On T-Stability of PicardIteration in

合集下载

固体地球物理学概论Snell定律课件

固体地球物理学概论Snell定律课件
In three dimension orthogonal coordinate system, we can define stress p as (pxx pxy pxz pyx pyy pyz pzx pzy pzz).
固体地球物理学概论
第七章
弹性概念——应力 (续)
The stresses are symmetrical(对称的), i.e. only six components of the stress tensor p are independent because
P = - (pxx+ pyy+ pzz)/3 This is a general definition of the “pressure”. In the special case of a liquid at rest, pxx= pyy= pzz = - P, this is the hydrostatic pressure. In geology, lithostatic pressure is often estimated by using
When the material in the mantle is heated, it expands and becomes lighter. In spite of its high viscosity(粘性), it rises more or less vertically in some places, especially under the oceanic ridges. With its losing pressure and heat during traveling upward, the material is forced to travel horizontally. They drag the lithosphere motion.

On the Evaluation of Images Complexity A Fuzzy Approach

On the Evaluation of Images Complexity A Fuzzy Approach

On the Evaluation of Images Complexity:A Fuzzy ApproachMaurizio Cardaci1,2,Vito Di Ges`u1,3,Maria Petrou4,and Marco Elio Tabacchi31Universit`a di Palermo,C.I.T.C.,Italy2Universit`a di Palermo,DDP,Italy3Universit`a di Palermo,DMA,Italy4University of Surrey,UKAbstract.The inherently multidimensional problem of evaluating thecomplexity of an image is of a certain relevance in both computer sci-ence and cognitive puter scientists usually analyze spa-tial dimensions,to deal with automatic vision problems,such as feature-extraction.Psychologists seem more interested in the temporal dimensionof complexity,to explore attentional models.Is it possible,by mergingboth approaches,to define an more general index of visual complexity?We have defined a fuzzy mathematical model of visual complexity,usinga specific entropy function;results obtained by applying this model topictorial images have a strong correlation with ones from an experimentwith human subjects based on variation of subjective temporal estima-tions associated with changes in visual attentional load,which is alsodescribed herein.Keywords:Fuzzy sets,image analysis,complexity,entropy,mentalclock,internal clock.1IntroductionThe problem of evaluating the complexity of an image is of a certain relevance to both cognitive and computer science studies,although in broader contexts the general problem of visual complexity measurement is ill-defined.The evaluation of visual complexity is useful in understanding relations among different levels of the recognition process and it is also of interest to real applications such as image compression and information theory.The Computer Science approach to visual complexity is generally space-based: local feature extraction and selection plus global statistical parameter estimation are employed to quantify complexity from the point of view of a rational agent. Nevertheless,complexity is not only relevant to the stimulus’spatial properties, but,as an emerging factor affecting the human perceiver’s cognitive operations, it can also involve the temporal dimension.The recent Mental Clock Model[1], developing an intuition by Ornstein[2],relies on the simple hypothesis that the subjective passing duration is affected by a hypothetical internal clock which tends to modify its speed according to the attentional load of the current task.I.Bloch,A.Petrosino,and A.G.B.Tettamanzi(Eds.):WILF2005,LNAI3849,pp.305–311,2006.c Springer-Verlag Berlin Heidelberg2006306M.Cardaci et al.These biases in the subjective time evaluation allow us to indirectly determine the visual complexity of stimuli,by comparing(under the same experimental conditions)their perceived durations.Is it possible to define a fuzzy model of complexity that encompasses both approaches,leading to the calculation of a true visual index of complexity?We have defined such a fuzzy model,based on local and global spatial fea-tures of the image and a definition of entropy.We chose an entropic measure because information theory and several branches of statistics have been proven to be powerful tools in quantifying the infinitesimal differences between two probability density functions.Entropic distances have been used successfully for image comparison and object matching problems in query by content applica-tions,showing their ability to grasp the pictorial visual content[3].For each image we computed its fuzzy index of complexity using the adopted model.The results from this fuzzy model were compared,using proven data analysis techniques,with the ones obtained by an experiment on subjective estimate of the perceived time,performed while the subjects were exposed to pictorial stimuli of increasing complexity.2The Entropic Model of Visual ComplexityTo create a mathematical model of the visual complexity based on spatial para-meters we have reviewed many of the local and global features from literature. Global features are suited to derive single values from the general properties of an image.Local features are needed to take into account classical verbal expla-nations for the meaning of complexity:many versus few,curved and/or detailed versus linear and planar,complex textures versusflat ing local features also helps reducing ambiguities in results.2.1Local Features ExtractionPoints of interest may be identified by using local operators.We chose two well-known local features:the image edges[4,5],and the local symmetries computed by the Discrete Symmetry Transform(DST)[6].Edge detectors highlight image zones with abrupt changes in luminosity level,associated with surface discon-tinuity.The rationale is straightforward:the more edges,the more objects(or the more surfaces),and a greater perceived complexity.DST extracts zones of the image in which the local gray levels show a high degree of radial symmetry (where the degree of locality depends on the radius of the local window used). It is interesting to note that points of interest detected by DST appear to be related with points to which shifts of gaze are directed performed by humans watching the same image.Apart from the natural attraction of symmetry,this also means that the more the points of interest in an image,the more complex the image is perceived as.More specifically,DST computes local symmetries of an image based on a measure of axial moments of a body around its center ofOn the Evaluation of Images Complexity:A Fuzzy Approach 307gravity.In the image case,the pixels inside a circular window are considered as point masses,with their mass expressed by their gray value g .Details of the algorithm can be found in [6]2.2An Entropic Measure of ComplexityWe are now interested in a global algorithm that can output a single value for each filtered images,while preserving its class of complexity.We decided to investigate the usefulness for this task of the fuzzy entropic distance func-tions detailed in [3].There are plenty of reasons for considering these functions among many others usually employed in this kind of task:first,a soft com-puting approach using fuzzy values seems appropriate when we are trying to describe a situation where binary logic is too strict.As for the entropic distance function,we can reformulate our main question from “How complex is this im-age?”to “What is the distance of this image from the simplest possible image in the defined feature space?”.This approach leads to the use of standard dis-tance functions,which respect the usual properties of identity,symmetry and triangular inequality,augmented by entropic functions.We chose the following functions:G 0(η)=−1log(2)×(ηlog(η)+(1−η)log(1−η).[7]z G 1(η)=2√e e −1 ηe 1−η−ηe η−1 ,G 2(η)=4η(1−η)where,η=1n n i =1|h i |,and h i are the gray levels of the image pixels normal-ized in the range [0,1].It can be easily shown that G j (j =0,1,2)satisfies the properties of a distance function,and it takes values in the interval [0,1].For each G j function,the first 15%of input values is mapped to more then half the range of output values.Most of the image would have a complexity index in the first quartile,so we expected to obtain a better classification through the expansion of exactly this range of input values.We considered three com-plexity classes of images,three different filters,and three entropy distance func-tions.In Table 1we show the mean value for each category.It is easy to see that all choices of functions and filters give results in line with our expecta-tions,with a slightly better performance obtained by the use of function G 2and the symmetry filter without thresholding.Note that the reported values of en-tropy are normalized in the interval [0,1]with the same procedure described in Section 3.Table 1.Results from filtering and applying entropy functions to all images in our pool,and then taking mean values for each category G 0G 1G 2I IIIII I II III I II III Edge detection 10.60010.39010.480Local Symmetry 10.45010.39010.490Thresholded Symmetry 10.42010.39010.340308M.Cardaci et al.3The Experiment on Perceived TimeWe devised an experiment to demonstrate that a subjective measure of the perceived time can be used as an indirect measure of the complexity of an image. We asked a number of volunteers to observe some images on a computer screen, and recorded their perceived duration of the observation.We run two sets of experiments,one in University of Surrey’s CVSSP group (50individuals)and the other in the Dipartimento di Psicologia dell’Universit`a di Palermo(15individuals).In order to minimize the cultural bias,all experi-mental subjects had university backgrounds.Participants were part of the staffand undergraduate students,on a volunteering basis,without any knowledge of our research’s aims.Privacy of the subjects was taken care of according to the Italian law on personal data;only initials,age and gender were recorded for each subject.Fig.1.Examples of test images,classified by intuitive complexity:high complexity (top);medium complexity(middle);low complexity(bottom)The experiments were held in a dim light room to reduce visual distraction, giving time to the participant for darkness adaptation.All the usual ergonomic precautions,such as using a quasi-soundproof room,were taken,and the subject was allowed to choose their own preferred position and visual angle.The images were presented full screen.The software used was home-made using the mul-timedia programming environment Macromedia Dreamweaver2004MX on an Apple Macintosh computer with a TFT LCD monitor.The chosen images were computer scans of paintings,divided in three categories representing different levels of visual complexity,based on the presence or absence of certain classes of features and cue points.Figure1shows examples of painting used in this study.Each image were presented for afixed period of time(90secs.),with no tempo-ral clues;the experiment also had a controlled design in order to minimize side effects:lights dimmed and uniform,subject alone in a soundproof room.The subject was alerted to focus their attention on the contents of the displayed im-ages.The images used for the experiments were chosen according to the intuitive hypothesis that the complexity of a scene increases with the number of objects and their relative position,and with its overall structure[8].The chosen images were paintings,divided in three categories representing different levels of visualOn the Evaluation of Images Complexity:A Fuzzy Approach309Table2.Mean and Normalized Time EstimationClass I Class II Class IIIˆµT61.7473.3885.15ˆσT33.0538.0941.14n T00.491complexity.Here the estimate time perceived by each subject is reported.We consider it as a subjective measures of complexity for the three categories of im-ages introduced above.In the following it will be denoted as{(t(i)T )|i=I,II,III}.The sample mean value and the variance of the perceived time(ˆµT,ˆσT)are re-ported in Table2.It is evident that the mean perceived time decreases with the complexity of the image.In order to compare the perceived time with the objec-tive measure of complexity,and to highlight our interest in relative differences between estimations made by the same subject when watching different images, the following normalized time measures are introduced:n(i) T =t(i)T−ˆµminˆµmax−ˆµminfor i=I,II,IIIwhereˆµmin=mini=I,II,III {ˆµ(i)T}andˆµmax=maxi=I,II,III{ˆµ(i)T},and0≤n(i)T≤1.The proposed normalization allows us a better comparison with the results obtained from the mathematical model,carried out in the next section.In this context,0and1have no strict numerical significance,but should be interpreted more like subjective degrees of complexity,which suits best with our fuzzy model.Results are in agreement with our model of time perception:complex images (category I)produce shorter time estimations than images in category II and the same is true for categories II and III.4Comparison of Measures and Data ValidationAs shown by comparing the entries of Tables1and2,our experimental data match those of the mathematical model.In fact,images with a high entropic complexity index generate,on average,a shorter estimation of the perceived time.Therefore,category I has the shortest evaluations and category III the longest.The strong anti-correlation between the entropic measure of complexity and the mental clock is shown in Figure2.The values of G s are the averages of those in Table1.We carried out a strict validation of the results using proven data analysis methods in order to ascertain the relation between data and model,minimizing the effects derived from the use of mean values and the cardinality of the dataset. To verify that a correlation between the experimental data and the mathematical results exists,we calculated the coefficient of correlation between the results310M.Cardaci et al.experimentFig.2.Anti-correlation between the estimate of the mental clock and the measure of complexity via entropy functionsfrom the experimental data and the complements of the entropic measures of complexity using Spearmans’ρ.Even in the worst case,the probability of data and model sequences being correlated is more than0.98.To confirm that the correlation is not due to the size of the data-set,we carried out many non-parametric bootstrap tests,using10,000virtual sets.In each test the difference between the mean obtained from the data and by the bootstrap method was under10−4.As we worked mostly with mean values,we also used the jackknife technique,re-calculating the results as many times as the number of images in our set,leaving out one image each time;all jackknife sets had the same distribution of values,with small numeric differences.5ConclusionsIn this paper we detailed an fuzzy model of visual complexity.This modelfits well with a subjective measure of complexity,based on perceived time.From a theoretical viewpoint we verified a strong correlation between spatial and tem-poral dimensions of complexity.Our results support the possibility to include the human information processing into the standard measure of visual complex-ity.Future work will be devoted to improve our experiment with more image categories.References1.M.Cardaci:The Mental Clock Model-Studies on the Structure of Time.Physicsto Psycho(patho)logy,(L.Buccheri,V.Di Ges`u,Saniga Eds.),Kluwer Acad-emic/Plenum Publishers,New York,2000.2.R.E.Ornstein:The Psychology of Consciousness.Freeman and Company,San Fran-cisco,1972.3.V.Di Ges`u,S.Roy:Fuzzy measures for image distance.Advances in Fuzzy Sys-tems and Intelligent Technologies(F.Masulli,R.Parenti,G.Pasi Eds.),Shaker Publishing(NL),2000.4.D.Marr,E.Hildreth:Theory of Edge Detection.Proc.R.Soc.Lond.B.,Vol.207,pp187–217,1980.On the Evaluation of Images Complexity:A Fuzzy Approach311 5.M.Petrou:The Differentiating Filter Approach to Edge Detection.Advances inElectronics and Electron Physics,Vol.88,pp297–345,1994.6.V.Di Ges`u,C.Valenti,“Symmetry operators in computer vision”,in Vistas inAstronomy,Pergamon,Vol.40,No4,pp461–468,1996.7.A.De Luca,S.Termini:Information and Control.Vol.20,p301,1972.8.A.Oliva,M.L.Mack,M.Shrestha,A.Peeper:Identifying the perceptual dimensionsof visual complexity of scenes.Proc.of the27th Annual Meeting of the Cognitive Science Society,Chicago,2004.。

美敦力起搏器使用寿命新篇章

美敦力起搏器使用寿命新篇章

预计使用寿命10.2年
延长 24 %
* Dual chamber, medium battery 60 bpm, 2,5V, 0,4 ms, 100 % pacing, 1000 Ohm
Improved longevities
7
起搏器寿命取决于哪些因素 ?
患者
起搏器
电池电量 起搏器诊断、治疗、EGM、 遥测、运算等功能 微处理芯片消耗电能
Improved longevities
5
Sigma/Kappa/EnPulse
预计使用寿命 8.2年
* Dual chamber, medium battery 60 bpm, 2,5V, 0,4 ms, 100 % pacing, 1000 Ohm
Improved longevities
6
Relia/Sensia/Versa/Adapta系列
25
既往产品的承诺与实践:
K700 系列 KDR701, 703, 706
Medtronic Product Performance Report 2009 1st edition Over 80 % lasted at least 7 years
Improved longevities
26
回顾7年样本量>5000台的起搏器存活率: MDT vs. SJM
12
最新的节电芯片
新的芯片功能优势有: 起搏器静态电流排放从 6.7uA减少到6.2uA 微处理器耗电量,从 9.14uA减少到 6.09uA 心室/心房起搏输出损 耗电流从1.0uA降到 0.06/0.3uA
起搏器的使用寿命显著延长了,但是并没有增加起搏器的体积
Improved longevities

Motion Estimation

Motion Estimation

. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
2 Introduction to Digital Video 2.1 Definitions and terminology . . . . . 2.1.1 Images . . . . . . . . . . . . . 2.1.2 Video sequences . . . . . . . 2.1.3 Video interlacing . . . . . . . 2.1.4 Contrast . . . . . . . . . . . . 2.1.5 Spatial frequency . . . . . . . 2.2 Digital image processing . . . . . . . 2.2.1 Fourier transform . . . . . . . 2.2.2 Convolution . . . . . . . . . . 2.2.3 Digital filters . . . . . . . . . 2.2.4 Correlation . . . . . . . . . . 2.3 MPEG-2 video compression . . . . . 2.3.1 The discrete cosine transform 2.3.2 Quantization . . . . . . . . . 2.3.3 Motion compensation . . . . 2.3.4 Efficient representation . . .
ii
c Philips Electronics N.V. 2001
no classification

