VLT narrow-band photometry in the Lyman continuum of two galaxies at z~3 Limits to the esca

合集下载

光动力治疗在动脉粥样硬化斑块稳定性方面的进展

3光动力治疗的优点光动力治疗简单、快捷、高效，且相对安全。它
相比于有创和微创治疗，减少了对正常动脉血管壁的损伤，从而降低手术并发症的风险;相比于低脂和分子靶向治疗，缩短治疗周期，减少并发症发生的可能。在治疗效果上，光动力治疗能通过抑制甚至消除炎症反应来稳定斑块，降低斑块易损性，减少并发症的发生。此外，光动力治疗不仅能稳定和缩小斑块，具有治疗的作用，同时还具有预防斑块治疗术后再狭窄的作用，对改善患者预后具有满意的效果3。研究证明，光动力治疗联合微创介入治疗和单纯应用介入治疗相比，在针对实验样本预后和避免再复发方面的效果要好得多“。
4光敏剂的种类 4.1初代光敏剂初代光敏剂的代表药物为血吓咻。在兔动脉粥样硬化模型中，能够清楚观察到斑块中血吓咻的浓度从内膜表面到中膜浓度呈梯度递减丽。利用血吓咻的光动力治疗,能够有效抑制兔模型平滑肌细胞生长，降低动脉粥样硬化内膜与介质的比值。但在临床实验中，血吓咻光敏剂会导致严重的皮肤光毒性，有急性红斑、水肿现象产生。
中，明显显示斑块的减少与消退㉔。 4.3第三代光敏剂第三代光敏剂目前正在研发之中，其目的在于使光敏剂具有更高的选择性，避免在血管内皮细胞中的积累，从而避免对血管壁的损害，减少并发症的发生。同时，新一代光敏剂应吸收更长波长的光,并与血管内介入技术、光纤技术、激光技术相适应，以保证对斑块病变细胞的均匀照射,从而减少不良反应。光敏剂还可以与其他纳米颗粒(如胶束、纳米粒子、蛋白纳米结构)形成纳米复合材料，从而具备更好的治疗效果阴。
光敏剂以Ml型巨噬细胞为靶点，光动力治疗能够有效减少斑块的形成影像学结果显示,大的坏死脂质核、炎性薄纤维帽和大量的巨噬细胞浸润是斑块易损性的主要特点，这表明斑块易损性与局部的炎症反应密切相关。如能抑制斑块的炎症反应，也就能够抑制斑块的进展，从而稳定斑块。而巨噬细胞在动脉粥样硬化斑块进展过程中发挥着重要作用。研究表明，使光敏剂被胶束包裹之后注射进入样本中，由于特殊的巨噬细胞相比于正常内皮细胞具有更高的脂肪酶活性，因而降解胶束的速度更快，比内皮细胞更早激活光敏剂，从而为选择性杀伤巨噬细胞创造了可能切。以此可实现稳定斑块，避免斑块破裂，避免致死性并发症发生的作用。尽管巨噬细胞在斑块易损性中发挥了巨大的作用，但

slow light in photonic crystal

Wideband slow light with ultralow dispersion in a W1photonic crystal waveguideJian Liang,1,2Li-Yong Ren,1,*Mao-Jin Yun,2and Xing-Jun Wang31State Key Laboratory of Transient Optics and Photonics,Xi’an Institute of Optics and Precision Mechanics,Chinese Academy of Sciences,Xi’an710119,China2College of Physics Science,Qingdao University,Qingdao266071,China3State Key Laboratory on Advanced Optical Communication Systems and Networks,Peking University,Beijing100871,China*Corresponding author:renliy@Received1July2011;revised21September2011;accepted21September2011;posted26September2011(Doc.ID150243);published20October2011A dispersion tailoring scheme for obtaining slow light in a silicon-on-insulator W1-type photonic crystalwaveguide,novel to our knowledge,is proposed in this paper.It is shown that,by simply shifting thefirst two rows of air holes adjacent to the waveguide to specific directions,slow light with large group-index,wideband,and low group-velocity dispersion can be realized.Defining a criterion of restricting thegroup-index variation within aÆ0:8%range as a flattened region,we obtain the ultraflat slow light withbandwidths over5.0,4.0,2.5,and1:0nm when keeping the group index at38.0,48.8,65.2,and100.4,respectively.Numerical simulations are performed utilizing the three-dimensional(3D)plane-wave ex-pansion method and the3D finite-difference time-domain method.©2011Optical Society of AmericaOCIS codes:230.3990,230.5298,230.7390,260.2030.1.IntroductionFor the past decade or more,slow light,meaning the group velocity v g of light is slowed down as compared with the vacuum velocity c,has attracted a lot of attention because of its many potential applications. These include tunable optical delay[1],spectral re-solution enhancement[2],all-optical signal buffer-ing,and regeneration[3].In particular,slow light in photonic crystals has recently become a hot topic because of its numerous merits,such as room-temperature operation and easy on-chip integration. One can realize any slow-light operating wavelength by suitably scaling the structural parameters of the photonic crystal.Utilizing the photonic bandgap of photonic crystals,the phenomenon of slow light can be easily observed close to the Brillouin zone edge in a photonic crystal waveguide(PCW)[4–6].But un-fortunately,the dispersion relationship around the slow-light operating point inherently shows a near-parabolic behavior,which means a large group-velocity dispersion(GVD)may always accompany the slow-light propagation.This disadvantage coun-ters most of the advantages of operating in the slow-light region and severely limits the bandwidth[7,8]. In2004,Baba and coworker reported that,by using a chirping PCW,it is possible to significantly depress the GVD while enlarging the bandwidth[9]. Since then,the GVD and the bandwidth in a slow-light PCW structure were found to strongly rely on its geometric parameters.Up to now,many theoreti-cal and experimental papers have been presented, based on the method of destroying the overall unifor-mity of the photonic crystal structure to tailor its dis-persion,that aim to realize wideband and low GVD slow light with high group index n g.These ways mainly include reducing the waveguide width[8,10], chirping the waveguide structure[9,11],perturbing the radius of air holes[12,13],introducing ring-shaped holes[14],simultaneously shifting the first0003-6935/11/310G98-06$15.00/0©2011Optical Society of AmericaG98APPLIED OPTICS/Vol.50,No.31/1November2011and second rows of air holes[15–17]adjacent to the waveguide,infiltrating microfluid into the air holes [18],embedding quantum dots[19],solely shifting the third row of air holes[20]as well as merging coupled cavities[21],etc.For a simpler fabrication process,shifting the location of the air holes is widely considered to be the most effective one[4].In pre-vious literature,authors usually determine the bandwidth of a flattened wideband slow-light region by restricting the variation of n g within aÆ10% range[10,14,17].However,in fact,such aÆ10%cri-terion may be not good enough to allow high-fidelity slow-light propagation for a short pulse,especially with a high n g.For a higher n g,there must be larger GVD and smaller accompanying bandwidth[9],so a slow-light PCW structure with a much stricter criter-ion of n g is needed.Here,we propose a W1-type slow-light PCW struc-ture with a line defect in the center.Such a PCW pos-sesses an acceptably wide and ultraflat band regionfor each given n g.In simulations,we focus on realiz-ing the transverse electriclike(TE)polarized evenmode slow light because a laterally odd mode canhardly couple with a monomode dielectric waveguidedue to the symmetry mismatch[8].Moreover,thelight–matter interaction of the even mode in thePCW is strong,which interacts with several rowsof air holes adjacent to the waveguide.This phenom-enon implies that shifting these air holes can drama-tically detune the dispersion[12,22].In this paper,we discover that the flattened wideband slow lightwith large n g can be generated by shifting the firstrow of air holes(above and below the waveguide)hor-izontally and the second row vertically.By using sucha novel lattice-shifting method,slow light with an ul-traflat band is easily achieved even for n g up to100.4;i.e.,our method permits us to achieve a simulta-neously wideband slow-light PCW with ultralow dis-persion at large n ing the three-dimensional(3D)plane-wave expansion(PWE)method and the3Dfinite-difference time-domain(FDTD)method,thedetailed numerical simulations are performed.2.Theoretical SimulationFigure1depicts the schematic diagram of the PCWstructure we propose,where the silica substrate isremoved to see clearly.The PCW above the substrateis a triangular lattice photonic crystal slab with aline defect in the center,usually called a W1-typePCW.The backbone material is silicon(n Si¼3:4). We denote the period of the lattice,the diameter ofthe air holes,and the thickness of the PCW by a,D,and h,respectively.Here h just indicates the thick-ness of the slab.In simulations,the thickness of thesilica substrate(n¼1:45)is set to1μm.The shifting method can be described as follows.Inaccordance with the definition of the x and z axesshown in Fig.1,we useΔx andΔz to denote theshifted distances of the air holes along the directionsof the x and z axes,respectively.For the lower tworows of air holes adjacent to the waveguide,the first row is shifted along the x direction and the second one along the z direction.The upper ones are shifted just opposite to the lower ones.We use the3D PWE method to calculate the dispersion of this slow-light waveguide[23].In the supercell calculation,we se-lect the plane waves’numbers of each axis as32and the eigenvalue tolerance as10−8for a sufficient cal-culation precision and a moderate time consumption and memory demand.The relationship between the dispersion and n g can be described asn g¼n effþωdn effdω;ð1Þwhere n eff is the effective index andωis the central angle frequency of the incident pulse.Here n eff andωcan be substituted by the propagation constantβ(β¼2πλn eff,whereλis the operating wavelength)and the normalized frequency u(u¼ωa2πc,where c is the light speed in vacuum),respectively.Considering the condition n g≫n eff in the case of slow light, Eq.(1)can be transformed ton g¼a2πdβdu:ð2ÞObviously,low-dispersion(equivalently,nearly con-stant n g)slow light can be obtained in specific fre-quency regions where u varies linearly withβ.Our purpose in this paper is to optimize the parametersΔx andΔz in the W1PCW to obtain a wideband and ultraflat slow light at large n g simul-taneously.In the numerical simulations,we fix D¼0:62a and h¼0:5a.Note that these parameters are referred theoretically and experimentally to most publications.Figure2(a)shows the relation uðβÞof the TE even mode whenΔx andΔz are varied for op-timization.As mentioned above,the linear region of each curve corresponds to a flatband slow-light re-gion with a specific n g(thusυg)determined byits Fig.1.(Color online)3D model diagram of the W1PCW structure we propose.It is a slab with air holes in silicon,which lies on a substrate made of silica.The two rows of air holes adjacent to the waveguide are shifted along the x and z arrows.a is the period of the lattice.D and h are the diameter and thickness of the slab, respectively.1November2011/Vol.50,No.31/APPLIED OPTICS G99slope.Such a linear region has been labeled in Fig.2(a).Figure 2(b)shows the calculated n g as a function of u .In order to evaluate the low-dispersion bandwidth of the slow light,some previous papers usually select a bandwidth criterion where n g varies within a Æ10%range.Bandwidths of 9.0,6.7,4.7,and 2:7nm are obtained for nominal group indices of 38.0,48.8,65.2,and 100.4,respectively .It should be noted that the dispersion relation-ships plotted in Fig.2remain unchanged provided that one enlarges or reduces the PCW structure pro-portionally (i.e.,by changing the lattice period a while maintaining the relative geometry).It is fasci-nating that one can obtain arbitrary operating wave-length of interest by only adjusting the value of a without renewing the simulation.This relationship can be expressed bya ¼u 0λ0;ð3Þwhere u 0denotes the specific normalized frequency for each given n g with a flat band [as shown in Fig.2(b)]and λ0denotes the specific operating wavelength.To further illustrate the dispersion property of the ultraflat band in Fig.2(b),we introduce the GVD parameter D defined as [14]D ¼1c ∂n g∂λ:ð4ÞAccording to Eq.(4),if the slope of n g ðλÞis near zero,the PCW structure must have a near-zero GVD.Figure 3shows dispersion curves as a function of wavelength for different n g .In Fig.3,we have fixed all the nominal operating wavelengths to be around 1550nm by setting a to different values according to Eq.(3).Taking n g ¼38:0as an example,the GVD stays within Æ1ps =ðmm ·nm Þover 3nm,which has been considered a low-dispersion region in [14].Simi-lar to the GVD feature under a small value of n g ,the GVD still keeps very low as expected even for a high n g ,which was quite difficult to obtain in previous models published [14,16].We can obtain this ultra-low dispersion region over 0:3nm in a high value of n g ¼100:4.Note that the dispersion region de-creases constantly as the value of n g increases,so the discussion of an even higher value of n g gradually becomes meaningless.If one considers n g to be prac-tically constant when it varies less than Æ0:8%,in fact,one proposes a much stricter criterion for flatness of the slow-light band region.In the sense of this criterion,we obtained an ultraflat bandwidth over 5.0,4.0,2.5,and 1:0nm for n g ¼38:0,48.8,65.2,and 100.4,respectively.The simulation results are listed in Table 1.It implies that,no matter whether this structure is used at a small n g or at a large n g ,an optical pulse will suffer almost no distortion because of the ultralow GVD.We also simulate the pulse propagation in the PCW using a 3D FDTD method with the perfectly matched layer as absorbing boundary conditions [24,25].Note that we have chosen 32unit cells per axis in our simulations,which has been verified to be good enough to ensure a reflectionless transmis-sion [26].We suppose that a sinusoidallymodulatedFig.2.(a)Normalized frequency as a function of the propaga-tion constant,four groups of selected Δx and Δz included with a ¼420nm and a diameter of air holes of D ¼0:62a ,h ¼0:5a .(b)Group index as a function of normalized frequency accord-ing to(a).Fig.3.GVD as a function of wavelength when the selected group indices are deemed to be constant with n g ¼100:4,65.2,48.8,and 38.0,respectively .We keep the nominal operating wavelength region at 1550nm by changing a .G100APPLIED OPTICS /Vol.50,No.31/1November 2011Gaussian pulse impinges upon the PCW.The 1=e half width of the input pulse was set at τ¼0:067ps.It should be pointed out that this choice of τ,as an ex-ample,corresponds to c τ¼20μm.The temporal pulse profiles in terms of the normalized field ampli-tude are then obtained as shown in Fig.4,corre-sponding to the two selected detection points at a and 10a ,respectively ,behind the light source.The normalized central frequency of the pulse is chosen to be u ¼0:269,which corresponds to the approxi-mately constant n g of 38.0as shown in Fig.2(b).Figure 4indicates that the shape of the pulse keeps almost unchanged even after a propagating distance of 9a [≈3:76μm].The 1=e half width of the pulse at 10a is 0:0672ps,meaning that the pulse experiences a broadening of only 0.3%.The value of n g can be cal-culated based on the relationship between the propa-gation time t and the propagation distance l ,i.e.,t ¼l =v g ¼ðn g =c Þl .Therefore,we obtain n g ¼36:6from Fig.4.Note that the value of n g ¼38:0we ob-tained above is determined by the 3D PWE method.We think that the slight discrepancy of n g is tolerant as far as the discretization in FDTD calculations and the certain number of plane waves adopted in PWE calculations are concerned.The nearly coinciding re-sult means that the dispersion relationship (thus n g )obtained with the 3D PWE method is correct.Hence,we can conclude that,in our proposed PCW struc-ture,an ultranarrow pulse can propagate with a slow group velocity without obvious distortion owing to its broad bandwidth and near-zero GVD.3.Results and DiscussionThe mechanism of the slow-light effect in the PCW can be classified into two different aspects,namely,coherently backscattering and omnidirectional re-flection [5].For an air-hole-based PCW ,the former one is usually the main mechanism.However,the op-posite shifting method along the x direction brings the latter one into the PCW due to the opposite change in refractive index.This mechanism avoids interference between the signal and the backscatter-ing pulse,so the GVD can be significantly depressed.We further discuss the independent impact of Δx and Δz on this model.Figure 5(a)indicates that,for a given Δz ,the group-index curve,as a function of the normalized frequency,has a left shift and changes from monotonous to nonmonotonous behavior with increasing Δx .For a given Δx ,similar characteristics happen with the increase of Δz as shown in Fig.5(b),except that the curve has a right shift.Besides,the impact of Δx is somewhat stronger.Therefore,by appropriately combining Δx and Δz ,it is possible to realize a nearly constant n g even in a broad frequency region.This is the reason why we can ob-tain wideband,ultraflattened slow light with a large group index at different normalizedfrequencyFig.5.n g as a function of normalized frequency .(a)Δx varied from 0:124a to 0:144a with a step of 0:01a ,where Δz is fixed at 0:015a .(b)Δz varied from 0:005a to 0:025a with a step of 0:01a ,where Δx is fixed at 0:134a.Fig.4.Temporal pulse propagation at two selected detection points at a and 10a ,respectively ,behind the light source.n g ¼38:0;the relevant parameters are the same as those in Table 1.Table 1.Group Index,Bandwidth under Different Optimized Shifting ParametersAdjusted Parameters n g ΔλðÆ10%;nm ÞΔλðÆ0:8%;nm ÞΔx Δz 0:134a 0:015a 38.09.0 5.00:120a 0:045a 48.8 6.7 4.00:103a 0:060a 65.2 4.7 2.50:084a0:070a100.42.71.01November 2011/Vol.50,No.31/APPLIED OPTICSG101regions,as shown in Fig.2,by purposely changing Δx and Δz in opposite directions by simply shifting the related air holes.Shifting the locations of the first two rows of air holes adjacent to the waveguide using our method has been verified a functional modification to detune the dispersion in the PCW as discussed in Section 2.However,there is another geometric parameter that exists that may be very sensitive to the slow-light ef-fect and is usually neglected in previous PCW simu-lations.It is the thickness of the waveguide slab,h .In the simulation above,we set this value as a constant to 0:5a ,which seems not important in the simulation.Here we take n g ¼38:0as an example for which the relevant parameters are listed in Table 1.Figure 6shows the dependence of n g on the wavelength as h varies from 200to 220nm by a step of 10nm,where a is set to 420nm.Note that,for the convenience of discussing,the abscissa was adopted to the wave-length instead of the normalized frequency.We can see from Fig.6that the thickness of the slab plays an important influence on the slow light in the PCW.With the increase of h ,the flattened region in-dicates a considerable redshift,while the shape of n g stays almost unchanged.Such a performance can be used to excellently explain the redshift between the simulative result and the experimental one in [27],where the shift is 20nm and the transmittance curve (there the transmittance curve is directly decided by n g )stays almost unchanged.Although shifting air holes is considered the sim-plest way in fabrication,there will also be some un-avoidable errors.So in the following,the acceptable fabrication tolerance of the model is discussed.We define δx and δz as the fabrication errors relative to the shifts Δx and Δz of interest,respectively.The signs of δx and δz are the same as that of Δx and Δz ,as shown in Fig.1.We take n g ¼38:0as an example with the optimized parameters being listed in Table 1.The results are shown in Fig.7.The solid curve denotes the ideal model without any fabrica-tion error.For simplicity ,we suppose that each airhole in the two first rows adjacent to the waveguide has the same error δx and that each one in the 2s rows has the same error δz .Of course this is an ex-treme scenario.In principle,if one of them has smal-ler absolute error,the fabrication error of the other parameter can be relaxed accordingly.We can see from Fig.7that the curve remains almost unchanged when δx is positive,while the curve becomes worse when δx is negative.This is easy to understand if one realizes that a positive (or negative)δx means a consistent (or reverse)shifting direction of air holes owing to δx with respect to that owing to Δx .Figure 7also indicates that the curve is more sensitive to δx than δz .Therefore,this wideband low-dispersion property can be easily reserved provided that we keep a positive δx .This is true for a positive δx even up to 1nm as shown in Fig.7.4.ConclusionBy simply shifting the four rows of air holes adjacent to the waveguide,a W1PCW structure,novel to our knowledge,is proposed to realize wideband slow light with ultralow GVD.This scheme includes a hor-izontal-opposite shift to the first upper and lower rows of air holes adjacent to the waveguide,as well as a vertical-opposite shift to the second two rows accordingly .Compared with some other schemes,such as introducing ring-shaped air holes in [14],this scheme is more convenient to fabricate.The most im-portant thing is that we realize ultraflat group-index curves varying only in a range of Æ0:8%even for the group index up to 100.4.This work was partially supported by the West Light Foundation of the Chinese Academy of Sciences (CAS)(2009LH01),the Innovation Founda-tion of CAS (CXJJ-11-M22),the National Natural Science Foundation of China (NSFC,60778020),the Open Research Fund of State Key Laboratory of Transient Optics and Photonics of CAS,and the Scientific Research Foundation for theReturnedFig.6.n g as a function of normalized frequency with h varied from 200to 220nm with a step of 10nm.n g ¼38:0;the relevant parameters are the same as those in Table 1.Fig.7.n g as a function of normalized frequency with several δx and δz .δx and δz denote fabrication errors relative to Δx and Δz ,respectively .n g ¼38:0,the relevant parameters are the same as those in Table 1.G102APPLIED OPTICS /Vol.50,No.31/1November 2011Overseas Chinese Scholars,State Education Minis-try.The authors gratefully acknowledge helpful discussions with Prof.Romano A.Rupp. References1.A.Talneau,“Slowing down the light for delay lines implemen-tation:design and performance,”C.R.Physique10,949–956 (2009).2.Z.Shi,R.W.Boyd,D.J.Gauthier,and C.C.Dudley,“Enhan-cing the spectral sensitivity of interferometers using slow-light media,”Opt.Lett.32,915–917(2007).3.F.G.Sedgwick,C.J.Chang-Hasnain,P.C.Ku,and R.S.Tucker,“Storage-bit-rate product in slow-light optical buf-fers,”Electron.Lett.41,1347–1348(2005).4.T.Baba,“Slow light in photonic crystals,”Nat.Photon.2,465–473(2008).5.T.F.Krauss,“Slow light in photonic crystal waveguides,”J.Phys.D40,2666–2670(2007).6.X.Letartre, C.Seassal, C.Grillet,P.Rojo-Romeo,P.Viktorovitch,M.Le V.d’Yerville,D.Cassagne,and C.Jouanin,“Group velocity and propagation losses measurement in a single-line photonic-crystal waveguide on InP membranes,”Appl.Phys.Lett.79,2312–2314(2001).7.R.J.P.Engelen,Y.Sugimoto,Y.Watanabe,J.P.Korterik,N.Ikeda,N.F.van Hulst,K.Asakawa,and L.Kuipers,“The ef-fect of higher-order dispersion on slow light propagation in photonic crystal waveguides,”Opt.Express14,1658–1672 (2006).8.A.Yu.Petrov and M.Eich,“Zero dispersion at small groupvelocities in photonic crystal waveguides,”Appl.Phys.Lett.85,4866–4868(2004).9.D.Mori and T.Baba,“Dispersion-controlled optical groupdelay device chirped photonic crystal waveguides,”Appl.Phys.Lett.85,1101–1103(2004).10.M.D.Settle,R.J.P.Engelen,M.Salib,A.Michaeli,L.Kuipers,and T.F.Krauss,“Flatband slow light in photonic crystals fea-turing spatial pulse compression and terahertz bandwidth,”Opt.Express15,219–226(2007).11.D.Mori and T.Baba,“Wideband and low dispersion slow lightby chirped photonic crystal coupled waveguide,”Opt.Express 13,9398–9408(2005).12.L.H.Frandsen,vrinenko,J.F.Pefersen,and P.I.Borel,“Photonic crystal waveguides with semi-slow light and tailored dispersion properties,”Opt.Express14,9444–9450(2006). 13.S.Kubo, D.Mori,and T.Baba,“Low-group-velocity andlow-dispersion slow light in photonic crystal waveguides,”Opt.Lett.32,2981–2983(2007).14.A.Säynätjoki,M.Mulot,J.Ahopelto,and H.Lipsanen,“Dispersion engineering of photonic crystal waveguides with ring-shaped holes,”Opt.Express15,8323–8328 (2007).15.J.T.Li,T.P.White,L.O’Faolain,A.Gomez-Iglesias,and T.F.Krauss,“Systematic design of flat band slow light in photonic crystal waveguides,”Opt.Express16,6227–6232(2008). 16.F.C.Leng,W.Y.Liang,B.Liu,T.B.Wang,and H.Z.Wang,“Wideband slow light and dispersion control in oblique lattice photonic crystal waveguides,”Opt.Express18,5707–5712 (2010).17.J.Wu,Y.P.Li,C.Peng,and Z.Y.Wang,“Wideband and lowdispersion slow light in slotted photonic crystal waveguide,”mun.283,2815–2819(2010).18.M. E.Heidari, C.Grillet, C.Monat,and B.J.Eggleton,“Dispersion engineering of slow light photonic crystal wave-guides using microfluidic infiltration,”Opt.Express17, 1628–1635(2009).19.T.R.Nielsen,J.Mørk,and vrinenko,“Slow light pulsepropagation on dispersive media,”C.R.Physique10,957–963 (2009).20.Y.Hamachi,S.Kubo,and T.Baba,“Slow light with lowdispersion and nonlinear enhancement in a lattice-shifted photonic crystal waveguide,”Opt.Lett.34,1072–1074 (2009).21.K.Üstün and H.Kurt,“Ultra slow light achievement in photo-nic crystals by merging coupled cavities with waveguides,”Opt.Express18,21155–21161(2010).22.A.Têtu,M.Kristensen,L.H.Frandsen,A.Harpøth,P.I.Jensen,and O.Sigmund,“Broadband topology-optimized photonic crystal components for both TE and TM polariza-tions,”Opt.Express13,8606–8611(2005).23.S.G.Johnson and J. D.Joannopoulos,“Block-iterativefrequency-domain methods for Maxwell’s equations in a pla-newave basis,”Opt.Express8,173–190(2001).24.K.S.Yee,“Numerical solution of initial boundary value pro-blems involving Maxwell’s equations in isotropic media,”IEEE Trans.Antennas Propag.14,302–307(1966).25.A.Taflove and S. C.Hagness,Computational Electrody-namics:The Finite-Difference Time-Domain Method,2nd ed.(Artech,2000).26.O.Hess, C.Hermann,and A.Klaedtke,“Finite-differencetime-domain simulations of photonic crystal defect struc-tures,”Phys.Status Solidi A197,605–619(2003).27.M.Zhang,N.Groothoff, A. C.Krüger,P.X.Shi,and M.Kristensen,“Direct slow-light excitation in photonic crystal waveguides forming ultra-compact splitters,”Opt.Express 19,7120–7126(2011).1November2011/Vol.50,No.31/APPLIED OPTICS G103。

