time t .Previous investigation indicated that NO could not be photolyzed under light irradiation.40It can be found in Figure 10a that NO could not be degraded without photocatalyst under light irradiation or with photocatalyst (CN-600°C)for lack of light irradiation.In the presence of photocatalyst,the NO reacted with the photogenerated reactive radicals to produce the ﬁnal product of HNO 3.Because g-C 3N 4has a suitable band gap that can be directly excited by visible light,all g-C 3N 4samples treated under di ﬀerent temperatures and for di ﬀerent times show decent visible light photocatalytic activity toward NO removal,as shown in Figure 10.Figure 10a indicates that the NO removal ratio of g-C 3N 4samples increases from 10.2%to 32.7%when the pyrolysis temperatures increase from 500to 600°C after 45min irradiation.Figure 10b implies that the NO removal ratio of g-C 3N 4samples increases from 7.7%to as high as 32.3%when the pyrolysis time increases from 0to 240min (Table 1).The visible light activity of CN-600°C and CN-240min samples exceeds that of C-doped TiO 2(21.8%)and BiOI (14.9%),suggestingthatFigure 10.Visible light photocatalytic activities of g-C 3N 4samples treated under di ﬀerent temperatures (a)and g-C 3N 4samples treated for di ﬀerent times (b)for removal NO in air (continuous reactor,NO concentration:600ppb).Monitoring of the fraction of NO 2intermediate over g-C 3N 4samples treated under di ﬀerent temperatures (c)and g-C 3N 4samples treated for di ﬀerent times (d)during photocatalytic reaction.variation of thermal treatment conditions is an e ﬀective approach to enhance the activity of g-C 3N 4.Under the optimized thermal conditions,the photocatalytic activity of g-C 3N 4from thiourea is higher than that of the sample from urea,demonstrating the advantage of thiourea as precursor.16The reaction intermediate of NO 2during photocatalytic oxidation of NO is monitored online as shown in Figures 10c and 10d.The fraction of NO 2generated over g-C 3N 4samples during irradiation decreases with increased pyrolysis temper-ature and prolonged pyrolysis time,which can be ascribed to the fact that the surface areas and pore volumes are increased accordingly.The di ﬀusion rate of reaction intermediate over g-C 3N 4samples with high surface areas and large pore volume is faster,thus promoting the oxidation of intermediate NO 2toﬁnal NO 3−,as shown in the following reactions.40The ﬁnal oxidation products (nitric acid or nitrate ions)can be simply washed away by water wash.Note that as the photocatalytic reaction was going on,the NO concentration in the outlet was decreased gradually due to the conversion of NO to NO 3−.The NO concentration would reach minima until the photocatalytic reaction reached equilibrium.The slight rising of NO concentration was due to the accumulation of NO 3−product on the catalyst surface.40,44After long-term irradiation,the NO concentration in the outlet would reach a steady state.+•→+NO 2OH NO H O 22(1)+•→+−+NO OH NO H 23(2)++→NO NO H O 2HNO 222(3)+•→−−NO O NO 23(4)Thermal treatment is a general process employed to crystallize catalytic materials.The e ﬀects of thermal treatment conditions on the microstructure and photocatalytic activity of di ﬀerent types of photocatalysts have been widely inves-tigated.23,30,31,41−43Yu et al.studied the e ﬀects of calcination temperature on the photocatalytic activity of TiO 2from titanate and found that activity of TiO 2deceased with an increase in calcination temperature in the range of 400to 900°C due to the sintering and crystallite growth and decrease of surface areas and pore volume.41In most cases,there was a medium calcination temperature (not too high and not too low)to make a balance between the surface areas and crystallinity in order to optimize the activity of photocatalysts.For example,Zaleska et al.42found that the optimal preparation temperature for boron-doped TiO 2with the highest activity was 400°C in the range of 300−600°C.However,in our case,the activity of the g-C 3N 4sample is enhanced progressively with continuous elevated temperature and prolonged pyrolysis time (Table 1).This unique variation of the activity should be related to theunusual change of texture property and band gap of g-C 3N4with di ﬀerent thermal treatment conditions (Figures 5,6,8,and 9).The remarkably improved photocatalytic activities of the g-C 3N 4samples with respect to elevated temperature and prolonged pyrolysis time demonstrated above can be explained as the synergistic e ﬀects of enhanced crystallinity,nanosheet-like morphology,large surface area,large pore volume,and increased band gap.First,for the g-C 3N 4sample treated at high temperature and for a long time,the enhancement of crystallization (Figure 2)is advantageous to reduce the recombination rate of photogenerated electrons and holes due to a decrease in the number of the defects.43Second,thenanosheet-like structure (Figures 5and 6)enhances thetransport of photogenerated electrons along the nanosheet,thus lowering the hole −electron recombination.44−47Third,the thin thickness and porous character result in a large surface area for pollutant adsorption.48,49Fourthly,large pore volume (Figure 8)provides more active site for quick reactantdi ﬀusion.31,49,50Lastly and importantly,the increase in theband gap increases the redox ability of charge carriers generatedunder irradiation (Figure 9).46All these favorable factorscocontribute to the signi ﬁcantly improved photocatalytic activities of g-C 3N 4samples synthesized at elevated temper-ature (600°C)and treated for a long time (240min).3.6.2.Photochemical Stability with Multiple Runs.To further test the stability of the optimized CN-600°C and CN-240min samples for practical application,repeated reaction tests were carried out.The sample after one run was useddirectly without further treatment for the next photocatalyticreaction run.As shown in Figure 11,the NO removal ratios ofCN-600°C and CN-240min samples could be well maintainedafter ﬁve cycles under visible light irradiation.Except for a slight drop in the activity during the third running,no further decrease in activity in the following runs can be observed.These results clearly demonstrate that nanostructured porous g-C3N 4photocatalysts with enhanced and durable activity canbe successfully synthesized and applied for e ﬃcient airpuri ﬁcation.Figure 11.Multiple photocatalytic reaction over the CN-600°C sample (a)and the CN-240min sample (b)for removal of NO in air.。

家具英语中英文对照

43 封边 base coating / edge banding
44 钻孔 boring
45 秘书桌 secretary desk
46 遗漏油漆 missing finish
47 油漆不良 finish defect
48 油漆未干 paint not dry
49 组立/结构 assembly / construction
parts broken/split missing/wrong hardware
parts
48 修补不良 improper repair / poor touch up
49
接缝不良/ 开裂
open joint
50 端头崩裂 split on the end
51 表面粗糙 rough surface
4
砂破/起砂（油漆）
over sanding
5 砂光变形 contour variation
6 面边不平直 gap between tops leaves
7 白身 white wood
8 前框线 front frame line
9 抽头线 drawer line
10 零件线 accessories line
41 L型套角 edge coner protector
42 孔径不对 improper hole diameter
43 孔距不对 improper kith betweem hole
44 孔深不对 improper hole depth
45 零配件松动 parts loose
零部件破裂
46 47
遗漏/裂或开放错配件
19 包装线 package line
20 涂装部 finish department

石墨炔的化学修饰及功能化

石墨炔的化学修饰及功能化李勇军;李玉良【摘要】石墨炔特殊的电子结构和孔洞结构使其在信息技术、电子、能源、催化以及光电等领域具有潜在、重要的应用前景.近几年石墨炔的基础和应用研究己取得了重要成果,并迅速成为了碳材料研究中的新领域.石墨炔中炔键单元的高活性为其化学修饰与掺杂提供了良好的平台.在这篇综述中,我们将重点介绍石墨炔的非金属杂原子掺杂、金属原子修饰以及表面改性,并深入探讨掺杂与衍生化对石墨炔材料的电子性质的影响及其对光电化学催化性能的协同增强.【期刊名称】《物理化学学报》【年(卷),期】2018(034)009【总页数】22页(P992-1013)【关键词】石墨炔;掺杂;非金属杂原子;金属原子;化学修饰【作者】李勇军;李玉良【作者单位】北京分子科学国家实验室,中国科学院分子科学科教融合卓越中心,中国科学院化学研究所有机固体院重点实验室,北京100190;中国科学院大学,北京100049;北京分子科学国家实验室,中国科学院分子科学科教融合卓越中心,中国科学院化学研究所有机固体院重点实验室,北京100190;中国科学院大学,北京100049【正文语种】中文【中图分类】O6491 引言石墨炔(graphdiyne，GDY，2010年第一次被李玉良等用汉语命名为石墨炔)，由sp和sp2杂化形成的一种新型碳的同素异形体，它是由 1,3-二炔键将苯环共轭连接形成二维平面网络结构，具有丰富的碳化学键，大的共轭体系、宽面间距、多孔、优良的化学和热稳定性和半导体性能、力学、催化和磁学等性能，是继富勒烯、碳纳米管、石墨烯之后，一种新的全碳二维平面结构材料1–5。

自2010年我们首次通过化学合成获得以来6，石墨炔吸引了来自化学、物理、材料、电子、微电子和半导体领域的科学家对其诱人的半导体、光学、储能、催化和机械性能进行了探索。

石墨炔特殊的电子结构和孔洞结构使其在信息技术、电子、能源、催化以及光电等领域具有潜在、重要的应用前景，近几年石墨炔的基础和应用研究已取得了重要成果，并迅速成为了碳材料研究中的新领域 7–14。

鞋类瑕疵常用中英文对照

一.港宝瑕疵（COUNTER DEFECTS）1.Counter sheet not lasted under insole（港宝没有拉邦至中底之下）2.Wrinkled counter caused either by poor lasting or by wrinkled chemical sheet（拉邦不良或港宝片皱折造成后套皱折）3.Soft chemical sheet due to wrong materials or wrong primer(港宝片品质不良或处理不良以致港宝太软)4.Crooked chemical sheet（港宝位置歪斜）5.Low positioned chemical sheet（港宝位置过低）6.Blisters on counter(后套部位有气泡现象)7.Back height not uniform within pair by 2 mm or more（后套高度不一致达2mm 或超过更多）8.Other counter defects （其他港宝瑕疵）二.Crooked Upper Defects （鞋面弯斜瑕疵）1.Topline crooked or wavy（鞋领口歪曲或成波浪状）2.Back seam off center, twisted or not vertical to heel seat, by 2mm（后合缝车线不正，偏离中线、扭曲或没与后跟垂直----偏离达2mm。

）3.Back heights not uniform within pair （左脚后套高度不一致）4.Different vamp length of left and right foot over 2 mm（左脚鞋面长度不同，相差超过2mm。

）5.other crooked upper defect （其他鞋面弯斜瑕疵）三.Color Variation defect（色差瑕疵）1.Poor antique treatment （锈色处理不良）2.Color variation between different piece of upper（鞋面不同部位的色差）3.Color of stitching and upper material not matching (as per confirmation sample) 车线材料与鞋面材料颜色不对（参照确认样品）4.Wrong color on ornaments and lace（装饰物与鞋带色差）5.Variation between left and right shoes（左右脚色差）6.Poor printing and embroidery（鞋面不良印刷或刺绣）7.Contrast color of PU coating and back fabric（PU面及布里成对比色）8.Other color variation defects（其它色差瑕疵）四.Damaged Upper Defects（鞋面破损瑕疵）1.Scratch mark on upper (鞋面擦痕)2.Visible roughing marks more than 1.5mm or conspicuous(看得见的磨粗过高,超过1.5MM 或很显眼.)3.Cut upper (鞋面破裂)4.Torn upper caused either by high stitch density or poor reinforcement. (因不良补强或车线针距太密造成的鞋面破裂)5.Poor heat embossing. (高温后压花不良)6.Toe box collapsing(鞋头下陷)7.Missing or incorrect material embossment(材料压纹消失或不正确)8.Loose or missing ornament parts.(装饰配件丢失或松动)9.Broken strap(前带或后带断裂)10.Other damaged upper defeats.(其它的鞋面破损瑕疵)五.Eyelet and lace defects(鞋眼和鞋带瑕疵)1.Punched hole bigger or smaller than eyelet.(鞋眼与打洞不合)2.Poor eyelet attachment. (鞋眼钉合不良)3.Loose eyelet. (鞋眼松落)4.Paint coat peeling off. (鞋眼的漆层剥落)5.Rust eyelet or other mental buckles. (鞋眼或其它的金属扣生锈)6.Wrong length. (错误的长度)7.Wrong color 、wrong material or wrong type of lace.(鞋带颜色、材质、型式不对)8.Broken yam（断纱）9.Eyelet row not uniform by 2.4mm.（鞋眼排列不整齐，偏离达2.4mm）10.Dirty marks on lace.(鞋带有不良记号)11.Shoelaces tied together. (左右脚鞋带系在一起)12.Other eyelet and lace defects. (其它的鞋带和鞋眼瑕疵)六.Lining Defects (includes vamp、sock and counter lining)（内里瑕疵，包括鞋面、鞋垫、后套里）1.Lining dirty or damaged or wrinkled (内里不洁、破损或皱折.)2.Trimming damage on topline stitching, causing lining to show gap(修内里口造成的破损导致开口。