热红外传感史

热红外传感史

History of infrared detectorsA.ROGALSKI*Institute of Applied Physics, Military University of Technology, 2 Kaliskiego Str.,00–908 Warsaw, PolandThis paper overviews the history of infrared detector materials starting with Herschel’s experiment with thermometer on February11th,1800.Infrared detectors are in general used to detect,image,and measure patterns of the thermal heat radia−tion which all objects emit.At the beginning,their development was connected with thermal detectors,such as ther−mocouples and bolometers,which are still used today and which are generally sensitive to all infrared wavelengths and op−erate at room temperature.The second kind of detectors,called the photon detectors,was mainly developed during the20th Century to improve sensitivity and response time.These detectors have been extensively developed since the1940’s.Lead sulphide(PbS)was the first practical IR detector with sensitivity to infrared wavelengths up to~3μm.After World War II infrared detector technology development was and continues to be primarily driven by military applications.Discovery of variable band gap HgCdTe ternary alloy by Lawson and co−workers in1959opened a new area in IR detector technology and has provided an unprecedented degree of freedom in infrared detector design.Many of these advances were transferred to IR astronomy from Departments of Defence ter on civilian applications of infrared technology are frequently called“dual−use technology applications.”One should point out the growing utilisation of IR technologies in the civilian sphere based on the use of new materials and technologies,as well as the noticeable price decrease in these high cost tech−nologies.In the last four decades different types of detectors are combined with electronic readouts to make detector focal plane arrays(FPAs).Development in FPA technology has revolutionized infrared imaging.Progress in integrated circuit design and fabrication techniques has resulted in continued rapid growth in the size and performance of these solid state arrays.Keywords:thermal and photon detectors, lead salt detectors, HgCdTe detectors, microbolometers, focal plane arrays.Contents1.Introduction2.Historical perspective3.Classification of infrared detectors3.1.Photon detectors3.2.Thermal detectors4.Post−War activity5.HgCdTe era6.Alternative material systems6.1.InSb and InGaAs6.2.GaAs/AlGaAs quantum well superlattices6.3.InAs/GaInSb strained layer superlattices6.4.Hg−based alternatives to HgCdTe7.New revolution in thermal detectors8.Focal plane arrays – revolution in imaging systems8.1.Cooled FPAs8.2.Uncooled FPAs8.3.Readiness level of LWIR detector technologies9.SummaryReferences 1.IntroductionLooking back over the past1000years we notice that infra−red radiation(IR)itself was unknown until212years ago when Herschel’s experiment with thermometer and prism was first reported.Frederick William Herschel(1738–1822) was born in Hanover,Germany but emigrated to Britain at age19,where he became well known as both a musician and an astronomer.Herschel became most famous for the discovery of Uranus in1781(the first new planet found since antiquity)in addition to two of its major moons,Tita−nia and Oberon.He also discovered two moons of Saturn and infrared radiation.Herschel is also known for the twenty−four symphonies that he composed.W.Herschel made another milestone discovery–discov−ery of infrared light on February11th,1800.He studied the spectrum of sunlight with a prism[see Fig.1in Ref.1],mea−suring temperature of each colour.The detector consisted of liquid in a glass thermometer with a specially blackened bulb to absorb radiation.Herschel built a crude monochromator that used a thermometer as a detector,so that he could mea−sure the distribution of energy in sunlight and found that the highest temperature was just beyond the red,what we now call the infrared(‘below the red’,from the Latin‘infra’–be−OPTO−ELECTRONICS REVIEW20(3),279–308DOI: 10.2478/s11772−012−0037−7*e−mail: rogan@.pllow)–see Fig.1(b)[2].In April 1800he reported it to the Royal Society as dark heat (Ref.1,pp.288–290):Here the thermometer No.1rose 7degrees,in 10minu−tes,by an exposure to the full red coloured rays.I drew back the stand,till the centre of the ball of No.1was just at the vanishing of the red colour,so that half its ball was within,and half without,the visible rays of theAnd here the thermometerin 16minutes,degrees,when its centre was inch out of the raysof the sun.as had a rising of 9de−grees,and here the difference is almost too trifling to suppose,that latter situation of the thermometer was much beyond the maximum of the heating power;while,at the same time,the experiment sufficiently indi−cates,that the place inquired after need not be looked for at a greater distance.Making further experiments on what Herschel called the ‘calorific rays’that existed beyond the red part of the spec−trum,he found that they were reflected,refracted,absorbed and transmitted just like visible light [1,3,4].The early history of IR was reviewed about 50years ago in three well−known monographs [5–7].Many historical information can be also found in four papers published by Barr [3,4,8,9]and in more recently published monograph [10].Table 1summarises the historical development of infrared physics and technology [11,12].2.Historical perspectiveFor thirty years following Herschel’s discovery,very little progress was made beyond establishing that the infrared ra−diation obeyed the simplest laws of optics.Slow progress inthe study of infrared was caused by the lack of sensitive and accurate detectors –the experimenters were handicapped by the ordinary thermometer.However,towards the second de−cade of the 19th century,Thomas Johann Seebeck began to examine the junction behaviour of electrically conductive materials.In 1821he discovered that a small electric current will flow in a closed circuit of two dissimilar metallic con−ductors,when their junctions are kept at different tempera−tures [13].During that time,most physicists thought that ra−diant heat and light were different phenomena,and the dis−covery of Seebeck indirectly contributed to a revival of the debate on the nature of heat.Due to small output vol−tage of Seebeck’s junctions,some μV/K,the measurement of very small temperature differences were prevented.In 1829L.Nobili made the first thermocouple and improved electrical thermometer based on the thermoelectric effect discovered by Seebeck in 1826.Four years later,M.Melloni introduced the idea of connecting several bismuth−copper thermocouples in series,generating a higher and,therefore,measurable output voltage.It was at least 40times more sensitive than the best thermometer available and could de−tect the heat from a person at a distance of 30ft [8].The out−put voltage of such a thermopile structure linearly increases with the number of connected thermocouples.An example of thermopile’s prototype invented by Nobili is shown in Fig.2(a).It consists of twelve large bismuth and antimony elements.The elements were placed upright in a brass ring secured to an adjustable support,and were screened by a wooden disk with a 15−mm central aperture.Incomplete version of the Nobili−Melloni thermopile originally fitted with the brass cone−shaped tubes to collect ra−diant heat is shown in Fig.2(b).This instrument was much more sensi−tive than the thermometers previously used and became the most widely used detector of IR radiation for the next half century.The third member of the trio,Langley’s bolometer appea−red in 1880[7].Samuel Pierpont Langley (1834–1906)used two thin ribbons of platinum foil connected so as to form two arms of a Wheatstone bridge (see Fig.3)[15].This instrument enabled him to study solar irradiance far into its infrared region and to measure theintensityof solar radia−tion at various wavelengths [9,16,17].The bolometer’s sen−History of infrared detectorsFig.1.Herschel’s first experiment:A,B –the small stand,1,2,3–the thermometers upon it,C,D –the prism at the window,E –the spec−trum thrown upon the table,so as to bring the last quarter of an inch of the read colour upon the stand (after Ref.1).InsideSir FrederickWilliam Herschel (1738–1822)measures infrared light from the sun– artist’s impression (after Ref. 2).Fig.2.The Nobili−Meloni thermopiles:(a)thermopile’s prototype invented by Nobili (ca.1829),(b)incomplete version of the Nobili−−Melloni thermopile (ca.1831).Museo Galileo –Institute and Museum of the History of Science,Piazza dei Giudici 1,50122Florence, Italy (after Ref. 14).Table 1. Milestones in the development of infrared physics and technology (up−dated after Refs. 11 and 12)Year Event1800Discovery of the existence of thermal radiation in the invisible beyond the red by W. HERSCHEL1821Discovery of the thermoelectric effects using an antimony−copper pair by T.J. SEEBECK1830Thermal element for thermal radiation measurement by L. NOBILI1833Thermopile consisting of 10 in−line Sb−Bi thermal pairs by L. NOBILI and M. MELLONI1834Discovery of the PELTIER effect on a current−fed pair of two different conductors by J.C. PELTIER1835Formulation of the hypothesis that light and electromagnetic radiation are of the same nature by A.M. AMPERE1839Solar absorption spectrum of the atmosphere and the role of water vapour by M. MELLONI1840Discovery of the three atmospheric windows by J. HERSCHEL (son of W. HERSCHEL)1857Harmonization of the three thermoelectric effects (SEEBECK, PELTIER, THOMSON) by W. THOMSON (Lord KELVIN)1859Relationship between absorption and emission by G. KIRCHHOFF1864Theory of electromagnetic radiation by J.C. MAXWELL1873Discovery of photoconductive effect in selenium by W. SMITH1876Discovery of photovoltaic effect in selenium (photopiles) by W.G. ADAMS and A.E. DAY1879Empirical relationship between radiation intensity and temperature of a blackbody by J. STEFAN1880Study of absorption characteristics of the atmosphere through a Pt bolometer resistance by S.P. LANGLEY1883Study of transmission characteristics of IR−transparent materials by M. MELLONI1884Thermodynamic derivation of the STEFAN law by L. BOLTZMANN1887Observation of photoelectric effect in the ultraviolet by H. HERTZ1890J. ELSTER and H. GEITEL constructed a photoemissive detector consisted of an alkali−metal cathode1894, 1900Derivation of the wavelength relation of blackbody radiation by J.W. RAYEIGH and W. WIEN1900Discovery of quantum properties of light by M. PLANCK1903Temperature measurements of stars and planets using IR radiometry and spectrometry by W.W. COBLENTZ1905 A. EINSTEIN established the theory of photoelectricity1911R. ROSLING made the first television image tube on the principle of cathode ray tubes constructed by F. Braun in 18971914Application of bolometers for the remote exploration of people and aircrafts ( a man at 200 m and a plane at 1000 m)1917T.W. CASE developed the first infrared photoconductor from substance composed of thallium and sulphur1923W. SCHOTTKY established the theory of dry rectifiers1925V.K. ZWORYKIN made a television image tube (kinescope) then between 1925 and 1933, the first electronic camera with the aid of converter tube (iconoscope)1928Proposal of the idea of the electro−optical converter (including the multistage one) by G. HOLST, J.H. DE BOER, M.C. TEVES, and C.F. VEENEMANS1929L.R. KOHLER made a converter tube with a photocathode (Ag/O/Cs) sensitive in the near infrared1930IR direction finders based on PbS quantum detectors in the wavelength range 1.5–3.0 μm for military applications (GUDDEN, GÖRLICH and KUTSCHER), increased range in World War II to 30 km for ships and 7 km for tanks (3–5 μm)1934First IR image converter1939Development of the first IR display unit in the United States (Sniperscope, Snooperscope)1941R.S. OHL observed the photovoltaic effect shown by a p−n junction in a silicon1942G. EASTMAN (Kodak) offered the first film sensitive to the infrared1947Pneumatically acting, high−detectivity radiation detector by M.J.E. GOLAY1954First imaging cameras based on thermopiles (exposure time of 20 min per image) and on bolometers (4 min)1955Mass production start of IR seeker heads for IR guided rockets in the US (PbS and PbTe detectors, later InSb detectors for Sidewinder rockets)1957Discovery of HgCdTe ternary alloy as infrared detector material by W.D. LAWSON, S. NELSON, and A.S. YOUNG1961Discovery of extrinsic Ge:Hg and its application (linear array) in the first LWIR FLIR systems1965Mass production start of IR cameras for civil applications in Sweden (single−element sensors with optomechanical scanner: AGA Thermografiesystem 660)1970Discovery of charge−couple device (CCD) by W.S. BOYLE and G.E. SMITH1970Production start of IR sensor arrays (monolithic Si−arrays: R.A. SOREF 1968; IR−CCD: 1970; SCHOTTKY diode arrays: F.D.SHEPHERD and A.C. YANG 1973; IR−CMOS: 1980; SPRITE: T. ELIOTT 1981)1975Lunch of national programmes for making spatially high resolution observation systems in the infrared from multielement detectors integrated in a mini cooler (so−called first generation systems): common module (CM) in the United States, thermal imaging commonmodule (TICM) in Great Britain, syteme modulaire termique (SMT) in France1975First In bump hybrid infrared focal plane array1977Discovery of the broken−gap type−II InAs/GaSb superlattices by G.A. SAI−HALASZ, R. TSU, and L. ESAKI1980Development and production of second generation systems [cameras fitted with hybrid HgCdTe(InSb)/Si(readout) FPAs].First demonstration of two−colour back−to−back SWIR GaInAsP detector by J.C. CAMPBELL, A.G. DENTAI, T.P. LEE,and C.A. BURRUS1985Development and mass production of cameras fitted with Schottky diode FPAs (platinum silicide)1990Development and production of quantum well infrared photoconductor (QWIP) hybrid second generation systems1995Production start of IR cameras with uncooled FPAs (focal plane arrays; microbolometer−based and pyroelectric)2000Development and production of third generation infrared systemssitivity was much greater than that of contemporary thermo−piles which were little improved since their use by Melloni. Langley continued to develop his bolometer for the next20 years(400times more sensitive than his first efforts).His latest bolometer could detect the heat from a cow at a dis−tance of quarter of mile [9].From the above information results that at the beginning the development of the IR detectors was connected with ther−mal detectors.The first photon effect,photoconductive ef−fect,was discovered by Smith in1873when he experimented with selenium as an insulator for submarine cables[18].This discovery provided a fertile field of investigation for several decades,though most of the efforts were of doubtful quality. By1927,over1500articles and100patents were listed on photosensitive selenium[19].It should be mentioned that the literature of the early1900’s shows increasing interest in the application of infrared as solution to numerous problems[7].A special contribution of William Coblenz(1873–1962)to infrared radiometry and spectroscopy is marked by huge bib−liography containing hundreds of scientific publications, talks,and abstracts to his credit[20,21].In1915,W.Cob−lentz at the US National Bureau of Standards develops ther−mopile detectors,which he uses to measure the infrared radi−ation from110stars.However,the low sensitivity of early in−frared instruments prevented the detection of other near−IR sources.Work in infrared astronomy remained at a low level until breakthroughs in the development of new,sensitive infrared detectors were achieved in the late1950’s.The principle of photoemission was first demonstrated in1887when Hertz discovered that negatively charged par−ticles were emitted from a conductor if it was irradiated with ultraviolet[22].Further studies revealed that this effect could be produced with visible radiation using an alkali metal electrode [23].Rectifying properties of semiconductor−metal contact were discovered by Ferdinand Braun in1874[24],when he probed a naturally−occurring lead sulphide(galena)crystal with the point of a thin metal wire and noted that current flowed freely in one direction only.Next,Jagadis Chandra Bose demonstrated the use of galena−metal point contact to detect millimetre electromagnetic waves.In1901he filed a U.S patent for a point−contact semiconductor rectifier for detecting radio signals[25].This type of contact called cat’s whisker detector(sometimes also as crystal detector)played serious role in the initial phase of radio development.How−ever,this contact was not used in a radiation detector for the next several decades.Although crystal rectifiers allowed to fabricate simple radio sets,however,by the mid−1920s the predictable performance of vacuum−tubes replaced them in most radio applications.The period between World Wars I and II is marked by the development of photon detectors and image converters and by emergence of infrared spectroscopy as one of the key analytical techniques available to chemists.The image con−verter,developed on the eve of World War II,was of tre−mendous interest to the military because it enabled man to see in the dark.The first IR photoconductor was developed by Theodore W.Case in1917[26].He discovered that a substance com−posed of thallium and sulphur(Tl2S)exhibited photocon−ductivity.Supported by the US Army between1917and 1918,Case adapted these relatively unreliable detectors for use as sensors in an infrared signalling device[27].The pro−totype signalling system,consisting of a60−inch diameter searchlight as the source of radiation and a thallous sulphide detector at the focus of a24−inch diameter paraboloid mir−ror,sent messages18miles through what was described as ‘smoky atmosphere’in1917.However,instability of resis−tance in the presence of light or polarizing voltage,loss of responsivity due to over−exposure to light,high noise,slug−gish response and lack of reproducibility seemed to be inhe−rent weaknesses.Work was discontinued in1918;commu−nication by the detection of infrared radiation appeared dis−tinctly ter Case found that the addition of oxygen greatly enhanced the response [28].The idea of the electro−optical converter,including the multistage one,was proposed by Holst et al.in1928[29]. The first attempt to make the converter was not successful.A working tube consisted of a photocathode in close proxi−mity to a fluorescent screen was made by the authors in 1934 in Philips firm.In about1930,the appearance of the Cs−O−Ag photo−tube,with stable characteristics,to great extent discouraged further development of photoconductive cells until about 1940.The Cs−O−Ag photocathode(also called S−1)elabo−History of infrared detectorsFig.3.Longley’s bolometer(a)composed of two sets of thin plati−num strips(b),a Wheatstone bridge,a battery,and a galvanometer measuring electrical current (after Ref. 15 and 16).rated by Koller and Campbell[30]had a quantum efficiency two orders of magnitude above anything previously studied, and consequently a new era in photoemissive devices was inaugurated[31].In the same year,the Japanese scientists S. Asao and M.Suzuki reported a method for enhancing the sensitivity of silver in the S−1photocathode[32].Consisted of a layer of caesium on oxidized silver,S−1is sensitive with useful response in the near infrared,out to approxi−mately1.2μm,and the visible and ultraviolet region,down to0.3μm.Probably the most significant IR development in the United States during1930’s was the Radio Corporation of America(RCA)IR image tube.During World War II, near−IR(NIR)cathodes were coupled to visible phosphors to provide a NIR image converter.With the establishment of the National Defence Research Committee,the develop−ment of this tube was accelerated.In1942,the tube went into production as the RCA1P25image converter(see Fig.4).This was one of the tubes used during World War II as a part of the”Snooperscope”and”Sniperscope,”which were used for night observation with infrared sources of illumination.Since then various photocathodes have been developed including bialkali photocathodes for the visible region,multialkali photocathodes with high sensitivity ex−tending to the infrared region and alkali halide photocatho−des intended for ultraviolet detection.The early concepts of image intensification were not basically different from those today.However,the early devices suffered from two major deficiencies:poor photo−cathodes and poor ter development of both cathode and coupling technologies changed the image in−tensifier into much more useful device.The concept of image intensification by cascading stages was suggested independently by number of workers.In Great Britain,the work was directed toward proximity focused tubes,while in the United State and in Germany–to electrostatically focused tubes.A history of night vision imaging devices is given by Biberman and Sendall in monograph Electro−Opti−cal Imaging:System Performance and Modelling,SPIE Press,2000[10].The Biberman’s monograph describes the basic trends of infrared optoelectronics development in the USA,Great Britain,France,and Germany.Seven years later Ponomarenko and Filachev completed this monograph writ−ing the book Infrared Techniques and Electro−Optics in Russia:A History1946−2006,SPIE Press,about achieve−ments of IR techniques and electrooptics in the former USSR and Russia [33].In the early1930’s,interest in improved detectors began in Germany[27,34,35].In1933,Edgar W.Kutzscher at the University of Berlin,discovered that lead sulphide(from natural galena found in Sardinia)was photoconductive and had response to about3μm.B.Gudden at the University of Prague used evaporation techniques to develop sensitive PbS films.Work directed by Kutzscher,initially at the Uni−versity of Berlin and later at the Electroacustic Company in Kiel,dealt primarily with the chemical deposition approach to film formation.This work ultimately lead to the fabrica−tion of the most sensitive German detectors.These works were,of course,done under great secrecy and the results were not generally known until after1945.Lead sulphide photoconductors were brought to the manufacturing stage of development in Germany in about1943.Lead sulphide was the first practical infrared detector deployed in a variety of applications during the war.The most notable was the Kiel IV,an airborne IR system that had excellent range and which was produced at Carl Zeiss in Jena under the direction of Werner K. Weihe [6].In1941,Robert J.Cashman improved the technology of thallous sulphide detectors,which led to successful produc−tion[36,37].Cashman,after success with thallous sulphide detectors,concentrated his efforts on lead sulphide detec−tors,which were first produced in the United States at Northwestern University in1944.After World War II Cash−man found that other semiconductors of the lead salt family (PbSe and PbTe)showed promise as infrared detectors[38]. The early detector cells manufactured by Cashman are shown in Fig. 5.Fig.4.The original1P25image converter tube developed by the RCA(a).This device measures115×38mm overall and has7pins.It opera−tion is indicated by the schematic drawing (b).After1945,the wide−ranging German trajectory of research was essentially the direction continued in the USA, Great Britain and Soviet Union under military sponsorship after the war[27,39].Kutzscher’s facilities were captured by the Russians,thus providing the basis for early Soviet detector development.From1946,detector technology was rapidly disseminated to firms such as Mullard Ltd.in Southampton,UK,as part of war reparations,and some−times was accompanied by the valuable tacit knowledge of technical experts.E.W.Kutzscher,for example,was flown to Britain from Kiel after the war,and subsequently had an important influence on American developments when he joined Lockheed Aircraft Co.in Burbank,California as a research scientist.Although the fabrication methods developed for lead salt photoconductors was usually not completely under−stood,their properties are well established and reproducibi−lity could only be achieved after following well−tried reci−pes.Unlike most other semiconductor IR detectors,lead salt photoconductive materials are used in the form of polycrys−talline films approximately1μm thick and with individual crystallites ranging in size from approximately0.1–1.0μm. They are usually prepared by chemical deposition using empirical recipes,which generally yields better uniformity of response and more stable results than the evaporative methods.In order to obtain high−performance detectors, lead chalcogenide films need to be sensitized by oxidation. The oxidation may be carried out by using additives in the deposition bath,by post−deposition heat treatment in the presence of oxygen,or by chemical oxidation of the film. The effect of the oxidant is to introduce sensitizing centres and additional states into the bandgap and thereby increase the lifetime of the photoexcited holes in the p−type material.3.Classification of infrared detectorsObserving a history of the development of the IR detector technology after World War II,many materials have been investigated.A simple theorem,after Norton[40],can be stated:”All physical phenomena in the range of about0.1–1 eV will be proposed for IR detectors”.Among these effects are:thermoelectric power(thermocouples),change in elec−trical conductivity(bolometers),gas expansion(Golay cell), pyroelectricity(pyroelectric detectors),photon drag,Jose−phson effect(Josephson junctions,SQUIDs),internal emis−sion(PtSi Schottky barriers),fundamental absorption(in−trinsic photodetectors),impurity absorption(extrinsic pho−todetectors),low dimensional solids[superlattice(SL), quantum well(QW)and quantum dot(QD)detectors], different type of phase transitions, etc.Figure6gives approximate dates of significant develop−ment efforts for the materials mentioned.The years during World War II saw the origins of modern IR detector tech−nology.Recent success in applying infrared technology to remote sensing problems has been made possible by the successful development of high−performance infrared de−tectors over the last six decades.Photon IR technology com−bined with semiconductor material science,photolithogra−phy technology developed for integrated circuits,and the impetus of Cold War military preparedness have propelled extraordinary advances in IR capabilities within a short time period during the last century [41].The majority of optical detectors can be classified in two broad categories:photon detectors(also called quantum detectors) and thermal detectors.3.1.Photon detectorsIn photon detectors the radiation is absorbed within the material by interaction with electrons either bound to lattice atoms or to impurity atoms or with free electrons.The observed electrical output signal results from the changed electronic energy distribution.The photon detectors show a selective wavelength dependence of response per unit incident radiation power(see Fig.8).They exhibit both a good signal−to−noise performance and a very fast res−ponse.But to achieve this,the photon IR detectors require cryogenic cooling.This is necessary to prevent the thermalHistory of infrared detectorsFig.5.Cashman’s detector cells:(a)Tl2S cell(ca.1943):a grid of two intermeshing comb−line sets of conducting paths were first pro−vided and next the T2S was evaporated over the grid structure;(b) PbS cell(ca.1945)the PbS layer was evaporated on the wall of the tube on which electrical leads had been drawn with aquadag(afterRef. 38).。

17.On the “Tulip Flame” Phenomenon

17.On the “Tulip Flame” Phenomenon
On the “Tulip Flame” Phenomenon
CHRISTOPHE CLANET and GEOFFREY
SEARBY’
Insitut de Recherche sur les Ph&omLnes Hors Equilibre, Laboratoire de Combustion et SystZmes Riactifs, U.M.R 138 du CNRS - Vniversit& d’ Aix-Marseille I et II, Service 252, Campus Universitaire de St. J@me, 13397 Marseille Cedex 20, France
axial coordinate of tip of flame axial coordinate of lower edge of skirt zskirt of flame Z wall axial position of flame tip when the flame skirt touches burner wall axial position at which tulip inversion Gulip occurs (flat front)
INTRODUCTION The propagation
* Corresponding author.
COMBUSTIONAND FLAME 105: 225-238 (1996)
of flames in tubes was first observed by Mallard and Le Chatelier in 1883 [l]. They noticed that if a flame is ignited at the closed end of a long half-open tube, the

法布里珀罗基模共振英文

法布里珀罗基模共振英文

法布里珀罗基模共振英文The Fabryperot ResonanceOptics, the study of light and its properties, has been a subject of fascination for scientists and researchers for centuries. One of the fundamental phenomena in optics is the Fabry-Perot resonance, named after the French physicists Charles Fabry and Alfred Perot, who first described it in the late 19th century. This resonance effect has numerous applications in various fields, ranging from telecommunications to quantum physics, and its understanding is crucial in the development of advanced optical technologies.The Fabry-Perot resonance occurs when light is reflected multiple times between two parallel, partially reflective surfaces, known as mirrors. This creates a standing wave pattern within the cavity formed by the mirrors, where the light waves interfere constructively and destructively to produce a series of sharp peaks and valleys in the transmitted and reflected light intensity. The specific wavelengths at which the constructive interference occurs are known as the resonant wavelengths of the Fabry-Perot cavity.The resonant wavelengths of a Fabry-Perot cavity are determined bythe distance between the mirrors, the refractive index of the material within the cavity, and the wavelength of the incident light. When the optical path length, which is the product of the refractive index and the physical distance between the mirrors, is an integer multiple of the wavelength of the incident light, the light waves interfere constructively, resulting in a high-intensity transmission through the cavity. Conversely, when the optical path length is not an integer multiple of the wavelength, the light waves interfere destructively, leading to a low-intensity transmission.The sharpness of the resonant peaks in a Fabry-Perot cavity is determined by the reflectivity of the mirrors. Highly reflective mirrors result in a higher finesse, which is a measure of the ratio of the spacing between the resonant peaks to their width. This high finesse allows for the creation of narrow-linewidth, high-resolution optical filters and laser cavities, which are essential components in various optical systems.One of the key applications of the Fabry-Perot resonance is in the field of optical telecommunications. Fiber-optic communication systems often utilize Fabry-Perot filters to select specific wavelength channels for data transmission, enabling the efficient use of the available bandwidth in fiber-optic networks. These filters can be tuned by adjusting the mirror separation or the refractive index of the cavity, allowing for dynamic wavelength selection andreconfiguration of the communication system.Another important application of the Fabry-Perot resonance is in the field of laser technology. Fabry-Perot cavities are commonly used as the optical resonator in various types of lasers, providing the necessary feedback to sustain the lasing process. The high finesse of the Fabry-Perot cavity allows for the generation of highly monochromatic and coherent light, which is crucial for applications such as spectroscopy, interferometry, and precision metrology.In the realm of quantum physics, the Fabry-Perot resonance plays a crucial role in the study of cavity quantum electrodynamics (cQED). In cQED, atoms or other quantum systems are placed inside a Fabry-Perot cavity, where the strong interaction between the atoms and the confined electromagnetic field can lead to the observation of fascinating quantum phenomena, such as the Purcell effect, vacuum Rabi oscillations, and the generation of nonclassical states of light.Furthermore, the Fabry-Perot resonance has found applications in the field of optical sensing, where it is used to detect small changes in physical parameters, such as displacement, pressure, or temperature. The high sensitivity and stability of Fabry-Perot interferometers make them valuable tools in various sensing and measurement applications, ranging from seismic monitoring to the detection of gravitational waves.The Fabry-Perot resonance is a fundamental concept in optics that has enabled the development of numerous advanced optical technologies. Its versatility and importance in various fields of science and engineering have made it a subject of continuous research and innovation. As the field of optics continues to advance, the Fabry-Perot resonance will undoubtedly play an increasingly crucial role in shaping the future of optical systems and applications.。

高等教育:《第2节解的延拓(english)》

高等教育:《第2节解的延拓(english)》

of (1) defined on an interval J which properly contains I and z
restricted to I equals y .
A solution is non-continuable or saturated if no such extension exists; i.e., I is the maximal interval on which a solution to (1) exists.
Consider the differential equation
dy f (t, y)
(1)
dt
If y(t) is a solution of (1) defined on an interval I, we say that
z(t) is a continuation or extension of y(t) if z(t) is itself a solution
Q: When can a given solution be continued?
Local Lipschitz condition
Definition A function f(t, y) (where U is an open set of R×R) satisfies a local Lipschitz condition if for any (t0 , y0 ) U there exist a neighborhood (t0 , y0 ) V U such that f satisfies a Lipschitz condition on V.
Remark: If the function f is of class C1 in U, then it satisfies a local Lipschitz condition.

Bistability and stationary gap solitons in quasiperiodic photonic crystals based on Thue–Morse sequ

Bistability and stationary gap solitons in quasiperiodic photonic crystals based on Thue–Morse sequ