A TiO study of the dwarf nova IP Pegasi

a r X i v :a s t r o -p h /0004199v 1 14 A p r 2000Mon.Not.R.Astron.Soc.000,1–10()Printed 1February 2008(MN L A T E X style ﬁle v1.4)A TiO study of the dwarf nova IP PegasiG.Beekman,M.Somers,T.Naylor and C.HellierDepartment of Physics,Keele University,Staﬀordshire ST55BG Accepted ???.Received ???;In original form ???.ABSTRACTWe present red spectra in the region ∼λ7000–8300˚A of the eclipsing dwarf nova IPPeg,with simultaneous narrow-band photometry centered at 7322˚A .We show that by placing a second star on the slit we can correct for the telluric absorption bands which have hitherto made the TiO features from the secondary star unusable.We use these TiO features to carry out a radial velocity study of the secondary star,and ﬁnd this gives an improvement in signal-to-noise of a factor two compared with using the NaI doublet.In contrast with previous results,we ﬁnd no apparent ellipticity in the radial velocity curve.As a result we revise the semi-amplitude to K 2=331.3±5.8km s −1,and thus the primary and secondary star masses to 1.05+0.14−0.07M ⊙and 0.33+0.14−0.05M ⊙respectively.Although this is the lowest mass yet derived for the secondary star,it is still over-massive for its observed spectral type.However,the revised mass and radius bring IP Peg into line with other CVs in the mass-radius-period relationships.By ﬁtting the phase resolved spectra around the TiO bands to a mean spectrum,we attempt to isolate the lightcurve of the secondary star.The resulting lightcurve has marked deviations from the expected ellipsoidal shape.The largest diﬀerence is at phase 0.5,and can be explained as an eclipse of the secondary star by the disc,indicating that the disc is optically thick when viewed at high inclination angles.Key words:novae,cataclysmic variables –binaries:close –binaries:spectroscopic –binaries:eclipsing –stars:fundamental parameters –stars:individual:IP Peg1INTRODUCTIONIP Peg is a dwarf nova which brightens by ∼2mag ev-ery 100days or so,with each outburst lasting ∼2weeks,and is important because it is the brightest known eclips-ing dwarf nova above the period gap.The eclipses give tight constraints on the geometry of the system,and in combina-tion with other data,such as radial velocity studies,allow the masses of the two stars to be determined.We should thus be able to compare the mass,radius and density of the secondary star with those of main sequence stars,and learn something of its structure and evolution.Unfortunately the radial velocity curve of the secondary star in IP Peg is prob-lematical,being apparently elliptical (Martin et al.1989,henceforth M89).The star is believed to be in a circular orbit,and the apparent ellipticity is thought to be due to irradiation of the secondary star (although we shall present an alternative explanation in Section 6.4),and an uncertain correction must be applied before the binary parameters can be determined.The result is acutely embarrassing.Smith &Dhillon (1998)collect together all the available masses,radii and spectral types for CVs.IP Peg is one of the few objects that does not rely on the dubious method of using the accre-tion disc lines to determine the white dwarf radial velocity,and so should give reliable parameters.However,in all therelationships IP Peg is a persistent oﬀender,lying well clearof the supposedly less reliable points in both the mass vs.orbital period and mass vs.spectral type diagrams.A further complication is presented by the lightcurve of the system.High-speed photometric observations show a light curve dominated by a bright spot and a deep eclipse.The eclipse is that of the bright spot and white dwarf by the secondary star,superposed on a more gradual disc eclipse.Unfortunately the bright spot ingress overlaps with that of the white dwarf,making the ephemeris determination more problematical.Wood &Crawford (1986)were able to use their high-speed photometric observations to derive an ephemeris for IP Peg based on white dwarf egress ter observations by Wood et al.(1989)and Wolf et al.(1993)show that the white dwarf egress varies markedly in strength and duration (10s to 300s).In this paper we present red spectroscopy and photom-etry of IP Peg.The observations and their reduction are explained in Section 2.After that the paper divides into two main threads.The ﬁrst is the radial velocity study and its results.The extraction of the velocities is described in Section 3after which we discuss the summed spectra and new ephemeris (Sections 4and 5).In Section 6we discuss the absence of apparent eccentricity in our data,and con-cRAS2Beekman et al.clude the system has changed in some way.We use our radialvelocity semi-amplitude to derive new system parameters in Section 7.The second thread,isolating the lightcurve of the secondary star,is explored in Section 8.2OBSER V ATIONSPhotometry and spectroscopy of IP Peg were obtained over four nights in 1995October from the 2.5-m INT and 1.0-m JKT on La Palma.Only two nights of photometry were taken –simultaneous with the spectroscopy (see Section 2.4).Table 1contains a log of the observations.Data from the AAVSO show that in the latter half of 1995IP Peg underwent two outbursts,one in mid-September and one in mid-December.The data presented in this paper were taken in the ﬁrst week of October,approximately two weeks after the end of the mid-September outburst,and are thus during quiescence.2.1PhotometryData were taken with the JKT on 1995October 4th &5th using an eev7-chip with a scale of 0.31arcsec per pixel.A narrow band (fwhm =46˚A )O ii ﬁlter with maximum trans-mission at a wavelength of λ7322˚A was used on both nights.The images were processed by subtracting oﬀa single bias image in the case of the Oct.4th data,and a median of three biases for the Oct 5th data.A single sky ﬂat taken at the beginning of each night was used to ﬂatﬁeld the im-ages.The photometry was then extracted using an opti-mal extraction method (Naylor 1998).The photometry of IP Peg was divided by that of a neighbouring star (the sec-ond star placed on the slit during the spectroscopic run—see below)to remove the eﬀects of sky transparency variations and,thus,the photometry is relative.The phase-folded light curves for both nights are shown in Figs.1and 2with the bright-spot hump at φ∼−0.2and the primary eclipse at φ=0.0.2.2SpectroscopyData were taken on 1995October 4th,5th,9th and 10th with the INT,using a tek4-chip with a scale of 0.33arcsec per pixel,but binned by a factor two in the spatial direction.A two-arcsec slit was used set at PA 58.7degrees to obtain the spectrum of a neighbouring star simultaneously (the star used to divide the IP Peg photometry by—see above).This allowed us to correct for slit losses (see Section 2.4).The R831R grating was used on the 4th and 5th,giving ∼1.22˚A per pixel;the R1200R grating was used on the 9th and 10th,giving ∼0.84˚A per pixel.An arc spectrum was taken each time the telescope’s position on the sky was changed (∼every hour).The nights of the 4th and 5th were generally clear but the nights of the 9th and 10th were aﬀected by cloud,especially so the data of the 9th.An exposure time of 600s was used on the 4th,5th and 9th:400s on the 10th.The images for each night were ﬁrst bias subtracted.Tungsten ﬂats taken each night were then used,after divid-ing by ﬁtted low-order polynomials,to ﬂatﬁeld the images.Extraction of the spectra was performed using an optimal extraction method (Horne 1986,Robertson 1986),and thenFigure 1.The lightcurves obtained from the photometry (dots)and spectroscopy (bars).The lower panel shows the percent-age residuals between the two,calculated as 100.0∗(R spec −R phot )/R phot .The point in the top left of the upper panel repre-sents the average size of a photometric error.Data taken on 1995October 4th.Figure 2.As previous ﬁgure but for the data taken on 1995October 5th.wavelength calibration was done using low-order polynomial ﬁts to the appropriate arc spectra.2.3Removing the eﬀects of Earth’s atmosphereNormally,one observes a standard star (usually a near-featureless O star)at regular intervals throughout the night to divide the program star by.Here,we use the second star placed on the slit and observed simultaneously with IP Peg.This has the advantage that it avoids the assumption of uni-formity of atmospheric absorption with position and time.Fig.3shows that the atmospheric absorption has been well corrected,including the large bands near 7600˚A .The disadvantage of this technique is that the star may not be as featureless as one would like,since its main selec-tion criterion is proximity to the target paring blue spectra of the star with those in Jacoby,Hunter &Chris-tian (1984)shows it to be a late F-star.Such stars are goodcRAS,MNRAS 000,1–10A TiO study of the dwarf nova IP Pegasi3Table1.Journal of observations of IP PegasiDate No.of Wavelength Exposure Phase Mean Seeing(Oct.’95)Frames Region(˚A)Time(sec)Covered(arcsec)Photometry:4th81Oiiﬁlter1200.51–1.82 2.175th100Oiiﬁlter1200.31–1.70 1.65Spectroscopy:4th187100–83006000.38–1.03 1.855th187100–83006000.38–1.70 1.429th227000–78006000.56–1.72 1.7810th397000–78004000.83–1.791.98Figure3.Thisﬁgure shows the correction for the atmospheric‘A’-band nearλ7600˚A.The left hand panel is the second star on the slit; the center panel is the uncorrected IP Peg spectrum;the right hand panel after correction.All the spectra have been normalised to one at7500˚A.Note how the drop of the TiO band heads(7550-7650˚A)and the KI lines(7650and7700˚A)become visible after correction.atmospheric calibrators at low resolution(see Mason et al 1995),but at the intermediate resolutions we used weak ab-sorption features are present.In Fig.4we have averaged all our higher-resolution divided spectra with no velocity shift, and can see these features as“emission”lines superimposed on a velocity smeared M-star spectrum(since we have not corrected to the rest frame of the secondary star).The fea-tures are weak(normally around a couple of percent,with a few at10percent)and as we shall show in Sections3and 4,have a negligible eﬀect our radial velocity study andﬁnal summed spectrum.2.4Relative spectrophotometryDuring our spectroscopic run with the INT,two nights of photometric data(simultaneous with the spectroscopicc RAS,MNRAS000,1–104Beekman etal.Figure 4.The result of a simple mean of the spectra which result from dividing the IP Peg spectrum by the second star on the slit.Only the higher-resolution spectra(1995Oct9and10) have been used.data)were taken with the JKT to test the two-star spec-trophotometry technique.The wavelength region of the spectroscopy corresponding to the photometry was isolated and summed according to the weights of the O iiﬁlter.After summing,the spectrophotometry of IP Peg was divided by the similarly summedﬂux from the second star on the slit.Figs.1and2show the photometric data over-plotted on the spectroscopic data.The lower panels in theﬁgures show the percentage residuals between the two,calculated as 100.0∗(R spec−R phot)/R phot.Where more than one photo-metric point is coincident with a spectroscopic one,a simple average of them is taken.For the spectroscopy point covering primary-eclipse egress in the data of1995October4th,the exposure was too large long sample the fast rate of change in brightness properly.This resulted in the relatively large residual seen here.Ignoring this point in the analysis,the average residual size for the nights of the4th and5th are 3.28%and2.80%,respectively.That is,the diﬀerence be-tween relative photometry and relative spectrophotometry is only at the few percent level and,for the most part,lies within the error of the photometry.Thus the real diﬀerence is probably well within the measurement errors.This shows that performing relative spectroscopy by the inclusion of a second star on the slit works well.This result is apparently in contradiction with the work of Webb et al.(2000).They performed a similar experi-ment to ours,obtaining spectroscopic observations of the low-mass X-ray binary J0422+32,again with a second star on the slit.When they compared the relative photometry from their spectroscopic observations with that from simul-taneous photometry they found diﬀerences of order15per-cent.The author in common between these two papers has no explanation as to why this should be so,though we note that IP Peg is much brighter,and that a broadbandﬁlter was used for J0422+32.3A RADIAL VELOCITY STUDYWe used an iterative procedure to determine the radial ve-locity curve of IP Peg,in which a summed spectrum(cre-ated after each iteration)is used as the template.This tech-nique is similar to that used by Mukai&Charles(1988)and Horne,Welsh&Wade(1993).First,the mid-white dwarf egress timings of Wood et al.(1989)and Wolf et al.(1993) wereﬁtted with a linear ephemeris,and this ephemeris was initially used to phase all of our spectra.The radial velocity semi-amplitude(K2)value from M89was then used to as-sign a stellar radial velocity component to each spectrum for a given phaseφby assuming a circular orbit for the red star of the form K2sin(2πφi),whereφi is the phase of spectrum i.Each spectrum was then Doppler-shifted by its orbital radial velocity;all such shifted spectra were then co-added into a grand-sum spectrum.This grand-sum spectrum(less the spectrum of interest—see later)was then used as a template spectrum against which to cross-correlate all the individual spectra to obtain a radial velocity study.The velocity extraction process is run iteratively.The velocities from the initial set of cross-correlations areﬁt-ted with a sine wave to return a new K2-value and cor-rection to phase zero;each individual spectrum is then as-signed a new phase and new radial velocity from this sec-ond generation ephemeris and K2-value;each spectrum is then shifted by these new velocities and combined to make a new orbit-averaged template spectrum;the spectra are then cross-correlated against this new average,new veloci-ties extracted and so on,and the whole process repeats until convergence.The process was considered to have converged when the correction to phase zero was less than0.001of the orbital pe-riod,corresponding to a correction of less than14seconds. This value was chosen because tests showed that beyond this,no signiﬁcant improvement within the errors of theﬁt was achieved.Convergence typically occurred within ten it-erations.A small number of spectra returned dubious ve-locities;these spectra were of lower signal-to-noise(mainly those of1995October9th,due to occasional cloud),and were removed from the analysis.Note that to avoid any danger of‘self’cross-correlation, the spectrum being correlated is not used in creating the average-spectrum template against which it is being cross-correlated.This results in a new template being created for each individual spectrum,but the large number of spectra which adequately sample the orbit ensures that the template spectrum is,to all intents and purposes,identical through-out.Theﬁrst two nights of data were taken at a slightly lower resolution than the last two,so the data were split into high and low resolution data sets and each one treated separately.For both sets of data,two regions were used in the cross-correlation but only one region was used at a time. These regions were a115˚A-wide region fromλ7105–7220˚A, and a170˚A–wide region fromλ7560–7730˚A.The slightly diﬀerent wavelength coverage between the two data sets al-lowed the shorter wavelength cross-correlation region in the slightly higher resolution data set to be extended blue-wards toλ7000˚A.Table2presents the results of this study for the two regions used in each data set.A weighted mean(labelledc RAS,MNRAS000,1–10A TiO study of the dwarf nova IP Pegasi5Table2.Radial velocity data using an orbital-average as templateλNotes N K2T0-2449998.02σcσe F p17105–7220TiO31353.8±14.10.51609(86)32.1431.670.4040.6717560–7730TiO,K32325.6±14.30.51722(95)33.9233.410.4240.6587000–7220TiO59310.6±9.40.51841(73)32.4231.78 1.1830.3147560–7730TiO,K59344.3±9.50.51612(64)33.0132.880.2200.803wΣ331.3±5.80.51692(39)7670–8319M89region32327.7±16.130.1130.110.00 1.001Orbit is consistent with being circular for p≥0.05.2Numbers in brackets are the errors on the same number ofﬁgures immediately to their left.wΣin the table)of these results provides us with a new mid-white dwarf eclipse(not egress)timing andK2.The individual radial velocity curves are shown in Fig.5with a circular orbitﬁt using the above weighted mean values over plotted as a full line;the individual orbitalﬁts of Table2 are over-plotted as dashed lines.The quoted error is the68% conﬁdence interval(‘1σ’)calculated for two free parameters (Lampton,Margon&Bowyer1976).To derive this error the radial velocity points were given equal weight,and aχ2νassumed for the bestﬁt.We compared the signal-to-noise we obtained using the TiO bands with that from the more traditional NaI dou-blet.The gain using the new technique is about a factor two in signal-to-noise if the results from the two TiO bands are averaged.We also tried cross-correlating each spectrum against an M-star template instead of the mean IP Peg spec-trum.There was an increase in noise for theﬁts to these data,presumably because the mean spectrum gives a better match to the spectral type than any template star.Finally we wished to ascertain whether the weak F-star features introduced by our technique for removing telluric absorption(see Section2.3)had aﬀected our values for the radial velocity.Speciﬁcally,there was a concern that it would introduce a signal at zero velocity shift,which might reduce the semi-amplitude.During the run we took several obser-vations of early-type stars to check our correction technique. We chose a pair of observations where an O-star corrected the observation of the second star on our slit well,yielding a telluric absorption corrected spectrum of this star.We then divided our corrected spectra of IP Peg by this spectrum, thus doubling the strength of the F-star features.We then repeated our cross correlation analysis for Oct.4and5.The derived semi-amplitude changed by less than the error(+7 km s−1for7105-7220˚A and-4km s−1for7560-7730˚A).It should be noted that this is an over-estimate of how much the F-star aﬀects our study,since we will also have intro-duced noise,and poor atmospheric correction,all of which might tend to decrease the semi-amplitude.Despite this,it is clear that the second star on the slit makes a negligible contribution to the cross-correlation function,for reasons we explain in the next Section.4THE SUMMED SPECTRUMFig.6shows theﬁnal Doppler-shifted weighted-mean spec-trum.The reasons for our choice of cross-correlation bands should now be clear.Theﬁrst corresponds to the TiO band heads around7100˚A,the second to the TiO band heads Figure6.The Doppler-shifted,averaged spectra of IP Peg taken during1995October.LR is the lower resolution data taken on the nights of the4th&5th(upper spectrum);HR is the higher resolution data taken on the9th&10th(lower spectrum).Lower limit symbols mark the boundaries of the cross-correlation regions used in the radial velocity study.and KI lines(λ7764.9,7699.0˚A)near7600˚A.Theﬁnal fea-ture obviously from the secondary star is the NaI doublet (λ8183.3,8194.8˚A).The emission features are harder to identify,so to aid this process we show the sums of spectra in four phase ranges near phase zero in Fig.7.In this,and Fig.6we can see emission features around7770˚A,7800˚A,7870˚A,7930˚A and 8200˚A.These features are all broad,and weaken or disap-pear when the disc is eclipsed in the phase0.95-0.05spec-trum of Fig.7.This suggests a disc origin for all these lines. The easiest to identify is the emission around8200˚A,which is where the Paschen series converges,but we can also expect disc emission from the NaI doublet,and perhaps CaII8203˚A and maybe even HeII8237˚A.Friend et al.(1988)identify the OI triplet atλλ7771.9,7774.2and7775.4˚A,which co-incides with our7770˚A feature.This leaves three features still unexplained,of which only the7800˚A feature appears in M89.In principle there is a contribution to this spectrum from the F-star used for telluric correction.However,as stated in Section2.3the strongest features are around10 percent of the continuum level.As the summed spectrum is velocity shifted,these features are smeared out over600km s−1,or15pixels,making them less than1percent in depth. The lines will be similarly smeared in the template spectrac RAS,MNRAS000,1–106Beekman etal.Figure 5.Extracted IP Peg radial velocity curves.(1)The region 7560–7330˚A for Oct.4th &5th;(2)the region 7105–7220˚A for Oct.4th &5th;(3)the region7560–7730˚A for Oct.9th &10th;(4)the region 7000–7220˚A for Oct.9th &10th.The dashed lines are the best ﬁt to each panel (see Table 2).The full lines are the weighted mean of the individual ﬁts.Figure 7.Phase-binned spectra of IP Peg from 1995October 4th &5th.Phase ranges are marked,as are the K i and Na i absorption features.used for the cross-correlation,which is presumably why they have no signiﬁcant eﬀect on the radial velocity study.5A NEW EPHEMERISOur spectroscopically derived eclipse timing-point and that of M89,both corrected to time of mid-white dwarf egress,were added to the photometric timings of Wood et al.(1989)and those of Wolf et al.(1993).All timings were converted to TDB.A linear ephemeris was then calculated to give JD(TDB)T egress =2447965.89144(5)+E ∗0.15820617(3)which is for the time of mid-white dwarf egress .Fig.8shows the O −C plot for this ephemeris.We note in passing that the O −C residuals calculated from the linear ephemeris can be ﬁtted by a sine wave with parameters close to those given in Wolf et al.(1993),which they interpreted as a third body in the system.cRAS,MNRAS 000,1–10A TiO study of the dwarf nova IP Pegasi7 Figure8.The O-C plot for the mid-white dwarf egressephemeris from the high-speed photometry timings of Wood etal.1989(circles),Wolf et al.1993(crosses)and the spectroscopicpoints from M89and this paper(ﬁlled triangles,also marked withM and B for clarity).6THE APPARENT ECCENTRICITY6.1Are the new data consistent with a circularorbit?M89found that an elliptical orbit was a betterﬁt to theirradial velocity data than a circular one.Here,we test ourspectra for evidence of eccentricity.Weﬁtted each red-star radial-velocity data set with acircular orbit of the formV=γ+K2sin(2π[φi−φ0])and then an elliptical orbit of the form(e.g.Smart1977,Chap.XIV;Petrie1962)V=γ+K2[cos(ν+ω)+e cos(ω)]recording the mean square deviations(S2)of the data fromtheﬁtted model in each case.In the elliptical orbitνis thetrue anomaly of the red star,ωis the longitude of periastronof the red star from the centre of gravity of the system,ande is the eccentricity of the red-star orbit.Following Lucy andSweeney(1971),we use the F-testF=N obs−f ellS2ell−1 (1)to determine the signiﬁcance of the ellipticalﬁt over the circular one.N obs is the number of observations,f ell is the number of degrees of freedom in the ellipticalﬁt,f crc those in the circularﬁt,and S2crc,S2ell are the variances of the data from the circular and ellipticalﬁts respectively,i.e. S2=Σ(y data−y f it)2.This test may be derived from the relationship between the F-andχ2-distributions(e.g.Bev-ington1969,Ch.10).For both the circular and elliptical ﬁts the systemic velocity,γ,was heldﬁxed at zero since our method of radial velocity extraction uses an orbital-average as a template;the orbital period was also heldﬁxed.The number of free parameters in eachﬁt is then f ell=4(K2,e,νandω)and f crc=2(K2andφ0).In statistical terms we are testing the hypothesis that the radial velocity curves are consistent with zero eccen-tricity.Table2shows for each radial velocity data set the probability(p)that this hypothesis is correct,and since p never falls below30percent,we must accept the hypothesis.6.2Are the new data consistent with thepreviously measured eccentricity?Although our data are consistent with a circular orbit,it is still possible that,were they noisy enough,they could also be consistent with an eccentricity at the level measured by M89.So we determined the maximum level of eccentricity that could be present in our data.For this exercise weﬁtted simultaneously the veloci-ties from both TiO bands,but retained the separate data sets for each resolution.For each data set we constructed an e-χ2space by freezing e at set values but allowing all other parameters(except orbital period,always heldﬁxed) to run free.Imposing the condition thatχ2ν=1for the e=0ﬁt,deﬁnes the error for each velocity point.From the90% conﬁdence interval forﬁve free parameters(Lampton,Mar-gon&Bowyer1976),both data sets yield an upper limit to any eccentricity in our data of around0.05.Thus we conclude that our upper limit to the eccentric-ity is smaller than M89’s detection(0.089±0.020).6.3Is the diﬀerence in eccentricity due to adiﬀerence in technique?There are two important diﬀerences between our measure-ments and those of M89.They used the NaI doublet,whilst we used TiO;and they cross correlated against an M-star template,whilst we used a velocity-shifted mean of the data. So,we performed a cross-correlation study using the M3.5V template GL27B,over the NaI region used by M89.For test-ing the signiﬁcance of the eccentricity we used f ell=5and f crc=3,sinceγis an extra free parameter in each case.The results are shown in Table2,and show that this study also returns zero eccentricity.The only remaining diﬀerence in technique is that M89 used a slightly diﬀerent formula for the radial velocity,which breaks down for e>0.1.Although this approximation should not create a problem,for completeness we also tested their table2data,labelled M89in our Table2.Our code ﬁnds a similar eccentricity(e=0.094),rejecting a circular orbit at greater than99.99%conﬁdence.6.4Orbital eccentricity;a discussionThe conclusion is thus inescapable,the apparent eccentric-ity of the orbit has changed between the observations of M89and our own.IP Peg is not the only system where an apparent eccentricity has been observed in the radial ve-locity data.AM Her,U Gem,CH UMa and YY Dra all show signiﬁcant eccentricities(Friend et al.1990a,b).The explanation that has been put forward to explain these de-viations is irradiation of the secondary star,either by the white dwarf,bright spot or a combination of both.Under this assumption,Davey&Smith(1992)were able to map the distribution of the NaI doublet across the surface of the secondary star.As Davey&Smith point out,there is noc RAS,MNRAS000,1–108Beekman et al.satisfactory theoretical explanation for the asymmetric dis-tributions this model predicts,nor is there any apparent correlation with other parameters which would indicate the presence of strong heating.As our spectra show the NaI dou-blet to be strongly contaminated by emission from the ac-cretion disc,we put forward an alternative hypothesis,that this contamination,even when it is much weaker,aﬀects the radial velocity curves of CVs.IP Peg is a stunning example of the random nature of deriving an eccentricity from radial velocity data:it has been observed twice,over the same wavelength region,both times in quiescence(although the M89data may have been on the decline from a small outburst),both use a similar procedure,both return the same value for the radial velocity of the secondary star and yet one shows an eccentricity while the other does not.The raison d’ˆe tre for M89’s correction to the radial velocity was the apparent eccentricity of the orbit. Signiﬁcant irradiation would introduce an eccentricity(see M89),so the the absence of such eﬀects(see also Section 8),means that we should not make any correction to our semi-amplitude.7COMPONENT MASSESThe mass functionP K32f(M)=;M2=qM1.(4) sin3(i)The mass ratio may be determined by combining our K2 with the rotational broadening measurement of Catal´a n, Smith&Jones(2000).Theyﬁnd V rot sin(i)=125±15km s−1whereV rot sin(i)。