国科大慕课英语考试试卷

英语A期末测试卷样题PART I VOCABULARYDirections：Choose the best word or phrase from the choices A, B, C, or D to complete the following sentences.词汇选择1.For most people in the study, the only thing that changed in their lives was that some researchers tried to ________ them to do something new.A) demonstrateB) convinceC) refuteD) satisfy2.The 7301 participants ________ individuals > 50 years of age and their spouses of any age.A) containedB) consistedC) comprisedD) concluded3.Marine construction technology is very complex, somewhat _____ to trying to building a bridge under water.A) specificB) analogousC) peculiarD) comparative4.If schooling is a training in expression and communication, college is ________ the establishment of broad convictions.A) abundantlyB) significantlyC) essentiallyD) possibly5.Although most Asian students have good grammatical competence, they complain of the inability to speak, whereas European students often have the ________ complaint.A) subjectiveB) upsetC) reverseD) optionalPART II CLOZE TEST完型填空(MCQ)_ 1Directions: choose the appropriate verbs from the following box to complete theFinally, the VAERS deaths （6）so far are for the very short term. We have no idea what the death numbers will be in the intermediate and long-term; the clinical trials did not test for those.In a large, population-based study on nearly 400,000 individuals, we （7）that higher baseline participation in physical exercise was associated with a significant lower risk of the development of depression, one of the most prevalent psychiatric disorders.The waste anaerobic digestion (AD) （8）to be an efficient technology for sewage sludge treatment that allows generation of biogas as renewable energy from the same process.It should also be （9）that the consensus varied from 0.40 to 0.56 implying that there was neither perfect disagreement nor perfect agreement between the respondents regarding the effectiveness of online learning.A second wave of the epidemic was （10）on September 24, 2018 and can be explained by the unpredictable violent attacks on health teams and community members in the Ebola affected areas, hampering epidemiological surveillance efforts.完型填空(MCQ)_ 2Directions: choose the appropriate opinion markers from the following box to complete the sentences.In art criticism, people must （11）that the artist has a secret message hidden within the work so as to explore the deeper interpretation of the masterpiece.No one seems to （12）that the American dollar will eventually crash and burn along with many other currencies, but no one seems to have any real idea of what “should” replace it. The （13）aim was to analyze the influence of the second-opinion system on guideline implementation with a view to improving the quality of care.The appropriate people to blame here are the people that allowed these security measures to be pushed through without （14）training.The influence of review valence has received considerable attention both from scholars and practitioners (e.g., Lee and Youn 2009); （15）, findings regarding the role of review valence on consumers' evaluations are not straightforward.PART III ORDERING词组替换_ 1(ID=20001682)Directions: The following paragraph describes how climate change policies affect human health. The first few sentences are given below, while the rest five sentences are not in the correct order. Rearrange them in an appropriate order.There is high scientific confidence that processes associated with climate change affect human health, exacerbating existing global health challenges and creating new ones. Climate change, which has been characterised as the biggest global health threat in the 21 century, negatively impacts health outcomes: from changing diseases patterns to food insecurity and mental health.（16）、（17）、（18）、（19）、（20）PART Ⅳ READING COMPREHENSIONPassage 1阅读理解_ 1(ID=20001683)Directions: Read the following excerpt taken from a research paper on ecology, and then choose the best answer for eachquestion.Para. 1 Many bird species, including those of early successional habitats and those of small tree-fall gaps within mature forest, select disturbed habitats during some portion of the year (Hunter et al. 2001). Several studies have documented greater bird abundance in forest canopy gaps created by natural treefalls (Willson et al. 1982, Blake and Hoppes 1986, Martin and Karr 1986) or group-selection harvest (Kilgo et al. 1999, Moorman and Guynn 2001) than in the mature forest surrounding gaps. Some mature-forest breeders shift into more densely vegetated habitats between breeding and post-breeding periods (Anders et al. 1998; Vega Rivera et al. 1998, 2003; Pagen et al.2000; Vitz and Rodewald 2006). Birds use a variety of forested habitats during migratory periods (Petit 2000, Rodewald and Brittingham 2002), but mature-forest edges and early-succession habitats may experience relatively greater use (Rodewald and Brittingham 2004). Reasons for greater use of disturbed habitats by birds during certain periodsremain speculative, but abundant food and protection from predators have been proposed (Marshall et al. 2003).Para. 2 Arthropod（节肢动物）populations also are influenced by season and habitat type (Johnson and Sherry 2001, Greenberg and Forrest 2003) as well as canopy gap size (Shure and Phillips 1991). It should be advantageous for birds to choose sites with the greatest resource availability (Martin and Karr 1986), and greater invertebrate （无脊椎动物）biomass has been positively correlated to bird abundance (Blake and Hoppes 1986, Holmes et al. 1986), daily nest survival rates, growth rates of nestlings (Duguay et al. 2000), and timing of warbler （鸣鸟）migration (Graber and Graber 1983). Studies of experimental prey removal have not linked decreased prey abundance with negative consequences for the local bird community (Nagy and Smith 1997, Marshall et al. 2002, Champlin et al. 2009).Para. 3 Bowen et al. (2007) documented seasonal shifts in relative use by birds of canopy gap and forest habitat. They speculated these shifts may be driven by seasonal changes in arthropod abundance in gaps. Previous studies have not investigated seasonal shifts in avian habitat use as related to resource availability over multiple periods.Para. 4 Our objectives were to: (1) investigate whether bird use of forest gaps was associated with arthropod abundance or vegetation structure, and (2) ascertain if shifts in relative use of gap and forest understory were related to spatial and temporal variation in arthropod abundance. We predicted positive relationships between avian habitat use and arthropod abundance (i.e., relative bird use of gap vs. forest understory will shift based on changes in local arthropod abundance) from spring migration through fall migration.21. According to Para. 1, when will many bird species use disturbed habitats?A) during any portion of a yearB) before breeding seasonsC) during migration periodsD) whenever food is insufficient22. Which statement is the best paraphrase of the underlined sentence in Para. 1?A) Reasons for birds’ greater use of disturbed habitats in some periods are various, among which food and protection from enemies are two important ones (Marshall et al. 2003).B) Marshall et al. (2003) proposed that food and protection from predators are two major reasons for birds to use disturbed habitats in some st periods.C) Many researchers, including Marshall et al. (2003), have investigated the reasons why birds sometimes use disturbed habitats more frequently.D) Food and protection from predators are two issues (Marshall et al. 2003) to explore why birds use more disturbed habitats in someperiods.23. Why does the author shift the topic from bird’s habitats (Para. 1) to arthropod populations (Para. 2)?A) Because arthro pod populations influence birds’ habitat type.B) Because arthropods are the basic natural food for birds.C) Because arthropod populations are linked with bird abundance.D) Because arthropod activities affect birds’ nestling and breeding.24. What is th e most possible meaning of the underlined word “avian” in Para. 3?A) related to birdsB) related to arthropodsC) related to seasonsD) related to forests25. What is the writing purpose of the underlined sentence in Para. 4?A) to justify research objectivesB) to propose research hypothesisC) to illustrate research methodsD) to summarize research findings。

公差配合与测量技术英文

公差配合与测量技术英文Title: The Essentials of Tolerance, Fit, and Measurement Techniques in Engineering.Tolerance, fit, and measurement techniques are fundamental concepts in engineering that play a pivotalrole in ensuring the precision and reliability of mechanical components. This article delves into the intricacies of these topics, discussing their significance, applications, and practical implications in the realm of manufacturing and quality control.Tolerance Analysis.Tolerance analysis is the process of determining the acceptable range of variations in the dimensions of a component or assembly. It ensures that parts will function as intended when assembled, accounting for manufacturing tolerances and wear over time. Understanding tolerance analysis is crucial for designers, engineers, and qualitycontrol personnel to ensure the interoperability and performance of mechanical systems.Types of Tolerances.There are several types of tolerances commonly used in engineering, including:1. Basic Size Tolerance: This tolerance is based on the nominal or basic size of a part and defines the maximum and minimum allowable deviations from this size.2. Limit Tolerance: It defines the upper and lowerlimits of acceptable dimensions for a part. Any dimension falling outside these limits is considered out of tolerance.3. Fit Tolerance: This tolerance governs therelationship between two mating parts and ensures that they fit together properly. There are various types of fits, including clearance fits (where there is a gap between the parts) and interference fits (where the parts are pressed together).Tolerance Stack-Up.Tolerance stack-up is the cumulative effect of individual tolerances on the overall performance of a system. It considers the tolerances of each component and assembly to predict the overall variation in the final product. Proper tolerance stack-up analysis is essential for ensuring that a system meets its performance requirements.Fit Analysis.Fit analysis is the study of how different components fit together within a mechanical system. It considers both the dimensional tolerances of the individual parts and the functional requirements of the system. Fit analysis helps designers choose appropriate fits for different applications, ensuring reliable performance and ease of assembly.Measurement Techniques.Accurate measurement is crucial for effective tolerance and fit analysis. There are various measurement techniques used in engineering, including:1. Dimensional Measurement: This involves the use of precision measuring tools such as calipers, micrometers, and gauges to measure the dimensions of parts.2. Surface Finish Measurement: Techniques such as profilometry and surface roughness meters are used to assess the surface texture and finish of components.3. Geometric Measurement: Geometric measurements involve the use of tools like coordinate measuring machines (CMMs) to determine the precise geometry and position of features on a part.The selection of appropriate measurement techniques depends on the specific requirements of the application and the accuracy needed for the measurements.Conclusion.Tolerance, fit, and measurement techniques are integral to ensuring the quality and performance of mechanical systems. A thorough understanding of these concepts, along with the ability to apply them effectively, is essentialfor engineers and designers working in the field of manufacturing and quality control. By mastering tolerance analysis, fit analysis, and precise measurement techniques, professionals can create reliable and efficient mechanical systems that meet the demands of modern engineering challenges.。

木质玩具常见不良品描述

Dust 灰尘Pin hole on surface – mi 表面有孔The surface has the wormhole-mi 表面有虫孔Mold mark on surface –ma(CR)霉斑表面Death knot on surface (看尺寸，比较大的记录为MA)死结Life knot on surface – ma 活结wooden knot –ma 结疤The surface not smooth (比较严重的记录为MA)不光滑Gap between joint of part (比较严重的记录为MA)贴合部分裂开wheel come off 轮子脱落axle twist apart 车轴断裂Stick together between the wheels and the train body粘在一起在轮子和车身之间iron wire deformation 铁丝变形Cracked of wood – ma 木材破裂poor quality of wood-ma 木材质量差Splinter on surface – ma 碎片,微小的东西Sawdust on the base –mi 锯屑在基座splinters stuck on the car –mi 碎片粘在车子上Burr on surface –ma 毛边芒刺Sawing mark on surface – mi (比较严重的记录为MA)锯齿Hi-Low of assembly parts – mi (装配高低)Color stain on surface – mi 色上有其他颜色（颜色污迹）Poor painting on surface –mi 油漆喷涂不良Dirty mark on surface – mi 表面脏污dirty mark on plastic containers 塑料容器上脏污Part line on surface –mi 拼缝Painting peeled off – mi 掉漆Chipped mark on surface – ma 缺口Slightly chipped mark on surface – mi 小缺口Severely painting peeled off – ma严重掉漆Broken of color box/color sheet- ma/mi 盒子破裂Scratch mark on surface – mi擦伤刮痕Pressing mark on surface –mi 压痕Cutting mark on surface –mi 切割印迹刀痕Rough surface –mi 表面粗糙big gap between blocks-ma 在积木之间有大缝隙air bubbles –mi 气泡the shape is out of position 形状积木不在应有的位置the size of shape is a bit large 形状积木的尺寸有点大the hand is too easy to move 指针移动起来太紧常见的备注：Delamination in plywood-ma夹板分层(Delamination inplywood was found on 30% inspection samples )在抽样中有30%的夹板有分层gap between joint-ma 连接处有缺口broken of shrink film 收缩薄膜破损The shrink film loose 收缩薄膜松散dog ear on shrink wrap –mi 收缩薄膜的角翘起Plywood夹半板。

znse的带隙 -回复

znse的带隙-回复Title: Band Gap of ZnSe: Understanding the Key AspectsIntroduction:In the realm of materials science and semiconductor physics, the band gap plays a crucial role in determining the electronic properties and potential applications of a material. ZnSe, or Zinc Selenide, is a particularly interesting compound with a wide range of applications due to its unique band structure. This article aims to provide a comprehensive understanding of the band gap of ZnSe, including its definition, significance, and factors influencing its value.1. Definition and Significance of Band Gap:The band gap refers to the energy difference between the valence band and the conduction band in a material's electronic band structure. It can be thought of as the energy threshold that electrons must overcome to transition from the valence band to the conduction band. The magnitude of the band gap determines a material's electronic and optical properties, such as conductivity, transparency, and light absorption.2. Basic Structure and Electronic Band Structure of ZnSe:ZnSe crystallizes in a zinc-blende structure, consisting of alternating layers of zinc (Zn) and selenium (Se) atoms. These atoms are arranged in a face-centered cubic lattice, forming a three-dimensional crystal structure.The electronic band structure of ZnSe is described by energy bands formed by the electronic states in the crystal lattice. Key bands include the valence band, the conduction band, and the band gap between them. The valence band is filled with electrons, while the conduction band is unoccupied. The band gap separates these two bands and influences the material's behavior as an insulator, semiconductor, or conductor.3. Determinants of ZnSe Band Gap:Several factors contribute to the band gap of ZnSe. Understanding these factors is crucial in controlling and manipulating the electronic properties of the material. The following subsections discuss the key determinants briefly:- Atomic Structure: ZnSe's band gap is largely determined by the atomic arrangement within its crystal lattice. The size of the Zn andSe atoms and the distance between them influence the electronic interactions and, subsequently, the band gap energy. The size and arrangement of other atoms, impurities, or dopants in the crystal lattice also affect the band gap.- Composition: The stoichiometry of the compound, i.e., the ratio of Zn to Se atoms, affects the band gap. Varying the composition allows for the tuning of the band gap, which can be advantageous for specific applications.- Strain and Defects: Mechanical strain and lattice defects can significantly impact the band gap of ZnSe. Stress or distortion in the crystal lattice can alter the electronic energy levels and, consequently, affect the band gap.- Temperature: The band gap of ZnSe is temperature-dependent. Changes in temperature can lead to thermal expansion, affecting the atomic arrangement and, in turn, the band gap.4. Applications of ZnSe's Band Gap:The tunable and wide band gap of ZnSe makes it highly versatile for various applications. Notable areas include:- Optoelectronics: ZnSe is widely used in light-emitting diodes (LEDs), laser diodes, and photodetectors. Its wide band gap enables emission in the blue and green regions of the electromagnetic spectrum, vital for these devices' efficient operation.- Solar Cells: ZnSe films are utilized in thin-film solar cells. By optimizing the band gap, these solar cells can effectively convert solar energy into electricity.- Optical Coating: Due to its transparency in the visible and infrared regions, ZnSe is employed in optical coatings, lenses, and windows for environments requiring wide spectral coverage.Conclusion:In conclusion, the band gap of ZnSe plays a pivotal role in determining its electronic, optical, and photoelectric properties. Understanding the various factors influencing the band gap allows for controlled manipulation of its value, enabling tailored applications in optoelectronics, solar cells, and optical coatings. The continued exploration of ZnSe's band gap characteristicspromises exciting possibilities for future technological advancements.。

家具英语(中英文对照).