Bistability and stationary gap solitons in quasiperiodic photonic crystals based on Thue–Morse sequenceV .V .Grigoriev *,F.BiancalanaNonlinear Photonic Nanostructures Research Group,Max Planck Institute for the Science of Light,Gu¨nther-Scharowsky-Str.1/Bau 26,91058Erlangen,Germany Received 15January 2010;received in revised form 5May 2010;accepted 5May 2010Available online 21May 2010AbstractThe nonlinear properties of quasiperiodic photonic crystals based on the Thue–Morse sequence are investigated.The intrinsic asymmetry of these one-dimensional structures for odd generation numbers results in bistability thresholds which are sensitive to propagation direction.Along with resonances of perfect transmission,this feature allows to obtain strongly nonreciprocal propagation and to create an all-optical diode.The efficiency of two schemes is compared:passive and active when an additional short pump signal is applied to the system.The existence of stationary gap solitons in quasiperiodic photonic crystals is shown numerically,and their difference from the Bragg case is emphasized.#2010Elsevier B.V .All rights reserved.Keywords:Quasicrystals;Multilayers;Bistability;Gap solitons1.IntroductionQuasiperiodic photonic crystals (QPCs)are deter-ministically generated dielectric structures with non-periodic modulation of refractive index.They represent an intermediate stage between random media and traditional periodic photonic crystals,effectively combining both localization properties as a result of short-range disorder and the presence of band gaps due to long-range correlations [1].In the one-dimensional case,QPCs can be formed by stacking together dielectric layers of several different types according to substitutional sequences (Fibonacci,Thue–Morse,Rudin–Shapiro,Cantor,etc.).The Fibo-nacci sequence is of particular importance,since it leads to the existence of two incommensurable periods in thespatial spectrum of the structure.This property is very valuable for satisfying the phase-matching condition in nonlinear optics and especially for the process of third harmonic generation [2].However,there are many other classes of quasicrystals,which can prove to be useful for potential applications.In this work,the nonlinear properties of the Thue–Morse (ThM)sequence are considered,with emphasis on bistability and all-optical switching.We show how to apply them to achieve a strongly nonreciprocal propagation and to create an all-optical diode.2.Thue –Morse sequenceTwo types of arbitrary dielectric layers are required to fabricate ThM quasicrystals.Indicated with letters A and B,they should be arranged in the same way as in the literal ThM sequence,which is governed by the following inflation rules:A !AB,B !BA.Thus,/locate/photonicsAvailable online at Photonics and Nanostructures –Fundamentals and Applications 8(2010)285–290*Corresponding author.E-mail address:victor.grigoriev@mpl.mpg.de (V .V .Grigoriev).1569-4410/$–see front matter #2010Elsevier B.V .All rights reserved.doi:10.1016/j.photonics.2010.05.002starting from a single layer S0¼A,which is defined to be the ThM quasicrystal of the0th generation,one obtains S1¼AB;S2¼ABBA;S3¼ABBABAAB and so forth,with each step giving sequence of generation number increased by one(Fig.1).It can be shown that an additional recurrence relation follows from this definition,which holds for ThM sequences as single blocks:S nþ1¼S n˜S n.In this notation˜S n means sequence conjugated to S n,where all letters are interchanged as in the rule A$B.The linear transmission spectrum of such photonic structures was shown to have a number of remarkable properties[3–7].Firstly,it demonstrates the fractal nature of transmission,which becomes apparent from self-similarity and characteristic trifurcation of resonances as generation number increases.Secondly,and more importantly for switching applications is that almost all of these resonances are of perfect(100%)transmis-sion,irrespective of the generation number(Fig.2).It is interesting to note that there is only a slight difference between the ThM and Bragg(periodic)sequences as faras inflation rules are ly,for the latter case they take the form:A!AB,B!AB,or being written as a recurrence relation:S nþ1¼S n S n.However, this slight difference leads to completely distinct transmission and localization properties.3.Bistability and folding of resonancesThe transfer matrix method was applied to calculate transmission spectra andfield profiles at resonant frequencies.The level of self-similarity in thefield profiles depends on the generation number when a particular resonancefirst appears and is maximal for normalized frequencies which belong to the following series:V=V0¼f1Às;1;1þs;2;3Às;3;3þs;...g, where s¼0:512082for the specific parameters used in Fig.1[8].However,the localization strength is relatively small at these frequencies,and it means that nonlinear response cannot be enhanced significantly. For this purpose,resonances located near the edges of pseudo band gaps are more suitable.The case of Kerr(cubic)nonlinearity was consid-ered,so that the nonlinear refractive index was additionally taken into account for each type of the layers.This leads to intensity-dependent self-phase modulation,which is able to shift resonance frequen-cies.The direction of this shift is determined by the sign of the Kerr coefficient,and it is possible tofind such resonances in the spectrum which bend into band gap regions(Fig.3).The advantage that ThM sequence gives is that for odd generation numbers the corresponding photonic structures are intrinsically asymmetric,and nonlinearity is capable of making transmission sensitive to the propagation direction[9].This feature is completely absent in nonlinear Bragg structures,where hysteresis curves are the same for the leftward and rightward incidence.Although a similar nonreciprocal behavior can be achieved in the framework of linear optics,it requires making use of magneto-optical media withV.V.Grigoriev,F.Biancalana/Photonics and Nanostructures–Fundamentals and Applications8(2010)285–290286Fig.1.(a)Application of inflation rules A!AB,B!BA for ThM sequence.Numbers near the notches designate the right boundaries of the sequences with corresponding generation numbers.(b)Field profile at one of the resonance frequencies(V=V0¼0:756042)of a ThM7quasicrystal,which demonstrates the self-similar pattern of ThM3sequence.The frequency V0corresponds to quarter wavelength condition taken at0:7m m.The linear refractive indices of the materi-als used are n A¼1:5(polydiacetylene9-BCMU)and n B¼2:3(rutile TiO2).Background shows alternation of layers inside thestructure.Fig.2.(a)Linear transmission spectrum as a function of normalized frequency V=V0for ThM7quasicrystal with the same set of para-meters as in Fig.1.Overlapped diagram shows trifurcation of resonances as generation number increases from3to7(see vertical axis on the right).(b)The same spectrum plotted in logarithmic scale, which emphasizes the location of band gap regions and their relative strength.externally applied static magneticfields[10]or chiral media such as cholesteric liquid crystals[11].The level of self-similarity in thefield profiles is also related to the type of hysteresis that can be observed near a particular resonance[8].If there exist several independent localization centers inside the structure, the interplay between them gives rise to multistability.It is related mainly to the fact that ThM structures of higher generation numbers can be decomposed into those of lower ones.Therefore,to single out a particular hysteresis in its pure form,it is necessary to check that the corresponding resonance does not occur in the previous generations.Two characteristic examples are shown in Fig.3a and b.Thefirst case(Fig.3a) corresponds to a resonance with no self-similarity in the field profile,whereas in the second case(Fig.3b)two independent localization centers can be distinguished. Notice that similar to the Bragg structures,the maxima of electricfield in both cases can be located mostly in the layers of one type,so that it is sufficient to use only one nonlinear material.Since typical nonlinear changes of refractive index are very small,it is possible to apply perturbation theory to derive the formulaT¼11þðdþI out=I resÞ2(1)which defines transmission T around some resonant frequency as a function of output intensity I out and detuning from this frequency d¼ðvÀv resÞ=g res. The parameters necessary to specify resonance are its frequency v res,spectral width g res,and characteristic intensity I res.In the case of single Dirac-delta-like nonlinear layer in the system,this formula coincides with a previously investigated exact analytical solution[12].However, taking into account distributed nonlinearity gives rise to a new feature:if the structure is spatially asymmetric, the characteristic intensity I res can depend on the direction of incidence[13,14].Since this formula is equivalent to the polynomial of the third power,it fails by definition to describe resonances with multistability, but it gives a good approximation for resonances with strictly bistable response(Fig.3c)and for small input intensities(Fig.3d).As the intensity increases,the agreement becomes worse,and the resonances not only undergo a shift in frequency,but also their maximal transmission lowers(Fig.3e).4.Optical diode actionThe difference in bistability thresholds can be small, but it creates favorable conditions for unidirectional propagation,so that such structures can be used as all-optical diodes.Similar to electronic circuits,these devices are indispensable,when there is a need to suppress theflow of light in one direction or to avoid problems caused by unwanted reflections.Various types of optical diodes have been proposed and realized, which can work both in linear[10,11]and nonlinear regime[15–18].The primaryfigure of merit,which determines the efficiency of this device,is the contrastV.V.Grigoriev,F.Biancalana/Photonics and Nanostructures–Fundamentals and Applications8(2010)285–290287 Fig.3.(a and b)Electricfield profiles at several resonance frequencies(V=V0¼0:739780and V=V0¼0:809977,respectively)for linear ThM7 structure with the same set of parameters as in Fig.1.Background shows alternation of layers inside the structure.(c)Hysteresis of transmission at normalized frequency V=V0¼0:7375.The nonlinear Kerr coefficients used are n2A¼2:5Â10À5cm2=MW and n2B¼1:0Â10À8cm2=MW.The switching thresholds are different for the forward(blue solid)and backward(red dotted)incidence.Vertical grid lines correspond to the intensities of up(I up¼22:0MW=cm2)and down(I down¼8:1MW=cm2)transitions for the case of forward incidence.(d and e)Nonlinear transmission spectra forfixed input intensities I down and I up,respectively.In addition to the nonlinear transmission spectra for the forward(blue solid)and backward(red dotted)incidence,the linear transmission spectrum(gray solid)is shown.(For interpretation of the references to color in thisfigure legend,the reader is referred to the web version of this article.)ratio C¼T f=T b between transmission along forward T f and backward directions T b.It can be very large in ThM quasicrystals due to the fact that this device can operate on two different hysteresis branches depending on the propagation direction.It is possible to use the intensities of both up I up and down transitions I down to achieve strongly nonreciprocal transmission.In the former case,the scheme works in passive mode,but the maximum value of transmission is limited[14].In the latter case,the transmission can be almost perfect,but the scheme requires an additional short pump signal in order to switch to the higher stable branch of hysteresis[19].The FDTD method was applied to demonstrate the switching dynamics in time domain(Fig.4).To maintain the second order accuracy of the Yee-scheme in case of multilayered structures,nonuniform spatial mesh was used with electricfield nodes aligned to the boundaries between layers.Since the structure can accumulate and release energy,the sum of reflected and transmitted intensities should not necessarily give the incident one.This constraint becomes valid only after the transient dynamics is over.The passive scheme was investigatedfirst.In this case the input intensity should be set to a value which is sufficiently large for the up switching in the forward direction I up¼22:0MW=cm2,but at the same time smaller than the corresponding value for the backward direction(Fig.3c).However,these thresholds are meaningful only for the signals the intensity of which varies adiabatically in time.The steepness of the signal can be the second factor which determines whether the switching occurs or not.In fact,it is advantageous to use steep signals in passive scheme not only because it increases the operating speed of the device but also because it forces diode to work at a more favorable point of hysteresis curve.We found that it is possible to observe switching even at20MW=cm2provided that the switching duration of the incident signal(when it changes from zero to a constant intensity)is shortened to0.5ps.The maximum transmission obtained was T f¼57:5%with the contrast ratio C¼8:8,which is quite close to the theoretical limit of C max¼9 following from the Eq.(1)in the assumption of passive scheme[14].To demonstrate the work of diode in active scheme, the input intensity was set just above the down switching threshold I down¼8:1MW=cm2for the forward direction,which ensures transmission of almost 100%in this direction,while in the opposite direction transmission will be strongly suppressed due to pseudo band gap.The contrast ratio achieved was C¼23:7,but it is not limited in principle.The major drawback of this scheme is that it is necessary to force system to switch to the upper branch of hysteresis whether by applying an auxiliary pump signal,or by temporarily increasing the input intensity.5.Stationary gap solitonsBesides folding of resonances,Kerr nonlinearity can lead also to the existence of solitons.The large group velocity dispersion inherent to multilayered structures even if they are composed of dispersionless media can be compensated by Kerr nonlinearity preventing the broadening of the signals[20].The properties of theseV.V.Grigoriev,F.Biancalana/Photonics and Nanostructures–Fundamentals and Applications8(2010)285–290288Fig.4.Numerical simulations of nonlinear ThM7structure acting as an optical diode at normalized frequency V=V0¼0:7375for the passive(a and b)and active(c and d)schemes.In the latter case an additional short pump signal is applied to facilitate switching.The intensities of incident(gray solid),reflected(red dotted),and transmitted(blue dashed)signals are given as a function of time.Upper part of thefigure refers to the forward incidence(a and c),and the lower one corresponds to the backward incidence(b and d).(For interpretation of the references to color in thisfigure legend,the reader is referred to the web version of this article.)objects have been studied in detail for the case of Bragg structures,and it was shown that they can exist not only in the pass bands(Bragg solitons)but also in the band gaps(gap solitons).The incident radiation coupled to gap solitons could be conveyed through the structure with no changes in the intensity even in the very center of band gap[21,22].A number of attempts has been made tofind similar gap solitons in quasiperiodic lattices.It was shown that quasiperiodic solitons do exist in optical lattices with refractive index profile described by two cosine functions of incommensurable period[23].Another particular case is to consider quasiperiodic structures, where nonlinearity is modeled by delta-functions,or so-called nonlinear Dirac comb lattices.As to structures withfinite width of nonlinear layers(in longitudinal direction),gap solitons were found only for Fibonacci sequence[24].However,a significant simplification was used in that work,because only several dominant peaks of the spatial Fourier spectrum were taken into account.In particular,this assumption cannot be applied to ThM sequence,which is characterized by singular-continuous spatial spectrum.On performing a scan over pseudo band gaps of ThM structure,several families of stationary gap solitons were identified(Fig.5).They look like L-shape curves on the contour plot and have similar distribution of electricfield at all points.In contrast to the Bragg case, these stationary gap solitons are not related to any linear resonances,although they exist throughout the full band gap region,making a kind of bridge from one pass band to another.Moreover,theirfield profiles inherit the quasiperiodicity of the underlying structure to a certain extent.The optimal conditions for the existence of these solitons are when nonlinearity effectivelyflattens the contrast in refractive index between the layers.In other words,material with lower(higher)linear refractive index should have positive(negative)Kerr coefficient.From numerical point of view,it should be noted that the nonlinear transfer matrix method is not efficient to handle steep modulations of refractive index typical for gap solitons.It is preferably to solve Maxwell’s equations inside each nonlinear layer directly by rewriting them as a system of ordinary differential equations and applying adaptive Runge–Kutta method or similar routine for solving initial value problem.6.ConclusionsThe nonlinear properties of QPCs based on ThM sequence were investigated.It was shown that the interplay between spatial asymmetry of these structures for odd generation numbers and Kerr nonlinearity can make switching thresholds sensitive to propagation direction.The role of self-similarity was emphasized to explain the shape of hysteresis curves observed near particular resonance,and conditions necessary to achieve highly nonreciprocal propagation were for-mulated.The FDTD simulation was used to confirm the results obtained by nonlinear transfer method and to show how ThM structure can work as optical diode both in passive and active mode.The existence of stationaryV.V.Grigoriev,F.Biancalana/Photonics and Nanostructures–Fundamentals and Applications8(2010)285–290289 Fig.5.Nonlinear transmission as a function of normalized frequency V=V0and amplitude of transmitted wave j E out j for Bragg7(a)and ThM8 structures(c).Stationary gap solitons tend to group into families with similarfield profiles and correspond to L-shape curves inside band gap region. Thefield profiles of these gap solitons inherit periodicity(b)or quasiperiodicity(d)of underlying structure.gap solitons for ThM structures withfinite width of the layers was shown numerically.AcknowledgmentsThis work was supported by the German Max Planck Society for the Advancement of Science(MPG). References[1]E.L.Albuquerque,M.G.Cottam,Polaritons in Periodic andQuasiperiodic Structures,Elsevier,Amsterdam,2004.[2]S.Zhu,Y.Y.Zhu,N.B.Ming,Quasi-phase-matched third-har-monic generation in a quasi-periodic optical superlattice,Sci-ence278(5339)(1997)843–846.[3]N.H.Liu,Propagation of light waves in Thue–Morse dielectricmultilayers,Physical Review B55(6)(1997)3543–3547. [4]X.G.Wang,U.Grimm,M.Schreiber,Trace and antitrace mapsfor aperiodic sequences:extensions and applications,Physical Review B62(21)(2000)14020–14031.[5]J.M.Luck,Cantorspectra and scalingof gapwidths indeterministicaperiodic systems,Physical Review B39(9)(1989)5834–5859.[6]C.S.Ryu,G.Y.Oh,M.H.Lee,Extended and critical wavefunctions in a Thue–Morse chain,Physical Review B46(9) (1992)5162–5168.[7]L.Dal Negro,M.Stolfi,Y.Yi,J.Michel,X.Duan,L.C.Kimerling,J.LeBlanc,J.Haavisto,Photon band gap properties and omnidirectional reflectance in Si=SiO2Thue–Morse quasi-crystals,Applied Physics Letters84(25)(2004)5186–5188.[8]V.Grigoriev,F.Biancalana,Bistability,multistability and non-reciprocal light propagation in Thue–Morse multilayered struc-tures,arXiv(2009)0912.3203v1.[9]F.Biancalana,All-optical diode action with quasiperiodic pho-tonic crystals,Journal of Applied Physics104(9)(2008)093113.[10]T.Amemiya,H.Shimizu,M.Yokoyama,P.N.Hai,M.Tanaka,Y.Nakano,1.54-mm TM-mode waveguide optical isolator based on the nonreciprocal-loss phenomenon:device design to reduce insertion loss,Applied Optics46(23)(2007)5784–5791. [11]J.Hwang,M.H.Song,B.Park,S.Nishimura,T.Toyooka,J.W.Wu,Y.Takanishi,K.Ishikawa,H.Takezoe,Electro-tunableoptical diode based on photonic bandgap liquid–crystal hetero-junctions,Nature Materials4(5)(2005)383–387.[12]E.Lidorikis,K.Busch,Q.M.Li,C.T.Chan,C.M.Soukoulis,Wave propagation in linear and nonlinear structures,Physica D 113(2–4)(1998)346–365.[13]M.Soljacic,M.Ibanescu,S.G.Johnson,Y.Fink,J.D.Joanno-poulos,Optimal bistable switching in nonlinear photonic crys-tals,Physical Review E66(5)(2002)055601.[14]X.S.Lin,n,Unidirectional transmission in asymmetricallyconfined photonic crystal defects with kerr nonlinearity,Chinese Physics Letters22(11)(2005)2847–2850.[15]M.D.Tocci,M.J.Bloemer,M.Scalora,J.P.Dowling,C.M.Bowden,Thin-film nonlinear optical diode,Applied Physics Letters66(18)(1995)2324–2326.[16]K.Gallo,G.Assanto,K.R.Parameswaran,M.M.Fejer,All-optical diode in a periodically poled lithium niobate waveguide, Applied Physics Letters79(3)(2001)314–316.[17]S.O.Konorov,D.A.Sidorov-Biryukov,I.Bugar,M.J.Bloe-mer,V.I.Beloglazov,N.B.Skibina,D.Chorvat,M.Scalora,A.M.Zheltikov,Experimental demonstration of a photonic-crystalfiber optical diode,Applied Physics B78(5)(2004) 547–550.[18]R.Philip,M.Anija,C.S.Yelleswarapu,D.Rao,Passive all-optical diode using asymmetric nonlinear absorption,Applied Physics Letters91(14)(2007)141118.[19]X.S.Lin,J.H.Yan,L.J.Wu,n,High transmission contrastfor single resonator based all-optical diodes with pumpassisting, Optics Express16(25)(2008)20949–20954.[20]J.E.Sipe,H.G.Winful,Nonlinear Schrodinger solitons in aperiodic structure,Optics Letters13(2)(1988)132–133,times cited:108.[21]C.M.de Sterke,B.J.Eggleton,J.E.Sipe,Bragg solitons:theoryand experiments,in:S.Trillo,W.Torruellas(Eds.),Spatial Solitons,Springer,Berlin,2001.[22]E.Lidorikis,C.M.Soukoulis,Pulse-driven switching in one-dimensional nonlinear photonic band gap materials:a numerical study,Physical Review E61(5)(2000)5825–5829.[23]H.Sakaguchi,B.A.Malomed,Gap solitons in quasiperiodicoptical lattices,Physical Review E74(2)(2006)026601. [24]C.B.Clausen,Y.S.Kivshar,O.Bang,P.L.Christiansen,Quasi-periodic envelope solitons,Physical Review Letters83(23) (1999)4740–4743.V.V.Grigoriev,F.Biancalana/Photonics and Nanostructures–Fundamentals and Applications8(2010)285–290 290。

Photosynthetic Characteristics of Three Fig Cultivars in Nyingchi, Tibet

Photosynthetic Characteristics of Three Fig Cultivars in Nyingchi, Tibet
Journal of Landscape Research 2019, 11(4): 128–130
Photosynthetic Characteristics of Three Fig Cultivars in Nyingchi, Tibet
LIU Xuesen, GAO Yi, XING Zhen* (Xizang Agriculture and Animal Husbandry College, Nyingchi, Tibet 860000, China)
1 Materials and methods 1.imental site
The test was conducted in campus practice base of Resource and Environmental College, XiZang Agriculture and Animal Husbandry College, in Bayi Town, Bayi District, Nyingchi City, Tibet Autonomous Region. The site locates on the west side of Nyang Qu, with the average elevation 2,960 m, the annual average temperature 8.6 , the annual sunshine hours 1,988.6–2,000.4 h, the frost-free period 170 d, and the annual precipitation 650–750 mm. The number of days with the annual average temperature above 10 is about 160d, and the percentage of sunshine is 46%. The climate is warm and moist, with distinct rainy and dry seasons. The dry season is mainly concentrated from October to December, and the rainy season is mainly concentrated from May to September. Due to high forest coverage rate, meteorological disasters such as drought, wind and hail are rare. The soil is loam, with general fertility. Shading shelters were set on the south side, the west side and above the site, with a transmittance of 60%[1-2]. 1.2 Materials

DTMCPack 包说明说明书

DTMCPack 包说明说明书

Package‘DTMCPack’October12,2022Type PackageTitle Suite of Functions Related to Discrete-Time Discrete-StateMarkov ChainsVersion0.1-3Date2022-04-10Author William NicholsonMaintainer William Nicholson<*********************>Description A series of functions which aid in both simulating and determining the properties offi-nite,discrete-time,discrete state markov chains.Two functions(DTMC,MultDTMC)pro-duce n iterations of a Markov Chain(s)based on transition probabilities and an initial distribu-tion.The function FPTime determines thefirst passage time into each state.The function stat-distr determines the stationary distribution of a Markov Chain.Imports statsLicense GPL(>=2)LazyLoad yesNeedsCompilation noRepository CRANDate/Publication2022-04-1102:12:30UTCR topics documented:DTMCPack-package (2)DTMC (3)FPTime (4)gr (5)hh (5)id (6)MultDTMC (6)statdistr (7)Index812DTMCPack-packageDTMCPack-package Suite of functions related to discrete-time discrete-state MarkovChainsDescriptionA series of functions which aid in both simulating and determining the properties offinite,discrete-time,discrete state markov chains.This package may be of use to practioners who need to simulate Markov Chains,but its primary intended audience is students of an introductory stochastic processes studying class properties and long run behavior patterns of Markov Chains.Two functions(DTMC, MultDTMC)produce n iterations of a Markov Chain(s)based on transition probabilities and an initial distribution.The function FPTime determines thefirst passage time into each state.The function statdistr determines the stationary distribution of a Markov Chain.Updated4/10/22to maintain compatibility with R.DetailsPackage:DTMCPackType:PackageVersion:0.1-2Date:2013-05-22License:GPL(>=2)LazyLoad:yesAuthor(s)Will NicholsonMaintainer:<****************>ReferencesSidney Resnick,"Adventures in Stochastic Processes"Examplesdata(gr)data(id)DTMC(gr,id,10,trace=FALSE)DTMC3 DTMC Simulation of Discrete-Time/State Markov ChainDescriptionThis function simulates iterations through a discrete time Markov Chain.A Markov Chain is a discrete Markov Process with a state space that usually consists of positive integers.The advantage of a Markov process in a stochastic modeling context is that conditional dependencies over time are manageable because the probabilistic future of the process depends only on the present state,not the past.Therefore,if we specify an initial distribution as well as a transition matrix,we can simulate many periods into the future without any further information.Future transition probabilities can be computed by raising the transition matrix to higher-and higher powers,but this method is not numerically tractable for large matrices.My method uses a uniform random variable to iterate a user-specified number of iterations of a Markov Chain based on the transition probabilities and the initital distribution.A graphical output is also available in the form of a trace plot.UsageDTMC(tmat,io,N,trace)Argumentstmat Transition matrix-rows must sum to1and the number of rows and columns must be equal.io Initial observation,1column,must sum to1,must be the same length as transi-tion matrix.N Number of simulations.trace Optional trace plot,specify as TRUE or FALSE.ValueTrace Trace-plot of the iterations through states(if selected)State An n x nrow(tmat)matrix detailing the iterations through each state of the Markov ChainAuthor(s)Will NicholsonReferences"Adventures in Stochastic Processes"by Sidney ResnickSee AlsoMultDTMC4FPTimeExamplesdata(gr)data(id)DTMC(gr,id,10,trace=TRUE)#10iterations through"Gambler s ruin"FPTime First Passage TimeDescriptionThis function uses the companion function multDTMC to simulate several Markov chains to deter-mine thefirst passage time into each state,i.e.thefirst time(after the initial iteration)that a specified state is reached in the Markov Process.First Passage Time can be useful for both determining class properties as well as the stationary/invariant distribution for large Markov Chains in which explicit matrix inversion is not computationally tractable.UsageFPTime(state,nchains,tmat,io,n)Argumentsstate State in which you want tofind thefirst passage time.nchains Number of chains you wish to simulate.tmat Transition Matrix,must be a square matrix,rows must sum to1.io Initial Distributionn Number of iterations to run for each Markov Chain.Valuefp1Vector of length(nchains)which givesfirst passage time into the specified state for each Markov Chain.Author(s)Will NicholsonSee AlsoDTMCExamplesdata(gr)data(id)FPTime(1,10,gr,id,10)#First passage time into first state on Gambler s ruingr5 gr Example Data Set:Gambler’s ruin on4statesDescriptionMotivating example,random walk with absorbing boundaries on4states.Analogous to a gambler ata casino.The4states represent a range of wealth.States1and4are absorbing with state1="Broke",state4="wealthy enough to walk away"and the intermediate states2and3are transitory.It is assumed that he bets of all his winnings in the intermediate states and has equal probability of winning and losingExamplesdata(gr)data(id)DTMC(gr,id,10,trace=FALSE)hh Harry the SemiProDescriptionExample Markov Chain from page139of Resnick.The protagonist,basketball player"Happy Harry’s"productivityfluctuates between three states(0-1points),(2-5points),(5or more points) and the transition between states can be modeled using a Markov ed as a motivating example to calculate the long run proportion of time spent in each state using the statdist function.SourceSidney Resnick"Adventures in Stochastic Processes"Examplesdata(hh)statdistr(hh)6MultDTMC id Initial distributionDescriptionA starting distribution for the gambler’s ruin example,which assigns equal probability of startingin each state.Examplesdata(id)data(gr)DTMC(gr,id,10,trace=FALSE)MultDTMC Multiple Discrete time Markov ChainsDescriptionAn extension of the DTMC package which enables multiple cocurrent Markov Chain simulations.At this time,plotting is not enabled.UsageMultDTMC(nchains,tmat,io,n)Argumentsnchains Number of chains to simulate(integer).tmat Transition Matrixio Initial distributionn Number of iterations to run each chain.Valuechains Returns nchains matrices of length nrow(tmat)by n which depict the transition of the Markov Chain.Author(s)Will NicholsonSee AlsoDTMCstatdistr7Examplesdata(gr)data(id)MultDTMC(20,gr,id,10)#20chains with10iterations using the Gambler s ruin example. statdistr Computing Stationary DistributionDescriptionThis function computes the stationary distribution of a markov chain(assuming one exists)using the formula from proposition2.14.1of Resnick:pi=(1,...1)(I-P+ONE)^(-1),where I is an mxm identity matrix,P is an mxm transition matrix,and ONE is an mxm matrix whose entries are all1.This formula works well if the number of states is small,but since it directly computes the inverse of the matrix,it is not tractable for larger matrices.For larger matrices1/E(FPTime(n))is a rough approximation for the long run proportion of time spent in a state n.Usagestatdistr(tmat)Argumentstmat Markov chain transition matrix,must be a square matrix and rows must sum to 1.ValueReturns a stationary distribution:mxm matrix which represents the long run percentage of time spent in each state.Author(s)Will NicholsonReferencesResnick,"Adventures in Stochastic Processes"Examplesdata(hh)statdistr(hh)Index∗Markov ChainsDTMCPack-package,2∗datasetsgr,5hh,5id,6DTMC,3,4,6DTMCPack(DTMCPack-package),2 DTMCPack-package,2FPTime,4gr,5hh,5id,6MultDTMC,3,6statdistr,78。