Evidence of a cyclotron feature in the spectrum of the anomalous X-ray pulsar 1RXS J170849-

a rXiv:as tr o-ph/32490v124Fe b23EVIDENCE FOR A CYCLOTRON FEATURE IN THE SPECTRUM OF THE ANOMALOUS X-RAY PULSAR 1RXS J170849−400910Nanda Rea 1,2,Gian Luca Israel 2,7,Luigi Stella 2,7,Tim Oosterbroek 3,Sandro Mereghetti 4,Lorella Angelini 5,Sergio Campana 6,7,Stefano Covino 6ABSTRACT We report the results of a long observation of the Anomalous X-ray Pulsar 1RXS J170849−400910obtained with the Beppo SAX satellite in August 2001.The best ﬁt phase-averaged spectrum was an absorbed power law plus black-body model,with photon index Γ∼2.4and a black body temperature of kT bb ∼0.4keV.We conﬁrm the presence of signiﬁcant spectral variations with the rotational phase of the pulsar.In the spectrum corresponding to the rising part of the pulse we found an absorption-like feature at ∼8.1keV (a signiﬁcance of 4σ),most likely due to cyclotron resonant scattering.The centroid energy converts to a magnetic ﬁeld of 9×1011G and 1.6×1015G in the case of electrons and protons,respectively.If conﬁrmed,this would be the ﬁrst detection of a cyclotron feature in the spectrum of an anomalous X-ray pulsar.Subject headings:stars:magnetic ﬁelds —stars:pulsars:general —pulsar:individual:1RXS J170849−400910—X–rays:stars1.INTRODUCTIONAnomalous X-ray pulsars(AXPs)are characterized by spin periods in the range of5-12s, steady spin down(∼10−11ss−1),steep and soft X–ray spectra with luminosities exceding by several orders of magnitude their spin–down luminosities(Mereghetti&Stella1995;van Paradijs,Taam&van den Heuvel1995).Allﬁve conﬁrmed AXPs lie in the“galactic”plane and two(or three),are associated with supernova remnants.AXPs show no evidence for a companion and are thus believed to be isolated neutron stars either having extremely strong magnetic dipoleﬁelds(∼1014−1015G;“magnetars”;Duncan&Thompson1992,Thompson &Duncan1995)or accreting from a residual disk(Li1999,Alpar2001,Perna et al.2000). Alternatively,they can accrete from a light companion.1E1048–5937and1E2259+586, recently displayed short and intense X–ray bursts(Graviil,Kaspi,&Woods2002;Kaspi& Graviil2002),strengthening a possible connection between AXPs and Softγ-Ray Repeaters (SGRs;Zhang2002).For recent rewiews see Israel,Mereghetti&Stella2002,Mereghetti et al.2002,and references therein.1RXS J170849−400910was discovered with ROSAT(Voges et al.1996);∼11s pulsa-tions were found in its X–rayﬂux with ASCA(Sugizaki et al.1997).Early measurements suggested that is a fairly stable rotator(Israel et al.1999).In October1999a sudden spin–up event occurred,which was interpreted as a glitch(Kaspi et al.2000).Searches for an optical counterpart ruled out the presence of a massive companion(Israel et al.1999); an IR counterpart has been recently proposed(Israel et al.2003).The diﬀuse(∼8′)radio emission at1.4GHz(Gaensler et al.2001)is due to the supernova remnant G346.5–0.1,the association of which to the AXP is still under debate.There is no evidence of pulsed radio emission from the AXP with an upper limit of70µJy on the pulsation amplitude(Israel et al.2002).Here we report the results of a long Beppo SAX observation of the AXP that conﬁrms signiﬁcant spectral variations with pulse phase(Israel et al.2001)and shows the presence of an absorption-like feature at∼8.1keV,probably due to cyclotron resonant scattering.2.OBSERVATIONThe source was observed by Beppo SAX on2001August17–22with the imaging Narrow Field Instruments(NFI):the Low Energy Concentrator Spectrometer(LECS:0.1–10keV; Parmar et al.1997;78.0ks of eﬀective exposure time)and the Medium Energy Concentrator Spectrometer(MECS:1.3–11keV;Boella et al.1997;200.0ks exposure).We extracted the source photons in the MECS and LECS from a circular region of6arcmin radius around theposition of the AXP.Photons extracted from a sourceless circular region of the same size were used for background subtraction.MECS and LECS photon arrival times were corrected to the barycenter of the Solar System.3.RESULTS3.1.Timing and Spectral AnalysisIn order to obtain a precise estimate of the pulse period,we divided the MECS obser-vation in nine time intervals and calculated the pulsation phase for each of them.Fitting these phases with a linear function gave a best period of11.000563±0.000005s.The folded light curve shows an energy-dependent proﬁle(Fig.1).Speciﬁcally,the pulse minimum shifts from a phase of∼0.0in the lowest energy light curve(0.1–2keV)to∼0.3in the6–10keV light curve.Correspondingly,the pulsed fraction decreases from∼30%to∼17%.We restricted the analysis of the LECS and MECS spectra to the0.4–4keV and1.65–10.8keV energy range,respectively.The spectra were binned so as to have about two bins per spectral resolution element(FWHM).Futhermore,the data at the extremes of the spectral range given above were further rebinned so as to have at least20source events per bin(such that minimum chisquare techniques could be reliably used in spectralﬁtting).The spectra were wellﬁt with an absorbed blackbody plus a power law model(see Table1).The bestﬁt of the phase-average spectrum gave a reducedχ2of0.95for298 degree of freedom(dof)for the following parameters:column density of N H=(1.36±0.06)×1022cm−2,a blackbody temperature of kT bb=0.44±0.01keV(blackbody radius of R bb=6.6±0.4km,assuming a distance of5kpc)and a photon index ofΓ=2.40±0.06(all error bars in the text are90%conﬁdence).The unabsorbedﬂux in the0.5–10keV range was1.87×10−10erg cm−2s−1corresponding to a luminosity of5.6×1035erg s−1(for a5kpc distance).In the0.5–10keV band the blackbody component accounts for∼30%of the total unabsorbedﬂux.We also tried other combinations of spectral models(cutoﬀpower law plus blackbody or two blackbodies with diﬀerent temperatures)but they all produced larger reducedχ2values.3.2.Pulse Phase SpectroscopyWe adopt for zero phase the minimum of the pulse proﬁle in the0.2–3keV folded light curve(see Fig.1).In order to carry out pulse–phase spectroscopy we accumulated spectrain six diﬀerent phase intervals.The boundaries of these(0.0,0.26,0.4,0.58,0.7,0.84,1.0) were designed so as to sample separately the minimum,rising,maximum,and decaying part of the pulse proﬁle,while maintaining a suﬃciently good statistics for a detailed spectral study(results in Tab.1,Fig.2,3and4).A signiﬁcant variation of the spectral parameters with pulse phase was clearly seen(especially forΓand R bb,see Fig.2;see also Israel et al. 2001).In all intervals but one,an acceptableﬁt was obtained with the absorbed power–law plus black body model(reducedχ2in the0.8–1.1range);a reducedχ2of1.2was instead obtained in the0.4–0.58phase interval.Speciﬁcally,the data were systematically below the bestﬁt model in the∼7.8–8.4keV range(see Fig.4b).We tried toﬁt three diﬀerent models: a Gaussian,an absorption edge and a cyclotron feature.While the inclusion of a Gaussian or an absorption edge did not lead to a signiﬁcant improvement of theﬁt,the cyclotron model (CYCLABS in the XSPEC package,see Mihara et al.1990for details)led to a reduced chisquare of1.1for86dof,corresponding to an F-test probability of3×10−3(3.1σ).In order to improve the statistics we added in phase the spectra from the1999Beppo SAX observation of the source(Israel et al.2001),increasing the total exposure time to∼250ks. For this we used the pulse period determination from the timing solution of a4year–long R XTE monitoring(Kaspi et al.2000)8.The determination of the zero phase was done again by using the pulse proﬁle minimum in the0.3–2keV energy range(note that no signiﬁcant shape diﬀerence was found with respect to the2001observation).We estimate that our procedure can introduce an uncertainty of up to0.01in the phasing of the two observations which is negligible for the aims of our study.For each of the six phase intervals,a spectrum was summed together with the cor-risponding spectrum from the2001observation.The spectral analysis was then repeated. In the MECS and LECS spectra from the0.4–0.58phase interval the inclusion of a Gaus-sian(E g=8.3keV,σ=0.4keV andχ2=1.1)and an absorption edge(E e=7.7keV,τ=0.4 andχ2=0.9)toﬁt the feature at∼8.1keV resulted in an improvement of the chisquare, which converted to a single-trial F-test probability of0.18(∼1σ)and4×10−4(∼3.7σ), respectively.A much more signiﬁcant improvement was obtained by adding instead a Res-onant Cyclotron Feature(RCF)model;the F-test probability in this case was1.8×10−5 corresponding to a single trial signiﬁcance of4.5σor4σafter correction for the six spectra that we analysed(see Fig.4and Tab1).4.DISCUSSIONDuring a Beppo SAX study of1RXS J170849−400910we discovered an absorption-like feature at an energy of∼8.1keV in a pulse phase interval corresponding to the rising part of the∼11s pulse.This feature was bestﬁt by a resonant cyclotron feature model with a centroid energy of∼8.1keV and an equivalent width of∼460eV.The detection of an RCF in a speciﬁc pulse-phase interval and superposed to an X-ray continuum that varies with the pulse phase is reminiscent of the behaviour seen in standard accreting pulsars in X-ray binaries(see Wheaton et al.1979,Santangelo et al.1999).If interpreted as an electron resonant feature at the base of the accretion column,the feature at∼8.1keV implies a neutron star surface magneticﬁeld of∼9.2×1011Gauss(using a gravitational redshift z=0.3).This value is just slightly lower than that measured for electron RCFs in typical accreting X-ray pulsars(see Fig.5);more interestingly it is close to that required by models for AXPs which involve residual disk accretion in the spin–down regime(Mereghetti&Stella1995;Alpar2001;Perna et al.2001).In this context,one can solve the torque equation(see e.g.Eq.11.35in Henrichs1983)by exploiting the measured value of˙P and range of accretion luminosity derived from plausible distances(5–10kpc). The surface magnetic(dipole)ﬁeld obtained in this way is0.6–1.1×1012G(corrisponding to a fastness parameter range ofωs=0.57−0.54,a typical value for the spin-down accretion regime;see Ghosh&Lamb1979and Henrics1983).The agreement of this estimate with the magneticﬁeld inferred from the electron RCF interpretation is intriguing,especially in consideration of the other analogies with the pulse-phase spectral dependence of conventional accreting X-ray pulsars.By contrast,if an electron RCF arose somehow at the polar caps of a rotation powered pulsar,a B–ﬁeld strength of 9.2×1011G would be in the range of many radio pulsars and yet much lower than that required to spin–down at the observed rate through magnetic dipole radiation(∼5×1014G, indeed this was one of the motivations for magnetar model,see below).Alternatively the RCF might be due to protons.For the magneticﬁeld strengths forseen in the“magnetar”scenario,proton cyclotron features(if any)are expected to lie in the classical X–ray band(0.1–10keV;Zane et al.2001;Lai&Ho2002).A proton RCF feature at∼8.1keV would correspond to surfaceﬁeld of1.6×1015G(z=0.3).The fact that this value is∼3times higher than the surfaceﬁeld derived from the usual magnetic dipole spin down formula should not be of concern.According to the magnetar model,the magneticﬁeld at the star surface and its vicinity is dominated by higher order multipoleﬁeld components (Thompson&Duncan1995).At large radii the dipole component,responsible for the secular spin–down,dominates.It is thus expected that a proton RCF feature,sampling the(total) surface magneticﬁeld strength,provides a higher value than the mere dipole component.There is a clear correlation between the width and centroid energy of the electron RCFs in accreting X–ray pulsars(see Fig.5extending the results of Orlandini&Dal Fiume2001). The values from1RXS J170849−400910and SGR1806–20are in good agreement with such a relation.The modest range of width to centroid energy ratio implied by this indicates that magneticﬁeld geometry eﬀects at the neutron star surface likely dominate the RCF width(on the contrary temperature and particle mass would alter this ratio).This,in turn, suggests that similar(relative)ranges of surface magneticﬁeld strength are“sampled”by RCFs in accreting X–ray pulsar and,by extension,RCFs in AXPs and SGRs.Other interpretations of the feature at∼8.1keV appear less likely.Firstly,ﬁtting an edge or line due to photo-electric absorption provides a less pronounced improvement of the ﬁt than the RCF model.Secondly,an edge by iron at a suﬃciently large distance from the neutron star that energy shifts are negligible would require a high overabundance of this element and intermediate ionisation stages(such as C–like iron).Yet it has long been known that the photoionisation equilibrium of such a plasma is unstable(Krolik&Kallman1984; Nagase1989).The energy of an ion feature forming in the neutron star atmosphere would be drastically altered by magneticﬁeld eﬀects(see Mori&Hailey2002and references therein). In this and the above interpretations,however,it would also be diﬃcult to explain why an ion feature is observed only over a restricted range of pulse phases.Exploring in detail these possibilities is beyond the scope of this letter.In conclusion,we found an absorption-like feature in the Beppo SAX X–ray spectrum of1RXS J170849−400910taken during the rising phase of the∼11s pulse,which is best ﬁt by a cyclotron resonant scattering model.If this interpretation is correct,the centroid energy translates into a magneticﬁeld strength of∼1.6×1015G and∼9.2×1011G depending on whether protons or electrons,respectively,are responsible for the feature.Nanda Rea acknowledges useful discussions with S.Dall’Osso.This work was partially supported through ASI and COFIN2000grants.Table1:Bestﬁt spectral parameters from selected pulse-phase intervals:that showing feature at∼8.1keV(0.4-0.58)and those characterized by the lowest and the highest power law photon indexΓ(see also Fig.3).Allﬂuxes are unabsorbed and calculated in0.5–10keV range;uncertainties are90%conﬁdence.Phase IntervalsParameters0.4–0.580.84–1.00.26–0.4Fig.1.—MECS light curves of1RXS J170849−400910folded at the best spin period(two pulse cycles are shown)for six energy bands:(a)0.1–2keV;(b)2–3keV;(c)3–4keV;(d)Fig. 2.—Spectral parameter variability from pulse-phase spectroscopy using an absorbedblack body plus power-law model.Filled squares represent the spectral parameters after theFig.3.—LECS and MECS spectra in the two phase intervals characterised by the highestand lowest value of the power law photon indexΓ;see also Table1.Fig.4.—MECS and LECS spectra from the0.4–0.58phase intervalﬁt with the“standardmodel”(the sum of a blackbody and power law with absorption)plus a cyclotron line.Fig. 5.—Line width vs.centroid energy from a sample of accreting X-ray pulsars withelectron RCFs(Orlandini&Dal Fiume2001),1RXS J170849−400910(this Letter)andREFERENCESAlpar,M.A.,2001,ApJ,554,12-45Boella,G.,et al.1997,A&AS,122,327Duncan,R.C.,&Thompson,C.1992,ApJ,392,L9Gaensler,B.M.,et al.2000,MNRAS,Volume318,Issue1,pp.58-66Ghosh,P.,Lamb,F.K.,1979,ApJ,Part1,vol.234,p.296-316Graviil,F.P.&Kaspi,V.M.2002,ApJ,567,1067GGraviil,F.P.,Kaspi V.M.,&Woods,P.M.2002,Nature,419,142GHenrichs,H.F.,1983Accretion-driven stellar X-ray sources P.419Ibrahim,A.I.,Swank,J.H.,Parke,W.2002,astro–ph,0210515Ibrahim,A.I.,Saﬁ-Harb,S.,Swank,J.H.,Parke,W.,Zane,S.,Turolla,R.2002,ApJ,574, L51Israel,G.L.,Covino,S.,Stella,L.,Campana,S.,Haberl,F.,Mereghetti,S.1999,ApJ,518, L107Israel,G.L.,Oosterbroek,T.,Stella,L.,Campana,S.,Mereghetti,S.,Parmar,A.,2001, ApJ,560,L65Israel,G.L.,Mereghetti,S.,&Stella,L.2002a,inγ–Ray Bursts in the Afterglow Era,ed.S.Mereghetti&M.Feroci,Mem.S.A.It.,Vol.73,N.2,pag.465Israel,G.L.,et al.2003,ApJ submittedKaspi,V.M.,Lackey,J.R.,Chakrabarty,D.2000,ApJ,537,L31Kaspi,V.M.,&Gravill,F.P.,2002,IAU Circ,No.7924Krolik,J.H.,&Kallman,T.R.1984,ApJ,286,366KLai,D.,&Ho,W.C.G.,2002,astro–ph0211315Li,X.-D.,1999,ApJ,520,271LMereghetti,S.,&Stella,L.1995,ApJ,442,L17Mereghetti,S.,Chiarlone,L.,Israel,G.L.,Stella,L.2002,MPE Rep,278;Garching:MPE, 29Mihara,T.,et al.,1990,Nature,346,250MMori,K.,&Hailey,C.J.2002,ApJ,564,914Nagase,F.1989,PASJ,41,1-79Orlandini&Dal Fiume2001,AIP conference proceedings1999,Vol.599p.283Parmar,A.N.,et al.1997,A&AS122,309Perna,R.,Heyl,J.,Hernquist,L.,Juette,A.,Chakrabarty,D.2001,ApJ,557,18P Perna,R.,et al.,2000,ApJ,541,344Santangelo,A.,et al.,1999,ApJ,523,L85Sugizaki,M.,et al,1997,PASJ,v.49,p.L25-L30Thompson,C.,&Duncan,R.C.1995,MNRAS,275,255van Paradijs,J.,Taam,R.E.,&van den Heuvel,E.P.J.,1995,A&A,299,L41Voges,W.,et al.1996,IAU Circ.,6420,2(1996).Edited by Green,D.W.E. Wheaton,W.A.,et al.,1979,Nature282,p.240Zane,S.,Turolla,R.,Stella,L.,&Treves,A.,2001,ApJ,560,384Zhang,B.,2002,astro-ph,0212016。

一种基于三轴正交接收线圈的磁感应透地通信方案

第 55 卷第 3 期2024 年 3 月中南大学学报(自然科学版)Journal of Central South University (Science and Technology)V ol.55 No.3Mar. 2024一种基于三轴正交接收线圈的磁感应透地通信方案王羿帆，杨维，戴蒙天(北京交通大学电子信息工程学院，北京，100044)摘要：为提高磁感应透地通信系统的灵活性和覆盖范围，提出了一种新型透地通信方案，该方案采用单轴空气线圈作为地面发射机天线，用3个相同参数的空气线圈作为接收机天线。

首先，对通信信号进行空间分集接收，使通信系统的井下接收机能够以任意姿态地摆放于巷道之中，并能实现更广范围的信号覆盖；其次，为消除因接收线圈间近场耦合作用而产生的信号干扰，将接收机3个天线设置成共用1个圆心且两两相互正交的空间结构；最后，以远距离通信场景下收发线圈间的弱耦合情况为依据，建立一套与接收方案相对应的等效电路模型，并搭建仿真实验平台，分析提出的方案的通信性能。

研究结果表明：三轴正交磁感应透地通信接收方案的信号接收强度可达到同等通信距离下理想信号强度的57.7%以上，接收功率可达到最优接收功率的33.33%以上。

本文提出的方案解决了单轴线圈接收方案中的通信盲区问题，且有效提高了远距离磁感应透地通信系统的覆盖范围和对应急救援场景的适应性。

关键词：透地通信；磁感应；远距离；应急通信中图分类号：TD65.5 文献标志码：A 开放科学(资源服务)标识码(OSID)文章编号：1672-7207（2024）03-0973-09A magnetic induction through-the-earth communication schemebased on three-axis orthogonal receiving coilWANG Yifan, YANG Wei, DAI Mengtian(School of Electronics and Information Engineering, Beijing Jiaotong University, Beijing 100044, China)Abstract: In order to improve the flexibility and coverage of the magnetic induction through-the-earth (TTE) communication system, a new TTE communication scheme was proposed. The scheme uses a single axis air coil as the ground transmitter antenna and three air coils with the same parameters as the receiving antenna. Firstly, the communication signal was received by space diversity, so that the underground receiver of the system can be placed in the tunnel in any posture and achieve a wider range of signal coverage. Secondly, in order to eliminate the signal interference caused by the near-field coupling between the receiving coils, the three antennas of thereceiver were set to a spatial structure with a common center and orthogonal to each other. Finally, based on the收稿日期： 2023 −08 −30；修回日期： 2023 −11 −20基金项目(Foundation item)：国家自然科学基金资助项目(62071032) (Project(62071032) supported by the National Natural ScienceFoundation of China)通信作者：杨维，博士，教授，从事无线通信与无线安全监测技术研究；E-mail ：**************.cnDOI: 10.11817/j.issn.1672-7207.2024.03.012引用格式：王羿帆, 杨维, 戴蒙天. 一种基于三轴正交接收线圈的磁感应透地通信方案[J]. 中南大学学报(自然科学版), 2024, 55(3): 973−981.Citation: WANG Yifan, YANG Wei, DAI Mengtian. A magnetic induction through-the-earth communication scheme based on three-axis orthogonal receiving coil[J]. Journal of Central South University(Science and Technology), 2024, 55(3): 973−981.第 55 卷中南大学学报(自然科学版)weak coupling between the transmitting and receiving coils in the long-distance communication scenario, a set of equivalent circuit models corresponding to the receiving scheme were established. A simulation experiment platform was built to analyze the communication performance of the proposed scheme. The results show that the signal receiving strength of the three-axis orthogonal magnetic induction TTE communication receiving scheme can reach more than 57.7% of the ideal signal strength under the same communication distance, and the receiving power can reach more than 33.33% of the optimal receiving power. The scheme proposed in this paper solves the problem of communication blind area in the single-axis coil receiving scheme, and effectively improves the coverage of the long-distance magnetic induction TTE communication system and the adaptability to emergency rescue scenarios.Key words: through-the-earth communication; magnetic induction; long-distance; emergency communication煤炭是我国能源结构中十分重要的一环，但我国近95%的煤矿需要进行井下开采[1]。