24 回粘 reviscid
59 裂缝 gap
25 ห้องสมุดไป่ตู้干 wetting
60 擦拭不良 poor wipping
26 白化 blushing
61 漆膜不足 hungry
27 光泽不良 poor gloss
62 薄片太窄 veneer too narrow
28 光泽不均 uneven gloss
63 三角木 cleat
49 皇后床 queen bed
13 前档 front rail
50 床头(板) headboard
14 侧档 side rail
51 床尾(板) foot board
15 右侧下档 right stretcher
52 床侧 side pail
16 左侧下档 left stretcher
53 色板 color panel
17 下横档 center stretcher
54 立水 apron
18 沙座 seat cushion
55 隔板 panel
19 面板 top panel
56 座垫 seat pad
20 左脚 left leg
57 床垫 mattress
21 右脚 right leg
58 床垫块 slats
22 脚上档 front apron
33 色分 flocculation
68 不能通过 unacceptable
34 修色 puiding
69 通过 acceptable
35 明暗对比 hai-lai
70 勉强通过 pass this curting
1
中英文对照
品质术语
序号中文名称英文名称

家具英语中英文对照

中英文对照
品质术语
序号 1
2
中文名称英文名称
中密度纤维粒片板板/刨
花板
M.D.F(medium P.B
density
fiberboard)
3 塑合板 particle board
4 5
调理板(刮空气刀喷)涂雾
枪
drawknife air sry gun
6 桔装涂料 bawel wating
7 薄膜流涂法 curtain flow coating
parts broken/split missing/wrong hardware
parts
48 修补不良 improper repair / poor touch up
49
接缝不良/ 开裂
open joint
50 端头崩裂 split on the end
51 表面粗糙 rough surface
19 包装线 package line
20 涂装部 finish department
21 白身部 white wood department
22 家俱 household furniture
23 配方 formule
24 一套、套房 suite
25
层 layer
26 色桨喷点 sarin booth
27 手砂 hand sanding
28 传送带 conveyor
29 使成型 shape
30 边垫 side pad
31 实木 solid wood
32 面板涂膜 the gel of top
33 硬度 hardness
34 恢复原装 return to former state

分子生物学词汇（G）_生物化学英语词汇

g418 一种氨基糖苷类抗生素gabaergic gama-氨基丁酸能的gaeumannomyces graminis virus 禾顶囊壳病毒gal operon 半乳糖操纵子galactan 半乳聚糖galactocerebroside 半乳糖脑苷脂galactofuranose 呋喃半乳糖galactofuranosidase 呋喃半乳糖苷酶galactofuranoside 呋喃半乳糖苷galactokinase 半乳糖激酶galactosamine 半乳糖胺galactose 半乳糖galactosidase 半乳糖苷酶galactosyl ceramide 半乳糖神经酰胺galactosyl diglyceride 半乳糖甘油二酯galactosylation 半乳糖基化galactosyltransferase 半乳糖基转移酶galline 鸡精蛋白galnac n-乙酰半乳糖胺gametangial copulation 配囊交配gamete 配子gamete lytic enzyme 配子溶酶，溶配子酶gametic ratio 配子（分离）比gametoclonal variation 配子克隆变异gametocyst 配子囊gametocyte 配子母细胞gametogamy 配子生殖gametogenesis 配子发生gametophyte 配子体gamma 微克ganciclovir 9-(1,3-二羟-2-丙氧甲基)鸟嘌呤ganglion 神经节ganglioside 神经节苷脂gap 缺口，空隙，间隙，裂隙gap junction 间隙连接gap junction protein 间隙连接蛋白gap repair 缺口修复gapped duplex （带）缺口（的）双链体gastric inhibitory polypeptide 肠抑胃肽gastrin 胃泌素gastrointestinal hormone 胃肠激素gastrointestinal tract 胃肠道gastrone 抑胃素gastrula 原肠胚gastrulation 原肠胚形成gated channel 门控通道gauche conformation 邻位交叉构象，扭曲构象gaucher disease 戈谢病，高雪病gaulin homogenizer gaulin匀浆器[美国apv gaulin公司生产的匀浆器]gaulin press gaulin压碎器[美国apv gaulin公司生产的高压匀浆器]gaussian distribution 高斯分布gc box gc框[真核生物结构基因上游的顺式作用元件]gc clamp gc封条gc content gc含量gc tailing gc对gc value gc值，gc百分比gdp dissociation inhibitor gdp解离抑制因子[一种g蛋白调节蛋白]gdp dissociation stimulator gdp解离刺激因子[一种g蛋白调节蛋白]gegenion 反离子，反荷离子geiger counter 盖革计数器gel 凝胶gel casting 凝胶灌制gel mobility shiftassay 凝胶迁移率变动分析，凝胶移位分析[dna结合蛋白的电泳分析技术之一，在该试验中标记核酸因与蛋白质结合而导致迁移率变动（下降），相应的条带发生位移（滞后）]gel mold （凝）胶模gel mould （凝）胶膜gelatin 明胶gelatin liquefaction 明胶液化gelatinous precipitate 胶状沉淀gelation 凝胶作用geldanamycin 格尔德霉素gelsolin [肌动蛋白]凝溶胶蛋白[可使肌动蛋白从凝胶状编委溶胶状]geminivirus 双粒病毒组[一组植物病毒]gemma 芽胞[从真菌菌丝上长出的一种适于在不良环境下生存的厚壁孢子]；胞芽genbank nucleotide sequence database genbank核酸序列资料库gender determination 性别决定gender difference 性别差异gene 基因gene flow 基因流[一个随机交配群体，由于合子或配子的散布而造成基因流动，从而引起等位基因频率的改变]gene redundancy 基因丰余[因所编码的rna或蛋白质用量很大，故基因的份数很多]genealogical tree 系统树general initiation factor 通用起始因子genescreen [商]基因筛[是nen dupont公司的商标，一种用于杂交筛选的特制尼龙膜]genetic drift 遗传漂变[由于遗传群体大小有限造成基因频率的随机波动]geneticextinction 遗传绝灭，遗传死亡[由于突变使某基因型的适合度降低，并使有关个体的繁殖力降低或不育，从而导致某等位基因从基因库中消失]geneticfingerprint 基因指纹，遗传指纹[例如不同个体的dna表现不同的限制性片段长度多态性，即可将这种限制酶切片段的电泳带型作为基因指纹]genetic immunization 基因免疫接种[采用基因疫苗进行接种]genetic imprinting 遗传印记genetic polarity 遗传极性[上游基因翻译终止，使下游基因表达下降]genetic system 遗传体系[主要指生物的交配方式，如自交、异交，或两者兼具]genetic transformation 遗传转化[例如一个品系的生物吸收另一品系生物的遗传物质，并获得后一品系某些遗传性状]genetic typing 遗传分型[例如根据人白细胞抗原（hla）基因的限制性片段长度多态性对个体进行分型]geneticin [商]遗传霉素[一种氨基糖苷类抗生素，life technologies公司（gibco）的商标]genistein 染料木黄酮，4,5,7-三羟异黄酮genital cell 生殖细胞genocopy 拟基因型genome 基因组，染色体组genomic imprinting 基因组印记[配子发生过程中基因的选择性差异表达]genomic walking 基因组步查，基因组步移genonema 基因线，基因带genophore 基因线，基因带genotoxicity 基因毒性[dna损伤]genotype 基因型gentamycin 庆大霉素，艮他霉素gentianose 龙胆三糖gentiobiose 龙胆二糖geometric mean 几何均数geosmin [放线菌]土臭味素geotropism 向地性geraniol 牻牛儿糖geranyl 牻牛儿基，香叶基geranylpyrophosphate 牻牛儿焦磷酸germ 胚芽；胚；微生物，（细）菌germ band 胚带germ cell 生殖细胞germ layer 胚层germ line 种系germ nucleus 生殖核germicide 杀菌剂germinal center 生发中心germinal disc 胚盘germinal vesicle 生发泡[初级卵母细胞的细胞核]germplasm 种质[经生殖细胞传递的遗传物质]；生殖质[决定性细胞分化的卵质成分]gestagen 孕激素ghost 血影，空（细）胞；菌蜕[例如丧失胞质仅余胞壁空壳的细菌光学显微图像]；假峰ghost band 假带，鬼带ghost cell 血影细胞ghost peak 假峰，鬼峰giant colony 巨大菌落[可用于观察微生物形态]giant nicelle 巨胶束gibberellic acid 赤霉酸gibberellin 赤霉素gibbs adsorption equation 吉布斯吸附公式gibbs free energy 吉布斯自由能gibbs free energy of activation 活化吉布斯自由能giemsa band 吉姆萨带，g带[中期染色体带]giemsa banding 吉姆萨带，g显带giemsa stain 吉姆萨染液gigaseal 吉伽（欧）封口，千兆（欧）封口[见于膜片箝术，在膜片与吸液管间形成]ginsengenin 人参皂苷配基，人参皂苷元ginsenoside 人参皂苷glacial acetic acid 冰乙酸，冰醋酸gland 腺体glasgow minimum essential medium 格拉斯哥极限必需培养基glcnac n-乙酰葡糖胺glia （神经）胶质glial fibrillary acidic protein 胶质（细胞）原纤维酸性蛋白glial filament acidic protein 胶质（细胞）纤丝酸性蛋白glial growth factor 胶质（细胞）生长因子gliding 滑行，滑移[可指滑行细菌和蓝细菌与固体表面接触时的运动方式，通常很缓慢，并有分泌粘液的轨迹]gliding growth 滑过生长[见于植物]gliotoxin 胶霉毒素global alignment 总体（序列）对比global regulation 全局调节[例如多个分属不同代谢途径的操纵子受控于同一调节物]global regulation circuit 全局调节回路global regulon 全局调节子globin 珠蛋白globoside 红细胞糖苷globular protein 球状蛋白质globulin 球蛋白glomerulus 小球glucagon 胰高血糖素glucan 葡聚糖glucanase 葡聚糖酶glucoamylase 葡糖淀粉酶glucocerebrosidase 葡糖脑苷脂酶glucocerebroside 葡糖脑苷脂glucocorticoid 糖皮质（激）素glucocorticosteroid 糖皮质类固醇glucogenesis 糖生成（作用）glucogenic amino acid 生糖氨基酸glucokinase 葡糖激酶glucomannan 葡甘露聚糖gluconeogenesis 糖异生（作用）gluconic acid 葡糖酸gluconolactone 葡糖酸内酯glucosamine 葡糖胺，氨基葡糖glucosaminoglycan 葡糖胺聚糖glucosan 葡聚糖glucose 葡萄糖[简称或在复合词中可用葡糖]glucosidase 葡糖苷酶glucoside 葡糖苷glucosylation 葡糖基化glucosylceramidase 葡糖神经酰胺酶glucosylceramide 葡糖神经酰胺glucoxyltransferase 葡糖基转移酶glucurnic acid 葡糖醛酸glucuronidase 葡糖醛酸酶glucuronolactone 葡糖醛酸内酯glucuronyl 葡糖醛酸基glutaconate 戊烯二酸；戊烯二酸根、酯、盐glutaconic acid 戊烯二酸glutamate 谷氨酸；谷氨酸盐、酯、根glutamic acid 谷氨酸glutamic semialdehyde 谷氨酸半醛glutaminase 谷氨酰胺酶glutamine 谷氨酰胺glutaraldehyde 戊二醛glutaredoxin 谷氧还蛋白glutathion 谷胱甘肽glutathion peroxidase 谷胱甘肽过氧化物酶glutelin 谷蛋白glutenin 麦谷蛋白glycan 聚糖glycation 糖化，加糖（作用）glyceollin 大豆抗毒素glyceraldehyde 甘油醛glyceride 甘油酯glycerol facilitator 甘油易化蛋白[见于细菌，与甘油的转运有关]glycerol shock 甘油休克glycinamide 甘氨酰胺glycinergic synapse 甘氨酸能突触glycinin 大豆球蛋白glycobiology 糖生物学glycocalyx 糖萼，多糖包被[如见于细菌细胞壁外] glycocholic acid 甘氨胆酸glycoconjugate 糖缀合物，缀合糖，复合糖glycogen 糖原glycogenesis 糖原生成glycogenolysis 糖原分解glycoglyceride 糖基甘油酯glycolipid 糖脂glycolysis 糖酵解glycopeptidase 糖肽酶glycopeptide 糖肽glycophorin 血型糖蛋白glycoprotein 糖蛋白glycosaminoglycan 糖胺聚糖glycosidase 糖苷酶glycoside （糖）苷glycosphingolipid 鞘糖脂glycosyl 糖基glycosylated 糖基化的glycosylation 糖基化glycosylsphingolipid 鞘糖脂glycosyltransferase 糖基转移酶glycyrrhizin 甘草皂苷glyoxal 乙二醛glyoxaline 咪唑glyoxysome 乙醛酸循环体glypiation 糖基磷脂酰肌醇化[在蛋白质的近c端加上g-pi锚]glypican 磷脂酰肌醇（蛋白）聚糖[带有g-pi锚的蛋白聚糖]gnotobiology 悉生生物学gnotobiote 悉生生物[在其体内外生存的微生物均属已知]goblet cell 杯状细胞golgi apparatus 高尔基体golgi body 高尔基体golgi complex 高尔基复合体golgi membrane 高尔基体膜golgi network 高尔基体网络golgi protease 高尔基体蛋白酶gonad 性腺gonadoliberin 促性腺素释放素gonadotrophic hormone 促性腺激素gonadotrophin 促性腺素gonadotropin 促性腺素[为促滤泡素及促黄体素的统称]gonidium 微生子gonium 性原细胞gonococcus 淋球菌gonocyte 性原细胞gonoplasm 精原质gossypol 棉酚gougerotin 谷氏菌素graafian follicle 囊状滤泡，格拉夫卵泡gracilicute 薄壁（细）菌[胞壁由薄层肽聚糖和脂多糖构成的细菌（一般为革兰氏阴性）]graft hybrid 嫁接杂种gram stain 革兰氏染液gramicidin 短杆菌肽grana （复）基粒grana lamella 基粒片层granule 颗粒体[杆状病毒的包含体]；粒，颗粒granuliberin 颗粒释放肽[一种蛙皮肽]granulin 颗粒体蛋白granulocrine 颗粒性分泌granulocyte 粒细胞granulocyte chemotactic peptide 粒细胞趋化肽[即白细胞介素-8]granulocyte colony stimulating factor 粒细胞集落刺激因子granulopoiesis 粒细胞生成granulose 细菌淀粉粒granulosis virus 颗粒性病毒granum 基粒granzyme 粒酶[由细胞毒性t细胞及大颗粒淋巴细胞通过颗粒胞吞的方式分泌的丝氨酸蛋白酶]gratuitous inducer 义务诱导物，安慰诱导物[能诱导酶的合成但不能作为该酶的底物，如iptg就是beta半乳糖苷酶的义务诱导物]grisein 灰霉素gtpase gtp酶guanase 鸟嘌呤酶guanidine hydrochloride 盐酸胍guanidinium isothiocyanate 异硫氰酸胍guanylin 鸟苷蛋白[从肠中分离的鸟苷酸环化酶配体]guanylyl 鸟苷酰基guessmer 猜测体[用于基因克隆的低简并性寡核苷酸探针，其序列按已知的氨基酸序列推导，但仅采纳据猜测最可能与目前的基因配对的密码子]guest 客体guide rna 指导rna[rna编辑的模板]guide sequence 指导序列[如见于rna编辑或rna剪接]gusducin 味（转）导素[见于味蕾的一种转导素（g蛋白）]gustin 味肽，味多肽gut hormone 胃肠激素guttation 吐水gymnoplast 裸质体gynandromorph 雌雄嵌合体gynandromorphism 雌雄嵌合体gynoecium 雌蕊群gynogenesis 雌核发育，单雌生殖gynomerogony 雌核卵块发育gynospore 雌孢子gynostemium 合蕊柱gynotermone 雌性决定素gyplure 类舞毒蛾醇gyrase 回旋酶，促旋酶[大肠杆菌的ii类拓扑异构酶，可在dna中引入负超螺旋]。