Modularity and community structure in networks

Modularity and community structure in networks

a r X i v :p h y s i c s /0602124v 1 [p h y s i c s .d a t a -a n ] 17 F eb 2006Modularity and community structure in networksM. E.J.NewmanDepartment of Physics and Center for the Study of Complex Systems,Randall Laboratory,University of Michigan,Ann Arbor,MI 48109–1040Many networks of interest in the sciences,including a variety of social and biological networks,are found to divide naturally into communities or modules.The problem of detecting and characterizing this community structure has attracted considerable recent attention.One of the most sensitive detection methods is optimization of the quality function known as “modularity”over the possible divisions of a network,but direct application of this method using,for instance,simulated annealing is computationally costly.Here we show that the modularity can be reformulated in terms of the eigenvectors of a new characteristic matrix for the network,which we call the modularity matrix,and that this reformulation leads to a spectral algorithm for community detection that returns results of better quality than competing methods in noticeably shorter running times.We demonstrate the algorithm with applications to several network data sets.IntroductionMany systems of scientific interest can be represented as networks—sets of nodes or vertices joined in pairs by lines or edges .Examples include the Internet and the worldwide web,metabolic networks,food webs,neural networks,communication and distribution networks,and social networks.The study of networked systems has a history stretching back several centuries,but it has expe-rienced a particular surge of interest in the last decade,especially in the mathematical sciences,partly as a result of the increasing availability of large-scale accurate data describing the topology of networks in the real world.Statistical analyses of these data have revealed some un-expected structural features,such as high network tran-sitivity [1],power-law degree distributions [2],and the existence of repeated local motifs [3];see [4,5,6]for reviews.One issue that has received a considerable amount of attention is the detection and characterization of com-munity structure in networks [7,8],meaning the appear-ance of densely connected groups of vertices,with only sparser connections between groups (Fig.1).The abil-ity to detect such groups could be of significant practical importance.For instance,groups within the worldwide web might correspond to sets of web pages on related top-ics [9];groups within social networks might correspond to social units or communities [10].Merely the finding that a network contains tightly-knit groups at all can convey useful information:if a metabolic network were divided into such groups,for instance,it could provide evidence for a modular view of the network’s dynamics,with dif-ferent groups of nodes performing different functions with some degree of independence [11,12].Past work on methods for discovering groups in net-works divides into two principal lines of research,both with long histories.The first,which goes by the name of graph partitioning ,has been pursued particularly in computer science and related fields,with applications in parallel computing and VLSI design,among other ar-eas [13,14].The second,identified by names such as blockFIG.1:The vertices in many networks fall naturally into groups or communities,sets of vertices (shaded)within which there are many edges,with only a smaller number of edges between vertices of different groups.modeling ,hierarchical clustering ,or community structure detection ,has been pursued by sociologists and more re-cently also by physicists and applied mathematicians,with applications especially to social and biological net-works [7,15,16].It is tempting to suggest that these two lines of re-search are really addressing the same question,albeit by somewhat different means.There are,however,impor-tant differences between the goals of the two camps that make quite different technical approaches desirable.A typical problem in graph partitioning is the division of a set of tasks between the processors of a parallel computer so as to minimize the necessary amount of interprocessor communication.In such an application the number of processors is usually known in advance and at least an approximate figure for the number of tasks that each pro-cessor can handle.Thus we know the number and size of the groups into which the network is to be split.Also,the goal is usually to find the best division of the network re-gardless of whether a good division even exists—there is little point in an algorithm or method that fails to divide the network in some cases.Community structure detection,by contrast,is per-2haps best thought of as a data analysis technique used to shed light on the structure of large-scale network datasets,such as social networks,Internet and web data, or biochemical munity structure meth-ods normally assume that the network of interest divides naturally into subgroups and the experimenter’s job is to find those groups.The number and size of the groups is thus determined by the network itself and not by the experimenter.Moreover,community structure methods may explicitly admit the possibility that no good division of the network exists,an outcome that is itself considered to be of interest for the light it sheds on the topology of the network.In this paper our focus is on community structure de-tection in network datasets representing real-world sys-tems of interest.However,both the similarities and differences between community structure methods and graph partitioning will motivate many of the develop-ments that follow.The method of optimal modularity Suppose then that we are given,or discover,the struc-ture of some network and that we wish to determine whether there exists any natural division of its vertices into nonoverlapping groups or communities,where these communities may be of any size.Let us approach this question in stages and focus ini-tially on the problem of whether any good division of the network exists into just two communities.Perhaps the most obvious way to tackle this problem is to look for divisions of the vertices into two groups so as to mini-mize the number of edges running between the groups. This“minimum cut”approach is the approach adopted, virtually without exception,in the algorithms studied in the graph partitioning literature.However,as discussed above,the community structure problem differs crucially from graph partitioning in that the sizes of the commu-nities are not normally known in advance.If community sizes are unconstrained then we are,for instance,at lib-erty to select the trivial division of the network that puts all the vertices in one of our two groups and none in the other,which guarantees we will have zero intergroup edges.This division is,in a sense,optimal,but clearly it does not tell us anything of any worth.We can,if we wish,artificially forbid this solution,but then a division that puts just one vertex in one group and the rest in the other will often be optimal,and so forth.The problem is that simply counting edges is not a good way to quantify the intuitive concept of commu-nity structure.A good division of a network into com-munities is not merely one in which there are few edges between communities;it is one in which there are fewer than expected edges between communities.If the num-ber of edges between two groups is only what one would expect on the basis of random chance,then few thought-ful observers would claim this constitutes evidence of meaningful community structure.On the other hand,if the number of edges between groups is significantly less than we expect by chance—or equivalently if the number within groups is significantly more—then it is reasonable to conclude that something interesting is going on. This idea,that true community structure in a network corresponds to a statistically surprising arrangement of edges,can be quantified using the measure known as modularity[17].The modularity is,up to a multiplicative constant,the number of edges falling within groups mi-nus the expected number in an equivalent network with edges placed at random.(A precise mathematical formu-lation is given below.)The modularity can be either positive or negative,with positive values indicating the possible presence of com-munity structure.Thus,one can search for community structure precisely by looking for the divisions of a net-work that have positive,and preferably large,values of the modularity[18].The evidence so far suggests that this is a highly effective way to tackle the problem.For instance, Guimer`a and Amaral[12]and later Danon et al.[8]op-timized modularity over possible partitions of computer-generated test networks using simulated annealing.In di-rect comparisons using standard measures,Danon et al. found that this method outperformed all other methods for community detection of which they were aware,in most cases by an impressive margin.On the basis of con-siderations such as these we consider maximization of the modularity to be perhaps the definitive current method of community detection,being at the same time based on sensible statistical principles and highly effective in practice.Unfortunately,optimization by simulated annealing is not a workable approach for the large network problems facing today’s scientists,because it demands too much computational effort.A number of alternative heuris-tic methods have been investigated,such as greedy algo-rithms[18]and extremal optimization[19].Here we take a different approach based on a reformulation of the mod-ularity in terms of the spectral properties of the network of interest.Suppose our network contains n vertices.For a par-ticular division of the network into two groups let s i=1 if vertex i belongs to group1and s i=−1if it belongs to group2.And let the number of edges between ver-tices i and j be A ij,which will normally be0or1,al-though larger values are possible in networks where mul-tiple edges are allowed.(The quantities A ij are the el-ements of the so-called adjacency matrix.)At the same time,the expected number of edges between vertices i and j if edges are placed at random is k i k j/2m,where k i and k j are the degrees of the vertices and m=14m ijA ij−k i k j4m s T Bs,(1)where s is the vector whose elements are the s i.The leading factor of1/4m is merely conventional:it is in-cluded for compatibility with the previous definition of modularity[17].We have here defined a new real symmetric matrix B with elementsk i k jB ij=A ij−FIG.2:Application of our eigenvector-based method to the “karate club”network of Ref.[23].Shapes of vertices indi-cate the membership of the corresponding individuals in the two known factions of the network while the dotted line indi-cates the split found by the algorithm,which matches the fac-tions exactly.The shades of the vertices indicate the strength of their membership,as measured by the value of the corre-sponding element of the eigenvector.groups,but to place them on a continuous scale of“how much”they belong to one group or the other.As an example of this algorithm we show in Fig.2the result of its application to a famous network from the so-cial science literature,which has become something of a standard test for community detection algorithms.The network is the“karate club”network of Zachary[23], which shows the pattern of friendships between the mem-bers of a karate club at a US university in the1970s. This example is of particular interest because,shortly after the observation and construction of the network, the club in question split in two as a result of an inter-nal dispute.Applying our eigenvector-based algorithm to the network,wefind the division indicated by the dotted line in thefigure,which coincides exactly with the known division of the club in real life.The vertices in Fig.2are shaded according to the val-ues of the elements in the leading eigenvector of the mod-ularity matrix,and these values seem also to accord well with known social structure within the club.In partic-ular,the three vertices with the heaviest weights,either positive or negative(black and white vertices in thefig-ure),correspond to the known ringleaders of the two fac-tions.Dividing networks into more than two communities In the preceding section we have given a simple matrix-based method forfinding a good division of a network into two parts.Many networks,however,contain more than two communities,so we would like to extend our method tofind good divisions of networks into larger numbers of parts.The standard approach to this prob-lem,and the one adopted here,is repeated division into two:we use the algorithm of the previous sectionfirst to divide the network into two parts,then divide those parts,and so forth.In doing this it is crucial to note that it is not correct, afterfirst dividing a network in two,to simply delete the edges falling between the two parts and then apply the algorithm again to each subgraph.This is because the degrees appearing in the definition,Eq.(1),of the mod-ularity will change if edges are deleted,and any subse-quent maximization of modularity would thus maximize the wrong quantity.Instead,the correct approach is to define for each subgraph g a new n g×n g modularity matrix B(g),where n g is the number of vertices in the subgraph.The correct definition of the element of this matrix for vertices i,j isB(g)ij=A ij−k i k j2m ,(4)where k(g)i is the degree of vertex i within subgraph g and d g is the sum of the(total)degrees k i of the vertices in the subgraph.Then the subgraph modularity Q g=s T B(g)s correctly gives the additional contribution to the total modularity made by the division of this subgraph.In particular,note that if the subgraph is undivided,Q g is correctly zero.Note also that for a complete network Eq.(4)reduces to the previous definition for the modu-larity matrix,Eq.(2),since k(g)i→k i and d g→2m in that case.In repeatedly subdividing our network,an important question we need to address is at what point to halt the subdivision process.A nice feature of our method is that it provides a clear answer to this question:if there exists no division of a subgraph that will increase the modular-ity of the network,or equivalently that gives a positive value for Q g,then there is nothing to be gained by divid-ing the subgraph and it should be left alone;it is indi-visible in the sense of the previous section.This happens when there are no positive eigenvalues to the matrix B(g), and thus our leading eigenvalue provides a simple check for the termination of the subdivision process:if the lead-ing eigenvalue is zero,which is the smallest value it can take,then the subgraph is indivisible.Note,however,that while the absence of positive eigen-values is a sufficient condition for indivisibility,it is not a necessary one.In particular,if there are only small positive eigenvalues and large negative ones,the terms in Eq.(3)for negativeβi may outweigh those for positive.It is straightforward to guard against this possibility,how-ever:we simply calculate the modularity contribution for each proposed split directly and confirm that it is greater than zero.Thus our algorithm is as follows.We construct the modularity matrix for our network andfind its leading (most positive)eigenvalue and eigenvector.We divide the network into two parts according to the signs of the elements of this vector,and then repeat for each of the parts.If at any stage wefind that the proposed split makes a zero or negative contribution to the total mod-5ularity,we leave the corresponding subgraph undivided. When the entire network has been decomposed into in-divisible subgraphs in this way,the algorithm ends. One immediate corollary of this approach is that all “communities”in the network are,by definition,indi-visible subgraphs.A number of authors have in the past proposed formal definitions of what a community is[9,16,24].The present method provides an alter-native,first-principles definition of a community as an indivisible subgraph.Further techniques for modularity maximization In this section we describe briefly another method we have investigated for dividing networks in two by mod-ularity optimization,which is entirely different from our spectral method.Although not of especial interest on its own,this second method is,as we will shortly show,very effective when combined with the spectral method.Let us start with some initial division of our vertices into two groups:the most obvious choice is simply to place all vertices in one of the groups and no vertices in the other.Then we proceed as follows.Wefind among the vertices the one that,when moved to the other group, will give the biggest increase in the modularity of the complete network,or the smallest decrease if no increase is possible.We make such moves repeatedly,with the constraint that each vertex is moved only once.When all n vertices have been moved,we search the set of in-termediate states occupied by the network during the operation of the algorithm tofind the state that has the greatest modularity.Starting again from this state,we repeat the entire process iteratively until no further im-provement in the modularity results.Those familiar with the literature on graph partitioning mayfind this algo-rithm reminiscent of the Kernighan–Lin algorithm[25], and indeed the Kernighan–Lin algorithm provided the inspiration for our method.Despite its simplicity,wefind that this method works moderately well.It is not competitive with the best pre-vious methods,but it gives respectable modularity val-ues in the trial applications we have made.However, the method really comes into its own when it is used in combination with the spectral method introduced ear-lier.It is a common approach in standard graph par-titioning problems to use spectral partitioning based on the graph Laplacian to give an initial broad division of a network into two parts,and then refine that division us-ing the Kernighan–Lin algorithm.For community struc-ture problems wefind that the equivalent joint strategy works very well.Our spectral approach based on the leading eigenvector of the modularity matrix gives an ex-cellent guide to the general form that the communities should take and this general form can then befine-tuned by our vertex moving method,to reach the best possible modularity value.The whole procedure is repeated to subdivide the network until every remaining subgraph is indivisible,and no further improvement in the modular-ity is possible.Typically,thefine-tuning stages of the algorithm add only a few percent to thefinal value of the modularity, but those few percent are enough to make the difference between a method that is merely good and one that is, as we will see,exceptional.Example applicationsIn practice,the algorithm developed here gives excel-lent results.For a quantitative comparison between our algorithm and others we follow Duch and Arenas[19] and compare values of the modularity for a variety of networks drawn from the literature.Results are shown in Table I for six different networks—the exact same six as used by Duch and Arenas.We compare mod-ularityfigures against three previously published algo-rithms:the betweenness-based algorithm of Girvan and Newman[10],which is widely used and has been incor-porated into some of the more popular network analysis programs(denoted GN in the table);the fast algorithm of Clauset et al.[26](CNM),which optimizes modularity using a greedy algorithm;and the extremal optimization algorithm of Duch and Arenas[19](DA),which is ar-guably the best previously existing method,by standard measures,if one discounts methods impractical for large networks,such as exhaustive enumeration of all parti-tions or simulated annealing.The table reveals some interesting patterns.Our al-gorithm clearly outperforms the methods of Girvan and Newman and of Clauset et al.for all the networks in the task of optimizing the modularity.The extremal opti-mization method on the other hand is more competitive. For the smaller networks,up to around a thousand ver-tices,there is essentially no difference in performance be-tween our method and extremal optimization;the mod-ularity values for the divisions found by the two algo-rithms differ by no more than a few parts in a thousand for any given network.For larger networks,however,our algorithm does better than extremal optimization,and furthermore the gap widens as network size increases, to a maximum modularity difference of about a6%for the largest network studied.For the very large networks that have been of particular interest in the last few years, therefore,it appears that our method for detecting com-munity structure may be the most effective of the meth-ods considered here.The modularity values given in Table I provide a use-ful quantitative measure of the success of our algorithm when applied to real-world problems.It is worthwhile, however,also to confirm that it returns sensible divisions of networks in practice.We have given one example demonstrating such a division in Fig.2.We have also checked our method against many of the example net-works used in previous studies[10,17].Here we give two more examples,both involving network representationsmodularity Q network GN CNM DA this paper3419845311331068027519maximal value of the quantity known as modularity over possible divisions of a network.We have shown that this problem can be rewritten in terms of the eigenval-ues and eigenvectors of a matrix we call the modularity matrix,and by exploiting this transformation we have created a new computer algorithm for community de-tection that demonstrably outperforms the best previ-ous general-purpose algorithms in terms of both quality of results and speed of execution.We have applied our algorithm to a variety of real-world network data sets, including social and biological examples,showing it to give both intuitively reasonable divisions of networks and quantitatively better results as measured by the modu-larity.AcknowledgmentsThe author would like to thank Lada Adamic,Alex Arenas,and Valdis Krebs for providing network data and for useful comments and suggestions.This work was funded in part by the National Science Foundation un-der grant number DMS–0234188and by the James S. McDonnell Foundation.[1]D.J.Watts and S.H.Strogatz,Collective dynamics of‘small-world’networks.Nature393,440–442(1998). [2]A.-L.Barab´a si and R.Albert,Emergence of scaling inrandom networks.Science286,509–512(1999).[3]o,S.Shen-Orr,S.Itzkovitz,N.Kashtan,D.Chklovskii,and U.Alon,Network motifs:Simplebuilding blocks of complex networks.Science298,824–827(2002).[4]R.Albert and A.-L.Barab´a si,Statistical mechanics ofcomplex networks.Rev.Mod.Phys.74,47–97(2002).[5]S.N.Dorogovtsev and J.F.F.Mendes,Evolution ofnetworks.Advances in Physics51,1079–1187(2002). [6]M.E.J.Newman,The structure and function of complexnetworks.SIAM Review45,167–256(2003).[7]M.E.J.Newman,Detecting community structure in net-works.Eur.Phys.J.B38,321–330(2004).[8]L.Danon,J.Duch, A.Diaz-Guilera,and A.Arenas,Comparing community structure identification.J.Stat.Mech.p.P09008(2005).[9]G.W.Flake,wrence,C.L.Giles,and F.M.Co-etzee,Self-organization and identification of Web com-munities.IEEE Computer35,66–71(2002).[10]M.Girvan and M.E.J.Newman,Community structurein social and biological networks.Proc.Natl.Acad.Sci.USA99,7821–7826(2002).[11]P.Holme,M.Huss,and H.Jeong,Subnetwork hierar-chies of biochemical pathways.Bioinformatics19,532–538(2003).[12]R.Guimer`a and L.A.N.Amaral,Functional cartogra-phy of complex metabolic networks.Nature433,895–900 (2005).[13]U.Elsner,Graph partitioning—a survey.Technical Re-port97-27,Technische Universit¨a t Chemnitz(1997). [14]P.-O.Fj¨a llstr¨o m,Algorithms for graph partitioning:Asurvey.Link¨o ping Electronic Articles in Computer and Information Science3(10)(1998).[15]H.C.White,S.A.Boorman,and R.L.Breiger,Socialstructure from multiple networks:I.Blockmodels of roles and positions.Am.J.Sociol.81,730–779(1976). [16]S.Wasserman and K.Faust,Social Network Analysis.Cambridge University Press,Cambridge(1994).[17]M.E.J.Newman and M.Girvan,Finding and evaluat-ing community structure in networks.Phys.Rev.E69, 026113(2004).[18]M.E.J.Newman,Fast algorithm for detecting com-munity structure in networks.Phys.Rev.E69,066133 (2004).[19]J.Duch and A.Arenas,Community detection in complexnetworks using extremal optimization.Phys.Rev.E72, 027104(2005).[20]F.R.K.Chung,Spectral Graph Theory.Number92in CBMS Regional Conference Series in Mathematics, American Mathematical Society,Providence,RI(1997).[21]M.Fiedler,Algebraic connectivity of graphs.Czech.Math.J.23,298–305(1973).[22]A.Pothen,H.Simon,and K.-P.Liou,Partitioning sparsematrices with eigenvectors of graphs.SIAM J.Matrix Anal.Appl.11,430–452(1990).[23]W.W.Zachary,An informationflow model for conflictandfission in small groups.Journal of Anthropological Research33,452–473(1977).[24]F.Radicchi,C.Castellano,F.Cecconi,V.Loreto,andD.Parisi,Defining and identifying communities in net-A101,2658–2663 (2004).[25]B.W.Kernighan and S.Lin,An efficient heuristic proce-dure for partitioning graphs.Bell System Technical Jour-nal49,291–307(1970).[26]A.Clauset,M.E.J.Newman,and C.Moore,Findingcommunity structure in very large networks.Phys.Rev.E70,066111(2004).[27]P.Gleiser and L.Danon,Community structure in jazz.Advances in Complex Systems6,565–573(2003). [28]H.Jeong,B.Tombor,R.Albert,Z.N.Oltvai,and A.-L.Barab´a si,The large-scale organization of metabolic networks.Nature407,651–654(2000).[29]H.Ebel,L.-I.Mielsch,and S.Bornholdt,Scale-free topol-ogy of e-mail networks.Phys.Rev.E66,035103(2002).[30]X.Guardiola,R.Guimer`a,A.Arenas,A.Diaz-Guilera,D.Streib,and L. A.N.Amaral,Macro-and micro-structure of trust networks.Preprint cond-mat/0206240 (2002).[31]M.E.J.Newman,The structure of scientific collabora-tion A98,404–409 (2001).[32]L.A.Adamic and N.Glance,The political blogosphereand the2004us election.In Proceedings of the WWW-2005Workshop on the Weblogging Ecosystem(2005).。