C4_旋转对称光子晶体平板中的对称保护连续谱束缚态

第40卷第3期Vol.40㊀No.3重庆工商大学学报(自然科学版)J Chongqing Technol &Business Univ(Nat Sci Ed)2023年6月Jun.2023C4旋转对称光子晶体平板中的对称保护连续谱束缚态张铭洋重庆工商大学数学与统计学院,重庆400067摘㊀要:在光子晶体平板中,连续谱束缚态关于C2和C6旋转对称的依赖性已经在数值上进行了广泛研究,但是缺少严格的理论分析过程,此外还缺少对C4旋转对称的研究,鉴于此,构建了系统分析连续谱束缚态关于所有旋转对称的依赖性的理论,并且重点研究了C4旋转对称的情况;首先,通过分析具有旋转对称的结构中麦克斯韦方程组特征解的性质,将连续谱束缚态的存在性问题转变为旋转矩阵的特征值是否与一个简单代数方程的解相同的问题;其次,给出了C4旋转对称的结构中连续谱束缚态存在时所对应的条件;然后,证明了破坏C4旋转对称保持C2旋转对称时,连续谱束缚态依然存在;最后,利用有限元软件FreeFEM 进行了大量的数值验证;上述理论可适用于所有旋转对称的情况,深入揭示了旋转对称对连续谱束缚态存在的重要性,深入揭示了高阶旋转对称性与低阶旋转对称性之间的依赖关系,为连续谱束缚态的实际应用提供了理论指导㊂关键词:光子晶体;旋转对称;连续谱束缚态中图分类号:O436㊀㊀文献标识码:A㊀㊀doi:10.16055/j.issn.1672-058X.2023.0003.09㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀㊀收稿日期:2022-05-13㊀修回日期:2022-06-20㊀文章编号:1672-058X(2023)03-0064-07基金项目:重庆市自然科学基金面上项目(CSTC2019JCYJ -MSXMX0717).作者简介:张铭洋(1997 ),女,重庆忠县人,硕士研究生,从事光子晶体㊁麦克斯韦方程组数值计算研究.引用格式:张铭洋.C4旋转对称光子晶体平板中的对称保护连续谱束缚态[J].重庆工商大学学报(自然科学版),2023,40(3):64 70.ZHANG Mingyang.Symmetry-protected bound states in the continuum in C4rotationally symmetric photonic crystal plates J .Journal of Chongqing Technology and Business University Natural Science Edition 2023 40 3 64 70.Symmetry-protected Bound States in the Continuum in C 4Rotationally Symmetric Photonic Crystal Plates ZHANG MingyangSchool of Mathematics and Statistics Chongqing Technology and Business University Chongqing 400067 ChinaAbstract The dependence of bound states in the continuum BICs on C2and C6rotational symmetry in photonic crystalplates has been extensively studied numerically.However a rigorous theoretical analysis process is lacking and there is alack of studies on C4rotational symmetries.In view of this a theory of systematic analysis of the dependence of BICs on all rotational symmetries was constructed and the case of C4rotational symmetry was mainly studied.Firstly by analyzingthe characteristic solutions of Maxwell s equations with rotationally symmetric structure the problem of the existence of BICs was transformed into the question of whether the eigenvalue of the rotation matrix was the same as the solution of asimple algebraic equation.Secondly the conditions for the existence of BICs in C4rotationally symmetric structures weregiven.Then it was proved that the BICs still existed when C4rotation symmetry was destroyed and C2rotatory symmetrywas maintained.Finally the finite element software FreeFEM was used to do a lot of numerical verifications.The abovetheory can be applied to all cases of rotational symmetries revealing the importance of rotational symmetry for the existenceof BICs.The dependence between high-order and low-order rotational symmetries was revealed providing theoretical guidance for applying BICs.Keywords photonic crystal rotational symmetry bound states in the continuum第3期张铭洋,等:C4旋转对称光子晶体平板中的对称保护连续谱束缚态1㊀引㊀言光学连续谱束缚态(bound states in the continuum, BIC)是指位于连续谱中的导模,其不能与辐射场耦合,没有能量辐射,被完美地束缚在结构中[1-3]㊂数学上,光学连续谱束缚态是指开结构中麦克斯韦方程组的一类频率位于连续谱内的平方可积特征解㊂通常,导模(即平方可积特征解)的特征频率位于连续谱外㊂1929年冯诺依曼等[4]从数学模型上发现在一些特殊的结构中存在特征频率位于连续谱内的导模㊂直到1985年文献[5]才构造出具有连续谱束缚态的真实物理系统㊂2008年,文献[3]研究了光子晶体结构中的连续谱束缚态㊂此后,连续谱束缚态受到广泛关注,与之有关的研究快速发展㊂目前,连续谱束缚态的概念和研究已推广到水波㊁声波等其他波动现象[1]㊂连续谱束缚态可看成为品质因子为无穷大的共振,只存在于若干离散的频率上㊂连续谱束缚态由共振模式所包围㊂通过扰动波矢,可在连续谱束缚态附近找到任意大小品质因子的共振模式[6]㊂此性质使得连续谱束缚态在光学㊁光子学等领域都拥有广阔的应用前景㊂目前,连续谱束缚态已在波导㊁光栅㊁光子晶体及超材料等结构中被广泛研究[7],光子晶体中的连续谱束缚态现在已经被用于传感器,激光器和滤波器的设计当中[8-10]㊂通常,共振模式的品质因子与波矢之差的平方成反比㊂文献[6]证明了存在特殊的连续谱束缚态使得附近共振模式的品质因子与波矢之差的四次方和六次方成反比,并给出了两类特殊连续谱束缚态的条件㊂从实际应用角度来讲,在这些特殊的连续谱束缚态附近更容易构造出高品质因子的共振模式㊂连续谱束缚态可以大致分为两类:对称保护的连续谱束缚态[11-15]和非对称保护的连续谱束缚态[2,16-19]㊂对称保护的连续谱束缚态(Symmetry Protected Bound states in the continuum,SPBIC)的机理是:在对称结构中,布洛赫模的对称性与结构中辐射场的对称性不相容,从而与辐射场不耦合,变成一个连续谱束缚态[3]㊂而非对称保护连续谱束缚态的存在机理是:共振模式的辐射场之间发生干涉相消现象,造成没有辐射,成为连续谱束缚态[3]㊂连续谱束缚态的存在性与结构的对称性具有密切联系㊂早期研究结果都是在对称结构中研究连续谱束缚态,学术界一度认为连续谱束缚态只存在于对称结构中㊂目前数学上还没有非对称保护连续谱束缚态的存在性理论㊂非对称保护连续谱束缚态关于结构对称性的依赖关系非常复杂㊂文献[20-23]从数值和实验上演示了破坏二维结构的C2旋转对称性后,连续谱束缚态演化为共振模式㊂这间接说明了结构的对称性对非对称保护连续谱束缚态的存在性具有重要影响㊂但是破坏对称性后连续谱束缚态是否一定会演化为共振模式并没有明确的结论㊂最近,文献[24-26]证明了只要引入足够多的结构扰动参数,连续谱束缚态可以连续存在于非对称的结构中,且对于不同类型的连续谱束缚态,所需要引入的最小参数的数量是不同的㊂上述结论表明,非对称保护连续束缚态可以存在于非对称结构中,只要结构的自由参数足够多㊂对称保护连续谱束缚态只存在于对称结构中㊂光子晶体平板可具有四类旋转对称性:C2㊁C3㊁C4和C6旋转对称性,即分别旋转180㊁120㊁90和60度后结构不变㊂文献[11-12]首先从数学理论上证明了在具有C2旋转对称的二维介质结构中对称保护连续谱束缚态的存在性,在非对称结构中一定不存在对称保护连续谱束缚态㊂研究对称保护连续谱束缚态对上述四种不同类型对称性的连续依赖性具有十分重要的意义㊂文献[21]研究了C6旋转对称性对具有拓扑电荷为q=-2的对称保护连续谱束缚态存在性的影响㊂通过数值计算发现破坏C6对称保持C2对称,对称保护连续谱束缚态依然存在,但是变成拓扑电荷q=-1;破坏C6对称保持C3对称,对称连续谱束缚态演化为共振模式,而且会产生两个非对称保护连续谱束缚态㊂上述研究结果给出了一种产生非对称保护连续谱束缚态的方法㊂目前,对于拓扑电荷为q=-1或q=1的对称保护连续谱束缚态关于结构对称性的依赖关系没有进行系统讨论,缺乏严格系统的依赖性理论㊂研究C4旋转对称结构中对称保护连续谱束缚态关于对称性的依赖关系㊂建立了严格数学理论证明破坏C4对称保持C2对称,对称保护连续谱依然存在㊂并利用有限元软件FreeFEM进行数值验证㊂相比于以前的研究,研究既有严格的数学理论,又有数值验证㊂研究成果具有一般性,可推广到分析对称保护连续谱束缚态关于C6对称性的依赖性,有利于深入理解连续谱束缚态关于对称性的依赖关系,为其实际应用提供理论指导㊂2㊀连续谱束缚态考虑一个在x与y方向为周期,在z方向上厚度有限的光子晶体平板㊂光子晶体平板通过在平板上构造正方形空气柱晶格所构成㊂设平板厚度为2D,晶格常数为L,平板的介电常数为ε1,空气的介电常数为ε0= 1㊂记整个结构的介电常数为ε(r),其中r=(x,y,z),则z>D时有ε(r)=1,且ε(r)满足ε(r)=ε(x+mL,y+nL,z)(1)其中,m与n为任意整数㊂设光子晶体平板是无磁性㊁各向同性的,由麦克斯56重庆工商大学学报(自然科学版)第40卷韦方程组可知,具有时间依赖e -iwt 的时谐波的电场E 满足如下的控制方程:∇ˑ∇ˑE -k 2εE =0∇㊃(εE )=0其中k =ωc为真空中的波速,ω为角频率,c 为真空中的光速㊂光子晶体平板中的布洛赫模(即麦克斯韦方程组的特征解)可写成:E (r )=Φ(r )e ik ㊃r其中,k =(α,β,0)为布洛赫波矢,实数α与β分别为x 与y 方向的布洛赫波速,Φ(r )满足周期条件式(1)㊂由于在z >D 时,结构是均匀的,由傅立叶展开式与分离变量法可知,满足向外辐射条件的布洛赫模可以展开为[18]E (r )=ð+ɕm ,n =-ɕd ʃm ,neik ʃm ,n㊃r ʃz >D (2)其中,常数向量d ʃm ,n满足d ʃm ,n ㊃k ʃm ,n =0,k ʃm ,n=(αm ,βn ,ʃγm ,n ),αm =α+2m πL ,βn =β+2n πL,γm ,n =k 2-α2m -β2n ㊂若结构是无耗散的,即ε(r )为非负实函数,则布洛赫模可以分为三类:导模㊁共振以及连续谱束缚态㊂(1)若k 为实数,则布洛赫模为导模㊂可以证明k 为实数等价于E (r )满足lim z ңɕE (r )=0,即没有能量辐射,没有能量损失㊂当波速和布洛赫波矢满足条件0<k <α2+β2时,导模关于α与β连续存在㊂在上述条件下,γm ,n 的虚部都大于0,所以展开式(2)中每一个平面波都在无穷远出衰退到0,即E (r )自动满足lim z ңɕE (r )=0,从而是一个导模㊂(2)若k 为复数,则布洛赫模为共振模式㊂共振模式的波速k 满足[Re(k )]2-[lm(k )]2>α2+β2㊂由于共振的波速k 为复数且满足向外辐射条件,共振满足条件lim z ңɕE (r )=ɕ,即在空间上是无限增大,但随时间指数衰退㊂共振波速k (或共振频率ω)的虚部小于零,即lm(k )<0,它表示共振随着时间衰退的速度㊂共振的品质因子Q 定义为Q =-12Re(k )lm(k ),表示共振模式的振幅衰退到原来的e -1时所需要的振荡周期㊂共振模式关于α与β也是连续存在的㊂(3)若k 为实数且满足k >α2+β2,则布洛赫模是一个连续谱束缚态㊂连续谱束缚态可以看成是一个Q 因子为无穷大的共振,只在离散的(α,β)点上存在,在连续谱束缚态的附近,通过调整α与β可以获得任意大小Q 因子的共振㊂由于k 为实数等价于条件limz ңɕE (r )=0,在展开式(2)中,若lm(γm ,n )ȡ0,则e ikm ,n㊃r可向z ңɕ辐射能量,(m ,n )是对应一个开放的辐射通道㊂若lm(γm ,n )<0,则eik ʃn ,m ㊃r在z ңɕ时衰退到零,对应一个关闭的辐射通道㊂若Re(k )>α2+β2,则至少有lm(γ0,0)ȡ0,即(0,0)处辐射通道是开放的㊂记Z 0表示所有开放的辐射通道,即Z 0=(m ,n )lm(γm ,n )ȡ0{},则条件lim z ңɕE (r )=0等价于d ʃm ,n=0,∀(m ,n )ɪZ 0(3)式(3)是布洛赫模的一个附加条件,在一般情况下,连续谱束缚态不容易存在㊂3　对称保护连续谱束缚态当光子晶体平板具有旋转对称性时,可能存在对称保护连续谱束缚态㊂下面给出具有C n 旋转对称的光子晶体平板中对称保护连续谱束缚态的定义,并分析其关于对称性的依赖关系㊂利用旋转对称性下布洛赫模的性质,将连续谱束缚态的存在性问题转变为旋转矩阵的特征值是否与一个简单代数方程的解相同的问题;其次,给出了C4旋转对称的结构中连续谱束缚态存在时所对应的条件;然后,证明了破坏C4旋转对称保持C2旋转对称时,连续谱束缚态依然存在;具有C n 旋转对称性结构的介电函数ε(r )满足条件ε(r )=ε(T -1r )其中,T =cos φ-sin φ0sin φcos φ0001éëêêêùûúúú表示旋转矩阵,φ=2πn ,只考虑n =2与n =4的情况㊂其理论可推广到n =3与n =6的情况㊂C n 旋转对称光子晶体平板中的布洛赫模具有以下性质[24]:若E (r )=Φ(r )e ik ㊃r 是一个对应于波矢k =(α,β,0)和波速k 的布洛赫模,则TE (T -1r )是一个对应于波矢Tk 和波速k 的布洛赫模㊂特别地,取α=β=0,即k =(0,0,0),有Tk =k ,此时E (r )与TE (T -1r )是对应同一个波矢与波速的两个布洛赫模㊂若特征值问题是非退化的,则E (r )与TE (T -1r )线性相关,即存在常数τ使得:TE (T -1r )=τE (r )(4)由于任何结构旋转n 次2πn角度(即2π)后都不变,有T n=I ,其中I 表示单位算子㊂所以有τn=1,即τ=e i 2πn j,j =0,1,2, ,n -1㊂更具体地,当n =2时,τ=ʃ1;当n =4时,τ=ʃ1,ʃi ㊂注意到τ的取值对应于C n 点群的不可约表示的特征㊂设布洛赫模的波矢为k =(0,0,0)且频率满足0<Re(k )ε0<2πL,即只有(0,0)处辐射通道是开放的,则当n =2时,对应于τ=1以及当n =4时,对66第3期张铭洋,等:C4旋转对称光子晶体平板中的对称保护连续谱束缚态应于τ=ʃ1的布洛赫模一定是连续谱束缚态,称为对称保护连续谱束缚态㊂在上述条件下,布洛赫模是一个连续谱束缚态的充要条件是d ʃ0,0=0㊂下面证明当n =2时,τ=1以及当n =4时,τ=ʃ1,有d ʃ0,0=0㊂将展开式(2)代入条件式(4)得:Td ʃ0,0=τd ʃ0,0(5)令,d ʃ0,0=d ʃx ,d ʃy ,d ʃz []T ,T ^=cos φ-sin φsin φcos φéëêêùûúú,d ʃʅ=d ʃx ,d ʃy []T 表示d ʃ0,0的x 与y 分量所构成的向量㊂由d ʃ0,0㊃k ʃ0,0=0且α=β=0,知d ʃz k =0,即d ʃz =0㊂所以d ʃ0,0=0等价于d ʃʅ=0㊂由式(5)得:T ^d ʃʅ=τd ʃʅ若τ不是T ^的特征值,则必有d ʃʅ=0㊂当n =2时,T ^=-100-1éëêêùûúú,T ^只有一个特征值-1㊂所以当τ=1时,有d ʃʅ=0㊂当n =4时,T ^=0-110éëêêùûúú,此时T ^的特征值为ʃi ㊂所以当τ=ʃ1时有d ʃʅ=0㊂由上面的证明过程可知,条件d ʃ0,0=0是由C n 对称性所保证的㊂在具有C4旋转对称的结构中,对称保护连续谱束缚态对应的τ=1或-1㊂注意到无论是τ=1还是-1,都有τ2=1,即这些连续谱束缚态也同时由C2旋转对称所保护㊂有以下结论:具有C4旋转对称结构中的对称保护连续谱束缚态都是由C2旋转对称所保护的,即破坏C4旋转对称,保持C2旋转对称,这些连续谱束缚态依然存在㊂4㊀拓扑电荷连续谱束缚态对应于动量空间中辐射场的漩涡,因此其存在性与拓扑性质有关㊂前面提到了布洛赫模是一个连续谱束缚态的充要条件是d ʃ0,0=0,通过d ʃ0,0的x 与y 分量可以计算得到辐射场的极化角㊂极化椭圆的长轴与y 轴的夹角称之为极化角度,记为θ㊂θ可以看成是α与β的函数,即θ=θ(α,β)㊂在αβ平面上,任意给定一条曲线Γ,让(α,β)沿着Γ绕一圈重新定义θ,使其为连续函数㊂拓扑电荷的定义为q =12πɥΓd θ=12πɥΓ∇θ㊃ n d s拓扑电荷q 表示αβ上的一点绕Γ走一圈后,极化角度旋转了q 圈,q 是一个整数㊂若Γ所围区域内无圆极化与连续谱束缚态,则q =0;若Γ所围区域内有且仅有一个连续谱束缚态则q =ʃ1,ʃ2,ʃ3, ;若Γ所围区域内只有一个圆极化,则q =ʃ12㊂需要注意的是圆极化和连续谱束缚态是αβ平面中的一个极化奇点㊂5㊀数值实验由于辐射边界条件下的特征值问题定义在无穷区间上,无法用数值方法来计算㊂所以在实际计算连续谱束缚态的时候,可以用完美匹配层的方法来将无穷区域截断为有限区域㊂用完美匹配层截断后的特征值问题是原特征值问题的一个近似,它们之间的误差关于完美匹配层的参数σ∗㊁H 2-H 1(即完美匹配层的厚度)指数衰退到零㊂所以只需要选择合适的σ∗与H 2-H 1,便可以得到足够精确的特征解,即可以计算得到连续谱束缚态的频率㊂相对于拟周期边界条件,在有限元方法中周期边界条件更容易实现㊂用有限元方法求解偏微分方程最重要的是弄清楚解空间和变分形式㊂在用有限元求解时,变分问题被近似为下列代数方程的特征值问题:A Φ=k 2B Φ其中,A 与B 为矩阵㊂考虑如图1所示的具有正方形晶格空气柱的光子晶体平板,其俯视图如图2所示㊂平板是由空气所包围的㊂平板的厚度为2D =0.5L ,介电常数为ε1=4,空气中的介电常数ε0=1㊂空气柱体横截面参数分别为w =0.2L ,a =w2㊂若h 1=h 2,则结构具有C4旋转对称性㊂若h 1ʂh 2,则结构只有C2旋转对称性㊂为了验证前面得到的理论,用完美匹配层[27]的方法将无穷区域上的特征值问题转化为有限区域上的特征值问题,并用有限元[28]的方法求解㊂数值计算时,需要用完美匹配层方法将z 方向截断为-H 2,H 2[],如图3所示㊂取完美匹配层的厚度为H 2-H 1=L ,σ∗=18ˑm +1β0(H 2-H 1)[29],β0=k 20ε0-α2-γ2,m =3㊂其中H 1-D 表示完美匹配层的远近㊂采用基于FreeFEM 软件的有限元方法来数值求解特征值问题,以计算对称保护连续谱束缚态的频率㊂计算时在平板的每个边界的离散点个数取N =10,PML 层的离散点个数也取N =10㊂考虑如此复杂结构的原因是为了避免其他对称性(例如镜面反射对称)对结果的影响㊂ε=ε0ε=ε0z =D z =-Dzxy图1㊀光子晶体平板结构图Fig.1㊀Structure diagram of photonic crystal plates76重庆工商大学学报(自然科学版)第40卷h 2h 1h 1h 2L Lαωωxy图2㊀光子晶体平板结构的俯视图Fig.2㊀Top view of the photonic crystal flat plate structurez=H2z=H1z=D z=-Dz=-H1z=-H2ε=εε=εzxy 图3㊀PML截断后的计算区域Fig.3㊀Computation region after PML truncation 若取h1=h2=0.15L,这时结构具有C4旋转对称性㊂通过数值计算,可以找到5个TM-Like模式下(即E z是z变量的奇函数)的对称保护连续谱束缚态,其频率如表1的第2列所示㊂图4(a) 图8(a)分别是SPBIC1-SPBIC5在具有C4旋转对称的结构中log10Q 关于α与β的值㊂可以通过观察得到当(α,β)ң(0, 0)时,Q的值趋近于无穷大㊂图4(c) 图8(c)分别是SPBIC1到SPBIC5在C4旋转对称结构中的磁场z分量H z在z=0时的场图㊂从下面的场图可以观察得到, SPBIC1与SPBIC5对应于τ=1,其他3个对称保护连续谱束缚态对应于τ=-1㊂表1的最后一列表示为对称保护连续谱束缚态的拓扑电荷㊂若保持h1=0.15L,令h2=0.1L,参数扰动后结构的C4旋转对称性被破坏,但保持了C2旋转对称性㊂通过数值计算表明,SPBIC1-SPBIC5在扰动后的结构中依然存在,其频率如表1的第三列所示,可以发现两种结构下连续谱束缚态的频率近乎相等㊂图4(b) 图8(b)分别是SPBIC1到SPBIC5在具有C2旋转对称的结构中log10Q关于α与β的值㊂可以通过观察得到当(α,β)ң(0,0)时,Q的值趋近于无穷大,并且可以发现两种结构下,log10Q关于α与β的值很相近㊂图4(d) 图8(d)分别代表的是扰动后SPBIC1-SPBIC5在C2旋转对称结构中的磁场z分量H z在z=0时的场图㊂从下面的场图可以观察得到,SPBIC1与SPBIC5仍然对应于τ=1,其他3个对称保护连续谱束缚态也依旧对应于τ=-1㊂通过对比,可以发现两种结构下的场图几乎一模一样,并且可以发现结构扰动不改变对称保护连续谱束缚态的拓扑电荷㊂表1㊀C4与C2旋转对称结构中对称保护连续谱束缚态的频率ωL2πc的值Table1㊀Value ofωL2πc the frequency of symmetrically protected bound states in the continuum in the rotationallysymmetric structure of C4and C2C4C2q SPBIC10.61590.6101+1 SPBIC20.63670.6282-1 SPBIC30.85690.8485-1 SPBIC40.93830.9338-1 SPBIC50.95140.9492+10.05-0.05-0.0500.05βL/(2π)Q f a c t o rαL/(2π)8640.05-0.05-0.0500.05αL/(2π)864βL/(2π)Q f a c t o r㊀㊀(a)(b )0.5-0.5-0.500.5y/LR e/H zx/L0.5-0.5-0.500.5R e/H zx/L㊀㊀(c)(d)图4㊀SPBIC1的Q因子图和场图Fig.4㊀Q factor diagram and field diagram of SPBIC10.05-0.05-0.0500.05βL/(2π)Q f a c t o rαL/(2π)8765430.05-0.05-0.0500.05αL/(2π)βL/(2π)Q f a c t o r876543㊀㊀(a)(b)86第3期张铭洋,等:C4旋转对称光子晶体平板中的对称保护连续谱束缚态0.50-0.5-0.50.5y /LR e /H zx /L0.5-0.5-0.50.5y /LR e /H zx /L㊀㊀(c )(d )图5㊀SPBIC2的Q 因子图和场图Fig.5㊀Q factor diagram and field diagram of SPBIC20.010-0.01-0.010.01βL /(2π)Q f a c t o rαL /(2π)8765430.010-0.01-0.0100.01αL /(2π)βL /(2π)Q f a c t o r76543㊀㊀(a )(b )0.5-0.5-0.50.5y /LR e /H zx /L0.5-0.5-0.50.5y /LR e /H zx /L㊀㊀(c )(d )图6㊀SPBIC3的Q 因子图和场图Fig.6㊀Q factor diagram and field diagram of SPBIC30.010-0.01-0.010.01βL /(2π)Q f a c t o rαL /(2π)76540.010-0.01-0.0100.01αL /(2π)βL /(2π)Q f a c t o r7654㊀㊀(a )(b )0.5-0.5-0.50.5y /LR e /H zx /L0.5-0.5-0.50.5y /LR e /H zx /L㊀㊀(c )(d )图7㊀SPBIC4的Q 因子图和场图Fig.7㊀Q factor diagram and field diagram of SPBIC40.020-0.02-0.0200.02βL /(2π)Q f a c t o rαL /(2π)8640.020-0.02-0.0200.02αL /(2π)βL /(2π)Q f a c t o r6543㊀㊀(a )(b )0.5-0.5-0.50.5y /LR e /H zx /L0.5-0.5-0.500.5y /LR e /H zx /L㊀㊀(c )(d )图8㊀SPBIC5的Q 因子图和场图Fig.8㊀Q factor diagram and field diagram of SPBIC5经过数值计算,从扰动前后不同结构下对称保护连续谱束缚态的频率以及对比分析它们的Q 因子图和场图可以观察得到具有C4旋转对称结构的光子晶体平板中的对称保护连续谱束缚态都是由C2旋转对称性所保护的㊂即若破坏C4旋转对称但保持C2旋转对称,原有的对称保护连续谱束缚态依然存在㊂进一步反映了C4旋转对称与C2旋转对称之间的依赖关系㊂6㊀结束语构建了系统分析连续谱束缚态关于旋转对称性的依赖理论,并且重点研究了C4旋转对称的情况,分别从理论和数值两个方面证明了具有C4旋转对称光子晶体平板中的对称保护连续谱束缚态都是由C2旋转对称性所保护的㊂即破坏C4旋转对称但是保持C2旋转对称性,原对称保护连续谱束缚态依然存在㊂虽然只考虑了C4旋转对称光子晶体平板中的对称保护连续谱束缚态,但提出的理论和数值分析方法都可以用于研究具有C6旋转对称的光子晶体平板,不过由于此结构同时具有C2与C3旋转对称性,对称保护连续谱束缚态与对称性的依赖关系可能会更加复杂㊂提出的理论分析方法也可以适用于所有旋转对称的情况㊂由于是从麦克斯韦方程组出发,没有引入模型近似,并且分析过程根据严格㊂研究结果有利于深入理解对称保护连续谱束缚态的性质,为其理论分析和实际应用提供指导㊂96重庆工商大学学报(自然科学版)第40卷参考文献References1 ㊀HSU C W ZHEN B STONE A D et al.Bound states in thecontinuum J .Nature Reviews Materials 2016 1 9 1 13.2 ㊀HSU C W ZHEN B LEE J et al.Observation of trappedlight within the radiation continuum J .Nature 2013 4997457 188 191.3 ㊀MARINICA D C BORISOV A G SHABANOV S V.Boundstates in the continuum in photonics J .Physical Review Letters 2008 100 18 1 4.4 ㊀NEUMANN J WIGNER E P.Über merkwürdige diskreteEigenwerte M .Berlin Heidelberg Springer 1993.5 ㊀FRIEDRICH H WINTGEN D.Interfering resonances andbound states in the continuum J .Physical Review A 198532 6 3231 3239.6 ㊀YUAN L LU Y Y.Bound states in the continuum on periodicstructures surrounded by strong resonances J .Physical Review A 2018 97 4 1 8.7 ㊀AZZAM S I KILDISHEV A V.Photonic bound states in thecontinuum from basics to applications J .Advanced Optical Materials 2020 9 1 1469 1477.8 ㊀KODIGALA A LEPETIT T GU Q et sing action fromphotonic bound states in continuum J .Nature 2017 5417636 196 199.9 ㊀JIN J YIN X NI L et al.Topologically enabled ultrahigh-Qguided resonances robust to out-of-plane scattering J .Nature 2019 574 7779 501 504.10 HAN S CONG L SRIVASTAVA Y K et al.All-dielectricactive terahertz photonics driven by bound states in the continuum J .Advanced Materials 2019 31 37 1 28.11 BONNET-BENDHIA A S STARLING F.Guided waves byelectromagnetic gratings and non-uniqueness examples for the diffraction problem J .Mathematical Methods in the Applied Sciences 1994 17 5 305 338.12 VENAKIDES S SHIPMAN S P.Resonance and bound statesin photonic crystal slabs J .SIAM Journal on Applied Mathematics 2003 64 1 322 342.13 SHIPMAN S VOLKOV D.Guided modes in periodic slabsexistence and nonexistence J .SIAM Journal on Applied Mathematics 2007 67 3 687 713.14 LEE J ZHEN B CHUA S L et al.Observation anddifferentiation of unique high-Q optical resonances near zero wave vector in macroscopic photonic crystal slabs J .Physical Review Letters 2012 109 6 1 5.15 HU Z LU Y Y.Standing waves on two-dimensional periodicdielectric waveguides J .Journal of Optics 2015 176065601 065608.16 YANG Y PENG C LIANG Y et al.Analytical perspective for bound states in the continuum in photonic crystal slabs J . Physical Review Letters 2014 113 3 1 5.17 YUAN L LU Y Y.Propagating bloch modes above the lightline on a periodic array of cylinders J .Journal of Physics B Atomic Molecular and Optical Physics 2017 505 1 5.18 KANG M ZHANG S XIAO M et al.Merging bound states in the continuum at off-high symmetry points J .Physical Review Letters 2021 126 11 1 7.19 YUAN L LU Y Y.Conditional robustness of propagating bound states in the continuum in structures with two-dimensional periodicity J .Physical Review A 2021 1034 1 10.20 OVERVIG A C MALEK S C CARTER M J et al.Selection rules for quasibound states in the continuum J .Physical Review B 2020 102 3 1 30.21 YODA T NOTOMI M.Generation and annihilation of topologically protected bound states in the continuum and circularly polarized states by symmetry breaking J .Physical Review Letters 2020 125 5 1 12.22 LI S ZHOU C LIU T et al.Symmetry-protected bound states in the continuum supported by all-dielectric metasurfaces J . Physical Review A 2019 100 6 1 6.23 LI L LI Y ZHU Y et al.Rotational symmetry of photonic bound states in the continuum J .Scientific Reports 2020 101 1 8.24 SAKODA K.Optical properties of photonic crystals M . Berlin Springer Science&Business Media 2004.25 YUAN L LU Y Y.Parametric dependence of bound states in the continuum on periodic structures J .Physical Review A 2020 102 3 1 9.26 YUAN L LUO X LU Y Y.Parametric dependence of bound states in the continuum in periodic structures vectorial cases J .Physical Review A 2021 104 2 1 11.27 LU Y Y.Minimizing the discrete reflectivity of perfectly matched layers J .IEEE Photonics Technology Letters 2006 18 3 487 489.28 LI Y J JIN J M.Fast full-wave analysis of large-scale three-dimensional photonic crystal devices J .JOSA B 2007 249 2406 2415.29 ZHEN B HSU C W LU L et al.Topological nature of optical bound states in the continuum J .Physical Review Letters 2014 113 25 1 20.责任编辑:田㊀静07。