211177870_基于深度学习的纹理布匹瑕疵检测方法

基于深度学习的纹理布匹瑕疵检测方法许玉格 1钟铭 1吴宗泽 2, 3任志刚 4刘伟生5摘要布匹瑕疵检测是纺织工业中产品质量评估的关键环节, 实现快速、准确、高效的布匹瑕疵检测对于提升纺织工业的产能具有重要意义. 在实际布匹生产过程中, 布匹瑕疵在形状、大小及数量分布上存在不平衡问题, 且纹理布匹复杂的纹理信息会掩盖瑕疵的特征, 加大布匹瑕疵检测难度. 本文提出基于深度卷积神经网络的分类不平衡纹理布匹瑕疵检测方法(Detecting defects in imbalanced texture fabric based on deep convolutional neural network, ITF-DCNN), 首先建立一种基于通道叠加的ResNet50卷积神经网络模型(ResNet50+)对布匹瑕疵特征进行优化提取; 其次提出一种冗余特征过滤的特征金字塔网络(Filter-feature pyramid network, F-FPN)对特征图中的背景特征进行过滤, 增强其中瑕疵特征的语义信息; 最后构造针对瑕疵数量进行加权的MFL (Multi focal loss)损失函数, 减轻数据集不平衡对模型的影响, 降低模型对于少数类瑕疵的不敏感性. 通过实验对比, 提出的方法能有效提升布匹瑕疵检测的准确率及定位精度, 同时降低了布匹瑕疵检测的误检率和漏检率, 明显优于当前主流的布匹瑕疵检测算法.关键词布匹瑕疵检测, 深度学习, 特征过滤, 深度卷积神经网络, 不平衡分类引用格式许玉格, 钟铭, 吴宗泽, 任志刚, 刘伟生. 基于深度学习的纹理布匹瑕疵检测方法. 自动化学报, 2023, 49(4):857−871DOI 10.16383/j.aas.c200148Detection of Detecting Textured Fabric Defects Based on Deep LearningXU Yu-Ge 1 ZHONG Ming 1 WU Zong-Ze 2, 3 REN Zhi-Gang 4 LIU Wei-Sheng 5Abstract Fabric defect detection is a key part of product quality assessment in the textile industry. Achieving fast,accurate and efficient fabric defect detection is of great significance for improving the productivity of the textile in-dustry. In the production process of fabric, imbalance exists in the shape, size and quantity distribution of fabric de-fects, and the complex texture information of the jacquard fabric will cover the characteristics of the defect, which makes it difficult to detect fabric defects. This paper proposes a method for detecting defects in imbalanced texture fabric based on deep convolutional neural network (ITF-DCNN). First, an improved ResNet50 convolutional neural network model (ResNet50+) based on channel concatenate is established to optimize the fabric defect features.Second, F-FPN (filter-feature pyramid network) method for filtering redundant feature is proposed to filter the background features in the feature maps and enhance the semantic information of defect features. Finally, a MFL (multi focal loss) function weighted with the number of defects is construct to reduce the impact of imbalance on the model, and reduce the model ＇s insensitivity to a small number of defects. Experiments shows the proposed method effectively improves the accuracy of fabric defect detection and the accuracy of defect positioning, while re-ducing the false detection rate and missed detection rate of defect detection, which is significantly higher than the mainstream fabric defect detection algorithm.Key words Fabric defect detection, deep learning, feature filtering, deep convolutional neural network (DCNN),imbalance classificationCitation Xu Yu-Ge, Zhong Ming, Wu Zong-Ze, Ren Zhi-Gang, Liu Wei-Sheng. Detection of detecting textured fabric defects based on deep learning. Acta Automatica Sinica , 2023, 49(4): 857−871随着人工智能的发展, 相关技术的不断成熟,人工智能的应用领域不断扩展, 以深度卷积神经网收稿日期 2020-03-20 录用日期 2020-06-19Manuscript received March 20, 2020; accepted June 19, 2020国家自然科学基金(61703114, 61673126, U1701261, 51675108)资助Supported by National Natural Science Foundation of China (61703114, 61673126, U1701261, 51675108)本文责任编委刘青山Recommended by Associate Editor LIU Qing-Shan1. 华南理工大学自动化科学与工程学院广州 5100062. 深圳大学机电与控制工程学院深圳 5180003. 人工智能与数字经济广东省实验室(深圳) 深圳5180004. 广东省离散制造知识自动化工程技术研究中心广州 5100065. 深圳禾思众成科技有限公司深圳 5180001. School of Automation Science and Engineering, South China University of Technology, Guangzhou 5100062. School of Elec-tromechanical and Control Engineering, Shenzhen University,Shenzhen 5180003. Guangdong Provincial Laboratory of Arti-ficial Intelligence and Digital Economy (Shenzhen), Shenzhen 5180004. Guangdong Discrete Manufacturing Knowledge Automation Engineering Technology Research Center, Guang-zhou 5100065. Shenzhen Hesi Zhongcheng Technology Co.,Ltd., Shenzhen 518000第 49 卷第 4 期自动化学报Vol. 49, No. 42023 年 4 月ACTA AUTOMATICA SINICAApril, 20230.18∼络(Deep convolutional neural network, DCNN),为代表的深度学习方法在大量的工业界实际应用场景中展示了优于传统方法的性能[1]. 在布匹瑕疵检测领域中, 传统的瑕疵检测手段主要是采用人工检测的方式, 这种方式要求检测工人站在流水线旁边使用人眼检测布匹瑕疵, 该方法存在诸多缺陷, 如检测效率低下, 检测速度仅仅达到 0.36 m/s [2],漏检率高, 瑕疵检出率只有70%左右[3], 检测效率随工作时长而下降等. 基于此, 研究自动化的检测系统取代人工检测在纺织工业中有着重要的意义.目前的布匹瑕疵自动化检测的算法主要可分为三大类[4]: 基于统计的方法[5−6]、基于频谱的方法[3, 7−9]和基于模型的方法[10−12]. 统计法通过对布匹瑕疵图片中的纹理特征进行统计分析, 判断输入图片有无瑕疵, 该方法目前只在分辨率较小的瑕疵图片中有所应用, 且其对瑕疵的位置计算不准确. 频谱法是在实际布匹瑕疵检测中使用较多的一类方法, 该方法将图片特征从空间域转换到频率域中, 在频率域中图片特征将得到增强, 由此完成布匹瑕疵的分类,但是该方法目前只能实现对输入图片有无瑕疵的判断, 无法定位瑕疵的位置. 模型法通过假设布匹图片纹理由规则纹理和随机纹理组成, 使用随机过程建模方法对图片特征进行描述从而完成分类, 但是该方法只在少量样本中取得好效果, 对于实际工业生产中大量样本的情况, 尚未有研究表明其可行性. 基于上述方法进行的研究对布匹瑕疵缺陷检测问题有很好的借鉴作用, 但是其研究往往局限于小规模的小分辨率布匹瑕疵图片数据集, 因此参考价值有限. 综上所述, 传统方法存在着瑕疵分割性能差、错检率高、噪声敏感等问题, 随着深度学习的发展, 采用卷积神经网络解决布匹瑕疵检测问题是当前研究的热点之一.基于深度卷积神经网络的方法在目前取得了极大的进展. 2012年Krizhevsky 等[13]提出的Alex-Net, 突破了大规模的视觉识别挑战赛(ImageNet large scale visual recognition challenge, ILSVRC)中的图片分类准确度记录. 2015年Simonyan 等[14]提出VGG (Visual geometry group)系列模型, 该模型通过深层卷积神经网络对图片特征进行提取,最后使用全连接层进行分类, 在当年的ILSVRC 上取得了巨大成功. 随后Google 提出InceptionNet [15−18]系列模型, 该模型通过对网络结构的精心设计, 从而在减小模型参数量的同时提升了模型的表达能力. 2016年He 等[19]提出ResNet, 通过引入残差模块, 解决了梯度消失的问题, 大大加深了网络的深度. 2017年Huang 等[20]提出DenseNet, 其采用密集块(Dense block)对每层的特征图进行复用, 加强网络中特征的传递, 提高网络表现的同时减少网络的参数量. 在此之后, 对网络结构的改进算法大量涌现, 如特征金字塔网络(Feature pyramid net-works, FPN), Cascade R-CNN, BiFPN 等[21−23]. 对网络结构的不断改进, 使得DCNN 在目标检测领域中的应用不断成熟, 已有大量算法实际应用于工业检测系统中[24−29]. 与此同时对于损失函数的改进,如FL (Focal loss)[30]和IOU loss [31]等进一步增强了深度学习模型在检测任务中的表现.基于以上工作基础, 本文提出一种基于深度卷积神经网络的分类不平衡纹理布匹瑕疵检测方法(Detecting defects in imbalanced texture fabric based on deep convolutional neural network, ITF-DCNN), 来解决布匹瑕疵检测中提取到的图片特征难以用于准确的瑕疵检测以及检测模型训练过程中瑕疵样本不平衡的问题. 本文的主要贡献在以下几方面: 1)针对原始的ResNet 中残差模块会造成信息流失, 导致特征图对瑕疵特征表达能力不足的问题, 提出了使用特征通道叠加后采用卷积融合的方式对残差模块进行改进; 2)为充分利用数据集中已有的布匹模板图片, 构造特征过滤金字塔网络(Filter-feature pyramid network, F-FPN), 通过布匹模板图片对布匹瑕疵图片中的冗余背景特征进行过滤;3)基于数据集中各类瑕疵样本分布存在着极度不平衡的问题, 提出改进的损失函数 MFL (Multi fo-cal loss)来解决模型对于少数类瑕疵样本检测准确率不高的问题.本文的组织架构如下: 第1节对布匹瑕疵检测的相关工作进行介绍. 第2节对布匹瑕疵检测中的问题进行描述, 阐明检测模型难以取得精确检测结果的原因. 在第3节中介绍本文采用的布匹瑕疵检测模型, 包括其中特征提取网络的构造, 特征过滤网络的建立, 以及数据集不平衡的解决方案. 第4节将基于构建的布匹数据集进行实验分析, 验证本文提出的模型在布匹瑕疵缺陷检测任务中的有效性, 同时为证明各改进结构对模型性能的提升, 将基于各改进结构进行单独的纵向对比实验. 第5节对全文的工作进行总结和展望.1 研究现状在瑕疵检测领域中, 被检测物体可分为有纹理材料和无纹理材料[2], 布匹作为一种典型的有纹理材料, 其瑕疵检测问题引起了大量专家学者的关注,目前针对布匹瑕疵检测过程中存在的布匹材料变形、背景问题复杂、类内瑕疵多样性大和类间瑕疵极度相似等问题, 已有许多方法尝试对其进行解决[32−36].2017年, Li 等[37]针对纹理变形的布匹瑕疵检测问题, 提出了一种基于Fisher 标准的深度学习方法,该方法首先构造一个基于Fisher 准则的堆叠式自动降噪编码器(Stacked denoise auto-encoder858自动化学报49 卷based on Fisher criterion, FCSDA), 然后将布匹图片分割成等大小的若干份, 分割后的瑕疵图片和无瑕疵图片都将参与FCSDA的训练, 编码器训练完成后, 将待检测图片分割成同样大小的若干份进行预测, 最后通过结合重建后图片与缺陷之间的残差和预先设定的阈值来对缺陷进行定位, 该方法虽然可以完成纹理图片的瑕疵检测, 但是检测前需要将图片进行分割, 且训练过程过于繁琐. 2019年, Jing等[38]提出一种基于卷积神经网络的纹理布匹瑕疵检测方法PTPI (Patches training and image testing), 该方法首先通过距离匹配函数计算出训练时对瑕疵进行分割的最佳尺寸, 随后利用该尺寸将样本图片中的瑕疵逐一分割出来并进行标注, 再将构建好的卷积神经网络模型在手写数字数据集上进行预训练,预训练完成后利用分割后的数据集进行学习, 最后模型在测试图片上进行滑动, 从而完成瑕疵检测.此方法的关键在于分割时尺寸的计算, 要求设计出的尺寸既能够使得分割后的图片完全具有缺陷的特征, 又不能包含太多的背景像素. 对于瑕疵尺寸变化不大的情况, 该方法能取得较好的检测效果, 但是该方法并没有对网络模型进行优化, 无法适应瑕疵分布复杂的情况. Raheja等[39]通过灰度共生矩阵对布匹图片特征进行提取, 辅以滑窗技术计算出待检测图片的纹理信息, 之后将所得信息和参考阈值进行对比, 若纹理信息超出阈值范围则认定为瑕疵图片, 该方法简单易行计算量小, 可在嵌入式数字信号处理器(Digital signal processor, DSP)系统中实现, 但是该方法依赖于手工设计的特征, 实际检测过程中可能会有很多有效信息流失. 