Non-Equilibrium in Adsorbed Polymer Layers

Non-Equilibrium in Adsorbed Polymer Layers

a r X i v :c o n d -m a t /0407246v 2 [c o n d -m a t .s o f t ] 1 N o v 2004TOPICAL REVIEWNon-Equilibrium in Adsorbed Polymer LayersBen O’Shaughnessy and Dimitrios Vavylonis Department of Chemical Engineering,Columbia University,New York,NY 10027,USA E-mail:bo8@,dv35@ Abstract.High molecular weight polymer solutions have a powerful tendency to deposit adsorbed layers when exposed to even mildly attractive surfaces.The equilibrium properties of these dense interfacial layers have been extensively studied theoretically.A large body of experimental evidence,however,indicates that non-equilibrium effects are dominant whenever monomer-surface sticking energies are somewhat larger than kT ,a common case.Polymer relaxation kinetics within the layer are then severely retarded,leading to non-equilibrium layers whose structure and dynamics depend on adsorption kinetics and layer ageing.Here we review experimental and theoretical work exploring these non-equilibrium effects,with emphasis on recent developments.The discussion addresses the structure and dynamics in non-equilibrium polymer layers adsorbed from dilute polymer solutions and from polymer melts and more concentrated solutions.Two distinct classes of behaviour arise,depending on whether physisorption or chemisorption is involved.A given adsorbed chain belonging to the layer has a certain fraction of its monomers bound to the surface,f ,and the remainder belonging to loops making bulk excursions.A natural classification scheme for layers adsorbed from solution is the distribution of single chain f values,P (f ),which may hold the key to quantifying the degree of irreversibility in adsorbed polymer layers.Here we calculate P (f )for equilibrium layers;we find its form is very different to the theoretical P (f )for non-equilibrium layers which are predicted to have infinitely many statistical classes of chain.Experimental measurements of P (f )are compared to these theoretical predictions.PACS numbers:82.35.-x,68.08.-p,05.40.-a Submitted to:J.Phys.:Condens.Matter1.IntroductionHigh molecular weight polymers are extremely surface-active molecules.Even a weak interaction between a single monomer and a surface tends to be magnified into a powerful attraction or repulsion when many of these monomers are linked together to form a polymeric chain.It is a remarkable fact that surfaces contacting even extremely dilute polymer solutions can readily develop very dense polymer layers.Technologies revolving around the properties of either synthetic or biological polymer layers are many and varied,including adhesion [1,2],coating[3],colloid stabilization[4],fiber reinforced thermoplastics[5,6],flocculation processes[7],DNA microaarrays[8]and biocompatibilization[9].Motivated both by fundamental questions and by technology, understanding and predicting the structure and formation kinetics of these soft layers is a major concern of polymer science[10].A crucial aspect of experimental studies is that adsorbed polymer dynamics are typically extremely slow for long chains:an individual chain has many surface stickers and interacts with numerous other chains impeding its motion.Irreversibility and non-equilibrium effects are therefore very common.The subject of this review is experimental and theoretical work on these non-equilibrium effects, and though we consider adsorption from dilute solutions,semi-dilute solutions and melts our main emphasis is neutral homopolymer adsorption from dilute solutions.This is the simplest and most widely studied class.Polymer-surface adsorption systems are naturally classified according to the mode of adsorption. Roughly,there are two classes:chemisorption and physisorption(seefigure1).The clearest example of irreversibility arises in chemisorption(figure1(b))where the covalent polymer-surface bonds which develop are often essentially irreversible on experimental timescales.Monomer sticking free energies,ǫ,have values typical of covalent bonds which are one to two orders of magnitude greater than kT.Chemical adsorption is employed in various technologies where polymers are attached by chemical reactions to solid surfaces either from a polymer melt as in the reinforcement of polymer-polymer or polymer-solid interfaces[2,6,11,12],or from solution as in colloid stabilization by chemically grafting polymers onto particle surfaces[13–15].What is less obvious is why non-equilibrium effects are so widely observed in physisorbing systems, even for rather weak sticking energies.Available experimental evidence suggests that irreversibility effects become important as soon asǫbecomes somewhat larger than kT.For example the experiments by Schneider et al.[16,17]for polymethylmethacrylate(PMMA)adsorption onto oxidized silica via hydrogen bonding in dilute CCl4solutions(ǫ≈4kT)show essentially frozen-in adsorbed chain confirge physisorption sticking energies(ǫ>kT)originate in hydrogen bonding or other dipolar forces,dispersion forces or attractions between charged groups.Metal and silicon-based surfaces are usually oxidized and many polymer species form strong hydrogen bonds with the surface oxygen or silanol groups[18,19].Biopolymers such as proteins and DNA attach tenaciously to many surfaces due to their many charged,polar and hydrophobic groups[8,9,20].Since hydrogen bonds,for instance,typically have energies of several kT[21,22]it is apparent that strong physical bonds are very common.This suggests that whether physical or chemical bonding is involved,for long chains irreversible effects may in practice be the rule rather than the exception.Figure 1.(a)Schematic of physisorption from afluid polymer phase onto a surface.Adsorbed chainsconsist of loops,tails and sequences of bound monomers(“trains”).When non-equilibrium effects becomeimportant,layer structure depends on kinetics of adsorption.This review addresses phyisorption from dilutesolutions in sections2,3and4and physisorption from melts in section6.(b)As in(a)but for chemisorption.In this case chains carry reactive groups which can form covalent bonds(shown black)with a functionalizedsurface.Chemisorption from dilute solutions is reviewed in section5and from concentrated solutions insection6.To understand non-equilibrium layers,one must identify how they differ from equilibrium layers.The theory of fully equilibrated layers is far more advanced,at both the mean-field[23]and scaling[24–31]level of description.A main result of these theories is expressions for the decay of the monomer density profile asFigure2.The two broad classes of polymer adsorption,physisorption and chemisorption,have very differentvalues of the parameter Q,the local monomer-surface association rate.Q can be thought of as the conditionalmonomer-surface sticking probablity per unit time,given the unattached monomer contacts the surface.Though many systems are in practice mixtures of chemisorption and physisorption,a simplified view ofmonomer free energy as a function of distance between monomer and surface is shown.(a)For physisorbingpolymers,the activation barrier is very small and and monomer-surface association is very likely upon contact,i.e.Qt a is of order unity,where t a the monomer relaxation time.When the sticking energyǫexceeds a fewkT,experiment indicates that chains need large amounts of time to escape the surface,presumably due tocomplex many-chain effects.(b)Chemisorption typically involves a large activation barrier,u≫kT.Manymonomer-surface collisions are needed to traverse this barrier,Qt a≪1.The adsorbed state is also usuallystrongly favored,ǫ≫kT.a function of the distance z from the surface.For adsorption from dilute solutions for example,in the scaling picture originally developed by de Gennes[24,25],Eisenriegler et al.[26,27],and de Gennes and Pincus[28], each adsorbed chain has sequences of surface-bound monomers(trains)interspersed with portions extending away from the surface(tails and loops of size s)with distributionΩ(s)∼s−11/5[29–31]leading to a self-similar density profile c(z)∼z−4/3.Experimentally,the existence of an extended diffuse layer is well established by a large number of neutron scattering[32–37]and neutron reflectivity[38–40]studies.However a universally accepted quantitative test of the predicted density profiles has been difficult to achieve,both due to intrinsic limitations of the experimental techniques[41]and to the asymptotic nature of many of the theoretical results which are valid in the limit of very long chains.Furthermore,for experiments specifically directed at equilibrium,ensuring that equilibrium conditions are realised is difficult when the very non-equilibrium effects one wishes to avoid are poorly identified.Understanding the origin of the observed deviations from equilibrium for weakly adsorbing systems in dilute solutions is a major unresolved puzzle in polymer science.At present it is unclear how universal these non-equilibrium effects may be.Various effects have been speculated to play important roles.Kremer[42] and de Gennes[43]have suggested that if the solution temperature is below the polymer melt glass transition temperature,glassy effects may onset at the interface where polymer densities are similar to those of melts. Interactions with the surface might further enhance glassiness.Chakraborty and coworkers[44–47]suggested flattened-down chains experience strong kinetic barriers due to local steric constraints which drastically slow down dynamics.Ideas related to slow cooperative motions,mutual pinning,development of entanglements at the interface and crystalization have been proposed by Johner and Semenov[48],Sommer[49],Granick[50] and Raviv et el.[51]in a series of comments following a recent theoretical work by Raviv et al.[52]which interpreted past experiments[53,54]exhibiting non-equilibrium effects.In this review we do not attempt an exhaustive review of the vast body of past research work involving strongly physisorbing or chemisorbing polymers at interfaces.Instead,with fundamental issues in mind,our aim is to(i)assemble and classify a number of theoretical ideas and numerical simulations which articulate the community’s current level of understanding of equilibrium and non-equilibrium polymer adsorption,and (ii)summarize a number of experimental results which we believe are particularly pertinent and which still demand a theoretical explanation.The emphasis is on the simplest case:adsorption of neutralflexible polymers from dilute bulk solutions.We also review work on irreversible adsorption from semi-dilute bulk solutions and melts,motivated by ideas initiated by Guiselin[55].Polyelectrolyte solutions,polymers with complex architectures and non-flat surfaces are beyond the scope of the present review.Physisorption and chemisorption will be carefully distinguished.These are characterized by very different values of the local monomer-surface association rate,Q(seefigure2).In physisorption,monomer attachment is usually essentially diffusion-limited,Q=C/t a,where t a is monomer relaxation time and C is a system-dependent constant of order unity[56].Chemisorption is normally much slower[56–58]with Q values typically 8or more orders of magnitude smaller than those of physisorption.The origin of this difference is that chemical bond formation usually involves a large activation barrier(seefigure2).Similarly,desorption ratesafter chemisorption are usually very small and can be ignored.The effect of desorption on physisorbing systems is more subtle and is discussed in section4.The above two classes naturally lead to very different adsorption kinetics.This is analogous to bulk polymer-polymer reaction kinetics where depending on Q, polymer length N,and solvent conditions,the kinetics are described by one of a range of“diffusion-controlled”and“mean-field”kinetic regimes[58–61].Such regimes also arise for end-adsorbing polymers[56,57,62–67].In section2we briefly review the equilibrium picture for dilute solutions and in section3we discuss experimental evidence for non-equilibrium departures from this picture.Theoretical work related to physisorbing non-equilibrium layers from dilute solution is reviewed in section4.We move to chemisorption, again from dilute solution,in section5.Section6addresses irreversibility effects involving melts and semi-dilute solutions.We conclude with a brief discussion of the experimental and theoretical outlook.2.Adsorption from Dilute Solutions:The Equilibrium Picture2.1.Structure of Equilibrium LayersThis section briefly outlines some central results of equilibrium theories of adsorbed polymer layers.Much more extensive reviews can be found in refs.[41,68–71].In the scaling picture developed mainly in the 1980’s[24–28],each adsorbed chain consists of surface-bound monomers and large loops and tails generating a monomer density profile c(z)as a function of distance from the surface,z.Eisenriegler et al.[26–28]showed that very close to the surface,in the“proximal”region,the density decays as a power law,c(z)∼z−m,where the critical exponent m≈1/3represents competion between surface-sticking energy gain,chain entropy,and excluded volume interactions.The proximal region decay law crosses over to de Gennes’“self-similar grid”regime[24,25,28],c(z)∼z−4/3,beyond a certain distance h prox.For z>h prox the polymer layer can be thought of as a semi-dilute solution with continously varying local concentration c(z).In this region the correlation length measuring the range of excluded-volume interactions,ξ=a−5/4c−3/4,is proportional to the distance from the surface,z,since this is the only relevant length scale:ξ≈z.Here a is defined to be the monomer size.Expressingξin terms of c leads to†a3c(z)≈ (a/h prox)(a/z)1/3,a<z<h proxh prox=a kT/ǫ,R F=aN3/5(1)(a/z)4/3,h prox<z<R FUnless the bulk polymer concentration,c,is extremely small[72],then the equilibrium layer height is of order the Flory bulk coil radius R F as indicated in equation(1).In this same range of c the adsorption isotherm exhibits a large plateau,i.e.surface coverageΓis weakly dependent on c.Even in weakly adsorbing polymer systems,e.g.adosrption through weak van der Waals interactions,the value ofǫis usually of order kT.By studying the adsorption/desorption transition in binary solvent mictures, van der Beek et al.[18]estimated the sticking energies per monomer of many typicalflexible polymers onto silica and alumina surfaces from organic solvents to lie in the range0.5to6kT.Hence the width of the proximal region is typically of order the monomer size,h prox≈a,and a clear signature of the proximal region is hard to probe experimentally.In the following we considerǫof order kT or larger.We remark that the net monomer free energy of adsorptionǫincludes both the“stickiness”a monomer feels for the surface, but also the entropic disadvantage due to constraining local orientational degrees of freedom upon contact with the surface.Thus,crudely speaking one can say the stickiness contribution must exceed a crtical value ǫc representing the entropic disadvantage before a monomer can puter simulations showǫc is of order kT and is lattice-dependent[27].The real situation is more complex,with various contributions from electronic and structural factors such as solvent molecule entropy effects,etc[21].The density decay law of equation(1)reflects a power law distribution of loop and tail sizes.Neglecting differences between loops and tails and the details associated with the proximal region,then the loop size distribution per surface site is[29–31]Ω(s)≈a−2s−11/5.(2) Beyond this,Semenov and Joanny[31]showed that the inner region of the layer,z<z∗≡aN1/2,is dominated by loops while the outer region,z>z∗,is dominated by tails;the resulting density profile obeys a z−4/3law above and below z∗,respectively,but with different numerical prefactors.Support for the scaling conclusions of equations(1)and(2)is provided by Monte-Carlo simulations of Zajac and Chakrabarti[73], de Joannis et al.[74,75],and Cifra[76].These produce a density decay consistent with the z−4/3law for long chains.Zajac and Chakrabarti[73]additionally report agreement with equation(2).†The cross-over distance h prox and the prefactor in the proximal region density law can be determined by demanding(i)a smooth cross-over at h prox and(ii)the osmotic free energy per unit area, R F a dzkT/ξ3,balances the sticking free energy per unit area,ǫa c(a).Complementary to the scaling approach outlined above has been the numerical lattice model of Scheutjens and Fleer(SF)[77,78].This is a self-consistent meanfield theory which averages excluded volume interactions and thus treats self-avoidance in an approximate manner.This approximation however allows numerical solutions for the density profile and for the loop and tail distributions and can additionally describe chains of finite length.The meanfield description becomes more accurate for solvents near the theta temperature(a common case)where self-avoidance is a weak perturbation except for the longest loops and tails.The existence of the loop-and tail-dominated regions of the layer was in factfirst established by the SF model[78,79].The layer height in the SF theory scales as h∼N1/2[77,78]while the density profile decays as c(z)∼z−2(for sufficiently long chains),different to the z−4/3decay predicted by the scaling approach,as shown by van der Linden and Leermakers[80].Analytical meanfield solutions for the density profile in the limit of very long chains were derived within the ground-state dominance approximation[81] by Jones and Richmond[82].Going beyond the ground state dominance approximation,Semenov et al.[23] subsequently generalized this approach to account forfinite length effects to leading order and analytically describe the different contributions of loops and tails to the density profile.They found that loops dominate for z<z∗MF≡aN1/3while tails dominate for z>z∗MF,similarly to the scaling approach of Semenov and Joanny[31].These new methods have revived interest in analytical and numerical meanfield approaches to polymer adsorption[83–89].Turning now to experiment,the fundamentals of polymer adsorption at the solid/liquid and air/liquid interface have been studied in a vast number of experiments.Research prior to1993is reviewed in the book by Fleer et al.[41].Given the strong evidence for nonequilibrium effects(see below),in general one should be very cautious when comparing experimentalfindings to equilibrium expectations.Overall,experiment is consistent with the general trend predicted by equilibrium theories regarding structure of the polymer layers which were studied,although thefine differences between the meanfield and scaling pictures are hard to distinguish.Very briefly,measurements of the layer’s surface bound monomer fraction as a function of total adsorbed amount and molecular weight(MW)by techniques such as NMR[90,91],ESR[92,93],or infrared spectroscopy[94]give results fairly consistent with the predictions of the SF theory[41].The thickness, h exp,of polymer layers has been probed as a function of chain length by hydrodynamic methods[41,95–97], ellipsometry[98],and the surface force apparatus[99].Depending on the method,h exp is proportional to a certain moment of the density profile and many existent measurements are compatible with a power law,h exp∼Nα.Certain studies have favored the SF theory predictions[41,96]while others support the scaling predictions[97,99].Forflexible polymer species the total surface coverageΓas a function of bulk concentration is found to be very weakly dependent on the polymer concentration in the bulk except for very dilute solutions,in qualitative agreement with both the scaling and the SF theories[41].For a given bulk concentration,meaurements ofΓas a function of N in good solvents typically show a weak dependence on chain length for large N[41].This is consistent with the SF and scaling theories which predictΓ∼ h a dz c(z) is dominated by the lower,N-independent limit.Small angle neutron scattering(SANS)and neutron reflectivity have been used to probe the density profile.These experiments established the existence of a diffuse extended layer but at present there is no general agreement as to the exact form of the density decay.A techical difficulty intrinsic to SANS,as dicussed in ref.[41],is its limited sensitivity to the more dilute regions of the layer.Neutron reflectivity experiments are also often ambiguous,since multiple density profiles can befitted to the same data.The SANS experiments of Auvray and Cotton[35]and Hone et al.[37]are consistent with the z−4/3scaling law.However the results of Hone et al.could also be described by an exponential profile(see also[32]). SANS experiments by Cosgrove et al.[33,34]do not support the scaling predictions,but are more consistent with the SF theory.Other SANS studies by Rennie et al.[36]are inconsistent with both scaling and SF predictions,while neutron reflectivity experiments of Lee et al.[38,39]and Sun et al.[40]have generated data consistent with the scaling predictions.2.2.Single Chain Statistics and the Equilibrium Distribution of Bound FractionsSo far this section has discussed many-chain layer properties.Equally important,and characteristic of the layer-forming processes,are properties of individual surface-adsorbed chains in the layer.What is the spectrum of configurations of individual chains?According to the scaling picture,a typical chain has ND(s) loops of length s or greater,where D(s)≡ ∞s ds′Ω(s′)∼s−6/5after using equation(2).Semenov and Joanny[100]argue that because of screening effects these are essentially independent blobs and their2D spatial extent parallel to the surface is[ND(s)]1/2as3/5=aN1/2.This occurs for all scales s;in particular, a given chain has of order one loop of length N5/6,also of size aN1/2.Hence a typical chain has a lateral size of order aN1/2,the ideal result(to within logarithmic corrections[100]).Figure3.Equilibrium probabilty distribution P eq of chain bound fraction,f,in good solvents.For verylong chains the distribution is sharply peaked at a value¯f of order unity.For realistic values of N thedistribution is rather broad.A special role is played by another single chain property,directly reflecting the degree to which individual chains are bound to the surface.This is the probability distribution P eq(f)that a chain has a fraction f of its monomers touching the surface.This property plays a central role in this review,since its features closely reflect whether irreversible effects matter or not.In two independent Monte Carlo studies by Wang and Rajagopalan[101]and Zajac and Chakrabarti[73]an equilibrium distribution was found with a single peak at a value of f of order unity.To our knowledge,P eq(f)has not been calculated analytically,at least at the scaling level.In order to compare equilibrium predictions with experimental measurements of bound fractions in non-equilibrium layers in later sections,we have calculated P eq(f)by relating an adsorbed chain to a1D unidirectional walk combining an initial tail,a sequence of intermediate steps,and afinal tail.The result,which is derived in the appendix and is shown infigure3,reads:N−1/5P eq(f)≈Figure4.Schematic of new chain adsorption in an equilibrium polymer layer(shown as a self-similar grid)as described in ref.[102].(a)Entry:a bulk chain reptates into the layer and makes afirst attachment to thesurface.(b)Spreading:the incoming chain establishes an increasing number of contacts with the surface.(c)A typical adsorbed chain configuration is adopted,consisting of trains,loops and tails.A similar picturewas used in ref.[100].Chain desorption follows the same path in the reverse order.3.1.Theories of Dynamics in Equilibrium LayersCompared to static properties,much less is established theoretically about equilibrium dynamics.These have been studied for good solvents by de Gennes[43,102–105],Semenov and Joanny[100],and Baschnagel et al.[106]for bidisperse solutions.The picture emerging from these works is that the layer of bound chains has a certain characteristic equilibration timeτeq.This can be thought of as the time after which the chains following the distribution P eq(f)offigure3are completely reshuffled among themselves.The exchange of chains between the bulk and the layer was predicted to be a slower process due to the fact that incoming and outgoing chains have to pass through unfavored configurations having a small fraction of bound monomers (seefigure4).de Gennes assumed reptation dynamics(i.e.entangled layers)and found the longest relaxation time of the layer to scale as[102,104,105]τeq≈t s N3(entangled layers).(4) Here t s is the relaxation time of an adsorbed monomer which,due to wall-polymer interactions,may be much larger than the corresponding time in the bulk,t a[107].Semenov and Joanny[100]assumed unentangled layers and Rouse-Zimm dynamics and obtainedτeq≈t a N2(in their work t s≈t a was assumed).In equilibrium,layer and bulk chains desorb and adsorb at the same average rate,respectively.In de Gennes’picture bulk chains adsorb in two stages(seefigure4).During thefirst“entry”stage,the bulk chain overcomes the exclude-volume barrier presented by the layer’s loops and tails and makes itsfirst contact with the surface,i.e.f=1/N.During a second“spreading”stage the chain continues to make an increasing number of surface-contacts,f increases up to f min,and the chain becomes part of the bound layer. When entry is rate-limiting he found that the mean lifetime of an adsorbed chain before its desorption is τex≈t a N3.7/φ,whereφis the volume fraction of polymer in the bulk.Semenov and Joanny[100]described the dynamics using a similar picture,but assuming unentangled layers and Rouse-Zimm dynamics.They obtained a slighlty different chain lifetime,τex≈t a N2.42/φ(to within logarithmic corrections).Note that the exchange timescale,τex,has a weak power law dependence on N rather than exponential because the incoming/outgoing barrier is small.The scalingτex∼1/φreflects the linear dependence on concentration of the rate of chain arrival at the surface.Note also that even for the highest dilute solution concentrations,φ=φ∗,whereφ∗≡N−4/5is the chain overlap threshold concentration[81],one still has τeq≪τex.A prediction of the above works is that chain desorption into pure solvent,φ→0,is extremely slow,which is well-established experimentally[108].Now suppose one starts with a layer of labeled chains in equilibrium and replaces the bulk solution with identical but unlabeled chains of the same concentration at t=0.An important prediction of the above theories is that the decay of the surface coverage of labeled chains,Γ,is a simple exponential for all times[43,100]:Γ(t)=Γ(0)e−t/τex(5) An implicit feature of equation(5)is that there is a single observed desorption rate sinceτex≫τeq,i.e.the desporption process is slow enough to sample an average over all equilibrium chain states in the layer.Note this result in fact assumes a given desorbed labeled chain does not readsorb,i.e.memory is lost instantly. Experimentally,this necessitates a mixing process in the bulk to dilute the desorbed chains.In the absence of such mixing diffusion returns a desorbed chain to the surface repeatedly,an effect which can lead to non-exponential decay[60,109,110].The kinetics of polymer layer build up starting from empty or“starved”surfaces is more complex and has been considered in refs.[43,100,102,111].3.2.Dynamics of Adsorbed Polymer Layers:Monte Carlo SimulationsThis sub-section provides a brief review of numerical Monte Carlo simulations of dynamics in many-chain polymer layers in contact with dilute solutions(for dynamics of single chains see refs.[112–116]).The simulations reported here did not include hydrodynamic interactions.The main results of the simulations by Wang et al.[101,117]are qualitatively in agreement with the theoretical picture of the previous subsection. They found that the lateral dynamics of adsorbed chains up to N=100are consistent with Rouse dynamics. For sufficiently sticky surfaces(0<ǫ≤1.5kT withǫc=0.5kT)the value ofτex was found to be much larger than the lateral relaxation time,even though the scaling dependence on N was the same.This should be contrasted with the Semenov and Joanny prediction that the two exponents differ by a small value,0.42. Wang et al.observed non-exponential exchange kinetics arising from readsorption of desorbed chains.Lai[113,118]studied the layer dynamics as a function ofǫfor N≤80and interestingly found that for ǫ>∼1kT(withǫc≈0.9kT)the lateral chain dynamics started to slow down and to approach apparently glassy dynamics atǫ≈4kT.This result was claimed to be valid despite an omission in the implemented algorithm[113,119].This report is important since it indicates that the value ofǫis crucial in polymer adsorption.Zajac and Chakrabarti[120]studied the dynamics for N=100and N=200andǫ+ǫc=1.8kT near and aboveφ=φ∗.Their algorithm involved unphysical reptation moves in order to speed up the dynamics. In equilibrium they found a distribution of bound fractions similar to the one offigure3and observed that the internal dynamics of reshuffling of chains between different f values is complex.The timescale for the exchange of adsorbed chains by bulk chains was found to be slower than internal equilibration processes. Simple exponential exchange kinetics were observed as in equation(5).Takeuchi[121]also used the Monte Carlo method with extra reptation moves.For surfaces withǫ≈1.6kT he observed exponential exchange kinetics while forǫ≈0.9kT readsorption effects were important leading to non-exponential exchange kinetics.3.3.Experiment:Departure from Equilibrium PictureThefirst experimental studies of exchange kinetics onflexible polymers were performed by Pefferkon et al.[122–125]using radioactive labeling techniques.One study[122,124,125]involved labeled polyacrylamide (PAM)in water adsorbed through hydrogen bonding onto aluminol-grafted glass beads.The beads were exposed to a dilute solution of labeled PAM for approximately30min until time-independent coverage was achieved.The labeled solution was then replaced by a dilute unlabeled PAM solution of varying concentration c and the amount of labeled PAM which remained adsorbed,Γ,was monitored as a function of time as shown infigure5(i).An interesting result of this experiment was that the exchange rate per labeled chain,shown infigure5(ii),was time-dependent and reached a constant value after a cross-over period of≈300min which was approximately the same for every c.This asymptotic rate was found to increase linearly with c,as shown infigure5(iii).The observed spectrum of exchange times disagrees with equation(5)and this can be interpreted in many ways as follows:(i)The observed non-exponential exchange kinetics is a signature of non-equilibrium.Pefferkorn et al.[122,124]argued that the interface is populated with a spectrum of different frozen or slowly changing configurations and,consequently,different kinetic properties.(They proposed that the layer consists of aflat sublayer of tightly bound chains which exchange slowly,plus a less tightly bound population which exchange more rapidly).(ii)The layer is not in equilibrium when the exchange experiment starts but it equilibrates afterτeq≈300min which is larger than the layer’s preparationg time.The asymptotic exchange rate then becomes constant and equal to1/τex.The fact that asymptoticallyτex∼1/c as seen infigure5(iii)and the fact that τex>300min as can be seen infigure5(ii),are consistent with this interpretation and the theories reviewed in subsection3.1.Assuming reptation dynamics,equation(4),and given N≈1400,this implies a relaxation time of adsorbed monomers of order t s≈10−5s.This is much larger than monomer relaxation times in the bulk,t a≈10−10s.(iii)The layer is in fact in equilibrium but its exchange kinetics and internal equilibration processes are much more complex than assumed by existent theories,at least for this system.For example,if the equilibrium P eq(f)is very broad and chains with different f values have very different exchange times,then the intial drop inΓwill be due mainly to the most rapidly desorbing chains if their desorption times are less than τeq.Issues related to surface density of aluminol groups,polydispersity,and effect of pH(this experiment was performed at pH=4where PAM is neutral while many of the surface aluminol groups were positively。