SCI科技论文写作常用句型大全

一．好的论文题目是成功的一半好的论文，每一个部分都需要精雕细琢。

我们先来看看Science, Nature 子刊上用的都是些什么题目，到底这些题目暗含哪些玄机？以下是从2016年发表的论文中随机挑选的一些题目，我将其做了一下简单的分类：?reduction1. Water splitting–biosynthetic system with CO2efficiencies?exceeding?photosynthesis. (Science, 2016, 352, 1210)类似: Scalable water splitting on particulate photocatalyst sheets with a solar-to-hydrogen energy conversion efficiency?exceeding?1%?(Nature Materials 2016,15,611-615)这种题目适合哪类文章？适合于那种性能极其显着的文章（可以创纪录的文章），比如说Nocera et al. 的?Science, 2016, 352, 1210，直接与光合作用进行对比，给人的震撼是非常强的。

这种对比效果能够一下子抓住人们的眼球，吸引着读者进行阅读。

要点：使用这样的题目首先你的实验结果得够牛，你对于实验结果要足够自信，对于背景知识的了解得要足够深。

因为取这样的题目意味着你要真正达到了某一个高度。

如果明明有很多大海在那里，你个小池塘和小水坑进行比较，那么会闹笑话的。

2.1 Quantifying?the promotion of Cu catalysts by ZnO for methanol synthesis (?Science,2016, 352, 969-974.)2.2?Exploring?the origin of high optical absorption in conjugated polymers. (Nature Materials 2016,?DOI:?10.1038/nmat4645?)2.3 Promoting?solution phase discharge in Li–O2 batteries containing weakly solvatingelectrolyte?solutions (Nature Materials 2016, DOI: 10.1038/nmat4629)2.4 Reconstructing?solute-induced phase transformations within individual nanocrystals (NatureMaterials 2016,?doi:10.1038/nmat4620)2.5 Tailoring?the nature and strength of electron–phonon interactions in the SrTiO3(001) 2D electron?liquid (Nature Material, 2016, doi:10.1038/nmat4623)我简单地检索了下Nature Materials上面的文章题目，发现这种类型的题目真的非常多。

Long-term in vivo biodistribution imaging and toxicity

Long-term in vivo biodistribution imaging and toxicity of polyacrylic acid-coated upconversion nanophosphorsLiqin Xiong a ,Tianshe Yang a ,Yang Yang a ,Congjian Xu b ,Fuyou Li a ,*a Department of Chemistry,Fudan University,220Handan Road,Shanghai 200433,PR ChinabThe Obstetrics and Gynecology Hospital,Fudan University,19Fangxie Road,Shanghai 200011,PR Chinaa r t i c l e i n f oArticle history:Received 20March 2010Accepted 25May 2010Available online 17June 2010Keywords:Rare-earth nanophosphors Upconversion luminescence In vivo imaging Biodistribution Toxicitya b s t r a c tRare-earth upconversion nanophosphors (UCNPs)have become one of the most promising classes of luminescent materials for bioimaging.However,there remain numerous unresolved issues with respect to the understanding of how these nanophosphors interact with biological systems and the environment.Herein,we report polyacrylic acid (PAA)-coated near-infrared to near-infrared (NIR-to-NIR)upconversion nanophosphors NaYF 4:Yb,Tm (PAA-UCNPs)as luminescence probes for long-term in vivo distribution and toxicity studies.Biodistribution results determined that PAA-UCNPs uptake and retention took place primarily in the liver and the spleen and that most of the PAA-UCNPs were excreted from the body of mice in a very slow manner.Body weight data of the mice indicated that mice intravenously injected with 15mg/kg of PAA-UCNPs survived for 115days without any apparent adverse effects to their health.In addition,histological,hematological and biochemical analysis were used to further quantify the potential toxicity of PAA-UCNPs,and results indicated that there was no overt toxicity of PAA-UCNPs in mice at long exposure times (up to 115days).The study suggests that PAA-UNCPs can potentially be used for long-term targeted imaging and therapy studies in vivo .Ó2010Elsevier Ltd.All rights reserved.1.IntroductionUpconversion luminescence (UCL)is a process in which low-energy light,usually near-infrared (NIR),is converted to higher-energy light (visible)through sequential absorption of multiple photons or energy transfers.By virtue of the f e f transition under continuous-wave (CW)excitation at 980nm,some rare-earth nanophosphors exhibit unique UCL properties (such as sharp emission lines,large anti-Stokes shift and high photostability)[1e 17].Compared with organic ﬂuorophores and semiconductor quantum dots,rare-earth upconversion nanophosphors (UCNPs)possess two important features as luminescent probes in biological labeling and imaging technology:the remarkable light penetration depth and the absence of background ﬂuorescence in biological samples under infrared excitation [18e 20].Recently,UCNPs have been proposed for use in many novel applications in bioimaging and medicine [20e 31].Zhang ’s group has developed biocompatible UCNPs as luminescent labels for in vitro imaging [21],and Prasad et al.have reported in vivo Maestro whole-body images of a Balb-c mouse injected with the UCNPs [22].We have demonstrateda high-contrast upconversion luminescence (UCL)imaging protocol for in vivo targeted imaging of tumors based on RGD/FA-labeled UCNPs [20,23].Also,we have reported dual-modality UCL imaging and magnetic resonance imaging (MRI)in vivo using NaGdF 4:Yb/Tm/Er [24].However,there remain numerous unresolved issues with respect to the understanding of how these nanoparticles interact with biological systems and the environment.To date,there are only a few studies concerning the biodistribution of UCNPs by ICP-MS analysis of rare-earth ions in organs [20e 24].It is worth noting that these studies do not address the in vivo biodistribution imaging of UCNPs and do not examine the in vivo sequestration and excretion of UCNPs.In addition to nanoparticle biodistribution,scientists and clinicians have serious concerns about the utilization of nanopartilces for in vivo applications due to their potential toxicity.Reported in vitro cytotoxicity studies suggested that UCNPs have no or low toxicity when used within a certain range of concentration and within a limited time period of incubation [20e 28].Lim et al.reported the in vivo toxicity of UCNPs in the Caenorhabditis elegans nematode and results indicated that UCNPs are not signi ﬁcantly toxic except at high concentrations of 10mg/mL or higher [29].To date,there are no long-term toxicological reports of UCNPs using animal models,the preferred system for toxicological evaluation of a novel agent,which should be used to*Corresponding author.Tel.:þ862155664185;fax:þ862155664621.E-mail address:fyli@ (F.Li).Contents lists available at ScienceDirectBiomaterialsjournal h omepage:/locate/biomaterials0142-9612/$e see front matter Ó2010Elsevier Ltd.All rights reserved.doi:10.1016/j.biomaterials.2010.05.065Biomaterials 31(2010)7078e 7085characterize the toxicity of UCNPs.Moreover,no histological or biochemical studies were conducted to evaluate the pathological damage of UCNPs to the major organs such as liver,spleen,heart, kidney and lung.In this present study,we report polyacrylic acid(PAA)-coated Yb,Er-codoped NaYF4nanocrystals with an average diameter of w11.5nm as NIR-to-NIR upconversion luminescence probes (denoted as PAA-UCNPs).The long-term in vivo biodistribution of these nanocrystals was performed to assess their uptake by the tissues and their clearance from the mice body.Moreover,to determine their potential in vivo toxicity,theﬂuctuation in body weight of mice,histological assessment,hematological and serum biochemistry assays were also conducted.2.Experimental section2.1.MaterialsAll of the chemicals used were of analytical grade and were used without further puriﬁcation.Deionized water was used throughout.NH4F,sodium oleate,ethanol, methanol,chloroform,and toluene were purchased from Sinopharm Chemical Reagent Co.(China).Oleic acid(OA)was obtained from Alfa Aesar.Polyacrylic acid (PAA M wt1800),diethylene glycol(DEG),and octadecene(ODE)were all purchased from Sigma e Aldrich.Rare earth chlorides(LnCl3,Ln:Y,Yb,Er,Tm)were purchased from Beijing Lansu Co.China.2.2.Preparation of PAA-UCNPsSynthesis of OA-UCNPs.A typical procedure is as follows:to a100mL three-neckedﬂask of6mL oleic acid(OA)and15mL octadecene(ODE)at room temper-ature were added given amounts of YCl3(0.79mmol,154.3mg),YbCl3(0.20mmol,55.9mg)and TmCl3(0.01mmol,2.8mg).The mixture was heated to160 C to forma pellucid solution,and then cooled down to room temperature.0.74g sodium oleate was added and10mL of methanol solution containing NH4F(4mmol)was slowly added drop-wise into theﬂask and stirred for30min.Subsequently,the solution was heated to300 C and maintained for1h under an argon atmosphere. After the solution was left to cool naturally,an excessive amount of ethanol was poured into it,the resultant mixture was centrifugally separated,and the products were collected and washed with ethanol three times.Synthesis of PAA-UCNPs.To aﬂask containing30mL diethylene glycol(DEG) was added PAA-1800(300mg).The mixture was heated to110 C to form a clear solution.Toluene solution containing100mg OA-UCNPs nanocrystals treated with chloroform was added slowly and the temperature maintained for1h under argon protection.The solution was then heated to240 C for1.5h.The resultant solution was cooled down to room temperature and ethanol was added to yield a precipitate. The PAA-UCNPs were recovered via centrifugation and washed three times with ethanol/water(1:1v/v).2.3.CharacterizationSizes and morphologies of UCNPs were determined at200kV using a JEOL JEM-2010F high-resolution transmission electron microscope(HR-TEM).Samples of the as-prepared UCNPs were prepared by placing a drop of dilute aqueous dispersions on the surface of a copper grid.Energy-dispersive X-ray analysis(EDXA)of the samples was also performed during HR-TEM measurements.X-ray diffraction(XRD) measurements were carried out on a Bruker D4X-ray diffractometer using Cu K a radiation(l¼0.15418nm).The size distribution of UCNPs in aqueous solution was measured by dynamic light scattering(DLS)carried out on a Malvern Zetasizer Nano ZS90with a He e Ne laser(633nm)and90 collecting optics.UCNPs samples were prepared in aqueous solution at a concentration of0.2mg/mL andﬁltered through a Millipore0.45m mﬁlter prior to measurements.All measurements were carried out at25 C,and data were analyzed by Malvern Dispersion Technology Software4.20. The zeta potential measurements were performed using a dip cell in automatic mode using Malvern Zetasizer Nano ZS90.UCL spectra were measured with an Edinburgh LFS-920ﬂuorescence spectrometer by using an external0e800mW adjustable laser(980nm,Beijing Hi-Tech Optoelectronic Co.,China)as the excitation source.2.4.Cytotoxicity assayA human nasopharyngeal epidermal carcinoma cell line(KB cell)was provided by Shanghai Institutes for Biological Sciences(SIBS),Chinese Academy of Sciences (CAS,China).The KB cells were cultured in RPMI1640(Roswell Park Memorial Institute’s Medium)supplemented with10%FBS(Fetal Bovine Serum)at37C and 5%CO2.The in vitro cytotoxicity was measured using3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT)assay.KB cells growing in log phase were seeded into a96-well cell-culture plate at5Â104/well and then incubated for24h at 37 C under5%CO2.RPMI1640solutions of UCNPs(100m L/well,containing1% HEPES)at concentrations of6,60,120,240,480m g/mL were added to the wells of the treatment group,and RPMI1640containing1%HEPES(100m L/well)to the negative control group,respectively.The cells were incubated for24h at37 C under5%CO2. Subsequently,10m L MTT(5mg/mL)was added to each well and incubated for an additional4h at37 C under5%CO2.After the addition of10%Sodium dodecyl sulfate(SDS,100m L/well),the assay plate was allowed to stand at room temperature for12h.A Tecan Inﬁnite M200monochromator-based multi-function microplate reader was used to measure the OD570(A value)of each well with background subtraction at690nm.The following formula was used to calculate the viability of cell growth:cell viability(%)¼(mean of A value of treatment group/mean of A value of control)Â100.2.5.In vivo toxicity studiesFour-toﬁve-week-old Kunming mice were purchased from the Second Military Medical University(Shanghai,China).Animal procedures were in agreement with the guidelines of the Institutional Animal Care and Use Committee.PAA-UCNPs at a total dose of15mg/kg were injected into Kunming mice(n¼3)via the tail vein; this group of mice constituted the test group.Kunming mice(n¼3)with no injection of particles were selected as the control group.The body weight of the mice in both groups was recorded every other day for115days.Histology and Hematology Studies:blood samples and tissues were harvested from mice injected with PAA-UCNPs115days post-injection and from mice receiving no injection.Blood was collected from the orbital sinus by quickly removing the eyeball from the socket with a pair of tissue forceps.Three important hepatic indicators(alanine aminotransferase(ALT),aspartate aminotransferase (AST),total bilirubin)and two indicators for kidney functions(creatinine and urea) were measured.Blood smears were prepared by placing a drop of blood on one end of a slide,and using another slide to disperse the blood along the length of the slide. The slide was left to air dry,after which the blood was stained with hematoxylin and eosin.Upon completion of the blood collection,mice were sacriﬁced.The liver, spleen,heart,lung,and kidney were removed,andﬁxed in ethanol,embedded in parafﬁn,sectioned,and stained with hematoxylin and eosin.The histological sections were observed under an optical microscope.2.6.In vivo imaging studiesFour-toﬁve-week-old athymic nude mice were purchased from the Second Military Medical University(Shanghai,China)and were used in the biodistribution imaging studies.In vivo and ex vivo upconversion luminescence imaging was per-formed with a modiﬁed upconversion luminescence in vivo imaging system designed by our group[20].Fig.S1shows the diagram of this upconversion lumi-nescence in vivo imaging system.In this system,two external0e5W adjustable CW 980nm lasers(Shanghai Connet Fiber Optics Co.,China)were used as the excitation sources and an Andor DU897EMCCD as the signal collector.Images of luminescent signals were analyzed with Kodak Molecular Imaging Software.UCL signals were collected at800Æ12nm.3.Results and discussion3.1.Synthesis and characterization of UCNPsOleic acid(OA)-capped NaYF4:20%Yb and1%Tm(OA-UCNPs) was synthesized by a modiﬁed solvothermal route[21].Due to the presence of oleic acid on the surface of the UCNPs,the OA-UCNPs sample can only dispersed in nonpolar solvents such as cyclo-hexane,chloroform and dichloromethane.Therefore,surface functionalization with hydrophilic oleic acid ligand is required prior to the biological applications.Herein,using PAA coating methods by a modiﬁed ligand exchange procedure[12],hydrophobic OA-UCNPs was easily converted into hydrophilic ones.Following the exchange with oleic acid,the resultant PAA-coated UCNPs(PAA-UCNPs)possessed two properties:(I)good dispersibility in aqueous solutions,and(II)the presence of carboxyl functional groups on the surface to allow conjugation with biologically active molecules (such as antibodies,peptides,and proteins)for further targeted in vitro and in vivo delivery studies.More importantly,surface charge plays an important part in particles’interactions with charged phospholipid head groups or protein domains on cell surfaces.L.Xiong et al./Biomaterials31(2010)7078e70857079Cationic particles are more cytotoxic and more likely to induce haemolysis and platelet aggregation than neutral or anionic parti-cles[32].Therefore,the negative surface charges arisen from the carboxyl groups are preferable for in vivo imaging.As shown in Fig.1,the transmission electron microscopy(TEM) image shows that the PAA-UCNPs sample was dispersed with an average diameter of w11.5nm.HR-TEM analysis of a single PAA-UCNPs nanoparticle provides more detailed structural information (Fig.1).The lattice fringes are indicative of the high crystallinity of these particles,and the distance between lattice fringes was measured to be0.517nm,which corresponds to the d spacing for the(110)lattice planes in the hexagonal NaYF4structure.This image also reveals the single-crystal nature of the product. Furthermore,the energy-dispersive X-ray analysis(EDXA)patterns (Fig.S2)conﬁrmed the presence of Na,Y,F,and Yb in the as-synthesized samples.A peak corresponding to the minor-doped Tm ion cannot usually be discerned due to its minute amount(only 1mol%Tm in the rare-earth elements).The powder X-ray diffrac-tion(XRD)patterns(Fig.1C)were in good agreement with the data for hexagonal NaYF4nanocrystals as reported in the JCPDS card(No. 16-0334).This indicates that high purity of the NaYF4nanocrystals with good crystallinity was obtained,which is very beneﬁcial for obtaining bright luminescence.As shown in Fig.2,under excitation of CW laser at980nm,the UCL spectrum of the PAA-UCNPs sample exhibited three distinct Tm3þemission bands.The UCL bands at475,695and800nm originated from1G4/3H6,3F3/3H6and3H4/3H6transitions of Tm3þ,respectively.We further tested the UCL spectrum of the PAA-UCNPs in fetal bovine serum,no obvious difference was observed, showing that PAA-UCNPs were stable in serum(Fig.2).The dynamic light scattering(DLS)measurement showed that the effective hydrodynamic diameter of PAA-UCNPs was w21nm (Fig.S3).The increase in hydrodynamic diameter is attributed to the linkage of PAA polymer to the surface of UCNPs.The zeta potential of PAA-UCNPs in fetal bovine serum was aboutÀ10mV.3.2.Cytotoxicity of PAA-UCNPsThe human nasopharyngeal epidermal carcinoma cell line KB is one of the most common and representative human cancer cell lines.Therefore,an MTT assay with KB cells was used to investigate the cytotoxicity of PAA-UCNPs.No signiﬁcant differences in the proliferation of the cells were observed in the absence or presence of6e480m g/mL PAA-UCNPs(Fig.3).After24h of incubation with PAA-UCNPs,the cellular viabilities were estimated to be greater than94%.Even after48h of incubation with PAA-UCNPs at a concentration as high as480m g/mL,KB cells maintained greater than80%cell viability.These results demonstrated that PAA-UCNPs prepared by the ligand exchange synthesis approach have good dispersibility,water solubility,serum solubility and low cytotox-icity,suggesting their potential for in vivoimaging.Fig.1.TEM images of NaYF4:Yb,Tm samples.A)OA-UCNPs,B)PAA-UCNPs.Inset:HR-TEM images of single PAA-UCNPs nanoparticle.C)XRD pattern of PAA-UCNPs samples and the standard pattern of hexagonal(JCPDS card16-0334)phase of NaYF4.L.Xiong et al./Biomaterials31(2010)7078e708570803.3.In vivo biodistribution imaging of PAA-UCNPsFor in vivo biodistribution imaging studies,athymic nude mice were injected with 15mg/kg of PAA-UCNPs through the tail vein.A concentration of 15mg/kg body weight was chosen for this study as this dose is high enough for the long-term imaging application.At different time points post-injection,mice were anesthetized,sacri ﬁced,and imaged using a modi ﬁed upconversion lumines-cence in vivo imaging system designed by our group [20].The overlays of UCL images and bright ﬁeld images of dissected mice con ﬁrmed that the signal originated predominantly from the liver and spleen (Fig.4).No signi ﬁcant uptake was observed in other organs.Blood analysis at 0.5h detected no PAA-UCNPs,indicating their rapid clearance from the systemic circulation.Early accumu-lation by the liver and spleen is expected and is related to theclearance of nanoparticles from the blood by cells of the mono-nuclear phagocytic system.Within 24h,uptake by the spleen increased,while uptake by the liver decreased.At 24h,much larger uptake by the spleen versus the liver was observed,which is partially due to the spleen being the largest organ of the immune system.In particular,region of interest (ROI)analysis of the UCL signal reveals a high signal-to-noise ratio (>11)in the liver and the spleen over the background (Fig.S4).At 7days post-injection,the UCL signal was signi ﬁcantly reduced in the liver and spleen.At 14days post-injection,almost no UCL signal was detected in the liver and spleen.However,the presence of UCL signals in the intestinal tract indicates a clearance of PAA-UCNPs via hepatobiliary trans-port.Similar clearance pathway has been reported in using lysine crosslinked mercaptoundecanoic acid CdSe 0.25Te 0.75/CdS QDs as optical probes for long-term in vivo imaging [33].At 21days post-injection,the UCL signal was only detected in the intestinal tract and remained unchanged up to 90days.At 115days post-injection,nearly no UCL signal was observed in the mice,showing that most of the PAA-UCNPs were excreted from the body of the mice.The results of the biodistribution studies were further con ﬁrmed by ex vivo UCL imaging of organs (Fig.5)and the measurement of the Y 3þconcentration in organs by inductively coupled plasma atomic emission spectroscopy (ICP-AES)(Fig.6).Ex vivo UCL imaging and ICP analysis showed that PAA-UCNPs uptake and retention took place primarily in the liver and spleen,with little PAA-UCNPs accumulation in the heart,the kidney or the lung.These data were consistent with the in vivo biodistribution imaging results.Recently,Zhang ’s group reported that silica coated NaYF 4nanocrystals were mostly cleared from mice ’s body by day 7post-injection at a dose of 10mg/kg [21].Compared with these two results,the possible reasons for different retention time in the body of mice are likely caused by differences in sample preparation,dosage,size,aggregation,surface coating,and animal type.3.4.Body weight measurements of PAA-UCNPsThe ﬂuctuation in body weight is a useful indicator for studying the toxicity effects of the PAA-UCNPs.In our study,15mg/kg PAA-UCNPs in phosphate buffered saline (PBS,pH 7.4)were administered to 3Kunming mice through tail vein injection.Another three Kunming mice with no injection of particles were selected as the control group.The body weight of the mice in the two groups was recorded every other day for 115days and the results are shown in Fig.7and Fig.S5.Over a period of 53days,the body weight of the mice injected with PAA-UCNPs increased quickly in a pattern similar to that of the control mice with no injection of PAA-UCNPs,suggesting that the mice continued to mature without any signi ﬁcant toxic effects.During the period from day 55to day 93,small weight differences between the mice of the two groups was observed,indicating the low toxicity of PAA-UCNPs in mice.After 95days,body weight of the test group mice was again approximating to that of the control group mice,showing that most of the PAA-UCNPs had been excreted from the body of mice.Recently,Prasad et al.reported the toxicity of CdTe/ZnTe QDs [34].Based on this report,no incre-ment in body weight was observed in mice injected with 5mg/kg of CdTe/ZnTe QDs for 15days,indicating that PAA-UCNPs have lower toxicity than CdTe/ZnTe QDs.Furthermore,mice injected with the PAA-UCNPs and mice receiving no injection underwent observational evaluations for 115days.No changes in feed intake (3e 7g/day/mouse),drinking water consumption (4e 7mL/day/mouse),fur color,exploratory behavior,activity and neurological status were observed (Fig.S6).Fig.3.Cell viability values (%)estimated by MTT proliferation tests versus incubation concentrations of PAA-UCNPs.Cells were incubated with 0e 480m g/mL PAA-UCNPs at 37 C for 24h and 48h.Fig.2.Luminescence spectrum of PAA-UCNPs samples in H 2O and fetal bovine serum under 980nm NIR excitation.Inset:the visual photograph of PAA-NaYF 4shows a blue color.L.Xiong et al./Biomaterials 31(2010)7078e 708570813.5.Histology and hematology results of PAA-UCNPsTo continue the investigation of toxicity,histological assessment of tissues was conducted to determine whether or not the PAA-UCNPs cause tissue damage,in ﬂammation,or lesions from toxic exposure.Analysis was performed on the tissues obtained from the harvested organs (heart,lung,liver,spleen,and kidney)to assess signs of potential toxicity.As seen in Fig.8,the structures of organs from the exposed mice were normal,hardly different from those of the control group.Cardiac muscle tissue in the heart samples showed no hydropic degeneration.Hepatocytes in the liver samples appeared normal,and there were no in ﬂammatory in ﬁltrates.No pulmonary ﬁbrosis was observed in the lung samples.The glomerulus structure could be distinguished easily in the kidney samples.No necrosis was found in any of the groups.However,the spleen was slightly affected by the injection of PAA-UCNPs,thereFig.4.Real-time in vivo upconversion luminescence (UCL)imaging of athymic nude mice with intravenous injection of PAA-UCNPs (15mg/kg)at different time points.Column 3:overlays of UCL and bright ﬁeld images of mice.Column 6:overlays of UCL and bright ﬁeld images of dissected mice.L.Xiong et al./Biomaterials 31(2010)7078e 70857082was slight hyperplasia in the periarteriolar lymphoid sheath (PALS)of white pulp.This may be caused by the nanotoxicity of PAA-UCNPs and this phenomenon has previously been observed in other nanoparticles-treated spleen tissues [35,36].Nanoparticles are particulate materials within the size regime of viruses and large proteins,and consequently they may induce an in ﬂammatory response and increase or decrease the activity of the immune system and alter related hematological factors such as white blood cell count [35,37].Therefore,an assessment of stan-dard hematological and biochemical markers was used to further quantify the potential toxicity of PAA-UCNPs.As shown in Fig.8K,blood smears indicated that the number and shape of red blood cells,platelet,and white blood cells was normal and did not indi-cate a trend associated with PAA-UCNPs treatment.Results did not indicate signi ﬁcant toxicity in the test mice compared to control mice (Fig.8L).Established serum biochemistry assays were used to evaluate more quantitatively the in ﬂuence of PAA-UCNPs on the exposed mice,especially those for potential hepatic injury and kidney functions.As shown in Fig.9,the three importanthepaticFig.5.Real-time ex vivo upconversion luminescence (UCL)imaging of athymic nude mice with intravenous injection of PAA-UCNPs (15mg/kg)at different time points.1:kidney;2:lung;3:heart;4:spleen;5:liver;6:stomach;7:intestines.Fig.6.Biodistribution of particles in organs of mice with intravenous injection of PAA-UCNPs (15mg/kg)at different time points.Error bars were based on tripletmeasurements.Fig.7.Change in body weight obtained from mice injected with PAA-UCNPs (n ¼3,dose ¼15mg/kg,Test)and without injection (n ¼3,Control).L.Xiong et al./Biomaterials 31(2010)7078e 70857083indicators,alanine aminotransferase (ALT),Aspartate aminotrans-ferase (AST)and total bilirubin,were at similar levels for the mice exposed to PAA-UCNPs and for the control mice.The two indica-tors for kidney functions,creatinine and urea,were also similar for the two groups of mice (Fig.9).These results suggest no toxicity of PAA-UCNPs in mice at exposure levels beyond those commonly used in luminescence imaging in vivo and at long exposure times (up to 115days).4.ConclusionsIn summary,we demonstrate the long-term in vivo bio-distribution and toxicity studies using polyacrylic acid-coated NaYF 4upconversion nanophosphors (PAA-UCNPs)as near infrared (NIR)-to-near infrared (NIR)luminescence probes.The PAA-UCNPs prepared by the modi ﬁed ligand exchange synthesis approach have a uniform shape and size distribution,water solubility and serum solubility.In vitro cytotoxicity results showed that PAA-UCNPs have no signi ﬁcant effects on the proliferation of KB cells and the cellular viabilities were estimated to be greater than 80%after 48h incu-bation with PAA-UCNPs ( 480m g/mL).Furthermore,biodistribution results determined that most of the amount of PAA-UCNPs in the liver and spleen were cleared from the body of mice in a very slow manner.In vivo toxicity studies results indicated that mice intra-venously injected with 15mg/kg of PAA-UCNPs survived for 115days without any evident (observational,histological,hematological and biochemical)toxic effects.These studies provide preliminary vali-dation for the use of PAA-UCNPs for long-term in vivo imaging.Biodistribution and toxicity studies of UCNPs of different size and surface coating are currently underway in our lab.AcknowledgementsThe authors thank National Natural Science Funds for Distin-guished Young Scholars (20825101),and Shanghai m.(1052nm03400),NCET-06-0353,Shanghai Leading Academic Discipline Project (B108),and the CAS/SAFEA International Part-nership Program for Creative Research Teams for ﬁnancial support.Appendix.Supplementary informationSupplementary information associated with this article can be found in the online version,at doi:10.1016/j.biomaterials.2010.05.065.Fig.8.H&E-stained tissue sections from mice injected with PAA-UCNPs 115days post-injection (A,C,E,G,I and K)and mice receiving no injection (B,D,F,H,J and L).Tissues were harvested from heart (A,B),spleen (C,D),liver (E,F),lung (G,H),kidney (I,J)and blood smear (K,L).Fig.9.Serum biochemistry results obtained from mice injected with PAA-UCNPs 115days post-injection (n ¼3,dose ¼15mg/kg,Test)and mice receiving no injection (n ¼3,Control).These ﬁndings did not indicate a trend of toxicity.L.Xiong et al./Biomaterials 31(2010)7078e 70857084。