此外, Tao等[40]提出使用一种两阶段的方法进行物体表面瑕疵的检测, 首先使用一种级联的自动编码器来对输入图片的每个像素进行预测, 随后只需再确定一个阈值便可实现对缺陷的分割. 该方法为Tao的团队首次提出[40], 在提取出缺陷部位后使用一个CNN模型对缺陷进行分类, 该方法能够实现对瑕疵特征的准确提取, 检测准确率高, 然而要求数据集具有缺陷的掩膜标注信息, 大部分缺陷检测的应用场景难以满足该要求, 且分割模型和分类模型要单独训练, 无法端到端的进行, 过程稍显繁复. He 等[41]使用卷积自动编码器和半监督的生成对抗网络组合来对缺陷进行检测, 该方法首先利用大量的无标签样本训练一个卷积自动编码器, 随后利用该编码器用于特征提取, 将提取到的特征输入到soft-max层进行分类, 最后结合半监督式生成对抗网络(Semi-supervised generative adversarial network, SGAN)模型提升网络的泛化性. 这种半监督的方式可以改善标签样本较少的情况下模型的分类性能, 但是对于样本类别不平衡的问题并没有进行考量. Arikan等[42] 提出一种基于卷积神经网络的方法对物体表面的瑕疵进行实时识别, 该卷积网络通过大卷积核以及跳级连接尽可能多地准确获取缺陷的信息, 且使用了一种新的数据增强方法, 并通过实验验证了该方法的有效性. 文献[42]的方法能够满足检测的实时性要求, 单张图片的检测时间仅为1.9 ms, 但是构造的卷积网络比较精简, 并不适用于背景复杂的样本.2 布匹瑕疵检测问题描述布匹瑕疵检测的自动化是一个集布匹瑕疵特征提取, 瑕疵分类和瑕疵定位的多任务过程. 具体来说, 检测算法要实现3个方面的功能: 1)判断输入的布匹图片有无瑕疵; 2)对有瑕疵布匹图片中的瑕疵进行定位; 3)对定位出来的布匹瑕疵进行分类.常见的布匹瑕疵样本如图1所示, 其主要存在以下3个特征: 1)布匹表面包含大量的花色, 且花色种类繁多, 同时其分布不具有规律性, 造成布匹图片纹理信息复杂, 瑕疵特征提取困难; 2)布匹中的一些瑕疵特征极其相似, 难以用肉眼进行判别, 如图中的“虫粘”类瑕疵和“破洞”类瑕疵; 3)布匹中的各类瑕疵形状各不相同, 尺寸跨度极大, 其形状分布如图2所示. 除此之外, 纹理布匹瑕疵检测还存在着大量其他问题, 如在纹理布匹瑕疵检测问题中,由于图片背景的复杂性, 为降低瑕疵检测的难度,通常会采用图片分割的手段, 得到分辨率较小的图片后再进行检测, 但是这无疑会加大瑕疵检测的工作量, 从而影响检测的速度, 且大部分经典的检测方法仅支持检测单张图片中只存在单个瑕疵的情形,对于单张图片中存在多个不同种类瑕疵的情况仍不能很好的判别. 此外目前大部分基于深度学习进行布匹瑕疵检测的相关工作中, 仅仅给出了对布匹图片是否含有瑕疵的判断结果, 并未对有瑕疵图片中含有沾污网折破洞虫粘花毛其他正常蜡斑图 1 纹理布匹瑕疵样本图片Fig. 1 Samples of jacquard fabric defects4 期许玉格等: 基于深度学习的纹理布匹瑕疵检测方法859的瑕疵进行具体的分类, 或者仅实现对易于检测的瑕疵类别的判断. 而对于布匹瑕疵数据集内各类样本的数量不均衡, 从而导致模型对于少数类样本的检测性能较差的问题, 目前尚未有文献对其进行研究.正是由于布匹瑕疵数据集的这些特性, 使得人工检测及传统的瑕疵检测方案难以达到理想的检测效果, 因此本文在对ResNet, Cascade R-CNN,FPN 等模型研究的基础之上, 提出一种基于深度卷积神经网络的纹理布匹瑕疵检测方法, 针对大分辨率的纹理布匹进行自动的瑕疵特征提取, 并将经过预处理后的瑕疵特征图输入到分类模型和回归模型中, 实现纹理布匹的瑕疵检测.3 布匹瑕疵检测模型构建考虑到布匹瑕疵检测算法对精度与速度的要求, 本文采用的特征提取网络框架为基于ResNet50改进后的ResNet50+. 利用ResNet50+ 对输入的布匹瑕疵图片进行特征的提取, 同时选用提取过程中产生的第2 ~ 6阶段的特征图构建F-FPN 网络模型. F-FPN 网络模型将对不同层级的特征图进行融合, 增强底层特征图的语义信息, 同时利用模板特征图对瑕疵特征图中的复杂背景信息进行过滤, 随后得到的特征图经由区域候选网络(Region proposal network, RPN)提取到约2 000个候选框, 将得到的2 000个候选框进行感兴趣区域 (Region of in-terest, ROI)池化得到输出大小一致的候选框. 在此基础上使用级联的分类器和回归器对候选框进行分类和回归, 提升检测模型对瑕疵的定位精度. 模C i B i i =123型的整体结构如图3所示, 其中 “FC” 代表全连接层,“ ”,“ ”, , , 分别代表分类网络和回归网络.x /像素x /像素x /像素x /像素x /像素x /像素x /像素y /像素y /像素y /像素y /像素y /像素y /像素图 2 纹理布匹瑕疵形状分布Fig. 2 Shape distribution of jacquard fabric defects瑕疵图片模板图片特征提取网络F-FPN 网络结构第 1 阶段预测C 1RPNROI FCB 1ROIFC B 2C 2ROIFC B 3C 3第 2 阶段预测第 3 阶段预测图 3 ITF-DCNN 模型的整体结构图Fig. 3 Structure of proposed model860自动化学报49 卷3.1 基于通道叠加的特征提取网络ResNet50+7×7本文采用的特征提取网络的原型为ResNet50,其网络结构如图4所示. 该网络可分为5个阶段,第1个阶段由卷积层、批归一化(Batch normaliza-tion, BN)层、ReLU 激活层和的池化层构成,其后的4个阶段都可视为由Identity block 和Conv block 组成的残差模块级联构成. 在原始的ResNet50中, Conv block 的结构如图5(a)所示,特征图输入后经过两个通道, 相减后输出, 在该过程中使用的ReLU 激活函数如式(1)所示:1×1之后经过一个卷积核为 , stride = 2的卷积层,输出后的特征图大小为Conv 特征图第 1阶段第 2 阶段第 3 阶段第 4 阶段第 5 阶段BN ReLU PoolingIdentity blockConv block图 4 ResNet50网络结构图Fig. 4 Model structure of ResNet50(a) 原始的 Conv block (a) Original Conv block (b) 改进的 Conv block (b) Modified Conv block (c) 原始的 Identity block (c) Original Identity block (d) 改进的 Identity block (d) Modified Identity block输入图 5 残差模块Fig. 5 Model structure of residual block4 期许玉格等: 基于深度学习的纹理布匹瑕疵检测方法861(W i ,H i )(W o ,H o )3/43×31×13×3 表示输入特征图的长宽, 表示输出特征图的长宽, 卷积核的步长为2, 卷积核尺寸为1将导致最多达的信息流失. 为避免这个问题, 本文对Conv block 的结构进行改进, 将其中的下采样卷积层后移, 并把的卷积核的stride 设置为2来达到下采样的目的, 同时将shortcut 中的卷积核替换成卷积核, 改进后的Conv block 结构如图5(b)所示. 与此同时, ResNet50采用级联的Identity block 模块来代替原本的卷积层,Identity block 模块的引入可以解决随着网络模型的加深而出现的梯度消失问题. 原始的Identity block 模块如图5(c)所示, 输入经过两通道后相加,再进入ReLU 做一个非线性激活, 这导致Identity block 的输出只能为正数, 极大制约模型的表达能力. 因此, 本文采用如图5(d)所示的改进结构, 将ReLU 层移到通道内部, 且特征图直接相加替换为两通道特征图叠加后使用卷积核来进行残差的提取, 从而提升模型对瑕疵特征的表达能力.3.2 冗余特征过滤网络F-FPNC in ={C in l 1,C in l 2,···,C in l n }C in l ii P out =f (C in )C in ={C in 2,C in 3,···,C in 6}1/2i 1×1在深度卷积神经网络中, 进行多尺度特征融合目的在于将不同分辨率的特征信息结合起来, 弥补高分辨率的特征图语义信息低的弱点. 给定输入特征层 , 其中, 代表第层的输入特征层, 融合的实质是找到一种由输入特征层到输出特征层的映射关系, 该过程可表示为:. 在传统的特征金字塔网络中, 取每张输入图片的5个特征层 ,每个特征层的分辨率为输入特征图尺寸的 , 各个特征层经过一个卷积层后将通道数进行统一, 最后采用一种自上而下的融合方式进行融合,该过程具体可见图6(a).2×其中, “ ” 表示进行两倍的上采样. 在布匹瑕疵检测中, 由于布匹图片存在大量的纹理特征,造成背景复杂, 使用传统的FPN 卷积网络生成瑕疵图片的特征图, 容易提取出大量的冗余特征, 增大瑕疵检测的难度. 为解决特征图中存在的特征冗余问题, 可使用特征过滤的方法, 消除无关特征对检测模型的影响. 一种简单的特征过滤方式是将瑕疵布匹图片像素值减去模板图片(模板图片中不含瑕疵, 但是包含对应的复杂背景信息)的像素均值,从而在一定程度上消除背景的影响, 即avg (p t )p d (i,j )p t (i,j )i <m,j <n m ×n 其中, 表示模板图片的像素均值, , 分别表示瑕疵图片和模板图片第i 行第j 列的像素值, 且满足 , 表示图片的分辨率. 但是该方法没有考虑纹理在布匹图片中的空间分布差异, 对于纹理单一的布匹图片可以取得较好的效果, 当布匹图片的纹理较复杂时, 往往难以奏效. 使用深度卷积神经网络对布匹图片进行卷积, 可以提取布匹图片一定空间范围内的特征, 进行多次重复卷积, 可以获取高语义信息的特征, 即在一定程度上消除布匹的纹理特征在空间分布上的差异. 因此使用卷积网络对瑕疵布匹图片和模板图片分别进行特征提取, 为缓解瑕疵图片和模板图片在纹理信息上的差异性, 将输入到网络中的模板图片进行随机1 ~ 10个像素上的抖动, 使得瑕疵图片和模板图片对齐, 得到相关特征图后, 利用模板图FPN 中瑕疵图片输入特征图FPN 中瑕疵图片输出特征图(a) 原始的 FPN (a) Original FPN (b) 加性过滤的 FPN (b) Added filtered FPN (c) 卷积过滤的 FPN(c) Convolutional filtered FPNFPN 中模板图片输入特征图加性过滤后的特征图卷积过滤后的特征图元素相加通道相加图 6 特征图过滤方式Fig. 6 Methods to filtering feature maps862自动化学报49 卷片特征对瑕疵布匹图片特征进行冗余特征过滤. 基于此, 本文结合FPN 网络提出两种过滤的方式.O i =I i d +I i t O i I i d I it i 1)加性过滤: , 其中, 代表过滤后的输出特征图, 和分别代表瑕疵布匹图片和模板图片的特征图, 代表特征图的通道数. 该过滤方式直接对应特征图中的各个特征进行操作, 并不改变特征图的维数, 但是可以增强每一维特征中瑕疵的语义信息, 过程具体可见图6(b). 加性过滤方式直接对特征图中的每一维特征进行相加, 并没有考虑每一维特征在输出特征图中应当占据的权重,这可能导致某些对检测有益的特征语义信息反而减弱, 基于此提出了卷积过滤.O i =concatenate (I i d,I i t )∗K i concatenate (I i d ,I i t )I i d ,I it∗K i 1×11×12)卷积过滤: , 其中表示对的特征图进行叠加, “ ” 表示卷积操作, 表示卷积核. 该过滤方式通过卷积核的引入, 使得网络可以自适应的对不同的特征信息进行加权求和, 避免了有效瑕疵特征语义信息的流失, 其具体过程如图6(c)所示.3.3 MFL 损失函数47:1在布匹瑕疵检测中, 不同种类的瑕疵在数量上分布极不平衡, 其中最常见的瑕疵为“沾污”, 数量最少的瑕疵为“其他”, 两者的不平衡比例高达. 瑕疵种类的不平衡性对于模型的检测效果会产生较大影响, 模型对于样本数量更多的瑕疵能取得更好的检测效果, 而对于样本数量较少的瑕疵,由于难以学得较好的特征表达, 使得模型对该类瑕疵的检测准确率不高, 定位精度较差.为缓解模型中参与训练的正负样本不平衡问题, 体现不同样本的难易分程度, Lin 等[30]在2017年提出了Focal loss 损失函数:p αt γαt γ其中, 表示对anchor 的预测置信度, 该损失函数针对正负类anchors 不均衡及不同anchors 的难易分程度进行特殊处理, 在损失函数中被当作平衡正负类anchors 的加权系数, 同时可通过值来调节难分anchors 在总损失中的占比, 在一定程度上解决样本不平衡的问题. 但是Focal loss 并没有考虑到数据集中各类样本的不平衡度, 仅仅依靠人工对参数的调节, 往往难以把控模型学习的方向, 最终导致模型的检测性能过于依赖参数和的选取, 想要依靠Focal loss 训练得到性能优良的检测模型, 往往需要花费大量的时间在超参数的调整上.为适应多分类不平衡数据集, 同时优化超参数αt βc βc βc 的选取, 本文结合Focal loss 提出了改进的MFL 损失函数. 为使该损失函数可对不同类别的瑕疵进行加权, 将原本的二分类加权系数移除, 并添加多类样本加权系数, 可针对不同瑕疵样本在数据集中的占比自动调整, 从而控制对应样本在损失函数中所占据的比重. 本文提出的计算方式为ωN c c βc 其中, 表示平衡系数, 表示类别为的样本数量. 的引入使得数据集中的每个大类样本的损失在总损失中所占的比例相对较小, 防止数据集的极度不平衡性导致大类样本完全主导梯度更新方向的情况发生. 改进后的损失函数可表示为p t 其中, 的取值如式(5)所示.4 实验分析4.1 数据集构建及实验设置4096×1810本文所采用的数据集来自广东省某纺织企业,数据集共包含有瑕疵图片2 259幅, 模板图片68幅,模板图片中只含有背景纹理信息, 不包含瑕疵, 图片的分辨率为像素, 在实验中, 将其中的2 015幅瑕疵图片划分为训练集, 剩下的244幅作为验证集. 在本数据集中包含7种布匹瑕疵, 分别为沾污、花毛、虫粘、破洞、蜡斑、网折和其他, 在训练过程中背景将单独作为一个类别参与训练, 因此实际是一个8分类的模型.2×2深度学习模型的训练需要大量的样本, 为达到理想的训练效果, 需对原始的数据集进行数据增强,本文采用的数据增强方法为: 1)对原始的瑕疵样本图片进行的切割; 2)对切割后的数据集分别进行水平和垂直方向上的翻转. 数据增强后的样本量为原始数据集的8倍. 增强前后数据集中的样本分布如表1所示. 表1真实反映了数据集中的各类瑕疵在数量上存在极度不平衡问题, 其中“沾污”类瑕疵数量最多,“其他”类瑕疵数量最少, 这种类别失衡问题将导致模型在学习过程中梯度被多数类瑕疵主导, 使得模型对少数类瑕疵的检测性能较差.本文实验采用的计算机配置为Intel(R) Xeon (R) CPU E5-2650 v4 @ 2.20 GHz, NVIDA Ge-Force 1080TI GPU, 操作系统为Ubuntu18.04. 本实验的网络模型基于Pytorch 框架搭建, 实验中设置batch size 为2, 训练步数为600 000, 初始学习率为0.0025, 动量为0.9, 衰减系数设置为0.0001.实验可分为6部分, 第1部分为数据集增强前后模4 期许玉格等: 基于深度学习的纹理布匹瑕疵检测方法863。