1-s2.0-S0956566314004308-main

1-s2.0-S0956566314004308-main

A novelfluorescent“turn-on”chemosensor for nanomolar detection ofFe(III)from aqueous solution and its application in living cells imagingJitendra Nandre a,Samadhan Patil a,Vijay Patil a,Fabiao Yu b,Lingxin Chen b,n,Suban Sahoo c,Timothy Prior d,Carl Redshaw d,Pramod Mahulikar a,n,Umesh Patil a,na School of Chemical Sciences,North Maharashtra University,P.B.No.80,Jalgaon425001,MS,Indiab Key Laboratory of Coastal Zone Environmental Processes and Ecological Remediation,Yantai Institute of Coastal Zone Research,Chinese Academy ofSciences,Yantai264003,Chinac Department of Applied Chemistry,S.V.National Institute Technology,Surat395007,Gujrat,Indiad Department of Chemistry,University of Hull,Cottingham Road,Hull HU67RX,UKa r t i c l e i n f oArticle history:Received11April2014Received in revised form5June2014Accepted10June2014Available online14June2014Keywords:Fluorescent“turn-on”sensorBenzo-thiazolo-pyrimidineFe3þICTPETLive cells imaging.a b s t r a c tAn electronically active and spectral sensitivefluorescent“turn-on”chemosensor(BTP-1)based on thebenzo-thiazolo-pyrimidine unit was designed and synthesized for the highly selective and sensitivedetection of Fe3þfrom aqueous medium.With Fe3þ,the sensor BTP-1showed a remarkablefluorescence enhancement at554nm(λex¼314nm)due to the inhibition of photo-induced electrontransfer.The sensor formed a host-guest complex in1:1stoichiometry with the detection limit down to0.74nM.Further,the sensor was successfully utilized for the qualitative and quantitative intracellulardetection of Fe3þin two liver cell lines i.e.,HepG2cells(human hepatocellular liver carcinoma cell line)and HL-7701cells(human normal liver cell line)by a confocal imaging technique.&2014Elsevier B.V.All rights reserved.1.IntroductionThe development of highly selective and sensitive chemosen-sors for bioactive metal ions has gained enormous importance(Au-Yeung et al.,2013;Callan et al.,2005;Chen et al.,2012;Duttaand Das,2012;Formica et al.,2012;Kim et al.,2012;Sahoo et al.,2012),as metal ions are well known to be involved in a variety offundamental biological processes,which are essential for main-taining the life of organisms and remain sustainable underenvironmental conditions.As an important physiologically rele-vant metal ion,Fe3þexhibits an obligatory role in many biochem-ical processes at the cellular level.Numerous enzymes use Fe3þasa catalyst for electron transfer,oxygen metabolism,and RNA andDNA synthesis(Cairo and Pietrangelo,2000;Crabtree,1994).However,both its deficiency(hypoferremia)and excess(hyperfer-remia)can induce a variety of diseases.As a result of theseconcerns,intense research efforts have been focused on thedevelopment of highly sensitive and selective receptors for thequalitative and quantitative detection of Fe3þ.However,becauseof the paramagnetic nature of the Fe3þion,recognition of Fe3þbyfluorescence response is mostly signaled by afluorescencequenching mechanism(Fan et al.,2006;Li et al.,2009;Lohaniet al.,2009).Also,the design of a“turn-on”fluorescent chemo-sensor for the Fe3þion remains a challenging task for researchersworking in thefield of chemosensing due to the need to overcomethe usualfluorescence quenching nature of Fe3þ.The lack ofsuitable“turn-on”fluorescent iron indicators is even more obviouswhen judged in terms of applications in bioimaging,althoughsignificant progress has been made onfluorescent molecularsensors for intracellular imaging of biologically important metalions.It is worth noting that many reportedfluorescent“turn-on”chemosensors for the Fe3þion utilize the rhodamine moiety dueto its advantageous photophysical properties(Sahoo et al.,2012).However,some rhodamine dyes are harmful if swallowed byhumans or animals,and cause irritation of the skin,eyes andrespiratory tract(Jain et al.,2007;Rochat et al.,1978).In addition,among the various sensing methods,sensors based on a naked-eyeresponse have many advantages because of their ability to providea simple,sensitive,selective,precise and economical method forthe detection of the target analyte without the use of sophisticatedinstrumentation.On the basis of previously reported data of Fe3þchemosensors,herein,we focused on the development of a simple“turn-on”fluorescent chemosensor for Fe3þions from purelyContents lists available at ScienceDirectjournal homepage:/locate/biosBiosensors and Bioelectronics/10.1016/j.bios.2014.06.0170956-5663/&2014Elsevier B.V.All rightsreserved.n Corresponding authors.Tel.:þ2572257432;fax:þ2572258403.E-mail addresses:lxchen@(L.Chen),mahulikarpp@(P.Mahulikar),udpatil.nmu@(U.Patil).Biosensors and Bioelectronics61(2014)612–617aqueous medium.To achieve this goal,we have designed and synthesized a novel benzo-thiazolo-pyrimidine basedfluorescent sensor(BTP-1)for the selective recognition of Fe3þ.The present sensor BTP-1is nontoxic and successfully used to study its in vitro glycosidase inhibitory activity(Patil et al.,2012).2.Material and methods2.1.Materials and instrumentationsAll the starting reagents and metal perchlorates were pur-chased either from S.D.Fine chemicals or Sigma-Aldrich depend-ing on their availability.All the reagents were used as received.All the solvents were of spectroscopic grade and were used without further treatment.The purity of the compounds and the progress of reactions were determined and monitored by means of analy-tical thin layer chromatography(TLC).Pre-coated silica gel60F254(Merck)on alumina plates(7Â3cm2)was used and visualized by using either an iodine chamber or a short UV–Visible lamp. Melting points were recorded on the Celsius scale by the open capillary method and were uncorrected.IR spectra were recorded on a Perkin-Elmer Spectrum One FT-IR spectrometer as potassium bromide pellets and nujol mulls,unless otherwise mentioned.IR bands are expressed in frequency(cmÀ1).NMR spectra were recorded in CDCl3on a Varian(Mercury Vx)SWBB Multinuclear probe spectrometer,operating at300MHz and75MHz for1H NMR and13C NMR,respectively and shifts are given in ppm downfield from TMS as an internal standard.UV–vis spectra were recorded on a U-3900spectrophotometer(Perkin-Elmer Co.,USA) with a quartz cuvette(path length¼1cm).Fluorescence spectra were recorded on a Fluoromax-4spectrofluorometer(HORIBA Jobin Yvon Co.,France).2.2.Spectroscopic studyThe receptor BTP-1was not soluble in water and therefore the stock solutions of BTP-1(1.0Â10À3M)was prepared in CH3CN. All cations(1.0Â10À2M)solutions were prepared in water. These solutions were used for all spectroscopic studies after appropriate dilution.For the spectroscopic(UV–vis andfluores-cence)titrations,the required amount of the diluted receptor BTP-1(2mL,2Â10À5M,in CH3CN)was taken directly into the cuvette and the spectra were recorded after each successive addition of cations(0–180m L,1Â10À3M,in H2O)by using a micropipette.putational methodsAll theoretical calculations were carried out using the Gaussian 09W computer program and the Gaussview5.0.9graphical inter-face(Frisch et al.,2009).Optimization of BTP-1and its complex with Fe3þwas carried out without symmetry constraints by applying the B3LYP/6-31G(d,p)method in the gas phase followed by the harmonic vibrational frequency which was calculated using the same methods to ascertain the presence of a local minimum. The basis set LANL2DZ was used for Fe3þatom.2.4.Living cells imagingThe solution of BTP-1(DMSO, 1.0mM)was prepared and maintained in a refrigerator at4°C.FeCl3is as the iron-supple-mented source.The confocalfluorescent images were acquired on an Olympus laser-scanning microscope with an objective lens (Â40).Excitation of the probe was carried out using a Spectra Physics InSightDeepSee ultrafast laser at700nm and emission was collected between500and600nm.Prior to imaging,the medium was removed.Cell imaging was carried out after washing cells with RPMI-1640for three times.Cell Culture:HepG2cells(Human hepatocellular liver carci-noma cell line)and HL-7701cells(human normal liver cell line) were purchased from the Committee on type Culture Collection of Chinese Academy of Sciences.Cells were seeded at a density of1Â106cells/mL for confocal imaging in RPMI1640medium supplemented with20%fetal bovine serum(FBS),NaHCO3(2g/ L),and1%antibiotics(penicillin/streptomycin,100U/ml).Cultures were maintained at37°C under a humidified atmosphere contain-ing5%CO2.The cells were sub-cultured by scraping and seeding on15mm petri-dishes according to the instructions from the manufacturer.2.5.Synthesis of BTP-1Synthesis of BTP-1was achieved by using a mild base through ring transformation of suitably functionalized4-(methylthio)-2-oxo-6-naphthyl-2H-pyran-3-carbonitriles(2)with2-amino-ben-zothiazole in DMF using DBU as the base(Scheme1)(Patil et al., 2012).The precursors,4-(methylthio)-2-oxo-6-naphthyl-2H-pyran-3-carbonitriles(2)was prepared by stirring an equimolar mixture of ethyl2-cyano-3.3-bis(methylthio)acrylate(1),substi-tuted acetophenone and KOH in DMF(Tominaga et al.,1984).Synthesis of ethyl2-cyano-3.3-bis(methylthio)acrylate(1):In a 250mL beaker,a mixture of water(10mL)and DMF(30mL)and KOH(9.87g,176mmol)were cooled to0°C.Ethyl2-cyanoacetate (10g,88mmol)was added dropwise to this cold solution for 30min,and stirred for10–15min.Then carbon disulfide(6.70g, 88mmol)was added dropwise for20min at0toÀ5°C.The mixture was stirred for60min at room temperature,and then reaction mixture was again cooled to0°C and dimethyl sulfate (22.20g,176mmol)was added dropwise for30min.The reaction mixture was allowed to remain at room temperature for12h and then it was poured into crushed ice cold water(400mL)and kept at room temperature with vigorous stirring for10–15min.The obtained yellowish solid wasfiltered,washed with cold water and dried.The crude product was recrystallized from methanol.Synthesis of6-naphthyl-3-cyano-4-methylthio-2H-pyran-2-ones (2):In a100mL round bottomflask,a mixture of2-acetyl naphthalene(1.7g,0.01mol)and ethyl2-cyano-3.3-bis (methylthio)acrylate(2.17g,0.01mol,1),powdered KOH(1.12g, 0.02mol)and50mL of dry DMF was stirred at room temperature for5–6h.Progress of the reaction was monitored by TLC(ethyl acetate:hexane,3:7).After completion of the reaction,the reaction mixture was poured onto crushed ice(500mL of ice-water)with vigorous stirring and then it was stirred at room temperature for 4–5h.The yellow precipitate formed wasfiltered,washed with NCMeS SMeOOC2H5CH3O OCNSCH3NSH2NNBTP-1rtNSNC12Scheme1.Synthesis of BTP-1.J.Nandre et al./Biosensors and Bioelectronics61(2014)612–617613cold water and dried.The crude product was recrystallized from methanol.Synthesis of BTP-1:A mixture of2-aminobenzothiazole (1mmol)and DBU(2mmol)and DMF(15mL)was stirred under nitrogenflux for10–15min at room temperature.Then4-(methylthio)-2-oxo-6-naphthalen-2H-pyran-3-carbonitrile(2, 1mmol)was added under nitrogenflux with constant stirring. The progress of the reaction was monitored by TLC(ethyl acetate: hexane,2:8).After completion of the reaction(about4–12h),the reaction mixture was poured onto crushed ice with vigorous stirring about30min.The reaction mixture was allowed to remain at room temperature for about20min to settle down the solid, which was the isolated byfiltration.The obtained crude product was dissolved in hot MeOH(70mL)andfiltered.The collected insoluble solid was again dissolved in to hot chloroform(40mL),filtered rapidly(impurity remains insoluble in hot chloroform while product is soluble)and cooled.Chloroform was removed to afford the pure product.Mol.Formula:C22H13N3S;Mol.Weight:351.42g;Physical Nature:Yellow solid;IR(cmÀ1) [KBr]:2923,2178,1599,1526,1267;Mass[ESI,70Ev]m/z(%): 353(30),352(100),291(45),153(25);1H NMR(300MHz,CDCl3,δppm):5.15(s,1H,CH¼C),7.43(s,1H,PhH),7.45–7.55(m,4H, PhH),7.69–7.71(dd,J¼1.3Hz,J¼7.5Hz,1H,PhH),7.86–8.06(m, 5H,PhH),8.58(s,1H,PhH);13C NMR(75MHz,CDCl3,δppm): 64.2,104.4,115.7,119.5,122.9,123.3,125.1,126.4,126.5,127.0,127.2, 127.6,128.2,128.4,128.9,132.2,133.1,134.2,135.5,150.1,151.9, 160.4;HRMS(ESI):m/z:Calculated for C22H14N3S1:[MþH]þ352.0920,Found:352.0903.Results and discussion3.1.Synthesis of BTP-1Synthesis of BTP-1was achieved by using a mild base DBU through ring transformation of a suitably functionalized4-(methylthio)-2-oxo-6-naphthyl-2H-pyran-3-carbonitriles(2)with 2-aminobenzothiazole in DMF(Scheme1).The structure of BTP-1 was characterized by IR,1H-NMR and13C-NMR spectroscopy and HRMS(Figs.S1,S2,S3,S4and S5†).Finally,suitable crystal of BTP-1for a single crystal X-ray diffraction‡study was obtained from chloroform,and the molecular structure is shown in Fig.S6†.3.2.Naked-eye selectivity study of BTP-1The recognition properties of BTP-1(2mL,2Â10À5M,in CH3CN)toward different metal ions(20m L,1Â10À2M,in H2O) were studied experimentally by naked-eye,UV–Visible,andfluor-escence methods.In the naked-eye experiments(Fig.S7†),no obvious visual color changes of BTP-1were observed in the presence of the tested metal ions.However,under UV light,sensor BTP-1showed a selective“turn-on”fluorescence upon addition of Fe3þover other tested metal ions(Agþ,Ca2þ,Cd2þ,Co2þ,Csþ, Cu2þ,Fe2þ,Hg2þ,Kþ,Liþ,Mg2þ,Mn2þ,Ni2þ,Pd2þand Zn2þ). The observedfluorescence intensity reveals that the receptor BTP-1shows higher recognition ability for Fe3þ.Interestingly,this fluorescence“turn-on”response becomes reversibly“turn-off”after the addition of an aq.solution of EDTA.Encouraged by the Fe3þselective and reversible response shown by BTP-1,the quantitative and qualitative metal ions sensing behavior of BTP-1was determined further by spectrophotometric methods.3.3.UV–Visible absorption study of BTP-1Sensor BTP-1exhibited two absorption bands:one at313nm and a weak narrow band at271nm.On addition of5equivalents of Fe3þions(20m L,1Â10À2M,in H2O)to the solution of BTP-1 (2mL,2Â10À5M,in CH3CN),significant spectral changes were observed(Figs.1and S8†).A hypochromic shift was observed at 271nm while the band at313nm was disappeared completely and a new broad band appeared between350nm and425nm.The red-shifted band was observed presumably due to the delocaliza-tion of electrons from the nitrile nitrogen on formation of complex species with Fe3þ(Fig.S9†),which was supported by comparing the variations in the bond lengths of the DFT optimized structure of BTP-1and Fe3þ.BTP-1complex(Fig.S10†).Moreover,the red-shift can also be explained due to the intramolecular charge transfer(ICT)process and the lowering of the band gap between HOMO and LUMO on complexation with Fe3þ(Fig.S11†).This effective ICT induced by the electron push–pull system(Lin et al., 2008).As in–CN group,the presented nitrogen is having sp hybridization so less willing to bind with metal cations as compare to sp2nitrogen and sulfur.Importantly,no distinguishable spectral changes were observed in the presence of other tested metal ions.The absorption titration of BTP-1(2mL,2Â10À5M,in CH3CN) was next performed with incremental addition of Fe3þ(Fig.2). The spectral changes with the formation of an isosbestic point at 328nm indicate the formation of a single complex species between sensor BTP-1and added Fe3þ.From the absorption titration of BTP-1,the limit of detection(LOD)and quantification (LOQ)of Fe3þwere calculated to be0.10m M and0.32m M, respectively based on the ICH Q2B recommendations by using the equations:LOD¼3.3s/S and LOQ¼10s/S,where S and s represent the slope and the standard deviation of the intercept of regression line of the calibration curve,respectively(Fig.S12†). This detection limit is acceptable within the US EPA limit(0.3mg/ L,equivalent to5.4μM)for the detection of Fe3þin drinking water. Further,the Jobs'plot(Fig.S13†)and LCMS(Fig.S14†)analysis was performed,which suggested that there was only one type of1:1 binding interaction between the sensor BTP-1and Fe3þ.3.4.Emission spectroscopic study of BTP-1The cation binding behavior of BTP-1was also investigated by fluorescence spectroscopy.We observed a remarkablefluores-cence enhancement of BTP-1(2mL,2Â10-5M,in CH3CN)at 554nm(λex¼314nm)upon addition of Fe3þ(20m L,1Â10-2M, in H2O)due to the inhibition of photo-induced electron transfer (PET)from electron-donating nitrogen to electron-receptor naphthalene ring(Li et al.,2011),while no significant changes were observed in the presence of other tested metal ions(Figs.3 and S15†).Thefluorescence titration experiment of BTP-1with Fe3þshowed a E10-foldfluorescence enhancement at554nm (Fig.4).The binding constant(K)of BTP-1with Fe3þwas Fig.1.Changes in the absorbance of BTP-1(2mL,2Â10À5M,in CH3CN)in the absence and presence of5equivalents of different metal ions(20m L,1Â10À2M,in H2O).J.Nandre et al./Biosensors and Bioelectronics61(2014)612–617 614Fig.2.Absorbance titration of BTP-1(2Â10À5M,in CH 3CN)upon the addition of incremental amount of Fe 3þ(0–180m L,1Â10À3M,in H 2O).Inset shows the Benesi –Hildebrandplot.Fig.3.Fluorescence emission change of sensor BTP-1(2mL,2Â10À5M,CH 3CN)upon the addition of a particular metal ions (20m L,1Â10À2M,in H 2O),λex ¼314nm.Fig.4.Changes in fluorescence emission intensity of BTP-1(2mL,2Â10À5M,in CH 3CN)upon the addition of incremental amount of Fe 3þ(0–180m L,1Â10À3M,in H 2O).J.Nandre et al./Biosensors and Bioelectronics 61(2014)612–617615determined by a Benesi –Hildebrand plot analysis of both absorp-tion (Fig.2,inset)and fluorescence titrations data (Fig.S16†);additional data was obtained via a Scatchard plot from fluores-cence titration data (Fig.S17†).The cation binding af finity of BTP-1was found to be E 4Â104M À1for Fe 3þ.Based on the fluorescence titration,the LOD and LOQ of BTP-1for Fe 3þwas found to be 0.74nM and 2.23nM,respectively,and these values are quite better than the reported sensors (Table S1†).Further,the effect of coexisting biologically relevant metal ions on the detection of Fe 3þby BTP-1was investigated.In a CH 3CN solution of BTP-1(2mL,2Â10À5M,in CH 3CN),the addition of 2equivalents of Fe 3þ(8m L,1Â10À2M,in water)in the presence of 2equivalents of other tested metal ions (8m L,1Â10À2M,in water)caused a dramatic enhancement in the fluorescence inten-sity of BTP-1with only very slight or no interference effects (Fig.S18†).Therefore,we conclude that BTP-1is a reliable,highly selective and sensitive “turn-on ”fluorescent sensor for Fe 3þ.3.5.Live cells imaging study of BTP-1The fluorescent behavior of BTP-1was applied for the intra-cellular detection and monitoring of Fe 3þin two liver cell lines i.e.,HepG2cells (human hepatocellular liver carcinoma cell line)and HL-7701cells (human normal liver cell line).The HepG2cells in Fig.4b and d were incubated with 0.01μM and 100μM Fe 3þfor 30min in RPMI 1640medium at 37°C,and then washed with RPMI 1640to remove excess Fe 3þ.After being incubated with BTP-1(10μM)in RPMI 1640for 10min,the cells were imaged by a confocal fluorescence microscope.As shown in Fig.5,there is a signi ficant intracellular fluorescence increase revealed in Fig.5b and d compared with the control cells in Fig.5a,which indicates the ability of BTP-1to detect intracellular Fe 3þ.Further,to con firm that the increase in the fluorescence depended on the Fe 3þchanges,the iron-supplemented cells in Fig.5b and 5d were treated with 50μM of iron chelatordesferoxamine (DFO)for 40min to remove the intracellular levels of Fe 3þ.Asexpected,Fig.5.Fluorescence confocal microscopic images of living HepG2cells incubated with Fe 3þ.(a)Cells loaded with 10μM BTP-1for 10min as control.(b)and (d)Cells loaded with 0.1μM and 100μM Fe 3þfor 30min,then 10μM BTP-1for 10min.(c)and (e)Cells were treated as (b),then loaded with 50μM of iron chelatordesferoxamine (DFO)for 40min.(f)and (g)Overlays of bright field images and fluorescence channels in (a)–(e).Scale bar is 20μm.Fig. 6.Fluorescence confocal microscopic images of living HL-7701cells incubated with various concentrations of Fe 3þ.(a)Cells loaded with 50μM of iron chelatordesferoxamine (DFO)for 40min.(b)Cells loaded with 10μM BTP-1for 10min as control.(c)–(h)Cells incubated with 0.01,0.1,1,10,100and 1000μM Fe 3þ,respectively.(i)Quanti fication of mean fluorescence intensity in Fig.S20a –h correspondingly.Scale bar is 20μm.J.Nandre et al./Biosensors and Bioelectronics 61(2014)612–617616the DFO chelation shrunk the cellularfluorescence(Fig.5c and e), indicating that the observedfluorescence enhancements(Fig.5b and d)were due to the changing levels in the Fe3þ-supplemented cells.After establishing that the sensor BTP-1can detect Fe3þwithin living HepG2cells,we turned our attention to quantify the cellular Fe3þlevel changes in HL-7701cells.The relativefluorescence intensity of the confocal microscopy images in Fig.6was evaluated by Image-Pro Plus software(Fig.6i).HL-7701cells were treated with different concentrations of Fe3þ,and then loaded with10μM BTP-1for10min following which the cells were washed three times with RPMI-1640before imaging(Fig.6).Cells in Fig.6b were as control,which showed almost nofluorescence after treated with50μM of DFO(Fig.6a).Our sensor might also detect Fe3þat basal,endogenous levels within cells.Therefore,we selected the cell-body regions in the visualfield(Fig.6a–h)as the region of interest to determine the averagefluorescence intensity.The confocalfluorescence images became gradually brighter as the concentration of Fe3þincreased from0.01μM to100μM Fe3þ(Fig.6c–g),and then thefluorescence intensity becomes saturated after100μM Fe3þ(Fig.6h).Taken together,these quantitative assays established that the sensor BTP-1can be used for the fluorescence detection of Fe3þlevel changes within living cells. The results also suggest that our sensor has good membrane permeability.4.ConclusionsIn conclusion,we have developed a simple benzo-thiazolo-pyrimidine based Fe3þ-selectivefluorescent“turn-on”sensor BTP-1.Sensor BTP-1showed an excellent selectivity for Fe3þover other interfering metal ions with the detection limit down to nanomolar concentration.Also,the sensor works well in the pH range of6–8.Confocal microscopy images indicate that BTP-1can be used for detecting changes in Fe3þlevels within living cells.To the best of our knowledge,this is thefirst example of a chromo-fluorogenic sensor based on benzo-thiazolo-pyrimidine that allows the selective detection of the Fe3þion by afluorescence “turn-on”mode in live cells.Based on the bioactive molecules like benzo-thiazolo-pyrimidine,sensor BTP-1with its low cost and easy preparation,its excellent selectivity and low detection limit, suggests that this approach could potentially lead to many more sensors being designed using the benzo-thiazolo-pyrimidine as a core skeleton.AcknowledgmentsThe author Dr.U.D.Patil is grateful for thefinancial support from the Department of Science&Technology,New Delhi,India (Reg.no.CS-088/2013).We thank the EPSRC National Crystal-lographic Service(Southampton,UK)for data.Appendix A.Supplementary informationSupplementary data associated with this article can be found in the online version at doi:10.1016/j.bios.2014.06.017.ReferencesAu-Yeung,H.Y.,Chan,J.,Chantarojsiri,T.,Chang,C.J.,2013.J.Am.Chem.Soc.135, 15165–15173.Cairo,G.,Pietrangelo,A.,2000.Biochem.J.352,241–250.Callan,J.F.,Silva,A.P.,Magri,D.C.,2005.Tetrahedron61,8551–8588.Chen,X.,Pradhan,T.,Wang,F.,Kim,J.S.,Yoon,J.,2012.Chem.Rev.112,1910–1956. Crabtree,H.H.,1994.Science266,1591–1592.Dutta,M.,Das,D.,2012.Trends Anal.Chem.32,113–132.Fan,L.F.,Wayne,J.,Jones,E.,2006.J.Am.Chem.Soc.128,6784–6785.Formica,M.,Fusi,V.,Giorgi,L.,Micheloni,M.,2012.Coord.Chem.Rev.256,170–192. Frisch,M.J.,et al.,2009.Gaussian09,G09W s.Gaussian Inc.,Wallingford,USA. Jain,R.,Mathur,M.,Sikarwar,S.,Mittal,A.,2007.J.Environ.Manag.85,956–964. Kim,H.N.,Ren,W.X.,Kim,J.S.,Yoon,J.,2012.Chem.Soc.Rev.41,3210–3244.Li,N.,Xu,Q.,Xia,X.,Wang,L.,Lu,J.,Men,X.,2009.Mater.Chem.Phys.114,339–343. Li,Z.-X.,Zhang,L.-F.,Zhao,W.-Y.,Li,X.-Y.,Guo,Y.-K.,Yu,M.-M.,Liu,J.-X.,2011.Inorg.mun.14,1656–1658.Lin,W.,Yuan,L.,Cao,X.,2008.Tetrahedron Lett.49,6585–6588.Lohani,C.R.,Kim,J.-M.,Lee,K.-H.,2009.Bioorgan.Med.Chem.Lett.19,6069–6073. Patil,V.S.,Nandre,K.P.,Ghosh,S.,Rao,V.J.,Chopade,B.A.,Bhosale,S.V.,Bhosale,S.V., 2012.Bioorgan.Med.Chem.Lett.22,7011–7014.Rochat,J.,Demenge,P.,Rerat,J.C.,1978.Toxicol.Eur.Res.1,23–26.Sahoo,S.K.,Sharma,D.,Bera,R.K.,Crisponic,G.,Callan,J.F.,2012.Chem.Soc.Rev.41, 7195–7227.Tominaga,Y.,Ushirogochi,A.,Matsuda,Y.,Kobayashi,G.,1984.Chem.Pharm.Bull.32,3384–3395.J.Nandre et al./Biosensors and Bioelectronics61(2014)612–617617。

FR_ICA1

FR_ICA1

Face Recognition by IndependentComponent AnalysisMarian Stewart Bartlett,Member,IEEE,Javier R.Movellan,Member,IEEE,and Terrence J.Sejnowski,Fellow,IEEEAbstract—A number of current face recognition algorithms use face representations found by unsupervised statistical methods. Typically these methods find a set of basis images and represent faces as a linear combination of those images.Principal compo-nent analysis(PCA)is a popular example of such methods.The basis images found by PCA depend only on pairwise relationships between pixels in the image database.In a task such as face recognition,in which important information may be contained in the high-order relationships among pixels,it seems reasonable to expect that better basis images may be found by methods sensitive to these high-order statistics.Independent component analysis (ICA),a generalization of PCA,is one such method.We used a version of ICA derived from the principle of optimal information transfer through sigmoidal neurons.ICA was performed on face images in the FERET database under two different architectures, one which treated the images as random variables and the pixels as outcomes,and a second which treated the pixels as random variables and the images as outcomes.The first architecture found spatially local basis images for the faces.The second architecture produced a factorial face code.Both ICA representations were superior to representations based on PCA for recognizing faces across days and changes in expression.A classifier that combined the two ICA representations gave the best performance.Index Terms—Eigenfaces,face recognition,independent com-ponent analysis(ICA),principal component analysis(PCA), unsupervised learning.I.I NTRODUCTIONR EDUNDANCY in the sensory input contains structural in-formation about the environment.Barlow has argued that such redundancy provides knowledge[5]and that the role of the sensory system is to develop factorial representations in which these dependencies are separated into independent componentsManuscript received May21,2001;revised May8,2002.This work was supported by University of California Digital Media Innovation Program D00-10084,the National Science Foundation under Grants0086107and IIT-0223052,the National Research Service Award MH-12417-02,the Lawrence Livermore National Laboratories ISCR agreement B291528,and the Howard Hughes Medical Institute.An abbreviated version of this paper appears in Proceedings of the SPIE Symposium on Electronic Imaging:Science and Technology;Human Vision and Electronic Imaging III,Vol.3299,B.Rogowitz and T.Pappas,Eds.,1998.Portions of this paper use the FERET database of facial images,collected under the FERET program of the Army Research Laboratory.The authors are with the University of California-San Diego,La Jolla, CA92093-0523USA(e-mail:marni@;javier@; terry@).T.J.Sejnowski is also with the Howard Hughes Medical Institute at the Salk Institute,La Jolla,CA92037USA.Digital Object Identifier10.1109/TNN.2002.804287(ICs).Barlow also argued that such representations are advan-tageous for encoding complex objects that are characterized by high-order dependencies.Atick and Redlich have also argued for such representations as a general coding strategy for the vi-sual system[3].Principal component analysis(PCA)is a popular unsuper-vised statistical method to find useful image representations. Consider a setofpixels as random variables and the images as outcomes.1Matlab code for the ICA representations is available at /~marni.Face recognition performance was tested using the FERET database [52].Face recognition performances using the ICA representations were benchmarked by comparing them to per-formances using PCA,which is equivalent to the “eigenfaces”representation [51],[57].The two ICA representations were then combined in a single classifier.II.ICAThere are a number of algorithms for performing ICA [11],[13],[14],[25].We chose the infomax algorithm proposed by Bell and Sejnowski [11],which was derived from the principle of optimal information transfer in neurons with sigmoidal transfer functions [27].The algorithm is motivated as follows:Letbeaninvertiblematrix,and-neurons.Each componentof is an invertible squashing function,mapping real numbers intothe(1)The-neurons.The.The goal in Belland Sejnowski’s algorithm is to maximize the mutual informa-tion between theenvironment.The gradient update rule for the weightmatrix,,the ratio between the second andfirst partial derivatives of the activationfunction,,andof this matrixis the derivativeofwith respectto,resulting in the following learning rule[12]:encourages the individual out-puts to move toward statistical independence.When the form1Preliminaryversions of this work appear in [7]and [9].A longer discussionof unsupervised learning for face recognition appears in [6].of the nonlinear transferfunction[12],[42].In practice,thelogistic transfer function has been found sufficient to separate mixtures of natural signals with sparse distributions including sound sources [11].The algorithm is speeded up by including a “sphering”step prior to learning [12].The row meansof(4)This removes the first and the second-order statistics of the data;both the mean and covariances are set to zero and the variances are equalized.When the inputs to ICA are the “sphered”data,the full transformmatrixfor the following generative model of thedata:,the inverse of the weight matrix in Bell and Sejnowski’s algo-rithm,can be interpreted as the source mixing matrix andthevariables can be interpreted as the maximum-likeli-hood (ML)estimates of the sources that generated the data.A.ICA and Other Statistical TechniquesICA and PCA:PCA can be derived as a special case of ICA which uses Gaussian source models.In such case the mixingmatrixis unidentifiable in the sense that there is an infinite number of equally good ML solutions.Among all possible ML solutions,PCA chooses an orthogonal matrix which is optimal in the following sense:1)Regardless of the distributionofwhich areuncorrelatedwith.If the sources are Gaussian,the likelihood of the data depends only on first-and second-order statistics (the covariance matrix).In PCA,the rowsofof natural images,we can scramble their phase spectrum while maintaining their power spectrum.This will dramatically alter the appearance of the images but will not change their second-order statistics.The phase spectrum,not the power spectrum,contains the structural information in images that drives human perception.For example,as illustrated in Fig.1,a face image synthesized from the amplitude spectrum of face A and the phase spectrum of face B will be perceived as an image of face B [45],[53].The fact that PCA is only sensitive to the power spectrum of images suggests that it might not be particularly well suited for representing natural images.The assumption of Gaussian sources implicit in PCA makes it inadequate when the true sources are non-Gaussian.In par-ticular,it has been empirically observed that many natural signals,including speech,natural images,and EEG are better described as linear combinations of sources with long tailed distributions [11],[19].These sources are called “high-kur-tosis,”“sparse,”or “super-Gaussian”sources.Logistic random variables are a special case of sparse source models.When sparse source models are appropriate,ICA has the following potential advantages over PCA:1)It provides a better proba-bilistic model of the data,which better identifies where the data concentratein.3)It finds a not-necessarily orthogonalbasis which may reconstruct the data better than PCA in the presence of noise.4)It is sensitive to high-order statistics in the data,not just the covariance matrix.Fig.2illustrates these points with an example.The figure shows samples from a three-dimensional (3-D)distribution constructed by linearly mixing two high-kurtosis sources.The figure shows the basis vectors found by PCA and by ICA on this problem.Since the three ICA basis vectors are nonorthogonal,they change the relative distance between data points.This change in metric may be potentially useful for classification algorithms,like nearest neighbor,that make decisions based on relative distances between points.The ICA basis also alters the angles between data points,which affects similarity measures such as cosines.Moreover,if an undercomplete basis set is chosen,PCA and ICA may span different subspaces.For example,in Fig.2,when only two dimensions are selected,PCA and ICA choose different subspaces.The metric induced by ICA is superior to PCA in the sense that it may provide a representation more robust to the effect of noise [42].It is,therefore,possible for ICA to be better than PCA for reconstruction in noisy or limited precision environ-ments.For example,in the problem presented in Fig.2,we found that if only 12bits are allowed to represent the PCA and ICA coefficients,linear reconstructions based on ICA are 3dB better than reconstructions based on PCA (the noise power is re-duced by more than half).A similar result was obtained for PCA and ICA subspaces.If only four bits are allowed to represent the first 2PCA and ICA coefficients,ICA reconstructions are 3dB better than PCA reconstructions.In some problems,one can think of the actual inputs as noisy versions of some canon-ical inputs.For example,variations in lighting and expressions can be seen as noisy versions of the canonical image of a person.Having input representations which are robust to noise may po-tentially give us representations that better reflect thedata.Fig.1.(left)Two face images.(Center)The two faces with scrambled phase.(right)Reconstructions with the amplitude of the original face and the phase of the other face.Faces images are from the FERET face database,reprinted with permission from J.Phillips.When the sources models are sparse,ICA is closely related to the so called nonorthogonal “rotation”methods in PCA and factor analysis.The goal of these rotation methods is to find di-rections with high concentrations of data,something very sim-ilar to what ICA does when the sources are sparse.In such cases,ICA can be seen as a theoretically sound probabilistic method to find interesting nonorthogonal “rotations.”ICA and Cluster Analysis:Cluster analysis is a technique for finding regionsinbe a data matrixwithas outcomes (independent trials)of a random experiment.We think of theas the specific value taken by a randomvariableFig.2.(top)Example3-D data distribution and corresponding PC and IC axes.Each axis is a column of the mixing matrix W found by PCA or ICA.Note the PC axes are orthogonal while the IC axes are not.If only two components are allowed,ICA chooses a different subspace than PCA.(bottom left)Distribution of the first PCA coordinates of the data.(bottom right)Distribution of the first ICA coordinates of the data.Note that since the ICA axes are nonorthogonal,relative distances between points are different in PCA than in ICA,as are the angles between points.a distribution.For example,we say that rows ofOur goal in this paper is to find a good set of basis imagesto represent a database of faces.We organize each image in thedatabase as a long vector with as many dimensions as numberof pixels in the image.There are at least two ways in which ICAcan be applied to this problem.1)We can organize our database into a matrixare independent if whenmoving across pixels,it is not possible to predict the valuetaken by the pixel on imageare in the columns ofsponding value taken by pixelFig.4.Image synthesis model for Architecture I.To find a set of IC images,the images in X are considered to be a linear combination of statistically independent basis images,S ,where A is an unknown mixing matrix.The basis images were estimated as the learned ICA output U.Fig.5.Image synthesis model for Architecture II,based on [43]and [44].Each image in the dataset was considered to be a linear combination of underlying basis images in the matrix A .The basis images were each associated with a set of independent “causes,”given by a vector of coefficients in S .The basisimages were estimated by A =Wis the learned ICA weight matrix.individual.The training set was comprised of 50%neutral ex-pression images and 50%change of expression images.The al-gorithms were tested for recognition under three different con-ditions:same session,different expression;different day,same expression;and different day,different expression (see Table I).Coordinates for eye and mouth locations were provided with the FERET database.These coordinates were used to center the face images,and then crop and scale them to60so that the images are in rows and the pixels arein columns,i.e.,such that the rowsof,as shown in Fig.7.These coordinatesare contained in the mixingmatrixlinear combinations of those images,where.Recall that the image synthesis model assumes that the imagesintions (pixels).The use of PCA vectors in the input did not throw away the high-order relationships.These relationships still ex-isted in the data but were not separated.Letindependent source images in the rowsofthat comprised the face imagesinis obtainedbythat comprised,wherewas a minimum squared error approximationofwas,therefore,given by the rows of thematrix3000eigenvec-torsin3000inputmatrixwere updated according to (3)for 1900iterations.The learning rate was initialized at 0.0005and annealed down3B.However,if ICA did not removeall of the second-order dependencies then U will not be precisely orthonormal.4In pilot work,we found that face recognition performance improved with the number of components separated.We chose 200components as the largest number to separate within our processing limitations.5Although PCA already removed the covariances in the data,the variances were not equalized.We,therefore,retained the spheringstep.Fig.8.Twenty-five ICs of the image set obtained by Architecture I,which provide a set of statistically independent basis images (rows of U in Fig.4).ICs are ordered by the class discriminability ratio,r (4).to 0.0001.Training took 90minutes on a Dec Alpha 2100a.Fol-lowing training,a set of statistically independent source images were contained in the rows of the outputmatrixfound by ICA represents a cluster of pixelsthat have similar behavior across images.Each row oftheby the nearest neighbor algorithm,using cosinesas the similarity measure.Coefficient vectors in each test set were assigned the class label of the coefficient vector in the training set that was most similar as evaluated by the cosine of the angle betweenthem.Fig.9.First25PC axes of the image set(columns of P),ordered left to right, top to bottom,by the magnitude of the corresponding eigenvalue.In experiments to date,ICA performs significantly better using cosines rather than Euclidean distance as the similarity measure,whereas PCA performs the same for both.A cosine similarity measure is equivalent to length-normalizing the vectors prior to measuring Euclidean distance when doing nearestneighbor,be the overall meanof acoefficient.For both the PCA and ICA representations,we calculated theratio of between-class to within-classvariabilityclassmeans,andwere calculated separately for each test set,ex-cluding the test images from the analysis.Both the PCA and ICAcoefficients were then ordered by the magnitudeofFig.11.Selection of components by class discriminability,Architecture II. Top:Discriminability of the ICA coefficients(solid lines)and discriminability of the PCA components(dotted lines)for the three test ponents were sorted by the magnitude of r.Bottom:Improvement in face recognition performance for the ICA and PCA representations using subsets of components selected by the class discriminability r.The improvement is indicated by the gray segments at the top of the bars.Face classification performance was compared usingthe,so thatrows represent different pixels and columns represent differentimages.[See(Fig.3right)].This corresponds to treating thecolumnsof.Eachcolumnof(Fig.12).ICA attempts tomake theoutputs,comprised the columns ofthe input data matrix,where each coefficient had zero mean.The Architecture II representation for the training images wastherefore contained in the columnsofwas200for each face image,consisting of the outputsof each of the ICA filters.7The architecture II representation fortest images was obtained in the columnsof.A sample of the basis images is shown6Here,each pixel has zero mean.7An image filter f(x)is defined as f(x)=w1x.Fig.13.Basis images for the ICA-factorial representation(columns of AAA(a)(b)Fig.16.Pairwise mutual information.(a)Mean mutual information between basis images.Mutual information was measured between pairs of gray-level images,PC images,and independent basis images obtained by Architecture I.(b)Mean mutual information between coding variables.Mutual information was measured between pairs of image pixels in gray-level images,PCA coefficients, and ICA coefficients obtained by Architecture II.obtained85%,56%,and44%correct,respectively.Again,as found for200separated components,selection of subsets of components by class discriminability improved the performance of ICA1to86%,78%,and65%,respectively,and had little ef-fect on the performances with the PCA and ICA2representa-tions.This suggests that the results were not simply an artifact due to small sample size.VI.E XAMINATION OF THE ICA R EPRESENTATIONSA.Mutual InformationA measure of the statistical dependencies of the face repre-sentations was obtained by calculating the mean mutual infor-mation between pairs of50basis images.Mutual information was calculatedas(18)where.Again,therewere considerable high-order dependencies remaining in thePCA representation that were reduced by more than50%by theinformation maximization algorithm.The ICA representationsobtained in these simulations are most accurately described notas“independent,”but as“redundancy reduced,”where the re-dundancy is less than half that in the PC representation.B.SparsenessField[19]has argued that sparse distributed representationsare advantageous for coding visual stimuli.Sparse representa-tions are characterized by highly kurtotic response distributions,in which a large concentration of values are near zero,with rareoccurrences of large positive or negative values in the tails.Insuch a code,the redundancy of the input is transformed intothe redundancy of the response patterns of the the individualoutputs.Maximizing sparseness without loss of information isequivalent to the minimum entropy codes discussed by Barlow[5].8Given the relationship between sparse codes and minimumentropy,the advantages for sparse codes as outlined by Field[19]mirror the arguments for independence presented byBarlow[5].Codes that minimize the number of active neuronscan be useful in the detection of suspicious coincidences.Because a nonzero response of each unit is relatively rare,high-order relations become increasingly rare,and therefore,more informative when they are present in the stimulus.Field8Information maximization is consistent with minimum entropy coding.Bymaximizing the joint entropy of the output,the entropies of the individual out-puts tend to be minimized.Fig.18.Recognition successes and failures.{left)Two face image pairs which both ICA algorithms correctly recognized.(right)Two face image pairs that were misidentified by both ICA algorithms.Images from the FERET face database were reprinted with permission from J.Phillips.contrasts this with a compact code such as PCs,in which a few units have a relatively high probability of response,and there-fore,high-order combinations among this group are relatively common.In a sparse distributed code,different objects are rep-resented by which units are active,rather than by how much they are active.These representations have an added advantage in signal-to-noise,since one need only determine which units are active without regard to the precise level of activity.An ad-ditional advantage of sparse coding for face representations is storage in associative memory works with sparse inputs can store more memories and provide more effective re-trieval with partial information[10],[47].The probability densities for the values of the coefficients of the two ICA representations and the PCA representation are shown in Fig.17.The sparseness of the face representations were examined by measuring the kurtosis of the distributions. Kurtosis is defined as the ratio of the fourth moment of the dis-tribution to the square of the second moment,normalized to zero for the Gaussian distribution by subtracting3kurtosis,,,correspond to the similaritymeasure.Performance of the combined classifier is shown in Fig.19.Thecombined classifier improved performance to91.0%,88.9%,and81.0%for the three test cases,respectively.The difference inperformance between the combined ICA classifier and PCA wassignificant for all three test sets(,,;auditory signals into independent sound sources.Under this ar-chitecture,ICA found a basis set of statistically independent im-ages.The images in this basis set were sparse and localized in space,resembling facial features.Architecture II treated pixels as random variables and images as random trials.Under this ar-chitecture,the image coefficients were approximately indepen-dent,resulting in a factorial face code.Both ICA representations outperformed the“eigenface”rep-resentation[57],which was based on PC analysis,for recog-nizing images of faces sampled on a different day from the training images.A classifier that combined the two ICA rep-resentations outperformed eigenfaces on all test sets.Since ICA allows the basis images to be nonorthogonal,the angles and dis-tances between images differ between ICA and PCA.Moreover, when subsets of axes are selected,ICA defines a different sub-space than PCA.We found that when selecting axes according to the criterion of class discriminability,ICA-defined subspaces encoded more information about facial identity than PCA-de-fined subspaces.ICA representations are designed to maximize information transmission in the presence of noise and,thus,they may be more robust to variations such as lighting conditions,changes in hair,make-up,and facial expression,which can be considered forms of noise with respect to the main source of information in our face database:the person’s identity.The robust recogni-tion across different days is particularly encouraging,since most applications of automated face recognition contain the noise in-herent to identifying images collected on a different day from the sample images.The purpose of the comparison in this paper was to examine ICA and PCA-based representations under identical conditions.A number of methods have been presented for enhancing recognition performance with eigenfaces(e.g.,[41]and[51]). ICA representations can be used in place of eigenfaces in these techniques.It is an open question as to whether these techniques would enhance performance with PCA and ICA equally,or whether there would be interactions between the type of enhancement and the representation.A number of research groups have independently tested the ICA representations presented here and in[9].Liu and Wech-sler[35],and Yuen and Lai[61]both supported our findings that ICA outperformed PCA.Moghaddam[41]employed Euclidean distance as the similarity measure instead of cosines.Consistent with our findings,there was no significant difference between PCA and ICA using Euclidean distance as the similarity mea-sure.Cosines were not tested in that paper.A thorough compar-ison of ICA and PCA using a large set of similarity measures was recently conducted in[17],and supported the advantage of ICA for face recognition.In Section V,ICA provided a set of statistically independent coefficients for coding the images.It has been argued that such a factorial code is advantageous for encoding complex objects that are characterized by high-order combinations of features, since the prior probability of any combination of features can be obtained from their individual probabilities[2],[5].According to the arguments of both Field[19]and Barlow[5],the ICA-fac-torial representation(Architecture II)is a more optimal object representation than the Architecture I representation given its sparse,factorial properties.Due to the difference in architec-ture,the ICA-factorial representation always had fewer training samples to estimate the same number of free parameters as the Architecture I representation.Fig.16shows that the residual de-pendencies in the ICA-factorial representation were higher than in the Architecture I representation.The ICA-factorial repre-sentation may prove to have a greater advantage given a much larger training set of images.Indeed,this prediction has born out in recent experiments with a larger set of FERET face im-ages[17].It also is possible that the factorial code representa-tion may prove advantageous with more powerful recognition engines than nearest neighbor on cosines,such as a Bayesian classifier.An image set containing many more frontal view im-ages of each subject collected on different days will be needed to test that hypothesis.In this paper,the number of sources was controlled by re-ducing the dimensionality of the data through PCA prior to per-forming ICA.There are two limitations to this approach[55]. The first is the reverse dimensionality problem.It may not be possible to linearly separate the independent sources in smaller subspaces.Since we retained200dimensions,this may not have been a serious limitation of this implementation.Second,it may not be desirable to throw away subspaces of the data with low power such as the higher PCs.Although low in power,these sub-spaces may contain ICs,and the property of the data we seek is independence,not amplitude.Techniques have been proposed for separating sources on projection planes without discarding any ICs of the data[55].Techniques for estimating the number of ICs in a dataset have also recently been proposed[26],[40]. The information maximization algorithm employed to per-form ICA in this paper assumed that the underlying“causes”of the pixel gray-levels in face images had a super-Gaussian (peaky)response distribution.Many natural signals,such as sound sources,have been shown to have a super-Gaussian distribution[11].We employed a logistic source model which has shown in practice to be sufficient to separate natural signals with super-Gaussian distributions[11].The under-lying“causes”of the pixel gray-levels in the face images are unknown,and it is possible that better results could have been obtained with other source models.In particular,any sub-Gaussian sources would have remained mixed.Methods for separating sub-Gaussian sources through information maximization have been developed[30].A future direction of this research is to examine sub-Gaussian components of face images.The information maximization algorithm employed in this work also assumed that the pixel values in face images were generated from a linear mixing process.This linear approxima-tion has been shown to hold true for the effect of lighting on face images[21].Other influences,such as changes in pose and ex-pression may be linearly approximated only to a limited extent. Nonlinear ICA in the absence of prior constraints is an ill-condi-tioned problem,but some progress has been made by assuming a linear mixing process followed by parametric nonlinear func-tions[31],[59].An algorithm for nonlinear ICA based on kernel methods has also recently been presented[4].Kernel methods have already shown to improve face recognition performance。