光学纯对映体英文

光学纯对映体英文## Enantiomers and Optical Purity.In the realm of chemistry, chirality refers to the property of a molecule that lacks mirror symmetry, muchlike our left and right hands. Chiral molecules exist in two distinct forms known as enantiomers, which are mirror images of each other but cannot be superimposed. Enantiomers are like two non-identical twins, sharing the same molecular formula and connectivity but differing in their spatial arrangement.Optical purity, a crucial concept in stereochemistry, quantifies the enantiomeric excess of a chiral compound. It measures the proportion of one enantiomer relative to the other in a mixture. A mixture containing equal amounts of both enantiomers is considered racemic and has an optical purity of 0%. Conversely, a mixture containing only one enantiomer is optically pure and has an optical purity of 100%.### Separation of Enantiomers.The separation of enantiomers is a challenging yet essential task in many fields, including pharmaceuticals, agrochemicals, and fragrances. Various techniques can be employed to achieve this, including:Chiral chromatography: This technique utilizes achiral stationary phase that interacts differently with different enantiomers, allowing for their separation.Chiral resolution: This involves converting a racemic mixture into a pair of diastereomers, which can then be separated by conventional methods.Enzymatic resolution: Enzymes, being chiral themselves, can selectively catalyze reactions with one enantiomer over the other, leading to the formation of optically pure products.### Optical Purity Measurement.Optical purity can be determined using various methods, such as:Polarimetry: This technique measures the rotation of plane-polarized light as it passes through a chiral sample. The magnitude and direction of rotation depend on the enantiomeric composition of the sample.NMR spectroscopy: Chiral solvents or chiral shift reagents can be used in NMR spectroscopy to differentiate between enantiomers based on their different chemical shifts.Chromatographic methods: Chiral chromatography or capillary electrophoresis can be used to separate enantiomers and determine their relative abundance.### Significance of Optical Purity.Optical purity is of paramount importance in several areas:Pharmacology: Many drugs are chiral, and their enantiomers can have different pharmacological properties, including efficacy, toxicity, and metabolism. Enantiopure drugs offer advantages in terms of safety and effectiveness.Agrochemicals: Herbicides and pesticides can be chiral, and their enantiomers may differ in their selectivity and environmental impact. Optical purity ensures the targeted control of pests and weeds.Fragrances and flavors: The fragrance and flavor of chiral compounds can depend on their enantiomeric composition. Optical purity control allows for the creation of specific scents and tastes.### Applications of Chiral Compounds.Chiral compounds find widespread applications invarious industries:Pharmaceuticals: Enantiopure drugs include ibuprofen,naproxen, and thalidomide.Agrochemicals: Herbicides such as glyphosate and pesticides like cypermethrin are chiral.Fragrances and flavors: Enantiopure compounds like menthol, camphor, and limonene contribute to thedistinctive scents and tastes of products.Materials science: Chiral polymers, liquid crystals, and self-assembling systems have unique properties and applications in optics, electronics, and nanotechnology.### Conclusion.The concept of enantiomers and optical purity is crucial for understanding the stereochemistry of chiral compounds. The ability to separate and determine the optical purity of enantiomers is essential in numerous fields, including pharmaceuticals, agrochemicals, and fragrances. The significance of optical purity lies in itsimplications for the safety, efficacy, and properties of chiral compounds in various applications.。

雷诺数和侧风对S型垂直轴风力机气动特性的影响

Chinese Journal of Turbomachinery Vol.66，2024,No.1Effects of Reynolds Number and Crosswind on AerodynamicCharacteristics of S-type Vertical Axis Wind Turbine *Yang Zhu 1Hang Zuo 1Jian-yong Zhu 1，*Xiu-yong Zhao 2(1.College of Aero-engine,Shenyang Aerospace University；2.China Energy Science and Technology Research Institute Co.,Ltd.)Abstract：Compared with the large wind turbine working in wild wind field,the small wind turbine is easily affected by low wind speed,high turbulence and time-varying wind speed and direction in urban wind environment.The effects of Reynolds number and inflow angle on the aerodynamic performance of small S-type vertical axis wind turbine are numerically studied.The results show that the aerodynamic performance of S-type wind turbine is not sensitive to Reynolds number,the power coefficient of which almost does not change with Reynolds number.The crosswind leads to the deterioration of the aerodynamic performance,mainly reflected in the decrease of the static torque coefficient and the power coefficient,and the higher the crosswind inflow angle is,the worse aerodynamic performance is.The flow mechanism of crosswind on the aerodynamic performance is revealed mainly in two aspects.On the one hand,the inflow angle obviously decreases the horizontal velocity component of the incoming flow.On the other hand,the vertical velocity component decreases the pressure difference of the advancing blades,while increases the pressure difference of the returning blades,thereby further decreasing the aerodynamic performance.Keywords：Wind Energy;S-type Vertical Axis Wind Turbine;Reynolds Number;Inflow Angle;Numerical Simulation摘要：相较于工作在郊外良好风场的大型风力机，小型风力机易受城市风环境如低风速、高湍流度以及时变风速风向的影响。

Observation of 2nd band vortex solitons in 2D photonic lattices

Abstract We demonstrate second-band bright vortex-array solitons in photonic lattices. This constitutes the first experimental observation of higher-band solitons in any 2D periodic system. These solitons possess complex intensity and phase structures, yet they can be excited by a simple highly-localized vortex-ring beam. Finally, we show that the linear diffraction of such beams exhibits preferential transport along the lattice axes.
Observation of 2nd band vortex solitons in 2D photonic lattices
Guy Bartal,1 Ofer Manela,1 Oren Cohen,1 Jason W. Fleischer2 and Mordechai Segev1 1 Physics Department, Technion - Israel Institute of Technology, Haifa 32000, Israel 2 Department of Electrical Engineering, Princeton University, Princeton , New Jersey, 08544
β ) of the defect they jointly induce (Fig. 1). One of these modes is associated with the 2nd-band

急性呼吸道感染1

鼻腔堵塞
鼻窦炎sinusitis
咽鼓管 Eustachian tube : broad, short, horizontal
中耳炎 otitis media
Anatomical characters
咽 pharynx: 扁桃体tonsil is mature when 4-10 years old tonsillitis is few in infants
Clinical Manifestation pharyngo-conjunctival fever Adenoviruses 3,7 fever pharyngitis 咽炎 conjunctivitis 结膜炎
Complication并发症 otitis media sinusitis laryngitis Bronchitis pneumonia
Anatomical characters
喉 larynx: narrow, soft and tender rich in larynx
Anatomical characters
Acute upper respiratory infection ( AURI) most common disease in children acute nasopharyngitis acute pharyngitis acute tonsillitis
病原体侵袭心肌心肌炎毒素
肺炎 pneumonia
肺炎的定义
不同病原体或其他因素所致肺部炎症
概述发病率、死亡率共同临床特征：发热、咳嗽、气促呼吸困难肺部固定湿啰音
mediastina displacemen
缺氧
Co2潴留

基于传输线型超材料的激活太赫兹波导说明书

10.1117/2.1201308.005081 Active terahertz waveguides based on transmission-line metamaterialsBenjamin S.Williams,Amir Ali Tavallaee,Philip Hon,and Tatsuo ItohThe demonstration of a1D left-handed metamaterial waveguide for terahertz quantum-cascade lasers opens the door to new techniques for beam steering and shaping.Electromagnetic metamaterials are artiﬁcial structures that can be engineered to exhibit customizable or conventionally unob-tainable electromagnetic properties,such as propagation with near-zero or even negative refractive index.In a material with a negative index,theﬂow of energy is opposite to the move-ment of the wavefronts,an effect known as backward-wave or left-handed propagation(so named because the electricﬁeld, magneticﬁeld,and wavevector form a left-handed triple).At IR and optical frequencies,left-handed materials can be made by incorporating plasmonic structures into a dielectric.Pro-vided the size and periodicity of the structures is sufﬁciently small compared to the wavelength,waves propagate as if the medium were uniform with new values for the refractive index (or other bulk properties).Current research in this area investi-gates electromagnetic metamaterials for novel antenna concepts, sub-wavelength resonators and waveguides,superlenses that beat the diffraction limit,and even cloaking from electromag-netic radiation.Our research group has been working on methods to apply metamaterial concepts to the terahertz(THz)frequency range, where the wavelength is approximately a hundred times longer than in the visible.The novelty in our work is the combina-tion of metamaterial-inspired waveguides with a THz quantum-cascade laser-gain medium.In this way,stimulated emission of THz photons from intraband transitions in the gallium-arsenide-based medium compensates for losses and allows active devices.1To design and describe the metamaterial waveguide,we adopt the transmission-line formalism,where negativeand Figure1.Calculated dispersion relation for a balanced terahertz(THz) metamaterial waveguide exhibiting left-handed(LH)and right-handed (RH)propagation.GaAs:Gallium arsenide.AlGaAs:Aluminum gal-lium arsenide.p:Unit cell size.zero-index propagation are modeled by the introduction of additional lumped element capacitance and inductance into the series and shunt branches of the transmission line.2Where a conventional transmission line has series inductance L R and shunt capacitance C R,a metamaterial line is modeled by adding series capacitance C L and shunt inductance L L(the subscripts L and R stand for left-handed and right-handed propagation, respectively).We can adapt this scheme to THz quantum-cascade devices,which are fabricated into a metal-dielectric-metal waveguide.Figure1shows the calculated dispersion relation for a typical design with left-handed propagation be-low about2.6THz and right-handed propagation above2.6THz: a composite right-/left-handed(CRLH)metamaterial wave-guide.At2.6THz,the dispersion relation crosses between the two types of propagation,without a stopband,while maintain-ing non-zero group velocity.Such a condition is referred to as balanced and results from proper engineering of the effective capacitance and inductance on the transmission line.The key advance in this recent work is the inclusion of 200nm-size gaps in the top metallization of the waveguide:see Figure2.These gaps play the role of a series capacitance in theContinued on next page10.1117/2.1201308.005081Page2/2Figure 2.Image of a composite right-/left-handed metamaterial wave-guide implemented in a THz quantum-cascade (QC)metal-metal waveguide.The dielectric of this ‘transmission line’is made up of active THz QC gain material grown in GaAs/AlGaAs quantum wells.The inset image shows a close-up of 200nm gaps in the metallization that create the series capacitance C L and enable left-handed propagation.C x ,L x :Capacitors,inductors (where x denotes R or L).Cu:Copper.Cr:Chromium.Au:Gold.transmission-line model for the waveguide,and are the key fea-ture that enables left-handed propagation.We demonstrated the existence of left-handed propagation indirectly by using a sec-tion of the CRLH metamaterial waveguide as a leaky-wave cou-pling antenna for a THz quantum-cascade laser.The laser feeds the antenna with the THz signal,which is then radiated into the far-ﬁeld at an angle that depends on the THz frequency and its location on the dispersion diagram.While propagation in the right-handed region will result in a beam angled in the for-ward direction,propagation in the left-handed region generates a beam angled in the backward direction (off normal).Propaga-tion with a zero effective index (ˇD 0)gives a beam directed in the surface normal direction.Therefore,by measuring the far-ﬁeld beam pattern and the radiation frequency,we can recon-struct the dispersion relation:see Figure 1.We recently observed a backward-directed beam for the ﬁrst time,demonstrating the existence of left-handed propagation.3Beyond this proof of principle,we now have access to a wide array of microwave circuit,antenna,and metamaterial design techniques that can be applied to THz lasers.For example,such a metamaterial antenna could be used to steer a beam between the forward and backward directions (depending on the exact frequency).Or,if we can develop dynamic control of the circuit elements,tunable resonators and phase shifters become possi-ble.Our future work focuses on using these design techniques to create a new class of lasers with ﬂexible and dynamic control of spectral and radiation properties,including beam shaping and steering,wavelength tuning,and polarization state.Author InformationBenjamin S.Williams,Amir Ali Tavallaee,Philip Hon,and Tatsuo ItohUniversity of California at Los Angeles Los Angeles,CAReferences1.B.S.Williams,Terahertz quantum-cascade lasers ,Nat.Photon.1,pp.517–525,2007.i,C.Caloz,and T.Itoh,Composite right/left-handed transmission line metama-terials ,IEEE Microw.Mag.5,pp.34–50,2004.3.A.A.Tavallaee,P .W.C.Hon,Q.-S.Chen,T.Itoh,and B.S.Williams,Active terahertz quantum-cascade composite right/left handed metamaterial ,Appl.Phys.Lett.102,p.021103,2013.c2013SPIE。

NARROW BAND WAVELENGTH DIVISION DEMULTIPLEXER AND

专利名称：NARROW BAND WAVELENGTH DIVISIONDEMULTIPLEXER AND METHOD OFDEMULTIPLEXING OPTICAL SIGNALS发明人：JAYMIN AMIN,DAVID L. WEIDMAN,LAURA A.WELLER-BROPHY申请号：US09450607申请日：19991130公开号：US20020075537A1公开日：20020620专利内容由知识产权出版社提供专利附图：摘要：A wavelength division demultiplexer includes a channel dropping componentfor receiving optical signals transmitted through a plurality of optical channels, defined by successively different light wavelength bands at intervals ranging between a first channel having the lowest wavelength band to a last channel having the highest wavelength band. The channel dropping component separates at least one channel having a wavelength band intermediate the lowest wavelength band and the highest wavelength band. The demultiplexer further includes an edge filter for separating optical signals received from the channel dropping component that have wavelengths below the intermediate wavelength band from optical signals having wavelengths above the intermediate wavelength band. The separated optical signals are transmitted from the edge filter in two different optical paths. The demultiplexer further includes a channel separator for separating optical signals transmitted in at least one of the optical paths from one another.申请人：AMIN JAYMIN,WEIDMAN DAVID L.,WELLER-BROPHY LAURA A.更多信息请下载全文后查看。