反蛋白石结构水凝胶光子晶体的自组装及其对化学物质的响应性研究

摘要响应性光子晶体的折光率和晶格可以随着外界环境的改变而变化，进而引起光子带隙的移动。

当光子带隙落在可见光区域时，宏观上表现为结构色的变化，可用于视觉传感器的研究开发中。

分子印迹技术可以在聚合物中引入与印迹分子形状、大小和结合位点相匹配的纳米空腔，可以实现对目标分子的专一性检测。

本文从响应性光子晶体和它与分子印迹相结合两个角度，构建了不同的视觉传感器，研究了它们对不同化学物质的响应性。

通过超声诱导自组装的方法，制备了具有鲜艳颜色的大面积聚-2-甲基丙烯酸羟乙酯反蛋白石水凝胶(IOHG PHEMA)薄膜。

该IOHG PHEMA薄膜对一系列水溶性醇和羰基类化合物均有响应。

通过改变有机化合物的种类和化学品的浓度，IOHG PHEMA薄膜鲜艳的结构色可以在整个可见光区迅速改变。

此外，IOHG PHEMA 薄膜不仅能被重复利用，而且它的响应速度很快,因此通过最大反射峰位置的移动可以实现对有机化合物浓度变化的半定量分析。

利用毛细力诱导方法，成功制备了具有大面积，颜色鲜艳的2-甲基丙烯酸羟乙酯和苯硼酸共聚的反蛋白石水凝胶(IOHG HEMA+3APBA)薄膜。

利用在一定pH 范围内，苯硼酸可以与多羟基相结合的特点，研究了IOHG HEMA+3APBA薄膜对单糖、多糖和多羟基醇类物质的响应性。

IOHG HEMA+3APBA薄膜的反射峰随着检测物质浓度的增大向长波方向移动，覆盖了整个可见光区域，可以通过颜色的变化来判断浓度范围，实现了对它们的微量检测。

值得一提的是，本章也实现了对生理浓度范围内葡萄糖的线性检测。

以D/L-核糖(D/L-Ri)为印迹分子，制备了反蛋白石结构水凝胶薄膜，它能够对不同浓度的D/L-Ri水溶液进行简单快速检测。

由于印迹分子、AA单体和交联剂的含量均会影响D-Ri x-IOHG w PHEMA y+AAz薄膜的制备、响应性和机械性能，所以通过综合调节各个因素，最后采用了各方面性能都比较优异的D-Ri0.01-IOHG3%PHEMA0.21+AA0.14薄膜作为最终研究对象。

Universalbandgap...