随机volterra方程的数值解法

随机volterra方程的数值解法

随机volterra方程的数值解法Numerical methods for solving Volterra equations have been a subject of intense research in recent years. Volterra equations, also known as integral equations of the second kind, arise in various fields of science and engineering, ranging from physics to biology. Their complexity often necessitates the use of numerical techniques for approximate solutions.近年来,Volterra方程的数值解法一直是研究的热点。

Volterra方程,也称为第二类积分方程,出现在科学和工程的各个领域,从物理学到生物学都有涉及。

由于其复杂性,通常需要采用数值技术来求解近似解。

One of the most common numerical methods for Volterra equations is the discretization technique. This method involves dividing the integration interval into smaller subintervals and approximating the integrand at each subinterval by a suitable function, such as a polynomial. By integrating these approximate functions over the subintervals, an approximate solution to the Volterra equation can be obtained.求解Volterra方程最常用的数值方法之一是离散化技术。

Tikhonov吉洪诺夫正则化

Tikhonov吉洪诺夫正则化

Tikhonov regularizationFrom Wikipedia, the free encyclopediaTikhonov regularization is the most commonly used method of of named for . In , the method is also known as ridge regression . It is related to the for problems.The standard approach to solve an of given as,b Ax =is known as and seeks to minimize the2bAx -where •is the . However, the matrix A may be or yielding a non-unique solution. In order to give preference to a particular solution with desirable properties, the regularization term is included in this minimization:22xb Ax Γ+-for some suitably chosen Tikhonov matrix , Γ. In many cases, this matrix is chosen as the Γ= I , giving preference to solutions with smaller norms. In other cases, operators ., a or a weighted ) may be used to enforce smoothness if the underlying vector is believed to be mostly continuous. This regularizationimproves the conditioning of the problem, thus enabling a numerical solution. An explicit solution, denoted by , is given by:()b A A A xTTT 1ˆ-ΓΓ+=The effect of regularization may be varied via the scale of matrix Γ. For Γ=αI , when α = 0 this reduces to the unregularized least squares solution providedthat (A T A)−1 exists.Contents••••••••Bayesian interpretationAlthough at first the choice of the solution to this regularized problem may look artificial, and indeed the matrix Γseems rather arbitrary, the process can be justified from a . Note that for an ill-posed problem one must necessarily introduce some additional assumptions in order to get a stable solution.Statistically we might assume that we know that x is a random variable with a . For simplicity we take the mean to be zero and assume that each component isindependent with σx. Our data is also subject to errors, and we take the errorsin b to be also with zero mean and standard deviation σb. Under these assumptions the Tikhonov-regularized solution is the solution given the dataand the a priori distribution of x, according to . The Tikhonov matrix is then Γ=αI for Tikhonov factor α = σb/ σx.If the assumption of is replaced by assumptions of and uncorrelatedness of , and still assume zero mean, then the entails that the solution is minimal . Generalized Tikhonov regularizationFor general multivariate normal distributions for x and the data error, one can apply a transformation of the variables to reduce to the case above. Equivalently,one can seek an x to minimize22Q P x x b Ax -+-where we have used 2P x to stand for the weighted norm x T Px (cf. the ). In the Bayesian interpretation P is the inverse of b , x 0 is the of x , and Q is the inverse covariance matrix of x . The Tikhonov matrix is then given as a factorization of the matrix Q = ΓT Γ. the ), and is considered a . This generalized problem can be solved explicitly using the formula()()010Ax b P A QPA A x T T-++-[] Regularization in Hilbert spaceTypically discrete linear ill-conditioned problems result as discretization of , and one can formulate Tikhonov regularization in the original infinite dimensional context. In the above we can interpret A as a on , and x and b as elements in the domain and range of A . The operator ΓΓ+T A A *is then a bounded invertible operator.Relation to singular value decomposition and Wiener filterWith Γ= αI , this least squares solution can be analyzed in a special way viathe . Given the singular value decomposition of AT V U A ∑=with singular values σi , the Tikhonov regularized solution can be expressed asb VDU xT =ˆ where D has diagonal values22ασσ+=i i ii Dand is zero elsewhere. This demonstrates the effect of the Tikhonov parameteron the of the regularized problem. For the generalized case a similar representation can be derived using a . Finally, it is related to the :∑==qi iiT i i v bu f x1ˆσwhere the Wiener weights are 222ασσ+=i i i f and q is the of A .Determination of the Tikhonov factorThe optimal regularization parameter α is usually unknown and often in practical problems is determined by an ad hoc method. A possible approach relies on the Bayesian interpretation described above. Other approaches include the , , , and . proved that the optimal parameter, in the sense of minimizes:()()[]21222ˆTTXIX XX I Tr y X RSSG -+--==αβτwhereis the and τ is the effective number .Using the previous SVD decomposition, we can simplify the above expression:()()21'22221'∑∑==++-=qi iiiqi iiub u ub u y RSS ασα()21'2220∑=++=qi iiiub u RSS RSS ασαand∑∑==++-=+-=qi iqi i i q m m 12221222ασαασστRelation to probabilistic formulationThe probabilistic formulation of an introduces (when all uncertainties are Gaussian) a covariance matrix C M representing the a priori uncertainties on the model parameters, and a covariance matrix C D representing the uncertainties on the observed parameters (see, for instance, Tarantola, 2004 ). In the special case when these two matrices are diagonal and isotropic,and, and, in this case, the equations of inverse theory reduce to theequations above, with α = σD / σM .HistoryTikhonov regularization has been invented independently in many differentcontexts. It became widely known from its application to integral equations from the work of and D. L. Phillips. Some authors use the term Tikhonov-Phillips regularization . The finite dimensional case was expounded by A. E. Hoerl, who took a statistical approach, and by M. Foster, who interpreted this method as a - filter. Following Hoerl, it is known in the statistical literature as ridge regression .[] References•(1943). "Об устойчивости обратных задач [On the stability of inverse problems]". 39 (5): 195–198.•Tychonoff, A. N. (1963). "О решении некорректно поставленных задач и методе регуляризации [Solution of incorrectly formulated problems and the regularization method]". Doklady Akademii Nauk SSSR151:501–504.. Translated in Soviet Mathematics4: 1035–1038. •Tychonoff, A. N.; V. Y. Arsenin (1977). Solution of Ill-posed Problems.Washington: Winston & Sons. .•Hansen, ., 1998, Rank-deficient and Discrete ill-posed problems, SIAM •Hoerl AE, 1962, Application of ridge analysis to regression problems, Chemical Engineering Progress, 58, 54-59.•Foster M, 1961, An application of the Wiener-Kolmogorov smoothing theory to matrix inversion, J. SIAM, 9, 387-392•Phillips DL, 1962, A technique for the numerical solution of certain integral equations of the first kind, J Assoc Comput Mach, 9, 84-97•Tarantola A, 2004, Inverse Problem Theory (), Society for Industrial and Applied Mathematics,•Wahba, G, 1990, Spline Models for Observational Data, Society for Industrial and Applied Mathematics。

210975034_改性钛基二氧化铅电极催化氧化降解水中四环素

210975034_改性钛基二氧化铅电极催化氧化降解水中四环素

第 43 卷第 3 期2023年 3 月Vol.43 No.3Mar.,2023工业水处理Industrial Water Treatment DOI :10.19965/ki.iwt.2022-0416改性钛基二氧化铅电极催化氧化降解水中四环素张志军1,成鹏2,谢智翔2,束蒋成2(1.常州大学怀德学院建筑与环境工程系,江苏靖江 214500;2.常州大学环境科学与工程学院,江苏常州 213164)[ 摘要 ] 使用NiO 做中间层,并掺杂稀土元素(La )对Ti/PbO 2电极进行改性,制备了Ti/NiO/PbO 2-La 电极。

采用扫描电子显微镜、X 射线能谱仪和X 光电子能谱对电极表面进行表征,结果表明Ti/NiO/PbO 2-La 电极表层形貌致密、规整,具有较大的活性比表面积,因此具有较Ti/PbO 2电极更多的活性位点,且La 以La 2O 3的状态被掺杂在PbO 2中。

通过循环伏安法(CV )、线性伏安法(LSV )和交流阻抗(EIS )一系列电化学分析,证实以NiO 修饰钛基、活性层掺杂La 均可以改善电极的电化学性能,且二者具有协同作用。

将Ti/NiO/PbO 2-La 电极用于对四环素(TC )的电催化氧化,初步分析了溶液pH 、电流密度和TC 初始浓度对电催化氧化反应的影响,结果表明,Ti/NiO/PbO 2-La 电极对TC 的电催化降解过程符合准一级动力学模型,当溶液pH 为4.5,电流密度为20 mA/cm 2,TC 初始质量浓度为10 mg/L 时,反应90 min 后,TC 去除率可高达98.4%。

[关键词] 电化学沉积;阳极;电催化氧化;准一级动力学模型[中图分类号] X703.1 [文献标识码]A [文章编号] 1005-829X (2023)03-0071-09Catalytic oxidation degradation of tetracycline in water bymodified titanium -based lead dioxide electrodeZHANG Zhijun 1,CHENG Peng 2,XIE Zhixiang 2,SHU Jiangcheng 2(1.Department of Architecture and Environmental Engineering ,Huaide College ,Changzhou University ,Jingjiang 214500,China ;2.School of Environmental Science and Engineering ,Changzhou University ,Changzhou 213164,China )Abstract :Ti/NiO/PbO 2-La electrode was prepared by modifying Ti/PbO 2 electrode with NiO as interlayer and doped with rare -earth element (La ). The surface of the electrode was characterized by scanning electron microscope ,X -ray spectrometer and X -ray photoelectron spectroscopy. The results showed that the surface morphology of the Ti/NiO/PbO 2-La electrode was compact and regular ,with a large active specific surface area ,so it had more active sites than the Ti/PbO 2 electrode ,and La was doped in PbO 2 as the state of La 2O 3. Through a series of electrochemical analysis by cyclic voltammetry (CV ),linear voltammetry (LSV ) and AC impedance (EIS ),it was confirmed that the electro⁃chemical performance of the electrode could be improved by NiO modified titanium base and the active layer doped with La ,which had synergistic effect.Ti/NiO/PbO 2-La electrode was used for the electro -catalytic oxidation of TC ,and the effects of solution pH ,current density and initial concentration of TC on the electro -catalytic oxidation reac⁃tion were preliminarily analyzed. The results showed that the electro -catalytic degradation of TC by Ti/NiO/PbO 2-La electrode was in accordance with the quasi -first -order kinetic model. Under the conditions of the solution pH 4.5,current density 20 mA/cm 2,initial mass concentration of TC 10 mg/L ,and the reaction time 90 minutes ,the removal rate of TC could reach 98.4%.Key words :electrochemical deposition ;anode ;electrocatalytic oxidation ;quasi -first -order dynamic model开放科学(资源服务)标识码(OSID ):试验研究工业水处理2023-03,43(3)抗生素作为一种能够抑制和杀灭细菌的药物,已经被广泛应用于医疗卫生和畜禽养殖等领域。

picard_iteration

picard_iteration

PICARD ITERATIONDAVID SEALThe differential equation we’re interested in studying is(1)y′=f(t,y),y(t0)=y0.Manyfirst order differential equations fall under this category and the following method is a new method for solving this differential equation.Thefirst idea is to transform the DE into an integral equation,and then apply a new method to the integral equation.Wefirst do a change of variables to transform the initial conditions to the origin. Explicitly,you can define w=y−y0and x=t−t0.With a new f,the differential equation we’ll study is given by(2)y′=f(t,y),y(0)=0.Note:it’s not necessary to do this substitution,but it makes life a lot easier if we do.Now,we integrate equation(2)from s=0to s=t to obtainty′(s)ds= t s=0f(s,y(s))ds.s=0Applying the fundamental theorem of calculus,we havety′(s)ds=y(t)−y(0)=y(t).s=0Hence we reduced the differential equation to an equivalent integral equation given by(3)y(t)= t s=0f(s,y(s))ds.Even though this looks like it’s‘solved’,it really isn’t because the function y is buried inside the integrand.To solve this,we attempt to use the following algo-rithm,known as Picard Iteration:(1)Choose an initial guess,y0(t)to equation(3).(2)For n=1,2,3,...,set y n+1(t)= t s=0f(s,y n(s))dsWhy does this make sense?If you take limits of both sides,and note that y(t)= lim n y n+1=lim n y n,then y(t)is a solution to the integral equation,and hence a solution to the differential equation.The next question you should ask is under what hypotheses on f does this limit exist?It turns out that sufficient hypotheses are the f and f y be continuous at(0,0).These are exactly the hypotheses given in your existence/uniqueness theorem2.Note:If we stop this algorithm at afinite value of n,we expect y n(t)to be a very good approximate solution to the differential equation.This makes this method of iteration an extremely powerful tool for solving differential equations!For a concrete example,I’ll show you how to solve problem#3from section 2−8.Use the method of picard iteration with an initial guess y0(t)=0to solve:y′=2(y+1),y(0)=0.Note that the initial condition is at the origin,so we just apply the iteration to this differential equation.y1(t)= t s=0f(s,y0(s))ds= t s=02(y0(s)+1)ds= t s=02ds=2t. Hence,we have thefirst guess is y1(t)=2t.Next,we iterate once more to get y2: y2(t)= t s=0f(s,y1(s))ds= t s=02(y1(s)+1)ds= t s=02(2s+1)ds=222!t2+2t.Iterate again to get y3:y3(t)= t s=02(y2(s)+1)ds= t s=02 223!+(2t)2k!and hence the exact solution is given byy(t)=limn→∞y n(t)=∞k=1(2t)k k!−1=e2t−1.If you plug this into the differential equation,you’ll see we hit this one on the money.To demonstrate this solution actually works,below is a graph of y5(t), y15(t)and y(t),the exact solution.Approximate vs. Exact Solution。