Spasersexplained

COMMENTARYNano-optics is now undergoing a period of explosive growth where new ideas, developments and impressive results appear literally on a dailybasis. It is concerned with the science of concentrating optical energy into regions with subwavelength dimensions (typically tens of nanometres). Yetdespite all this progress, there is still the need for a coherent, intense, ultrafast (with pulse durations down to a few femtoseconds), source of optical energy concentrated to nanoscale areas, similar to the laser but on a much smaller scale. In 2003, David Bergman and I proposed such a source that is based on surface plasmons — the so-called spaser 1 (short for surface plasmon amplification by stimulated emission of radiation) — and researchers are now working to develop and exploit this idea. For example,Nikolay Zheludev and colleagues present their latest ideas regarding spasers on page 351 of this issue 2.Surface plaSmonSWhen introducing the concept of a spaser, it is first useful to explain how it is possible to beat the diffraction limit and focus electromagnetic waves to spots much smaller than a wavelength. The answer lies in the fact that on the nanoscale, optical fields are almost purely electric oscillations at optical frequencies, where the magnetic field component is small and does not significantlyparticipate in the nano-optical physics. The ability of a nanostructured material to support and concentrate such fields is due to the existence of optical modes that are localized on dimensions much smaller than the optical wavelength.Spasers explainedmark I. Stockmanis in the Department of Physics and Astronomy, Georgia State University, Atlanta, Georgia 30302, USA.e-mail:*****************The spaser is a proposed nanoscale source of optical fields that is being investigated in anumber of leading laboratories around the world. If realized, spasers could find a wide range of applications, including nanoscale lithography, probing and microscopy.e–h pairs–40–40–20–20020204040z (n m )x (nm)–40–40–20–20020204040z (n m )x (nm)Figure 1 the spasing mechanism. a , Schematic of a spaser made from a silver nanoshell on a dielectric core (with a radius of 10–20 nm), and surrounded by two dense monolayers of nanocrystal quantum dots (nQDs). b , Schematic of levels and transitions in a spaser. the external radiation excites a transition into electron–hole (e–h) pairs (vertical black arrow). the e–h pairs relax to excitonic levels (green arrow). the exciton recombines and its energy is transferred (without radiation) to the plasmon excitation of the metal nanoparticle (nanoshell) through resonant coupled transitions (red arrows). c ,d , field amplitudes, φ, around the nanoshell excited in two different plasmon modes.COMMENTARYThese energy-concentrating modes are surface plasmons (SPs). They are in essence the eigenmodes of a material system that correspond to oscillations, at optical frequencies, of the electron liquid with respect to the crystal lattice. It can be shown that for a system to support SPs, it should contain componentswith positive and negative dielectric permittivities. A material with negative dielectric permittivity does not support propagation of electromagnetic waves; instead the electromagnetic field decays inside the material within a certain skin depth and most of the incident radiation energy is reflected back. This behaviour is characteristic of a metal, where the skin depth is typically around 25 nm.But for a nanoparticle with a size smaller than the skin depth, optical fields are able to penetrate its entire volume and drive SP oscillations. It is the skin depth that determines the characteristic length scale in nano-optics and makes the nanoscale (from a few nanometres to a few tens of nanometres) so important.An SP mode is characterized by its quality factor, Q, which is the numberof electron oscillations that occur coherently, and during which the mode is able to sustain its phase and accumulate energy from the external excitation field. The best plasmonic metals, for which Q typically ranges from 10 to 100, are the noble metals (silver, gold and platinum), aluminium and the alkaline metals. In response to an external excitation, a plasmonic nanoparticle can generate local fields that are enhanced by a factor Q with respect to the external field. The region occupied by these enhanced fieldsis determined solely by the size of thenanoparticle. Another route to enhancingthe local field is through geometrictricks. For instance, the sharp tip of acone can create intense local fields (theso-called lightning-rod effect), and isexploited in scanning near-field opticalmicroscopy3 (SNOM).SpaSIng actIonIt is critically important that sourcesof concentrated, intense optical fieldslocalized on the nanoscale are available.Such sources will have benefits forfundamental nanoplasmonics andits numerous existing and potentialapplications, including: ultramicroscopy,(such as SNOM); ultrasensitivedetection and spectroscopy of chemicaland biological objects based onsurface-enhanced Raman scattering(SERS); fluorescence imaging withsingle-molecule sensitivity; hyper-Raman(or two-photon-Raman) capabilities;coupling of light to semiconductornano- and microstructures; andnumerous biomedical applications. Theproblem of delivering optical energy tothe nanoscale is a formidable one becausethe optical radiation (light) is limited bydiffraction, and thus can only be focusedinto micrometre-sized regions.Despite the availability of a widevariety of nanoscale optical sources(Box 1), none are ideal. Typically thesesources have a halo of backgroundscattered and delocalized optical fieldsaround them, and do not offer theintensity needed to induce nonlinearprocesses, or the ultrafast speeds requiredfor femtosecond spectroscopy. An idealnanoscale source would also generate‘dark’ optical modes that do not coupleto far-field zones. None of the existingnanosources of localized optical fieldspossess all of these properties, but thespaser that was proposed in 2003 may bejust such a source1.A spaser is the nanoplasmoniccounterpart of a laser, but it (ideally)does not emit photons. It is analogousto the conventional laser, but in aspaser photons are replaced by SPsand the resonant cavity is replaced bya nanoparticle, which supports theplasmonic modes. Similarly to a laser, theenergy source for the spasing mechanismis an active (gain) medium that is excitedexternally. This excitation field may beoptical and unrelated to the spaser’soperating frequency; for instance, aspaser can operate in the near-infraredbut the excitation of the gain mediumcan be achieved using a UV pulse.The reason that SPs in a spaser canwork analogously to photons in a laser isbecause their relevant physical propertiesare the same. First, SPs are bosons: theyare vector excitations and have spin 1, justas photons do. Second, SPs are electricallyneutral excitations. And third, SPs are themost collective material oscillations knownin nature, which implies they are themost harmonic (that is, they interact veryweakly with one another). As such SPs canundergo stimulated emission, accumulatingin a single mode in large numbers, whichis the physical foundation of both the laserand the spaser.One of the simplest and potentiallymost promising types of nanoparticlesto function as a spaser resonator isa metal–dielectric nanoshell. Suchnanoshells have been introduced byNaomi Halas and collaborators andhave since found a very wide range ofapplications4,5. A possible design of ananoshell-based spaser is illustrated inFig. 1a. It consists of a silver nanoshellsurrounded by a few monolayers ofnanocrystal quantum dots6–8 (NQDs).A schematic of the energy levels andtransitions in this spaser is shown inFig. 1b. An electron–hole pair excitedby an initial photon from the externalexcitation field (black arrow in Fig. 1b)relaxes to an excitonic (or possiblymulti-excitonic) state due to carriermultiplication8 (green arrow in Fig. 1b).In a free NQD the excitons wouldrecombine to form photons. However,when the NQD is sitting on the surfaceof a resonant nanoparticle, the excitonicCOMMENTARY energy is transferred, without anysignificant emission of radiation, to theresonant SPs of the nanoparticle (coupledred arrows in Fig. 1b), a process that hasa much larger probability by orders ofmagnitude. The SPs stimulate furthertransitions in the gain medium, leadingto the excitation of more identical SPsin the same SP mode, driving the actionof the spaser. Examples of two SP modesthat can be excited in the nanoshell areshown in Fig. 1c,d.Dark moDeSOne of the advantages of the spasercompared with existing sources of localfields, which also sets it apart fromthe laser, is that it can generate darkmodes that do not couple to the far-zoneoptical fields. In other words, a spasergenerates coherent, strong local fields,but does not necessarily emit photons.This is a potentially great technologicaladvantage because it offers a source of nanolocalized optical fields that does not emit any background radiation. The source can still act on molecules in its near field and excite their radiation (such as fluorescence and Raman effects) as conventional nano-optical sources do. This is because a small moleculeis affected by the field at the point ofits position and is blind to the overall (global) symmetry of the entire local field that determines whether the mode is dark or luminous (bright).Another important advantage of using dark SP modes is that they do not undergo radiative losses. Such lossesfor the luminous modes would lead to a higher threshold for the spaser operation. Interestingly enough, if the symmetryof the spaser’s nanoplasmonic particle is slightly broken, a dark spasing SP mode will become luminous. A collectionof such spasers may then start to emit light, just like a laser. This idea has been proposed by Nikolay Zheludev and colleagues on page 351 of this issue2, where they suggest using a metamaterial containing a planar array of spasers, each of which has a slightly perturbed symmetry. This array then becomes a very efficient planar laser emitting light normally to its plane.So what are the ideal operating conditions for a spaser? Well, theNQD packing density around the metal nanoparticle should be as highas possible. And the spectral widthof the spasing SP mode should be as small as possible (that is, its dephasing time should be as long as possible). These conditions are best met in the near-infrared, where the losses of noblemetals are at a minimum.In sharp contrast to the laser, thespecific shape of the nanoparticle atthe heart of the spasing mechanismdoes not affect its gain at a givenfrequency, provided that the spasing SPmode overlaps with the gain medium.This seriously limits the possibilitiesfor engineering favourable spasingconditions. On the other hand, thespaser gain depends on its operatingfrequency, which can be shifted into thedesired spectral range by engineering thegeometry of the nanoplasmonic particle,for instance as a nanoshell, as in Fig. 1a.Spasers can generate pulses oflocalized optical fields with durationson the femtosecond scale that can beas short as 5 fs. The magnitude of theelectric field in the pulse at the spasersurface is of quantum origin1 and anexample is illustrated in Fig. 2. The fieldamplitude is shown for a spaser whose‘nanoparticle’ consists of two silvernanorods connected at one end to forma V-shape. It can be seen that the fieldforms a hot spot at the tip of the V-shapewhere its amplitude reaches 108 V cm–1.For a spasing mode containing say 104 SPquanta, the resulting field is extremelystrong, approaching atomic values.The spaser is a nanoplasmonicdevice with dimensions that are less thanthe skin depth, and it generates onlynanolocalized SPs. There is, however,a spectrum of related effects based onsurface plasmon polaritons (SPPs), whichare waves localized in the directionnormal to a surface that propagate alongit for distances much greater than thewavelength. Researchers have reportedthe stimulated emission of SPPs (ref. 9).Although this is an important steptowards the realization of a spaser, incontrast to the spaser, SPPs on a flatmetal surface propagate freely and arenot subjected to the cavity feedbacknecessary for lasing (spasing).In other work, scientists havereported a quantum cascade laser thatoperates at a frequency of 17 THz,whose resonator is based on surfaceelectromagnetic waves10. There havealso been studies into the differentgain media to compensate for opticallosses, which could be used in spasers.Researchers have suggested using theeffect of gain on the scattering of lightfrom metal nanostructures11. But theseexperiments used dye molecules thatmay not be suitable for spasers, owing tothe low dipole oscillator strength of thefluorescent transition and problems withachieving sufficiently large molecularconcentrations without excited-statequenching. The most promising gainmedia for spasers remain NQDsat present.Very intense, ultrafast, temporarilycoherent pulses of nanolocalized opticalfields will find numerous applicationsin both fundamental science andengineering. It is one of the uniqueproperties of spasers that the spasingmay be dark, producing strong localoptical fields that do not emit light bythemselves into the far field unless thereare excitable molecules at the surfaceof the spaser. The big advantage ofspasers is the possibility of using them toperform background-free spectroscopy.A spaser that is electrically pumpedwould be particularly valuable, althoughthis possibility still needs to be explored.It may be no exaggeration to say that thespaser, when finally operational, will dofor nano-optics what the laser has donefor conventional optics.references1. Bergman, D. J. & Stockman, M. I. Phys. Rev. Lett.90,027402 (2003).2. Zheludev, N., Prosvirnin, S., Papasimakis, N. & Fedotov, V.Nature Photon.2, 351–354 (2008).3. Novotny, L. & Hecht, B. Principles of Nano-Optics (CambridgeUniv. Press, Cambridge, New Y ork, 2006).4. Hirsch, L. R. et al.Proc. Natl Acad. Sci. USA100,13549–13554 (2003).5. Averitt, R. D., Sarkar, D. & Halas, N. J. Phys. Rev. Lett.78,4217–4220 (1997).6. Klimov, V. I. et al.Science 290, 314–317 (2000).7. Klimov, V. I. et al.Nature 447, 441–446 (2007).8. Schaller, R. D., Pietryga, J. M. & Klimov, V. I. Nano Lett.7,3469–3476 (2007).9. Seidel, J., Grafstroem, S. & Eng, L. Phys. Rev. Lett.94,177401 (2005).10. Tredicucci, A. et al.Opt. Mater.17, 211–217 (2001).11. Noginov, M. A. et al.Appl. Phys. B86, 455–460 (2007).Figure 2 the electric field provided by the spaserdescribed in ref. 1. mode electric-field amplitude, E, isplotted over the surface of a V-shaped silver nanorodconfiguration that acts as a spaser resonator.。

卫星侧摆角英语术语 -回复

卫星侧摆角英语术语-回复Satellite Attitude AnglesSatellite Attitude Angles refer to the orientations and positions of a satellite in three-dimensional space. These angles play a significant role in determining the satellite's position, orientation, and stability during its mission. In this article, we will explore the various aspects and terminologies related to satellite attitude angles.Firstly, it is essential to understand the three primary angles that describe a satellite's attitude: Roll, Pitch, and Yaw. These angles are closely related to the aircraft's attitude angles, but they differ in terms of reference frames. The terms "roll," "pitch," and "yaw" stem from aviation, where they describe the three rotational movements of an aircraft along its longitudinal, lateral, and vertical axes, respectively.In satellite dynamics, however, the roll, pitch, and yaw angles are defined with respect to an Earth-centered inertial frame (ECI) or a satellite-centered frame. Let's delve into each of these angles individually.The Roll Angle:The roll angle, denoted by the symbol φ, describes the rotation of a satellite around its longitudinal axis. The longitudinal axis is an imaginary line passing through the satellite from the front to the back. The roll angle measures the satellite's rotation in degrees or radians from its initial position or reference frame. It affects the satellite's stability and its ability to perform specific tasks, such as imaging or communication.To understand the roll angle better, imagine a satellite as a cylindrical object spinning around an axis. The roll angle would determine the angular displacement of the satellite as it rotates. This angle is crucial for satellites that require a stable or fixed orientation, such as geo-stationary satellites that must always face the Earth.The Pitch Angle:The pitch angle, denoted by the symbol θ, describes the rotation of a satellite around its lateral axis. The lateral axis is an imaginary line passing through the satellite from one side to the other. Thepitch angle measures the satellite's rotation in degrees or radians with respect to its reference frame. The pitch angle affects the satellite's pointing direction, payload operation, and maneuverability.Similar to the roll angle, the pitch angle determines the satellite's stability and its ability to perform specific tasks. This angle plays a critical role in satellites used for Earth observation, remote sensing, and photography, as variations in the pitch angle can significantly influence the images acquired by the satellite.The Yaw Angle:The yaw angle, denoted by the symbol ψ, describes the rotation of a satellite around its vertical axis. The vertical axis is an imaginary line passing through the satellite from the top to the bottom. The yaw angle measures the satellite's rotation in degrees or radians from its reference frame. The yaw angle affects the satellite's directional control and ability to maintain a specific orientation or position.The yaw angle is particularly vital for satellites involved intelecommunications and direct broadcasting applications. Satellites in highly inclined or polar orbits often require continuous yaw maneuvers to maintain proper Earth coverage or to align antennas with the desired targets.To summarize, the satellite attitude angles, including roll, pitch, and yaw, are essential parameters that define a satellite's position and orientation in three-dimensional space. These angles directly impact a satellite's stability, maneuverability, pointing, and payload operations. Understanding and control of these angles are critical for the success of satellite missions, especially those involving Earth observation, communication, and remote sensing.。

超声纳米晶表面改性对选区激光熔化316L_不锈钢微观结构和力学性能的影响

第53卷第5期表面技术2024年3月SURFACE TECHNOLOGY·137·表面强化技术超声纳米晶表面改性对选区激光熔化316L不锈钢微观结构和力学性能的影响彭兰，张宇，高乐，叶一璇，叶畅*（华中科技大学，武汉 430074）摘要：目的改善选区激光熔化（Selective laser melting，SLM）316L不锈钢的表面完整性和力学性能。

方法采用超声纳米晶表面改性（Ultrasonic Nanocrystal Surface Modification，UNSM）这一新兴表面塑性变形方法对SLM 316L不锈钢进行超声冲击强化，利用维氏硬度计、扫描电镜、白光干涉仪、EBSD、XRD等对处理前后材料的表面完整性、微观组织演变和塑性变形行为进行表征和分析。

结果经过UNSM处理后，SLM 316L不锈钢的微观缺陷明显减少，初始未熔合孔隙发生闭合，表面粗糙度Ra由5.374 μm降至0.510 μm，表面硬度从230HV增至461.16HV；同时，材料表层发生了剧烈的塑性变形，形变诱导材料微观组织从γ相向α相转变，微观结构由初始不规则柱状粗晶转变为等轴状细晶。

从EBSD表征结果可知，在材料表面形成了深度约为20 μm的梯度纳米晶，材料内部存在明显的不均匀变形；与初始SLM试样相比，通过UNSM 处理在材料表面引入了最大为932 MPa的残余压应力。

结论超声纳米晶表面改性能够显著改善SLM 316L 不锈钢的表面完整性，形成较深的晶粒细化层和残余应力硬化层，从而有效提高其耐腐蚀性和疲劳抗性，是一项有前景的SLM后处理技术。

关键词：超声纳米晶表面改性；SLM 316L不锈钢；微观结构；残余应力；塑性变形机理中图分类号：TG665；TG142.71 文献标志码：A 文章编号：1001-3660(2024)05-0137-12DOI：10.16490/ki.issn.1001-3660.2024.05.014Effect of Ultrasonic Nanocrystal Surface Modification on Microstructure and Mechanical Properties of SLM 316L Stainless SteelPENG Lan, ZHANG Yu, GAO Le, YE Yixuan, YE Chang*(Huazhong University of Science and Technology, Wuhan 430074, China)ABSTRACT: Metal powder additive manufacturing (AM) technologies, such as selective laser melting (SLM), have attracted considerable interest owing to their near-net forming characteristic and layer-by-layer building-up strategy, which allows overcoming the constraints of traditional manufacturing technology, achieving complex components in a short time of mass customization. However, the SLM process-induced micro-defects (i.e. pores, lack-of-fusion, and undesired microstructures) will result in not only poor surface finish and interior thermal cracks but also more dispersion收稿日期：2023-03-24；修订日期：2023-06-14Received：2023-03-24；Revised：2023-06-14基金项目：国家自然科学基金（52075200）Fund：National Natural Science Foundation of China (52075200)引文格式：彭兰, 张宇, 高乐, 等. 超声纳米晶表面改性对选区激光熔化316L不锈钢微观结构和力学性能的影响[J]. 表面技术, 2024, 53(5): 137-148.PENG Lan, ZHANG Yu, GAO Le, et al. Effect of Ultrasonic Nanocrystal Surface Modification on Microstructure and Mechanical Properties of SLM 316L Stainless Steel[J]. Surface Technology, 2024, 53(5): 137-148.*通信作者（Corresponding author）·138·表面技术 2024年3月of mechanical properties. Therefore, for a more homogenized microstructure and smaller material anisotropy, a novel surface strengthening method of severe surface plastic deformation, ultrasonic nanocrystalline surface modification (UNSM) was applied to improve the surface integrity and mechanical properties of SLM 316L stainless steel in this study.A medium size laser powder bed fusion (LPBF 271 Series device from Farsoon Technologies Tech Co., Ltd.) wasused to fabricate the plate specimens with 316L stainless steel powder. The SLM-processed samples were fabricated using the optimized processing parameters with a laser power of 400 W, hatch spacing of 0.11 μm, a laser scan speed of 1 250 mm/s, and a layer thickness of 60 μm. Bidirectional laser scanning with a scan rotation of 67° for every layer was performed during building. An annealing heat treatment at 900 ℃for 2 h was conducted on the as-received SLM 316L stainless steel plate, followed by furnace cooling to room temperature. The SLM 316L stainless steel plate used in this investigation was 40 mm×20 mm×4 mm in size. For improving the strengthening efficacy, optimized UNSM process parameters were used in the current work: an ultrasonic frequency of 20 kHz, a WC (tungsten carbide) tip with a diameter of 2.4 mm, a static load of 50 N, an ultrasonic amplitude of 30 µm, a scanning speed of 500 mm/min, and a feed rate of 10 µm.The surface integrity, microstructure evolution, and plastic deformation behavior of the material before and after UNSM treatment were systematically characterized and analyzed through Vickers indentation, a scanning electron microscope (SEM), a white light interferometer, electron backscatter diffraction (EBSD), and x-ray diffraction (XRD).The result showed that the micro defects of SLM 316L stainless steel were significantly reduced. SLM's initial LOF defects were diminished under the high-frequency ultrasonic load. The surface roughness Ra decreased from 5.374 μm to 0.510 μm, and the surface hardness increased from 230HV to 461.16HV. Severe plastic deformation (SPD) occurred on the surface layer of the material, which induced the transformation of the microscopic structure from γ to α phase. The crystal microstructure was also refined from the initial irregular columnar coarse crystal to fine equiaxed crystal. As a result of local uneven plastic deformation in the UNSM process, a depth 20 μm gradient nanocrystal was captured through the result of EBSD analysis. Thus, compared with the initial SLM specimen, the UNSM treatment produced a maximum residual compressive stress of 932 MPa on the surface of the material. The improvement of the surface integrity, formation of deeper grain refinement layer, and residual stress hardening layer of SLM 316L after UNSM treatment successfully demonstrates that UNSM is a promising post-processing surface treatment technology for SLM metallic materials.KEY WORDS: UNSM; SLM 316L stainless steel; microstructure; residual stress; plastic deformation mechanism在各牌号不锈钢中，316L不锈钢的含碳量较低，且其中添加了质量分数为2%~3%的Mo元素，因此它在耐腐蚀性能方面表现优异，可以应用于金属易腐蚀的恶劣环境。

Radialarteriallines-UCL

P ractical P roceduresBritish Journal of Hospital Medicine, May 2006, Vol 67, No 5 M3IntroductionIntra-arterial cannulae in the radial artery are used for invasive arterial blood pressure (IABP) measurement and for collection of blood for analysis. The radial artery is the preferred site for insertion because of low complication rates. Arterial lines are the gold standard for accurate blood pres-sure measurement. They may be used in intensive care and high dependency units and in anaesthetized patients undergoing surgical procedures. An understanding of basic principles enables arterial lines to be used safely in these settings.IndicationsThe indications for a radial arterial line are:1. Continuous, beat-to-beat blood pres-sure measurement. Examples include patients on the intensive care unit (ICU) requiring inotropic support, or patients with severe cardiovascular dis-ease undergoing surgery.2. Frequent arterial blood gas analysis in patients with respiratory failure, or severe acid/base disturbance.Choice of arterial siteThe radial artery has low complication rates compared with other sites. It is a superficial artery which aids insertion, and also makes it compressible for haemostasis (AQ This sentaence has been rephrased, is it okay?).The ulnar, brachial, axillary, dorsalis pedis, posterial tibial, femoral arteries are alternatives.PreparationAllen’s test is recommended by many textbooks before the insertion of a radialarterial line. This is used to determine collateral perfusion between the ulnar and radial arteries to the hand: poor collateral perfusion is said to be present in 12% of people. If ulnar perfusion is poor and a cannula occludes the radial artery, blood flow to the hand may be reduced. The test is performed by asking the patient to clench their hand. The ulnar and radial arteries are occluded with digital pressure. The hand is unclenched and pressure over the ulnar artery is released. If there is good collateral perfusion, the palm should flush in less than 6seconds. In practice the use-fulness of this test is questionable.Equipmentn Arterial cannulae. Made from poly-tetrafluoroethylene (‘Teflon’) to minimize the risk of clot formation (Figure 1) they are short, with parallel sides to minimize the effect on blood flow distally. A 20G (pink) cannula is used in adult patients, a 22G (blue) for paediatrics, and a 24G (yellow) for neonates and small babies. Larger gauge cannulae increase the risk of thrombo-sis, smaller cannulae cause damping ofRadial arterial linesthe signal. The cannula is connected to an arterial giving set.n Arterial giving set. Specialized plas-tic tubing, short and stiff to reduce resonance (see below), connected to a 500ml bag of saline.n 500ml bag of saline. This is pressurized to 300mmHg using a pressure bag, i.e. a pressure higher than arterial systolic pressure to prevent backflow from the cannula into the giving set. The arte-rial giving set and pressurized saline incorporate a continuous slow flushing system of 3–4ml per hour to keep the line free from clots. The arterial givingDr Rachel Hignettis Specialist Registrar in Anaesthetics, Nuffield Department of Anaesthetics, The John Radcliffe Hospital, Oxford andDr Robert Stephens isAcademy of Medical Sciences/the Healthcare Foundation Clinical Research TrainingFellow, Institute of Child Health, UCL, London WC1N 1EHCorrespondence to: Dr R StephensFigure 1. Two arterial cannulae.Figure 2. Radial arterial line. (Please state where Figure 2 should be cited in the text)Figure 3. Direct cannulation of the radial artery.British Journal of Hospital Medicine, May 2006, Vol 67, No 5。