Universal Bandgap Bowing in Group III Nitride AlloysJ. Wu,Applied Science and Technology Graduate Group, University of California, Berkeley, and Materials Sciences Division, Lawrence Berkeley National Laboratory,Berkeley, California 94720,W. Walukiewicz, K.M. Yu, J.W. Ager III,Materials Sciences Division, Lawrence Berkeley National Laboratory,Berkeley, California 94720,S. X. Li, E.E. Haller,Department of Materials Science and Engineering, University of California, Berkeley, and Materials Sciences Division, Lawrence Berkeley National Laboratory,Berkeley, California 94720,Hai Lu, William J. Schaff,Department of Electrical and Computer Engineering, Cornell University, Ithaca, NewYork 14853The energy gaps of MBE-grown wurtzite-structure In1-x Al x N alloys with x≤ 0.25 have been measured by absorption and photoluminescence experiments. The results are consistent with the recent discovery of a narrow bandgap of ~ 0.8 eV for InN. A bowing parameter of 3 eV was determined from the composition dependen ce of these bandgaps. Combined with previously reported data of InGaN and AlGaN, these results show a universal relationship between the bandgap variations of group III nitride alloys and their compositions.PACS numbers: 78.66.Fd, 72.80.EyElectronicMail:*********************Group III nitrides and their alloys are attracting much research attention in part because of applications in optoelectronic devices [1, 2]. Particularly, GaAlN ternary alloys have room-temperature bandgap energies tunable continuously from ~ 3.42 eV for GaN to ~ 6.2 eV for AlN [3]. Ga-rich InGaN alloys are under intense investigations as a highly efficient blue light emitting material [1]. It has been discovered recently that wurtzite-structure InN is actually a narrow gap semiconductor, with a minimum bandgap energy equal to ~ 0.77 eV [4, 5]. The studies of In-rich InGaN have shown that the optical properties of these alloys are consistent with the observed narrow gap of InN, and with the relatively high luminescence efficiency as it is found in the Ga-rich InGaN alloys [5, 6]. The bandgap of the InGaN ternary system thus covers a very wide range of the optical spectrum from the infrared for InN to the ultraviolet for GaN.Similar to the case of InGaN, studies of InAlN alloys have concentrated on the wide-gap, Al-rich side due to the difficulties in the epitaxial grow th of high-quality In-rich materials [7, 8]. However, th e optical properties of InAlN over the entire composition range are of great interest as this alloy system offers an exceptionally large bandgap tunability and a wide range of lattice constant variation. In this letter, we report studies of the optical properties of In-rich In1-x Al x N with x up to 0.25. It is shown that the energy range covered by the bandgap of the InAlN ternary system extends continuously from the infrared for InN to the deep ultraviolet for AlN. The bandgap bowing parameter has been determined to be 3.0 eV. Combining these results with recently reported data on the bandgap energies of other group III-nitride alloys, we have found that their bandgap bowing parameters scale with the energy gap difference between the end-point compounds. This allows expressing the composition dependen ce of the bandgap in terms of a universal dependence applicable to all group III-nitride ternaries. Band edge offsets in group III-N alloys are also discussed in the context of these findings.The In1-x Al x N films (250 ~ 300 nm, 0 ≤x≤ 0.25) were grown on (0001) sapphire substrates by molecular beam epitaxy using an AlN buffer layer (~ 200 nm). The growth temperature was about 470 o C. The growth method and conditions were similar to those described in Ref. [9]. The Al atomic fraction was determined by X-ray diffraction (XRD) using the AlN (0002) peak as the reference peak. The XRD analysis showed that high-quality wurtzite-structure In1-x Al x N epitaxial layers formed with their c-axisperpendicular to the substrate surface. Typical Hall mobility of these samples was several hundred cm 2/Vs, and the sheet free electron concentration was slightly below 1014 cm -2. The samples were characterized by conventional optical absorption and photoluminescence (PL) spectroscopy. The optical absorption measurements were performed on a CARY-2390 NIR-VIS-UV spectrophotometer. The PL signals were generated in the backscattering geometry by excitation with the 515 nm line of an argon laser. The signals were then dispersed by a SPEX 1680B monochromator and detected by a liquid-nitrogen cooled Ge photodiode.As a typical example, the inset of Fig. 1 shows the PL and absorption curves of an In 0.75Al 0.25N sample obtained at room temperature. A photo-modulated reflectance (PR) curve is also shown for comparison. The sample emits an intense PL signal with the peak position observed at the edge of the absorption tail at 1.15 eV. The extrapolation of the linear part of the squared absorption leads to a bandgap energy at 1.55 eV, which is in reasonable agreement with the critical energy determined by the PR spectrum. Similar to the behavior of In -rich InGaN alloys [6], a large Stokes shift is also observed here and is mainly attributed to composition fluctuations. Figure 1 shows the absorption curves for the In 1-x Al x N samples with the Al fraction x ranging from 0 to 0.25. A blue shift of the energy gap with increasing AlN fraction is clearly visible, a result expected from the effect of alloying with wide-gap AlN. Considering the thickness of these films, an absorption coefficient as large as 6×104 cm -1 at photon energy of ~ 0.5 eV above the bandgap is estimated for these films, which is a typical absorption intensity for direct-gap semiconductors.The PL peak energies and the energy gaps derived from the absorption experiments are plotted as a function of x in the inset of Fig. 2. Some previously reported data on the Al-rich side are also shown on the same plot [7]. It can be seen from this plot that the energy gap of the InAlN ternary alloy covers a wide portion of the optical spectrum, ranging from the infrared for InN to the deep ultraviolet for AlN. The composition dependence of the bandgap can be described by the standard bowing equation,()()()x x b x E x E x E InN g AlN g InAlN g −⋅⋅−−⋅+⋅=11. (1)A best fit leads to a bowing parameter of b = 3.0 eV.It is interesting to compare the values of bowing parameter of different group III nitride ternaries. A bowing parameter of 1.43 eV has been found for InGaN [6]. We have also measured the composition dependence of the bandgap of wurtzite GaAlN alloys grown by the same method. Our results show that the bandgap energy as a function of composition can be well fit with a bowing parameter of 1.4 eV. This value is in good agreement with the value of 1.33 eV reported by Shan et. al. [10]. Considering the bandgap difference between the end-point nitrides for each ternary, we note a proportional relationship between the bowing parameter and the bandgap difference. To illustrate this relationship, we define, for the alloy of the form AB, a normalized bowing parameter B g A g E E b −≡β, and a dimensionless bandgap variation as()()B g A g B g A AB g A E E E x E x −−≡α. The standard bowing equation , Eq.(1), can berewritten as()()x x x x −⋅⋅−=1βα. (2)In this equation, the dimensionless parameter β describes the degree of the bandgap bowing relative to the bandgap difference of end-point materials. It is found that the value of β is essentially the same for these three group III-N alloys. It only varies from 0.50 for AlGaN to 0.55 for InAlN alloys. Shown in Fig. 2 is the data of α plotted as a function of x for In 1-x Al x N measured in this work and from a previous report [7], for Ga 1-x Al x N measured in this work, and for In 1-x Ga x N [6, 11, 12] adopted from the literature. It can be seen that although these gap energies were measured on different alloy systems and reported by different groups, they all fall into one single curve when expressed in the reduced form of Eq. (2). A common normalized bowing parameter of β = 0.54 well describes the universal composition dependence, as depicted by the curve in Fig.2. This scaling relationship is not surprising, though, because the main contribution to the bandgap bowing is due to the effects of composition disorder on the conduction and valence band edges [13]. Given the similar degree of disorder in space, for a larger bandgap difference between alloy constituents, the potential perturbation caused by the composition fluctuations is larger; consequently the bandgap bowing effect is expected to be proportionally stronger.The origin of the universal relationship describing the composition dependence of the bandgaps of group III-nitride alloys strongly suggests that similar arguments may be also used in the considerations of the composition dependence of the band offsets. Since the total change of the bandgap is a sum of shifts of the conduction and the valence band edges, it could be argued that the relationship given by Eq. (2) is also a proper scaling function for the band offsets. Namely, for any group III-nitride alloy system, the composition dependence of the conduction or the valence band offset is given by the band offsets of the end-point compounds multiplied by the universal scaling function in Eq. (2). This formula provides a method to estimate the band edge offsets between different group III-nitride alloys, which is an important issue in the design of heterostructure devices. Figure 3 shows the dependence of the bandgaps on the in-plane lattice constant obtained assuming a linear relationship between the lattice constant and the composition according to Vegard’s law. The inset in Fig. 3 shows the conduction and the valence band offsets calculated using the scaling function given by Eq. (2) and the experimentally determined valence band offsets of 1.05 eV for InN/GaN, and 0.70 eV for GaN/AlN [14].The results shown in Fig. 3 suggest that a large gap difference is expected between GaN and the lattice matched In0.18Al0.82N. Also, it is important to note that most of the bandgap difference is accommodated by a large conduction band offset of almost 1 eV. This offers an interesting possibility of using In0.18Al0.82N/GaN heterostructures to confine the two-dimensional electron gas in lattice-matched GaN quantum wells. Such a heterostructure design would eliminate strain-induced polarization effects that are known to be partially responsible for the transfer of electrons from surface defects into the GaN quantum well in standard AlGaN/GaN high electron mobility transistors [15]. A reduction of the piezoelectric field induced charge transfer could provide a better con trol of the heterostructure characteristics by enhancing the role of intentional doping of the barrier.In conclusion, the bandgap energy of In-rich InAlN alloys has been measured as a function of composition. The bowing parameter is found to be 3.0 eV. The bandgap bowing in group III nitride alloys was found to be proportional to the bandgap difference between the end-point compounds.This work is supported by the Director, Office of Science, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098. The work at Cornell University is supported by ONR under Contract No. N000149910936. J. Wu acknowledges support from US NSF Grant No. DMR-0109844.References[1] S. Nakamura, J. Cryst. Growth 202, 290 (1999)[2] S. Strite and H. Morkoc, J. Vac. Sci. Technol. B 10, 1237 (1992)[3] X. Zhang, et. al., Appl. Phys. Lett. 67, 1745 (1995)[4] J. Wu, et. al., Appl. Phys. Lett. 80, 3967 (2002)[5] V. Yu. Davydov, et. al., phys. stat. sol. (b) 230, R4 (2002)[6] J. Wu, et. al., Appl. Phys. Lett. 80, 4741 (2002)[7] K. S. Kim, A. Saxler, P. Kung, M. Razeghi and K. Y. Lim, Appl. Phys. Lett. 71, 800 (1997)[8] Shigeo Yamaguchi, et. al., Appl. Phys. Lett. 76, 876 (2000)[9] H. Lu, et. al., Appl. Phys. Lett. 79, 1489 (2001)[10] W. Shan, et. al., J. Appl. Phys. 85, 8507 (1999)[11]S. Pereira, M. R. Correia, T. Monteiro, E. Pereira, E. Alves, A. D. Sequeira and N. Franco, Appl. Phys. Lett. 78, 2137 (2001)[12] W. Shan, et. al., J. App. Phys., 84, 4452 (1998)[13] J. A. Van Vechten and T. K. Bergstresser, Phys. Rev. B 1, 3351 (1970)[14] G. Martin, A. Botchkarev, A. Rockett and H. Morkoc, Appl. Phys. Lett. 68, 2541 (1996)[15] see, e. g., L. Hsu and W. Walukiewicz, J. Appl. Phys., 89, 1783 (2001)Figure CaptionsFig. 1 Representative absorption curves of In1-x Al x N with x in the range of 0 to 0.25. The inset shows the squared absorption (abs), the PL and PR signals of In0.75Al0.25N taken at room temperature.Fig.2 Normalized bandgap variation s shown as a function of x for In1-x Ga x N, In1-x Al x N and Ga1-x Al x N. The curve is a fit based on Eq.(2) using β = 0.54. Inset, the measured bandgap and PL peak position as a function of Al fraction for In1-x Al x N.Fig. 3 Bandgaps of group III-nitride alloys as a function of in-plane lattice constant. Each curve between two end-points is the quadratic dependence of the bandgap of corresponding ternary alloy described by Eq. (1). Inset, the calculated valence and conduction band edges of group III-N ternary alloys as a function of lattice constant. The points at lattice constants of 3.11, 3.19, and 3.54 Å represent AlN, GaN and InN respectively. All the band edge energies are referenced to the top of the valence band of InN.Figures0123412345ab s o r b a nc e E (eV)Fig. 1 of 3J. Wu et. al.00.20.40.60.8100.20.40.60.81αx Fig. 2 of 3 J. Wu et. al.11012345673.1 3.2 3.3 3.4 3.5E g(e V ) a (A)oFig. 3 of 3J. Wu et. al.。

ZnO禁带宽度

ZnO禁带宽度（band gap）随粒径的变化《纳米材料导论》（习题与文献阅读之一）参考文献：1、Sreetama Dutta, S. Chattopadhyay, A. Sarkar, et al., Role of defects in tailoring structural,electrical andoptical properties of ZnO, Progress in Materials Science 54 (2009) 89–136阅读其中“4.3.4. Experimental results on band gap”一节。

2、自己查询相关文献资料。

分析与讨论：为什么论文中涉及ZnO禁带宽度（band gap）随粒径的减小而变窄（即出现“红移”现象）？什么条件下出现“蓝移”现象？附：参考文献[1]中的主要结果：The variation of shift in band gap of the entire milled sample with grain size (milling time varying from 10 min to 32 h) has been depicted in Fig. 19 and Fig. 20. All the data have been nicely fitted by the following equation:where D is expressed in nm and D0 = 11.6 nm.Fig.19. V ariation of band gap, Eg, of milled Fig. 20. V ariation of band gap, Eg, of differently (batches I and II) ZnO samples, fitted exponentially. treated ZnO samples with inverse of grain size.。

实验说明书- energy band gap experiment

LABORATORY MANUAL FORENERGY BAND GAP EXPERIMENTBackground Semiconductors, PN junction diode, Forward and reverse biasing, Band gap,Fermi level.Aim: To determine the Energy Band Gap of a Semiconductor by using PN Junction Diode.Apparatus: Energy band gap kit containing a PN junction diode placed inside the temperature controlled electric oven, microammeter, voltmeter and connections brought out at the socket, a mercury thermometer to mount on the front panel to measure the temperature of oven.Formula Used: The reverse saturation current, I s is the function of temperature (T) of the junction diode. For a small range of temperatures, the relation is expressed as,log10I s = Constant−5.036 X 103E gTWhere, T is temperature in Kelvin (K) and E g is the band gap in electron volts (eV).Graph between 103/T as abscissa and log10 I s as ordinate will be a straight line having slope = 5.036 E gHence band gap,E g=Slope of line5.036Theory: A semi-conductor (either doped or intrinsic) always possesses an energy gap between its valence and conduction bands (fig.1). For the conduction of electricity, a certain amount of energy is to be given to the electron so that it can jump from the valence band to the conduction band. The energy so needed is the measure of the energy gap (E g) between the top and bottom of valence and conduction bands respectively. In case of insulators, the value of E g varies from 3 to 7 eV. However, for semiconductors, it is quite small. For example, in case of germanium, E g = 0.72 eV and in case of silicon, E g = 1.1 eV.Fig.1. Energy Gap in Metals, Semi-conductors and InsulatorsIn semi-conductors at low temperatures, there are few charge carriers to move, so conductivity is quite low. However, with increase in temperature, more number of charge carriers get sufficient energy to be excited to the conduction band. This lead to increase in the number of free charge carriers and hence increase in conductivity. In addition to the dependence of the electrical conductivity on the number of free charges, it also depends on their mobility. The mobility of the charge carriers, however decreases with increasing temperature. But on the average, the conductivity of the semiconductors rises with rise in temperature.To determine the energy band gap of a semi-conducting material, we study the variation of its conductance with temperature. In reverse bias, the current flowing through the PN junction is quite small and internal heating of the junction does not take place.When PN junction is placed in reverse bias as shown in fig.2(a), the current flows through the junction due to minority charge carriers only. The concentration of these charge carriers depend on band gap E g.The saturation value, I s of reverse current depends on the temperature of junction diode and it is given by the following equation,I s = A (N n e v n + N p e v p ) e −E g kTWhere, N n (N p) is the concentration of electrons (holes) in N(P)-type region,v n and v p are the drift velocities of electrons and holes respectively,A is the area of junction,k = 1.38 x 10-23J/K, is Boltzman’s constant and T is absolute temperature of junction. Taking log of both sides of above equation, we havelog e I s = log e A (N n e v n + N p e v p ) - E gkTOr 2.303 log10I s = 2.303 log10A (N n e v n + N p e v p ) - E gkTOr log10I s = C−E g2.303 kTWhere C is a constant, which is equal to the first term of RHS of above equation. On substituting the value of k and converting the units of E g from eV to Joule, we getlog10I s = C− 1.6 x 10−19 E g2.303 x 1.38 x 10−23 TOr log10I s = C−5.036 X 103E gTWhich can be expressed as,log10I s = C+(−5.036 E g) 103TThis represents the equation of straight line having negative slope (5.036 E g) for graph drawn between log10 I s and 103/T. Therefore, by knowing the slope of the line, E g can be determined through following formula,Slope = 5.036 E gE g=Slope of grapℎ drawn between log10 I s and103T5.036Procedure:The experimental setup is shown in fig.2(b).1.Insert the thermometer in the hole of the oven.2.Switch ON the instrument using ON/OFF toggle switch provided on the front panel.3.Keep the temperature control switch to the high side.4.Adjust the voltage at 1V DC.5.Switch ON the oven using ON/OFF toggle switch provided on the front panel.Temperature starts increasing and the reading of microammeter also starts increasing.6.When temperature reaches to 90℃ or 100℃, switch OFF the oven and note down thereading of microammeter (µA).7.As the temperature starts falling, note down the readings of microammeter afterevery 5℃ or 10℃ drop in temperature.8.Repeat the whole procedure for 2V and 3V DC.9.Plot graph between log10 I s and 103/ T for different voltages.Fig. 2 (a) Reverse biased PN junction Diode (b) Experimental SetupCalculations:Taking 103/ T along X-axis and log 10 I s along Y-axis, plot a graph between log 10 I s and 103/ T for three different voltages. The graph will be a straight line as shown in fig.3. Determine the slope of straight line from this graph and then calculate band gap using formula,Band gap (E g ) = Slope5.036 = _______ eV.Take average of three values of band gap.Fig.3. Variation of log 10 I s v/s 103/TResult:The band gap (E g ) of the given semiconductor is found to be ______ eV.The following precautions should be taken while performing the experiment:1.The diode must be reverse biased.2.Do not exceed the temperature of the oven above 100℃ to avoid over heating of thediode.3.The voltmeter and ammeter reading should initially be at zero mark.4.Bulb of the thermometer should be inserted well in the oven.5.Readings of microammeter should be taken when the temperature is decreasing.6.Readings of current and temperature must be taken simultaneously.Sample viva voce questions:1.What is PN junction diode?2.What do you understand by band gap of a semi-conductor?3.What do you mean by valence band, conduction band and forbidden band?4.How many types of semi-conductors are there?5.What are P-type and N-type semi-conductors?6.Define doping and dopant.7.Why P-type (N-type) semi-conductor is called Acceptor (Donor)?8.What do you mean by Fermi energy level?9.What is the position of Fermi level in an intrinsic semi-conductor and in a p-type or n-type semi-conductor with respect to the positions of valence and conduction bands?10.What do you mean by forward biasing and reverse biasing?11.Why diode is reverse biased in determining the band gap of semi-conductor?12.What is the shape of graph between log10 I s and 103/ T? How do you find band gapenergy from this graph?13.Why conductivity of metals decreases with increase in temperature?14.Why conductivity of a semi-conductor increases with increase in temperature?References:∙Solid State Electronic Devices by Streetman and Banerjee∙ B.sc Practical Physics by Geeta SanonNote: Soft copy of this manual will be available on http://www.nitj.ac.in/physics/。