Abstract Structure of lipid bilayers

Abstract Structure of lipid bilayers

Structure of lipid bilayersJohn F.Nagle a Y b Y*,Stephanie Tristram-Nagle ba Department of Physics,Carnegie Mellon University,Pittsburgh,PA15213,USAb Department of Biological Sciences,Carnegie Mellon University,Pittsburgh,PA15213,USAReceived10May2000;received in revised form22August2000;accepted22August2000AbstractThe quantitative experimental uncertainty in the structure of fully hydrated,biologically relevant,fluid(L K)phase lipid bilayers has been too large to provide a firm base for applications or for comparison with simulations.Many structural methods are reviewed including modern liquid crystallography of lipid bilayers that deals with the fully developed undulation fluctuations that occur in the L K phase.These fluctuations degrade the higher order diffraction data in a way that,if unrecognized,leads to erroneous conclusions regarding bilayer structure.Diffraction measurements at high instrumental resolution provide a measure of these fluctuations.In addition to providing better structural determination,this opens a new window on interactions between bilayers,so the experimental determination of interbilayer interaction parameters is reviewed briefly.We introduce a new structural correction based on fluctuations that has not been included in any previous studies.Updated measurements,such as for the area compressibility modulus,are used to provide adjustments to many of the literature values of structural quantities.Since the gel(L L P)phase is valuable as a stepping stone for obtaining fluid phase results,a brief review is given of the lower temperature phases.The uncertainty in structural results for lipid bilayers is being reduced and best current values are provided for bilayers of five lipids.ß2000Elsevier Science B.V.All rights reserved. Keywords:Lipid bilayer;X-ray di¡raction;Structure determination;Fluctuation;Hydration;Interaction1.IntroductionThis is a review of the venerable,but still active, topic of lipid bilayer structure.Lipid bilayer struc-tural data are used for a variety of purposes in bio-physics,such as consideration of hydrophobic matching of intrinsic membrane proteins.We shall not attempt to review all the applications,but will concentrate instead on providing reliable data for general use.This project deserves considerable dis-cussion and analysis.However,the user in a hurry can¢nd our current bottom-line values in Table6in Section12as well as comparison values in Tables3 and5.This review is closest in content to the in£uential BBA review of Rand and Parsegian published over 10years ago[1].Although that review emphasized bilayer interactions,extensive tables of structural data for many bilayers were given.In comparison, the present review includes fewer lipid bilayers.We emphasize and compare the di¡erent results obtained by di¡erent methods for some of the most popular lipids,DPPC,DMPC,DOPC,EPC and DLPE. Much of the di¤culty in obtaining good quantita-tive structure for the biologically relevant,fully hy-drated,£uid(L K)phase is due to the intrinsic pres-0304-4157/00/$^see front matterß2000Elsevier Science B.V.All rights reserved. PII:S0304-4157(00)00016-2ence of £uctuations.A related topic is the interac-tions between bilayers.Interactions are connected with structure determination because interactions are present in the most useful,multilamellar vesicle (MLV),samples which are used to determine struc-ture.On the other hand,quantitative structure is a precursor to quantitative evaluation of interactions.Structure determination and interactions are also connected because £uctuations play a central role in both.However,to avoid undue length,this review will focus primarily on structure.This review focuses on experimental methods for obtaining bilayer structure.An alternative is com-puter simulations.This alternative is becoming in-creasingly attractive with the rapid progress in simu-lations because the level of detail is so much greater than can be obtained experimentally.This detail can even be a guide to the interpretation of experimental results [2,3].Of course,simulations are no better than the models (force ¢elds)that are simulated,and sometimes worse because of limitations to small systems and short times.Reliable experimental data,though incomplete,provides a guide to modeling and a necessary check on the reliability of simulations.At this point some readers may challenge our as-sertion that lipid bilayer structure should still be con-sidered an active area.It has a long and rich history.Many prominent biophysicists have published in it and moved ers of bilayer structural data have many references to choose from and each user has a favorite.Such a reader should examine Fig.1which shows literature values for a particularly cen-tral quantity,namely,the average interfacial area A per lipid molecule for DPPC bilayers at 50³C in the biologically relevant,fully hydrated,£uid (F ,synon-ymously,the L K or liquid crystalline)phase.Such scatter cannot be attributed to sample di¡erence since DPPC has been synthesized to high purity for 25years.The scatter in these A F DPPC values,all for the same state of the same lipid,is unacceptably large for guiding computer simulations,which are sensitiveto di¡erences of about 1A î2.The scatter in A F DPPC iseven larger when viewed from the perspective of comparing to the gel (G)phase,for whichA G DPPC =47.9A î2[4].The di¡erence,A F DPPC 3A G DPPC ,measures the e¡ect of £uidization which is what makes the bilayer biologically relevant.If one em-ploys the intellectually impoverished method of ob-taining a value by uncritically averaging all literature values,one would still face an uncertainty inA F DPPC 3A GDPPC at the 50%level.The mean thickness of the bilayer is also inversely proportional to A and is therefore subject to comparable scatter that de-grades important quantitative discussions of hydro-phobic matching [5^8].This review will hopefully convince the reader that structural quantities are no longer so poorly determined as indicated by Fig.1.This will involve a critical review of many of the methods that gave those results.In addition,in Sec-tion 7,we introduce a new correction based on £uc-tuations that has not been included in any of the previous analyses,including our own;using this cor-rection we provide adjustments to literature values of A .We also use new values of material moduli [17]to revise some of the earlier structural values given by ourselves and by Rand and Parsegian [1].Although everyone agrees that the £uid L K phase is the most important one for biology,the so-called gel (L L P )phase is valuable as a stepping stone for obtaining £uid phase results,so results for other,more ordered,lamellar phases are brie£y reviewed in Section 11.In Section 6a brief survey is given of recent work on the e¡ects of £uctuations on the determination of interbilayer interactions.First,we turn in the next section to what one can hope to achieve for the structure of lipid bilayers and we de¢ne some of the terms that are used.2.What is meant by lipid bilayer structure?It is often supposed that determining bilayer struc-ture by di¡raction means doingcrystallography.Fig.1.Summary of published areas for £uid phase DPPC at 50³C (black)and gel phase DPPC (grey)at 20³C.References:aSun et al.[4],b Pace and Chan [9],c Bu «ldt et al.[10],d Schindler and Seelig [11],e Nagle et al.[3],f Lewis and Engelman [12],gRand and Parsegian [1]and Janiak et al.[13],h DeYoung and Dill [14],i Lis et al.[15],j Thurmond et al.[16].J.F.Nagle,S.Tristram-Nagle /Biochimica et Biophysica Acta 1469(2000)159^195160While lipid crystallography has been pursued and is illuminating [18],it is important to recognize that fully hydrated lipid bilayers are not even close to being in a crystalline state.The contrast is strongest for bilayers that are in the £uid,L K phase where the hydrocarbon chains are conformationally disordered in contrast to the nearly all-trans chains in lipid crys-tals.Even for the conformationally ordered gel and subgel bilayer phases,there are substantial di¡eren-ces compared to the crystal structures.These di¡er-ences are not surprising since there is much more water in fully hydrated lipid bilayers,which substan-tially alters the balance of interaction energies of the bilayers compared to the nearly dry crystalline state and which also allows for increased £uctuations.Be-cause of the £uctuations,it makes no sense to con-template an atomic level structure for biologically relevant lipid bilayers [19].The absence of such struc-tures should not be blamed on poor di¡raction tech-nique or on sample preparation;rather,such struc-tures simply do not exist in the biologically relevant state.The appropriate description for the positions of atoms in the lipid molecule is that of broad statistical distribution functions.Fig.2a shows simulations for distribution functions for the component groups of DPPC along the direction of the bilayer normal [20].Most users of such information focus on the peak positions of the distributions.Equally impor-tant are the shapes of the distributions.At ¢rst glance,one would simply describe the shapes by their widths;in Fig.2a the full widths at half max-imum are of order 5Aî.However,one should also realize that such distributions are only Gaussians if the potential of mean force happens to be harmonic,and this would be strictly accidental.Non-Gaussian and skewed distributions occur most certainly for the terminal methyl distribution for methyls limited to lipids in one monolayer [21^23](the distribution in Fig.2a is automatically symmetric because it in-cludes methyls from both monolayers).Skewness warns one that the average position of a component group is not necessarily the position of the maximum in the distribution.Of course,if one is trying to¢tFig.2.Three representations of structure of DPPC bilayers in the L K £uid phase.(a)Probability distribution functions p for di¡erent component groups from simulations [20]and the downward pointing arrows show the peak locations determined by neutron di¡rac-tion with 25%water [10].The equality of the areas denoted K and L locates the Gibbs dividing surface for the hydrocarbon region de-termined by the simulation.(b)Electron density pro¢le b *from X-ray studies (solid line)[3]and from simulations (dots)(contributed by Scott Feller).(c)Two volumetric pictures.The version on the left monolayer is a simple three compartment representation.The version on the right monolayer is a more realistic representation of the interfacial headgroup region [26].D C is the experimentallydetermined Gibbs dividing surface for the hydrocarbon region.The x -axis is in Aîalong the bilayer normal with the same scale for a,b and c.The y -axis in c shows a lateral dimension along the surface of the bilayer.Values for the parameters in c are taken from Table 6.J.F.Nagle,S.Tristram-Nagle /Biochimica et Biophysica Acta 1469(2000)159^195161limited amounts of data,it is convenient to limit the ¢tting functions to Gaussians that are parameterized just by a mean position and a width.The errors in making this approximation have been assessed and improvements are indicated when the Gaussian as-sumption is not made,although for volumetric ap-plications the improvements are not large[22].How-ever,there is a di¡erent application,namely,for the positions of methylenes as a function of carbon num-ber,where using the most probable(peak)value in the non-Gaussian distribution gives di¡erent values and a di¡erent qualitative picture than using true ing averages shows that the mean dis-tance between successive methylenes decreases to-wards the methyl end[2];this is consistent with the usual picture of increasing disorder towards the bi-layer center.In contrast,using peak values in the distribution suggests wrongly that the successive distances are nearly constant(we are indebted to R.G.Snyder for bringing this example to our atten-tion).So far,the description has been exclusively along the spatial direction of the bilayer normal.In con-trast,in the lateral direction along the bilayer,the distribution functions for the L K phase are just con-stants because the lipid molecules are in a two-di-mensional£uid phase.For the lower temperature phases,however,there is interesting and valuable in-plane structure[4,24]which is reviewed in Section 12.Fluctuations in fully hydrated£uid phase bilayers mean that X-ray di¡raction data from multilamellar arrays of lipid bilayers can only yield electron density pro¢les(EDP)such as the one shown in Fig.2b.The peaks in this DPPC electron density pro¢le are asso-ciated with the electron dense phosphate group and the lower electron density in the center is associated with the hydrocarbon region and especially with the low electron density of terminal methyl groups of the fatty acids.Therefore,electron density pro¢les con-¢rm the usual picture of bilayer structure and they give a measure of the bilayer thickness,namely,the head^head separation,D HH.However,electron den-sity pro¢les only provide a good measure for the location of the phosphate rmation about the z-coordinates of other groups has been obtained using neutron di¡raction,reviewed in Section8,ei-ther with selective deuteration of various component groups(see the arrows in Fig.2a)([10],see p.689), or combined with X-ray di¡raction[25].The transverse description of the bilayer as a set of distribution functions along the z-axis is valuable, but it does not include other important information, such as A in the lateral direction,or the volumes of component groups of the lipid molecule.Therefore,a complementary description of bilayer structure is ap-propriate[26].The simplest such description,due to Luzzati[27]is shown on the left half of Fig.2c.For multilamellar arrays with repeat spacing D the vol-ume is divided into two regions.The¢rst region consists of the volume V L of the lipid and the second region consists of the volume n W V W of the water where V W is the volume of one water molecule. The full thickness of the bilayer region is de¢ned to be D B=2V L/A and the full thickness of the water region is then D W=2n W V W/A=D3D B.The volume V L of the lipid molecule is further divided into two regions,a hydrocarbon chain region and a headgroup region.This division emphasizes another important aspect of bilayer thickness, namely,the thickness2D C of the hydrophobic core. We include in D C all the hydrocarbon chain carbons except for the carbonyl carbon which has substantial hydrophilic character.For DPPC the hydrophobic core therefore consists of14methylenes and one ter-minal methyl on each of the two chains.With this convention the headgroup is then de¢ned to consist of the remainder of the more hydrophilic part of the lipid,which can be subdivided into the carbonyls, glycerol,phosphate and choline.(Another conven-tion is to de¢ne the headgroup to be just the phos-phate and the choline.)The half thickness of the hydrocarbon region is related to the hydrocarbon volume of the lipid V C by D C=V C/A.In view of the broad distributions shown in Fig. 2a,the boundaries drawn in Fig.2c are clearly arti-¢cially sharp,but Fig.2c is an appropriate average description in the sense that the sharp lines can be justi¢ed as Gibbs dividing surfaces[28].For example, the D C line cuts the methylene distribution at a prob-ability near0.5in Fig.2a.(The actual dividing sur-face criterion is that the integrated probability of methylenes outside D C,indicated by the region marked L in Fig.2a,should be equal to the inte-grated de¢cit probability inside D C,indicated by the region marked K.)It may also be noted that,J.F.Nagle,S.Tristram-Nagle/Biochimica et Biophysica Acta1469(2000)159^195 162even ignoring£uctuations,there are methylenes on the sn-2chain and carbonyls on the sn-1chain that are on the wrong side of D C because of the inequi-valence of the two chains in DPPC;this again is included in the spirit of Gibbs dividing surfaces[28]. To obtain a more realistic picture of the interface region,it is useful to consider a re¢nement to the simple description on the left side of Fig.2c.This re¢nement,shown on the right side of Fig.2c,ex-plicitly mixes the heads and water in the polar,in-terfacial region.This gives better correspondence with the simulated distribution functions for the headgroup components in Fig.2a.In particular, the steric bilayer thickness,de¢ned to be D B P,lies in the tails of the distribution function of the choline component in Fig.2a,whereas the volume delimited by D B includes less than half of the choline compo-nent.It is appropriate for structural studies to obtain values for all four of these membrane thicknesses (D HH,D B,D B P and D C)and to determine what rela-tions exist between them.It may be helpful to the reader to note that a glossary of terms along with simple relations between them is included in the Ap-pendix.3.Some precise structural quantities3.1.VolumesThe preceding section emphasizes that volumes are the pivotal quantity to relate lateral structure,such as A,to transverse structure,such as the bilayer thickness D B,using relations like AD B=2V L.Mea-surements of total lipid volume V L have been per-formed using a variety of techniques.Our favorite method employs neutral£otation in which the den-sity of the aqueous solvent is varied by mixing D2O with H2O,combined with dilatometry which mea-sures volume changes as a function of temperature [29,30].The density of the lipid is then given by the density of the aqueous mixture in which the bilayers neither sink nor£oat.However,this method is re-stricted to lipids that have densities intermediate be-tween D2O and pletely di¡erent methods employ a di¡erential vibrating tube densimeter [31,32],di¡erential weighing[33]or buoyant forces [34].Values of V L for di¡erent lipids are given in Table1.Agreement between the di¡erent methods is about3parts in1000and the errors in each meth-od alone is of order2parts in1000.It may be noted that many papers in the literature have assumed that the partial speci¢c volume of the lipid equals that of water and have simply used v L=1ml/g.As can be seen from Table1,this is not a bad approximation for many phospholipids in the L K phase,but it is considerably poorer for the gel phase.The volumes of the chains,V C,and the head-groups,V H,have been obtained for the gel phase of DPPC[4].As is reviewed in Section11,the lateral packing dimensions of the all-trans hydrocarbon chains in the gel phase of DPPC can be obtained. Multiplying the lateral area by the longitudinal dis-tance per methylene(1.27Aî)along the chains gives the volume of the methylenes V CH2.Analysis of the methyl trough in the X-ray low-angle data gave V CH31X93V CH2[39].The reason for the much larg-er volume of a terminal methyl,despite having only one additional hydrogen,is due to its having an extra hemispherical endcap of steric excluded volume com-pared to a methylene that is covalently bonded in both directions along the chain.Thence the total hydrocarbon volume,V C,and the headgroup volume V H=V L3V C follow for the gel phase of DPPC.Our best value of V H is319Aî3[4],which is quite close toTable1Comparison of literature values for volume/lipidLipid Temperature(³C)Ref.v L(ml/g)V L(Aî3/molecule) DPPC20[30]0.9391144[33]0.9391144[29]0.9371142[35]0.9401145DPPC50[30] 1.0111232[29] 1.0091230[29] 1.0081228[35] 1.0091230[32] 1.0061226DMPC30[29]0.9771100[36]0.9781101[35]0.9781101[32]0.9721094EPC30[36]0.988126120[37]0.9811252DOPC30[38]0.999130322[25]0.9931296J.F.Nagle,S.Tristram-Nagle/Biochimica et Biophysica Acta1469(2000)159^195163the value of 325Aî3suggested by Small [40].Some earlier values from our lab that were in the range340^348Aî3[29,39,41]used less well determined val-ues for the wide-angle packing.For £uid phases of phosphatidylcholines the vol-ume of the heads has been estimated based on the argument that V H is the same in the fully hydrated £uid phase as in the fully hydrated gel phase because the headgroup is fully immersed in water in both phases.This assumption also implies that V H is the same for all lipids with the same PC headgroup.The measured change in lipid volume [29]is then equated to the change in V C .The volume V CH 3of a terminal methyl is often assumed to be about 2V CH 2[29,42],although it was once suggested that a ratio closer to 1.2applies for the £uid phase [40].Analysis of com-bined neutron and X-ray data for £uid phase DOPC obtains a ratio of 2.1[43]and computer simulations yield a ratio in the range 1.9^2.1[3]with later simu-lations favoring 1.9[22].Using a ratio near 2then allows one to estimate the average V CH 2and V CH 3from V C ,as shown in Table 2.Estimates of the volumes of all the component groups on the lipid molecule have been obtained from computer simulations for £uid phase DPPC [20]and £uid phase DOPC and POPC [22].The method assumes that the average volume of each group is independent of its transverse distance from the center of the bilayer.The resulting volumes must satisfy an independent check that suggests that this assumption is a good approximation.There are only fairly minor variations in the component volumes for the di¡erent PC lipids studied and a composite set ofvolumes,reproduced in Table 2,was given [22].It is noteworthy that the simulation results in Table 2give V H =321Aî3,in very good agreement with the experimental value for the gel phase [4].The simula-tions also suggest that the component volumes do not change signi¢cantly with hydration level,which is consistent with the experimental result that total lipid volume does not change measurably with hy-dration [44].mellar repeat spacings D3.2.1.AccuracyMost di¡raction studies have been performed on stacks of bilayers,especially on the easily prepared dispersions consisting of multilamellar vesicles (MLVs).The easiest di¡raction result to obtain ac-curately is the repeat spacing D ,which is always given to at least two signi¢cant ¢gures and often tothree signi¢cant ¢gures,such as D =67.2Aîfor fully hydrated DPPC at 50³C [3].In fact,with the bestinstrumental resolution (0.0001Aî31)and by ¢tting line shapes to ¢nd the center of the di¡raction peak,it is possible to obtain nearly four signi¢cant ¢gures,such as 55.06Aî[3].Such high accuracy is not used in structure determination ^two signi¢cant ¢gures usu-ally su¤ce ^but it leads into an interesting discus-sion regarding the nature of the samples.The MLVs in a random dispersion presumably come in a variety of sizes.Once formed,each bilayer is in£uenced by its neighbors.It is usually assumed that such MLVs are `onion-like',consisting of closed concentric spheres,at least in the topological,if not the strict geometric,sense.Since lipid exchange be-tween bilayers and solvent is slow,it is likely that the number of lipids in each bilayer remains constant over fairly long times.The swelling of such MLVs with temperature changes might be expected to be non-uniform depending upon their original degree of £accidness.There are therefore many reasons to imagine that the D spacing might be di¡erent be-tween di¡erent MLVs in the same sample,or even within the same MLV -the inner bilayers versus the outer bilayers.It is therefore remarkable that highly precise X-ray di¡raction,which detects many MLVs simultaneously,almost always sees lamellar di¡rac-tion peaks that are very narrow.If we suppose,for the sake of discussion,that a lamellar peak is a com-Table 2Volumes of component groups for general L K phase lecithins ignoring temperature dependence Group Volume (A î3)CH 352.7þ1.2b53.9þ0.8a CH 228.1þ0.1b 28.4þ0.4a HC N CH 45.0þ1.6b Carbonyl 39.0þ1.4b Glycerol 68.8þ9.9b Phosphate 53.7þ2.4b Choline120.4þ5.0ba This work.bArmen et al.[22].J.F.Nagle,S.Tristram-Nagle /Biochimica et Biophysica Acta 1469(2000)159^195164posite of many peaks,each with a di¡erent D spac-ing,then the observed peak widths would correspond to a distribution of D spacings in the sample with a width less thanþ0.05Aî.Only once,with a damaged sample,did we observe heterogeneous D spacings. Indeed,there is excellent reason to believe that the observed narrow widths are not even due to polydis-persity in the D spacings in the above sense because ¢nite size e¡ects and£uctuations that will be dis-cussed in Section5fully account for the shapes of the observed di¡raction peaks with a single D[45]. Why then,are di¡ractionists loathe to quote four signi¢cant¢gures for D?Although there is at most only one narrow distribution of D values in a single data set,nominally identical samples often have dif-ferent values of D,equally narrowly determined.As usual,the most egregious example is the fully hy-drated£uid phase DPPC for which the same study [45]reported four di¡erent values of D from64.5Aîto67.2Aî;other studies reported values ranging down to60Aî[46].Another type of variation in D that occurs in a single sample was¢rst noted by Peter Rand(private communication)and con¢rmed by us.When there is an air bubble in a sample,D becomes smaller as the beam is positioned closer to the bubble.Although all this irreproducibility might appear to be devastating,it is not.Near full hydra-tion the balance of interbilayer forces is rather deli-cate and the free energy di¡erence caused by varia-tions of a few Aîin D is small[1,47].Basically,all that varies is a small amount of water between the bilayers which does not a¡ect structure and is easily dealt with by considering the continuous Fourier transform of the electron density pro¢le to be dis-cussed in Section5.3.2.2.Oriented samples and the vapor pressureparadoxBilayers in MLVs are isotropically oriented in space and therefore give so-called powder patterns (even though they may be thoroughly hydrated).It is convenient that such samples do not have to be (indeed,they cannot be)especially oriented in an X-ray beam.Furthermore,there is no concern with mosaic spread that occurs in any real aligned sample and that involves another experimental parameter. However,only a small fraction of the lipid in a pow-der sample di¡racts from a given beam,so intensities are weak.There is also the potential irregularity in the MLVs discussed in the previous paragraph which apparently does not a¡ect D but which certainly re-duces the correlation length of the domains within which the sample scatters coherently.For all these reasons it would be valuable to orient the stacks of bilayers.The simplest alignment procedure is to squeeze lipid between two£at substrates,but the strong absorption of X-rays by a substrate has led many researchers to try to orient the lipid on a single substrate and to hydrate the lipid from the vapor [44,48^51].Another preparation uses free-standing ¢lms[52].An important section of the Rand and Parsegian review[1]concerned the vapor pressure paradox (VPP).The result for all preparations of oriented samples since the1970s until quite recently was that the measured D was consistently smaller,by more than5Aî,than for fully hydrated samples. Rand and Parsegian[1]noted that a reduction of relative humidity to99%would su¤ce to explain this reduction in D,but the experimental care and concern for maintaining the relative humidity of the vapor at100%was emphasized[1,44].Since the chemical potential of water is the same for liquid as for saturated vapor,such a reduction in D was inconsistent thermodynamically,so this was aptly named the vapor pressure paradox[1]and it was suggested that there was an intrinsic physical reason for it[1,44].The¢rst paradigm shift regarding the VPP was that it was overcome,though with some e¡ort,for the gel phase of DPPC[41,53].This sug-gested that the VPP was associated with the excess £uctuations that occur in the£uid phase and an el-egant theory that involved suppression of these£uc-tuations by the substrate was developed mathemati-cally[55].We also interpreted some indirect experimental evidence in support of this explanation [54].Recently,however,Katsaras has reported that there really is no VPP[56].This breakthrough oc-curred using neutron di¡raction which has the ad-vantage that aluminum is fairly transparent to neu-trons,so the sample chamber has no need for special windows upon which vapor can condense as in X-ray chambers.Katsaras produced a massive aluminum sample chamber with excellent temperature and hu-midity control[57].Then the fully hydrated D spac-J.F.Nagle,S.Tristram-Nagle/Biochimica et Biophysica Acta1469(2000)159^195165。

  1. 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. 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. 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)。
Definition 2.1 see 4 . The iteration procedure xn 1 f T, xn is said to be T -stable with respect to T if {xn} converges to a fixed point q of T and whenever {yn} is a sequence in X with limn → ∞ d yn 1, f T, yn 0 we have limn → ∞ yn q.
1. Introduction and Preliminary
Let E be a real Banach space. A nonempty convex closed subset P ⊂ E is called a cone in E if it satisfies the following:
Correspondence should be addressed to S. M. Vaezpour, vaez@aut.ac.ir
Received 28 March 2009; Revised 28 September 2009; Accepted 19 October 2009
Recommended by Brailey Sims
i P is closed, nonempty, and P / {0}, ii a, b ∈ R, a, b ≥ 0, and x, y ∈ P imply that ax iii x ∈ P and −x ∈ P imply that x 0.
by ∈ P,
The space E can be partially ordered by the cone P ⊂ E; by defining, x ≤ y if and only if y − x ∈ P . Also, we write x y if y − x ∈ int P , where int P denotes the interior of P .
A cone P is called normal if there exists a constant K > 0 such that 0 ≤ x ≤ y implies x ≤K y .
In the following we always suppose that E is a real Banach space, P is a cone in E, and ≤ is the partial ordering with respect to P .
1.5
Fixed Point Theory and Applications
3
so
an 1 ≤ K bn hbn−1 · · · hmbn−m hm 1K an−m ,
1.6
and then by Remark 1.4 an → 0. Therefore an → 0.
2. T -Stability in Cone Metric Spaces
1 d T x, T y ≤ ad x, y , 2 d Tx, Ty ≤ b d x, Tx 3 d Tx, Ty ≤ c d x, Ty
d y, Ty , d y, Tx .
Every Zamfirescu operator T satisfies the inequality:
d T x, T y ≤ δd x.3
where λ ∈ 0, 1 , for all n ≥ n0 and for some positive integer number m. If bn → 0 as n → ∞. Then limn → ∞ an 0.
Lemma 1.5. Let P be a normal cone with constant K, and let {an} and {bn} be sequences in E satisfying the following inequality:
Recently Qing and Rhoades 5 established a result for the T -stability of Picard’s iteration in metric spaces. Here we are going to generalize their result to cone metric spaces and present an application.
Hindawi Publishing Corporation Fixed Point Theory and Applications Volume 2009, Article ID 751090, 6 pages doi:10.1155/2009/751090
Research Article On T -Stability of Picard Iteration in Cone Metric Spaces
Then d is called a cone metric on X, and X, d is called a cone metric space.
Definition 1.2. Let T : X → X be a map for which there exist real numbers a, b, c satisfying 0 < a < 1, 0 < b < 1/2, 0 < c < 1/2. Then T is called a Zamfirescu operator if, for each pair x, y ∈ X, T satisfies at least one of the following conditions:
The aim of this work is to investigate the T -stability of Picard’s iteration procedures in cone metric spaces and give an application. Copyright q 2009 M. Asadi et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Definition 1.1 see 1 . Let X be a nonempty set. Assume that the mapping d : X × X → E satisfies the following:
2
Fixed Point Theory and Applications
i 0 ≤ d x, y for all x, y ∈ X and d x, y 0 if and only if x y, ii d x, y d y, x for all x, y ∈ X, iii d x, y ≤ d x, z d z, y for all x, y, z ∈ X.
1.1
for all x, y ∈ X, where δ max{a, b/ 1 − b , c/ 1 − c }, with 0 < δ < 1. For normed spaces see 2.
Lemma 1.3 see 3 . Let {an} and {bn} be nonnegative real sequences satisfying the following inequality:
2 Department of Mathematics, Amirkabir University of Technology, Tehran 15916 34311, Iran 3 Department of Mathematics, Newcastle University, Newcastle, NSW 2308, Australia 4 Department of Mathematics, Indiana University, Bloomington, IN 46205, USA
One of the most popular iteration procedures for approximating a fixed point of T is Picard’s iteration defined by xn 1 T xn. If the conditions of Definition 2.1 hold for xn 1 T xn, then we will say that Picard’s iteration is T -stable.
an 1 ≤ 1 − λn an bn,
1.2
where λn ∈ 0, 1 , for all n ≥ n0,
∞ n
1
λn
∞, and bn/λn → 0 as n → ∞. Then limn → ∞ an
0.
Remark 1.4. Let {an} and {bn} be nonnegative real sequences satisfying the following inequality:
M. Asadi,1 H. Soleimani,1 S. M. Vaezpour,2, 3 and B. E. Rhoades4
1 Department of Mathematics, Science and Research Branch, Islamic Azad University (IAU), Tehran 14778 93855, Iran
an 1 ≤ han bn,
1.4
相关文档
最新文档