1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。

Band Gap Variation of Size-andShape-Controlled Colloidal CdSeQuantum RodsLiang-shi Li,Jiangtao Hu,Weidong Yang,and A.Paul Alivisatos*Department of Chemistry,Uni V ersity of California,Berkeley,Berkeley,California94720,and Materials Science Di V ision,Lawrence BerkeleyNational Laboratory,1Cyclotron Road,Berkeley,California94720Received May3,2001W This paper contains enhanced objects available on the Internet at /nano.ABSTRACTWe report the band gaps of rodlike CdSe quantum dots with diameter varying from3.0to6.5nm and length from7.5to40nm.A qualitative explanation for the dependence of band gap on width and length is presented.Semiconductor nanocrystals,also known as“quantum dots”,have been intensely investigated because of their size-dependent optical and electrical properties.1-4The confine-ment of photogenerated electrons and holes in the nano-crystals can be tuned by adjusting the shape and height ofthe potential.Extensive work has been devoted to detailedcomparison of theory of quantum-confined electronic statesin these nanocrystals with experiments in which the diameterof spherical dots is varied.5-13Early phenomenologicalmodels are based on the effective mass approximation;7-10later developments include tight-binding models11,12andempirical pseudo-potential calculations.13Each of thesemodels can provide an adequate description of the band-gap variation vs diameter for spherical or nearly sphericaldots,and they also provide varying levels of success indescribing higher electronic excited states.The advent of newmethods to precisely control the diameter and length ofrodlike CdSe nanocrystals provides a new set of experimentaldata against which the theories can be tested.14-16Here wereport initial measurements of the band gap(photolumines-cence energy)of CdSe quantum rods vs their diameter andlength.CdSe quantum rods are synthesized via pyrolysis ofdimethylcadmium and Se/tributylphosphine solution in a hotmixture of trioctylphosphine oxide,hexylphosphonic acid, and tetradecylphosphonic acid under inert gases.16Figure1 shows transmission electron micrographs of several CdSe rod samples.The size distribution is typically5%in diameter and10%in length.To understand the relationship between the band gap and the dimensions of these quantum rods,we measured photo-luminescence spectra of CdSe rod samples with different widths and lengths.Samples were dispersed in toluene,and their photoluminescence spectra were measured on a com-mercial Spex16820.22m fluorometer at room temperature. All the samples were excited at wavelengths far shorter than*To whom correspondence should be addressed.Figure1.TEM images of four CdSe rod samples.The scale bar is50nm.20017349351their absorption edges to avoid spectral size selection,and the spectral resolution for the photoluminescence spectra is 4nm.The room temperature quantum efficiency of the rods is typically 5-10%.Figure 2shows the emission spectra of 3.7nm wide rod samples with four different lengths.By only changing the length of CdSe quantum rods,we can tune the wavelength of their emission over the same range as in CdSe spherical dots,while the emission from each individual quantum rod is highly linearly polarized,16in contrast to the plane-polarized emission from spherical dots.Although the electronic structures of both zero-dimensional and one-dimensional quantum-confined systems 17,18have been well described by current models,little work has been done on the size regime intermediate between them,i.e.,quantum rods with large aspect ratio.Quantum rods provide an opportunity to study the evolution of properties fromquantum dots to quantum wires.In Figure 3,we show the band gap variation vs width and length for quantum rod samples at room temperature,with the aspect ratio ranging from 1to 12.The original data are also given in Table 1.As expected from quantum confinement considerations,the general tendency is that the emission shifts to lower energy with an increase in either width or length.The data are fit with a polynomial in 1/length (1/L ),1/width (1/W ),and aspect ratio (L /W ).The surface of best fit obtained is 1.8563-Figure 2.Photoluminescence spectra of 3.7((0.2)nm wide CdSe quantum rods with lengths of 9.2,11.5,28.0,and 37.2nm,respectively (from left to right),excited at 450nm.Figure 3.Band gap of CdSe quantum rods vs length and width viewed from two different angles.The mesh is the best fit described in the text.W A three-dimensional graph in wrl format is available.Table 1.Band Gap Energy of CdSe Quantum Rods with Different Widths and Lengths at 295and 7K alength (nm)width (nm)PL (eV)295K PL(eV)7K length (nm)width (nm)PL (eV)295K PL(eV)7K 11.0((0.7) 3.2 2.208.7((0.9) 4.3 2.0737.8((3.0) 3.3 2.0816.4((2.0) 4.3 2.0843.1((3.2) 3.4 2.038.6((1.0) 4.4 2.1028.0((2.2) 3.5 2.1131.5((3.3) 4.4 1.9838.5((4.4) 3.5 2.0515.3((0.8) 4.5 2.1011.5((0.8) 3.6 2.17 2.2019.8((2.0) 4.6 2.0222.1((1.8) 3.6 2.1619.8((1.2) 4.6 1.977.6((0.8) 3.7 2.1612.4((1.3) 4.8 2.039.2((0.7) 3.7 2.1919.4((1.4) 4.8 2.0326.1((3.2) 3.7 2.12 2.1740.4((3.7) 4.8 1.9328.8((4.8) 3.7 2.1218.4((2.0) 4.9 2.068.6((0.8) 3.8 2.1218.9((2.1) 4.9 2.06 2.0937.2((4.0) 3.9 2.07 2.1012.0((1.4) 5.1 1.9944.3((4.0) 3.9 2.06 2.1011.4((1.2) 5.2 2.009.7((0.7) 4.0 2.1222.2((2.2) 5.2 2.0011.6((1.0) 4.0 2.18 2.2340.8((4.2) 5.3 1.9041.3((6.8) 4.0 2.0218.2((1.5) 5.4 2.0012.7((1.0) 4.1 2.158.5((1.0) 5.5 1.9513.4((1.1) 4.1 2.1318.4((1.8) 5.5 1.96 2.0216.9((2.3) 4.1 2.0423.6((2.9) 5.5 1.9740.2((4.0) 4.1 2.0014.0((1.2) 6.2 1.9416.5((2.4) 4.2 2.0617.6((1.0)6.41.9320.2((2.0)4.22.022.05aThe numbers in parentheses are standard deviations.For rod width the standard deviation is 0.2-0.3nm.2.0835/L2+4.5507/W2-0.0018(L/W)2+0.0001(L/W)3+ 10.5824/L3-0.3833/W3.The standard deviation of this fit is∼30meV,which is only slightly greater than the thermal energy at room temperature.The polynomial fit is provided for the convenience of the reader but has no direct physical signifinance.In the surface we see a slight increase in band gap(<20meV)for rods shorte than10nm.This is due to uncertainty in both the experiments and the fitting and cannot be extrapolated to disklike nanocrystals.Recently,we reported an empirical pseudopotential cal-culation performed on CdSe quantum dot nanorods.16The results revealed a level crossing of the two highest occupied electronic states with increasing aspect ratio,which success-fully explains the polarization of the emission.A quantitative description of band gap variation for CdSe quantum rods, however,is still unavailable,though we can qualitatively understand it using the concepts of quantum confinement and“band-mixing”10in CdSe nanocrystals.From the fitting surface,we see that for all the widths and lengths of quantum rods we have studied here,the emission peak positions depend more sensitively on width than on length,as indicated by the slopes(Figure3).This suggests that the band gap is mainly determined by the lateral confinement,which plays an important role even when rods are very long.This can be further justified by the fact that the slope of the peak position with respect to width is almost the same for the rods when the widths are the same,even though the lengths may be very different.It is well-known that the mixing of the valence bands in spherical CdSe nanocrystals is a very important factor controlling their optical properties.10For example,the coupling of the“heavy hole”,“light hole”,and the“split-off”bands under confinement determines the oscillator strength and polarization of optical transitions.In CdSe quantum rods,the symmetry breaking caused by the elonga-tion modifies the band-mixing in such a way that each eigenstate has a definite component of total angular mo-mentum along the long axis.16As shown for3.0nm wide rods,for example,when the aspect ratio is greater than1.3, the hole eigenstates with the lowest energy have the greatest absolute value of translational momentum along the long axis of the crystal.Because the translational momentum projected onto the long axis depends on the length of the quantum rods,changing length will affect the band gap.Especially when the width of the rods is much smaller than the exciton Bohr radius,i.e.,5.6nm in bulk CdSe,the confinement on the carriers in radial directions are in the“strong confinement regime”.10The band mixing is substantial so that the band gap very sensitively depends on the length.For rods with width greater than∼5.6nm,however,where band mixing is less significant,increasing the length only slightly lowers the band gap.The modification of band mixing due to the elongation changes the transition selection rules and oscillator strengths as well.The geometric anisotropy of the rods facilitates their alignment,so that the polarized emission with tunable wavelength could possibly be used on a macroscopic scale. We hope that the measurements provided here will be of use to theorists interested in testing the models for quantum-confined structures.Future work will include studying of the dependence of higher excited states on the rod width and length.Acknowledgment.We thank Kathy Durkin for her help with the3D plot.This work is supported by the Director, Office of Energy Research,Office of Science,Division of Materials Sciences,of the U.S.Department of Energy under Contract No.DE-AC03-76SF00098and by the NIH National Center for Research Resources,Grant No.1R01RR-14891-01under the same DOE contract number.References(1)Efros,Al.L.;Efros,A.L.S o V.Phys.Semicond.1982,16,772.(2)Brus,L.E.J.Chem.Phys.1984,80,4403.(3)Brus,L.E.Appl.Phy s.A1991,53,465.(4)Alivisatos,A.P.Science1996,271,933.(5)Ekimov,A.I.;Hache,F.;Schanne-Klein,M.C.;Ricard,D.;Flytzanis,C.;Kudryavtsev,I.A.;Yazeva,T.V.;Rodina,A.V.;Efros,Al.L.J.Opt.Soc.Am.B1993,10,100.(6)Norris,D.J.;Bawendi,M.G.Phys.Re V.B1996,53,16338.(7)Efros,Al.L.;Rosen,M.;Kuno,M.;Nirmal,M.;Norris,D.J.;Bawendi,M.G.Phys.Re v.B1996,54,4843.(8)Xia,J.-B.Phys.Re V.B1989,40,8500.(9)Efros,Al.L.Phys.Re v.B1992,46,7448.(10)Efros,Al.L.;Rosen,M.Annu.Re V.Mater.Sci.2000,30,475.(11)Martin,E.;Delerue,C.;Allan,G.;Lannoo Phys.Re V.B1994,50,18258.(12)Leung,K.;Pokrant,S.;Whaley,K.B.Phys.Re V.B1998,57,12291.(13)Wang,L.-W.;Zunger,A.Phys.Re v.B1996,53,9579.(14)Peng,X.;Manna,L.;Yang,W.;Wickham,J.;Scher,E.;Kadavanich,A.;Alivisatos,A.P.Nature2000,404,59.(15)Peng,Z.A.;Peng,X.G.J.Am.Chem.So c.2001,123,1389.(16)Hu,J.;Li,L.-S.;Yang,W.;Manna,L.;Wang,L.-W.;Alivisatos,A.P.Science2001,292,2060.(17)Sercel,P.C.;Vahala,K.J.Phys.Re V.B1990,42,3690.(18)Sercel,P.C.;Vahala,K.J.Phys.Re V.B1991,44,5681.NL015559R。

Band Gap Variation of Size- and Shape-Controlled Colloidal CdSe Quantum Rods

PVP保护下纳米银颗粒的液相化学还原法制备及表征

多尺度特征融合的脊柱X线图像分割方法

lecture_16

硫脲热解制备G-C3N4分解NO如何获得更佳G-C3N4

家具英语 中英文对照

石墨炔的化学修饰及功能化

鞋类瑕疵常用中英文对照

国科大慕课英语考试试卷

公差配合与测量技术英文

木质玩具常见不良品描述

znse的带隙 -回复

家具英语(中英文对照).

家具英语 中英文对照

分子生物学词汇（G）_生物化学英语词汇

211177870_基于深度学习的纹理布匹瑕疵检测方法

反蛋白石结构水凝胶光子晶体的自组装及其对化学物质的响应性研究

Universalbandgap...

ZnO禁带宽度

实验说明书- energy band gap experiment

家具英语中英文对照

家具英语中英文对照