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

a r X i v :a s t r o -p h /0501382v 2 14 F eb 2005Astronomy &Astrophysics manuscript no.ms˙twocol February 2,2008(DOI:will be inserted by hand later)VLT narrow-band photometry in the Lyman continuum of twogalaxies at z ∼3⋆Limits to the escape of ionizing ﬂuxA.K.Inoue ⋆⋆1,2,3,I.Iwata 3,4⋆⋆⋆,J.-M.Deharveng 1,V.Buat 1,and D.Burgarella 11Laboratoire d’Astrophysique de Marseille,Traverse du Siphon,BP 8,13376Marseille,CEDEX 12,Francee-mail:akio.inoue@oamp.fr,veronique.buat@oamp.fr,denis.burgarella@oamp.fr,jean-michel.deharveng@oamp.fr 2Department of Physics,Kyoto University,Sakyo-ku,Kyoto 606-8502,Japan 3Department of Astronomy,Kyoto University,Sakyo-ku,Kyoto 606-8502,Japan4Subaru Mitaka Oﬃce (Subaru Telescope),National Astronomical Observatory of Japan,2-21-1Osawa,Mitaka,Tokyo 181-8588,Japane-mail:iwata@optik.mtk.nao.ac.jpSubmitted on August 2,2004Abstract.We have performed narrow-band imaging observations with the Very Large Telescope,aimed at detecting the Lyman continuum (LC)ﬂux escaping from galaxies at z ∼3.We do not ﬁnd any signiﬁcant LC ﬂux from our sample of two galaxies in the Hubble Deep Field South,at z =3.170and 3.275.The corresponding lower limits on the F 1400/F 900ﬂux density (per Hz)ratio are 15.6and 10.2(3-σconﬁdence level).After correction for the intergalactic hydrogen absorption,the resulting limits on the relative escape fraction of the LC are compared with those obtained by diﬀerent approaches,at similar or lower redshifts.One of our two objects has a relative escape fraction lower than the detection reported by Steidel et al.in a composite spectrum of z ∼3galaxies.A larger number of objects is required to reach a signiﬁcant conclusion.Our comparison shows the potential of narrow-band imaging for obtaining the best limit on the relative escape fraction at z ∼3.Stacking a signiﬁcant number of galaxies observed through a narrow-band ﬁlter would provide constraint on the galactic contribution to the cosmic reionization.Key words.cosmology:observations —diﬀuse radiation —intergalactic medium —galaxies:photometry —ultraviolet:galaxies1.IntroductionThe observations of the ﬁnishing and the beginning reionization of intergalactic hydrogen at z ∼6and 20(Becker et al.,2001;Kogut et al.,2003)combined with the later reionization of intergalactic He ii (Kriss et al.,2001;Zheng et al.,2004)at z ∼3are now placing con-straints on the nature and the evolution of the ultra-violet (UV)background radiation.A dominant contribution of star-forming galaxies is suggested at z >3before the den-sity of luminous quasars become signiﬁcant.This picture2Inoue et al.:Upper limit of Lyman continuum escape FUSE inﬁve other starburst galaxies(Heckman et al.,2001).At high redshifts,the increasing IGM opacity(absorp-tion of the ionizing radiation by the neutral hydrogen),notto speak of galaxies becoming fainter,makes the observa-tions diﬃcult.Nevertheless,large ground-based telescopescan enter the competition at z>∼3.Steidel et al.(2001)(hereafter S01)derived a1500˚A/900˚A observedﬂux den-sity1ratio,(F1500/F900)obs=17.7±3.8from a compositespectrum of29Lyman break galaxies(LBGs)at a meanredshift of3.4with Keck/LRIS.By comparison with mod-els of the UV spectral energy distribution of star-forminggalaxies,theﬂux density ratio,corrected for the IGMopacity,leads to the fraction of escaping LC(900˚A)pho-tons relative to the fraction of escaping non-ionizing UV(1500˚A)photons.This is called the relative escape frac-tion(f esc,rel)and is diﬀerent from theﬁrst deﬁnition ofthe escape fraction(f esc)used above for nearby galaxies.S01interpreted their results as implying f esc,rel>∼50%.However,not all z∼3galaxies emit a signiﬁcant LCing the FORS2spectrograph on the VLT,Giallongo et al.(2002)(hereafter G02)obtained1-σlowerlimits on(F1500/F900)obs four times larger than the valueof S01for two bright LBGs.Heckman et al.(2001)also de-duced a very low escape fraction for a gravitational lensedgalaxies,MS1512-cB58at z=2.7from the detailed anal-ysis of interstellar absorption lines.Observations have not been only spectroscopic.Imaging of galaxies at z∼1with the FUV solar-blinddetector of Space Telescope Imaging Spectrograph(STIS)have also provided constraints on theﬂux below theLyman limit(Ferguson,2001;Malkan et al.,2003).Inparticular,Malkan et al.(2003)(hereafter M03)have ob-tained lower limits of(F1500/F700)obs>∼150–1000(1-σ),implying much lower LC escape fraction than in the galax-ies of S01.Fern´a ndez-Soto et al.(2003)(hereafter FS03)have used the deep U300images of the Hubble Deep FieldNorth(HDFN)and reported an average LC escape frac-tion of no more than4%for27galaxies at redshifts1.9<z<3.5.As a number of galaxies in their samplehave their U300ﬂuxes contaminated by nonionizing UVphotons,their analysis is based on models and indirect.These conﬂicting results,as well as the diﬃculties ofobservations at high redshift,have led us to search for pos-sibilities of improvement.We haveﬁrst examined how thedetection advantage of imaging over spectroscopy is actu-ally working in the speciﬁc context of measuring faint LCradiation.On one hand,spectroscopy allows us to measurethe LCﬂux close to the redshifted Lyman limit,wherethe average IGM opacity is not yet as large as it would beat shorter wavelength because of the Lyman valley(e.g.,Møller&Jakobsen,1990).On the other hand,broad-bandmeasurements require a very large correction for the IGMopacity as we show in the current paper,because the ef-fective wavelength becomes very short for objects selectedInoue et al.:Upper limit of Lyman continuum escape 30.20.40.60.8124002800320036004000r e l a t i v e e f f i c i e n c y wavelength [A]Fig.1.Transmission curves of the FORS OII +44ﬁlter (solid)and the HST U 300(dashed).VLT/ISAAC and HST/WFPC2optical images.Thus,we selected the HDFS/WFPC2ﬁeld as our target ﬁeld.There are 24galaxies with I 814≤25.0mag(AB)2and z ph ≥2.5in the ﬁeld.We set this lower limit of z ph ,taking into account the 2-σuncertainty of z ph for z ∼3(σ≈0.08(1+z )noted in Labb´e et al.2003).Eight of them have z ph larger than 3.18.On the other hand,six out of 24have z sp ,but all of them are less than 3.18.We note that all the known z sp are smaller than the corresponding z ph .An accurate redshift is essential to avoid contamina-tion from nonionizing photons.We need a redshift accu-racy of the order of 0.01.Since the accuracy of z ph is not suﬃcient,we performed spectroscopic observations with FORS2on VLT using the MXU mode.A full description of the observation will be published elsewhere (Iwata et al.in preparation).Here,we summarize the results brieﬂy.Because of the limitation of the slit-let conﬁguration,we were able to take spectra of only 15galaxies out of the above 24galaxies.Five out of the 15galaxies have known z sp and 7of them have z ph >3.18.The choice of galaxies was made with a priority given to more luminous galaxies and galaxies with z ph >3.18.Additionally,we took spectra of galaxies with z ph ≥2.5but I 814>25.0mag(AB),for which we could conﬁgure the slit-let.Among these galaxies,one has z sp previously measured.We ended up with obtaining 13z sp :seven of them are new mea-surements and the remaining six are conﬁrmations of the previous results.Among the new redshifts,we got only one galaxy with a high enough redshift,z sp =3.275(HDFS 1825),in contrast to the expectation from photometric red-shifts.Indeed,for the 13galaxies with z sp ,we always ﬁndz sp <z ph ,again.There is another galaxy with z sp =3.170(HDFS 85)which was conﬁrmed by our spectroscopy.This redshift is slightly smaller than the limiting redshift,how-ever,the transmission eﬃciency of the OII +44ﬁlter at the Lyman limit of the galaxy is as small as a few per-cent of the peak eﬃciency (see Fig.1).The contamination of nonionizing photons is small enough,especially in the context of upper limit measurements.Therefore,our ﬁnal sample consists of the two galaxies.We have no other ex-planations for this small number of target galaxies than a systematic error in the estimation of z ph .2.3.Imaging observationsThe HDFS ﬁeld was observed through the OII +44narrow-band ﬁlter with FORS1and the TK2048EB4-1detector chip in service mode from 29June 2003to 28August 2003.The standard resolution collimator was adopted,giving a pixel scale of 0.′′2per pixel.The ﬁeld of view is 6.′8×6.′8and covers the whole area of the HDFS/WFPC2ﬁeld.Each image was taken with a small dithering (typically 10′′)and a typical exposure time of 1,080sec.The seeing size of each image was 0.′′39–1.′′23,typically ∼0.′′8.We secured 38images of the ﬁeld.The eﬀective exposure time is 40,636sec (i.e.about 11h).Table 1shows a summary of the observations.3.Data reduction and results 3.1.Final imageThe image data reduction was carried out in a standard manner,using IRAF 3.The bias subtraction was made us-ing over-scanned regions.Normalized twilight sky frames in each observing night were used for ﬂat ﬁelding.We se-lected 25stars in the ﬁeld and used them to register the 38frames.After registration,the rms of residual shifts was 0.03pixel.These stars were also used for an airmass correction.We found a slope of magnitude dependence on airmass of −0.45,and we corrected observed counts as4Inoue et al.:Upper limit of Lyman continuumescapeFig.2.Close-up of sample galaxies through the OII+44ﬁlter(VLT/FORS,this work),and the I814ﬁlter (HST/WFPC2,Williams et al.,2000).The FOVs are20′′×20′′.For the object85,the image center is6′′shifted from the object position,because it is close to an edge of I814image.it would have been observed at the zenith.The stability of the observing conditions is conﬁrmed by measuring the corrected counts of these stars;the rms errors in counts are less than0.05mag for19stars brighter than22mag(AB) through the OII+44ﬁlter among the above25stars. Then,the IRAF task IMCOMBINE was used to sum the frames.We took averages of each pixel adopting a3-σclipping.FWHM of stellar objects in theﬁnal image is ∼1′′.Fig.2shows the close-up images of the two sample galaxies through the OII+44ﬁlter and through the I814ﬁlter of HST/WFPC2.Neither galaxy seems to be seen through the OII+44ﬁlter.3.2.PhotometryThree standard stars,Feige110,G93-48,and LTT9491, were also observed through the sameﬁlter.These frames were processed in the same way as the HDFS frames.The absolute AB magnitude of these stars were calculated from the spectral table provided on ESO web page.We found that the airmass dependence for the standard stars was −0.58which is somewhat larger than that from stars in the HDFS frames.We derived the photometric zero point for the OII+44ﬁlter at the zenith as22.98from the airmass slope of the standard stars.The zero point changes less than0.02mag if we adopt the airmass slope−0.45which we used for correction of the HDFS frames.To reduce backgroundﬂuctuations,we apply a Gaussian smoothing to theﬁnal image adapted to the ex-pected size of the sample galaxies through the OII+44ﬁlter.The size through the HST/WFPC2B450ﬁlter is a reasonable approximation because the central wavelength is close to that of the OII+44ﬁlter and we are seeing the light from massive stars in bothﬁlters.Diameters ofInoue et al.:Upper limit of Lyman continuum escape5 Table2.Photometric properties of sample galaxies.ID a851825L int LC =exp(−τISMLC),(1)where L int LC is the intrinsic LC luminosity density(per Hz)of a galaxy,L outLCis the LC luminosity density just outsideof the galaxy(not the observed one,see below),andτISMLCis the opacity of the interstellar medium(ISM)for LCphotons in the galaxy.The other isf esc,rel≡f esc L out UV(F UV/F LC)obsexp(τIGMLC),(3)where we have replaced the observed luminosity densityratio into the observedﬂux density ratio.Hence,therelative escape fraction can be estimated from the ob-served UV-to-LCﬂux density ratio if we know the intrinsicUV-to-LC luminosity density ratio and the IGM opacity.Moreover,the absolute escape fraction can be estimatedfrom the relative escape fraction via equation(2)if weknow the ISM opacity for nonionizing UV photons.4.2.Intrinsic luminosity density ratiosThe intrinsic UV-to-LC luminosity density ratio is stillvery uncertain observationally.As noted by S01,we mustrely exclusively on models.Here we adopt the Starburst99model(Leitherer et al.,1999).From this model,weobtain(L1500/L900)int=1.5–5.5as shown in Fig.3,as-suming a constant star formation rate,the Salpeter initialmass function with the mass range of0.1–100M⊙,andthe metallicity of0.001–0.02(the solar value is0.02).Theluminosity density ratios mainly depend on the durationsince the onset of star formation;the ratio starts from asmall value,monotonically increases with time,and satu-rates at a larger value after several108yrs.In this paper,we adopt(L1500/L900)int=3.0according to the valueadopted by S01.46Inoue et al.:Upper limit of Lyman continuum escape Table3.Adopted intrinsic ratios of luminosity densitiesrelative to1500˚A.x(L1500/L x)int a(˚A)Inoue et al.:Upper limit of Lyman continuum escape7 Table4.Summary of related observations and escape fractions.Redshift range/method/redshift(F UV/F LC)obs aτIGMLC f esc,rel aλUV,λLC b L UV c A UV d ref.e f esc agalaxy name(%)(1029erg s−1Hz−1)(mag)(%) (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)z∼1Broadband photometryCl J0023+0423:[LPO98a]022 1.1074>1870.53<3.050.967 1.33<0.91 CFRS03.1140 1.1818>3080.59<1.76 2.56......<1.7 CFRS10.1887 1.2370>1190.63<4.560.969......<4.5 CFRS10.0239 1.2919>2280.68<2.56 2.28......<2.5 CFRS10.1168 1.1592>1630.57<3.160.802......<3.1 LDSS210.288 1.108>1020.53<5.550.859......<5.5 HDF:iw410021353 1.221>67.80.62<10.040.513 2.64<0.91 CFRS14.0547 1.160>3480.57<1.84 1.55 2.24<0.24 CFRS14.0154 1.1583>1920.57<3.440.610 2.94<0.24 SSA22-16 1.36>1650.74<4.64 1.050.204<3.8 CFRS22.1153 1.3118>1250.54<5.04 1.00 1.64<1.15There is one more galaxy(HDFN:[FLY99]688)satisfyingthe redshift criterion in the list of FS03,but it is detectedby Chandra as a X-ray source(Hornschemeier et al.,2001).Hence,we remove it from our discussions because it may be anAGN.8Inoue et al.:Upper limit of Lyman continuum escaperelative escape fraction(f esc,rel,column5),using the in-trinsic luminosity density ratios(Sect.4.2)and the IGM opacity model(Sect.4.3,column5),which are tailored to the speciﬁc wavelength and theﬁlter bandpass(column6) for each individual galaxy.5.1.1.z∼3galaxiesThe column3in the top of Table4shows that our two observed lower limits of F UV/F LC are comparable to those already obtained for galaxies at z∼3.G02have reached constraints about two times better than ours,because their two objects are more than two times intrinsically brighter.Our objects,as those of S01,are comparable in luminosity to the L∗of z∼3LBGs(see column7; L∗(z=3)∼1×1029erg s−1Hz−1;Steidel et al.,1999; Adelberger&Steidel,2000).In terms of relative escape fraction(column5),we add one more object(HDFS85)to the two galaxies of G02that have been reported with a relative escape fraction lower than the detection of S01.Because of the small sample size,these lower limits remain consistent with the rela-tive escape fraction of S01.Although we have applied a mean IGM opacity for simplicity,a line of sight with larger than average opacity cannot be ruled out for an individual galaxy(not to speak of an unusual object as the escape is probably a random phenomenon).The broad-band U300photometry in the HDFN reaches similar lower limits on F UV/F LC(it would have reached slightly better values at the same luminosity level as ours).This,however,does not result in interesting con-straints on the relative escape fraction because of the IGM opacity.The IGM opacity through the U300ﬁlter,with an eﬀective wavelength of700˚A in the source rest-frame,is much larger than those at wavelengths closer to the rest 900˚A(column4).Since meaningful constraints on the rel-ative escape fraction are not reached,this sample will be removed from discussions in terms of the absolute escape fraction presented below.The top of Table4is also interesting for a comparison of methods of observations and a validation of future ap-proaches.As expected,the narrow-band photometry can provide better constraints than spectroscopy.The better constraints of G02are actually due to the high luminos-ity of the objects.If our sample had a similar luminosity to the galaxies of G02,say,L=4×1029erg s−1Hz−1, we would have obtained F UV/F LC>∼40(3-σlevel).The broad-band photometry is disqualiﬁed by the IGM opac-ity contribution(or requires the help of models that lead to less direct constraints).The narrow-band photometry can provide,even for a single L∗LBG,a lower limit on F UV/F LC comparable with that measured in the compos-ite spectrum of29galaxies by S01.Stacking a signiﬁcant number of galaxies observed through a narrow-bandﬁl-ter will be able to go much deeper.We can also average the IGM opacity against any unusual line of sight in the stacking process.Such an approach would allow a signif-icant comparison with the result of S01and reveal which fraction of high-z galaxies have high relative LC escape fraction.5.1.2.z∼1galaxiesThanks to the high sensitivity of the HST/STIS,M03have obtained very good lower limits on theﬂux density ratio of11galaxies at z∼1,F UV/F LC>70–350at3-σcon-ﬁdence level(column3in the middle of Table4).With the correction for the average IGM opacity described in section4.3,these ratios are converted to upper limits on the relative escape fraction of2–10%(column5).The lu-minosities of this sample are similar to L∗of z∼3LBGs (column7).The average IGM opacity(column4)is much smaller than those at high redshift,but not completely negligible as assumed by M03,if we consider the Lyman limit sys-tems(LLSs)which dominate the IGM opacity for z∼1 galaxies.The rarity of LLSs favors a statistical treatment rather than the average opacity adopted above.Based on the number distribution function of the IGM clouds as-sumed here(see Appendix),the expected number of LLSs within the wavelength range observed by M03(1300˚A <∼λobs<∼1900˚A)is about0.3.Roughly speaking,for one-third of the objects of M03,the LLSs would loosen the constraints obtained under the assumption of no IGM opacity.In this sense,the number of galaxies in M03is large enough to conclude a very small relative escape frac-tion for their objects.5.1.3.Nearby galaxiesAlthough the upper limits on the LC from nearby galax-ies have been estimated from their Hαﬂuxes directly in terms of absolute escape fraction(Leitherer et al.,1995; Hurwitz et al.,1997;Deharveng et al.,2001),it is pos-sible to evaluate the observed lower limits of F UV/F LC for comparison with higher z objects(the bottom part of Table4;Table1in Deharveng et al.,2001).Except for Mrk54,these limits,corrected for the foreground absorp-tion by the Galactic HI and H2gas,are not better than those obtained at high-z(column3).The low luminos-ity of the sample galaxies is one of the causes(column 7).Nevertheless,moderate upper limits are obtained for the relative escape fraction,<25–40%(3-σconﬁdence level;column5),because no correction for the IGM opac-ity is necessary.For Mrk54,whose luminosity is similar to those of our sample(column7),the observed lower limit of Deharveng et al.(2001)translates into an upper limit on the relative escape fraction of3%comparable to those for z∼1galaxies of M03(column5).5.2.Absolute escape fractionAs discussed above and shown by equation(2),estimat-ing the absolute escape fraction from the relative escapeInoue et al.:Upper limit of Lyman continuum escape9 fraction requires an evaluation of the dust attenuationwithin each galaxy which is often diﬃcult and uncertain(e.g.,Buat et al.,2002;Pettini et al.,1998;Meurer et al.,1999).Although the infrared to UVﬂux ratio is agood estimator of the UV attenuation(Buat et al.,1999;Gordon et al.,2000),the infraredﬂuxes for high-z galaxiesare not available yet.Here,we adopt a calibration betweenthe nonionizing UV slope and the UV attenuation pro-posed by Meurer et al.(1999)for simplicity.However,weshould keep in mind that the calibration depends on thetype of galaxies,starburst or not(Bell,2002;Kong et al.,2004).The UV slope of the galaxies listed in Table4havebeen searched in the literature,or,if not available,esti-mated from broad-band measurements with the assump-tion of a power-low spectrum(fλ∝λβ)6(references incolumn9of Table4).This was not possible for somegalaxies of M03because their colors in the rest-frameλ<3000˚A are not available.For the composite spec-trum of S01,we estimated the UV attenuation from thereported mean E(B−V)via the Calzetti’s attenuationlaw(Calzetti et al.,2000).The UV slope of the compositespectrum shown in Fig.1of S01is consistent with the slopecorresponding to the estimated attenuation.We estimatedUV slopes of our two galaxies from their broad-band colorsalthough we have UV spectra of our two galaxies becausethe data quality is not good.7Since the calibration gives the attenuation at the rest-frame1600˚A,we convert it into the attenuation at theappropriate UV wavelength by the Calzetti’s attenuationlaw(Calzetti et al.,2000)if the UV wavelength is diﬀerentfrom1600˚A(column8).We note here that the uncertaintyresulting from those on the UV slope and colors is verylarge;for example,the uncertainty of about0.05mag inV606−I814for our two galaxies translates into∆A UV∼2∆β∼6∆(V606−I814)∼0.3mag(see Meurer et al.,1999).For galaxies at z∼3,weﬁnd that(1)the absoluteescape fraction from the detection by S01is60±13%(1-σobservational uncertainty),(2)the absolute escapefraction of the two brightest LBGs observed by G02isless than5%(3-σ),(3)the absolute escape fraction of L∗LBGs observed by us is less than20–40%(3-σ).For z∼1galaxies,weﬁnd very small upper limits onthe absolute escape fraction,typically,less than a few per-cent(3-σ).Since small upper limits were obtained even forthe relative escape fraction,the conclusion of very smallabsolute escape fractions for the observed galaxies seemsrobust against the uncertainty of estimating the dust at-tenuation.10Inoue et al.:Upper limit of Lyman continuum escapeopacity.Stacking a signiﬁcant number of deep narrow-band images of drop-out galaxies has,therefore,the potential to conﬁrm or not the high relative escape fraction reported by Steidel et al.(2001).In addition to increasing sensitivity,such a method would average the IGM opacity and the randomness of the LC escape. Acknowledgements.We thank Tsutomu T.Takeuchi for a lot of valuable comments,Matthew A.Bershady for kindly providing us with his opacity model as a machine-readable form,Alberto Fern´a ndez-Soto for helpful discussions,and ESO support astronomers for their cooperation during the phase 2submission and observations.AKI also thanks Hiroyuki Hirashita,Masayuki Akiyama,Hideyuki Kamaya,and Shu-ichiro Inutsuka for their continuous encouragements.In the middle of this work,AKI was invited to the Laroratoire d’Astrophysique de Marseille andﬁnancially supported by the Observatoire Astronomique de Marseille-Provence.AKI is also supported by the JSPS Postdoctoral Fellowships for Research Abroad.ReferencesAdelberger,K.L.,&Steidel,C.C.2000,ApJ,544,218 Becker,R.H.,et al.2001,AJ,122,2850Bell,E.F.2002,ApJ,577,150Bershady,M.A.,Charlton,J.C.,&Geoﬀroy,J.M.1999, ApJ,518,103Buat,V.,Donas,J.,Milliard,B.,&Xu,C.Buat,V.,Boselli,A.,Gavazzi,G.,&Bonfanti,C.2002, A&A,383,801Calzetti,D.,Armus,L.,Bohlin,R.C.,Kinney,A.L., Koornneef,J.,&Storchi-Bergmann,T.2000,ApJ,533, 682Casertano,S.,et al.2000,AJ120,2747Deharveng,J.-M.,Buat,V.,Le Brun,V.,Milliard,B., Kunth,D.,Shull,J.M.,&Gry,C.2001,A&A,375,805 Ferguson,H. C.2001,in Deep Fields,Proceedings of the ESO/ECF/STScI Workshop,eds.,Cristiani, C.,Renzini, A.,Williams,R. E.(Springer)p.54 (astro-ph/0101356)Ferland,G.J.1996,Hazy,a Brief Introduction to Cloudy, Vol.II,University of Kentucky Department of Physics and Astronomy Internal Report.Fern´a ndez-Soto,A.,Lanzetta,K.M.,&Chen,H.-W.2003, MNRAS,342,1215(FS03)Giallongo,E.,Cristiani,S.,D’Odorico,S.,&Fontana,A. 2002,ApJ,568,L9(G02)Gordon,K.D.,Clayton,G.C.,Witt,A.N.,&Misselt,K.A.2000,ApJ,533,236Heckman,T.M.,Sembach,K.R.,Meurer,G.R., Leitherer,C.,Calzetti,D.,&Martin,C.L.2001,ApJ, 558,56Hornschemeier,A.E.,et al.2001,ApJ,554,742 Hurwitz,M.,Jelinsky,P.,&Dixon,W.V.1997,ApJ,481, L31Inoue,A.K.2001,AJ,122,1788Inoue,A.K.,Hirashita,H.,&Kamaya,H.2001,ApJ,555, 613Inoue,A.K.,&Kamaya,H.2004,MNRAS,350,729 Kim,T.-S.,et al.2002,MNRAS,335,555Kogut,A.,et al.2003,ApJS,148,161Kong,X.,Charlot,S.,Brinchmann,J.,&Fall,S.M.2004, MNRAS,349,769Kriss,G.A.,et al.2001,Science,293,1112Labb´e,I.,et al.2003,AJ,125,1107Leitherer, C.,Ferguson,H. C.,Heckman,T.M.,& Lowenthal,J.D.1995,ApJ,454,L19Leitherer,et al.1999,ApJS,123,3Leitherer,C.,Li,I.-H.,Calzetti,D.,&Heckman,T.M. 2002,ApJS,140,303Madau,P.1995,ApJ,441,18Malkan,M.,Webb,W.,&Konopacky,Q.2003,ApJ,598, 878(M03)Meurer,G.R.,Heckman,T.M.,&Calzetti,D.1999,ApJ, 521,64Møller,P.,&Jakobsen,P.1990,A&A,228,299 Murdoch,H.S.,Hunstead,R.W.,Pettini,M.,&Blades, J.C.1986,ApJ,309,19Osterbrock,D.E.1989,Astrophysics of Gaseous Nebulae and Active Galactic Nuclei(Mill Valley:University Science Books)P´e roux,C.,McMahon,R.G.,Storrie-Lombardi,L.J..& Irwin,M.J.2003,MNRAS,346,1103Petitjean,P.,Webb,J.K.,Rauch,M.,Carswell,R.F.,& Lanzetta,K.1993,MNRAS,262,499Pettini,M.,Kellogg,M.,Steidel,C.C.,Dickinson,M., Adelberger,K.L.,&Giavalisco,M.1998,ApJ,508, 539Sargent,W.L.W.,Steidel,C.C.,&Boksenberg,A.1989, ApJS,69,703Schlegel,D.J.Finkbeiner,D.P.,&Davis,M.1998,ApJ, 500,525Steidel, C. C.,Adelberger,K.L.,Giavalisco,M., Dickinson,M.,&Pettini,M.1999,ApJ,519,1 Steidel,C.C.,Pettini,M.,&Adelberger,K.L.2001,ApJ, 546,665(S01)Stengler-Larrea,E.A.,et al.1995,ApJ,444,64 Weymann,et al.1998,ApJ,506,1Whittet,D.C.B.2003,Dust in the galactic environment, 2nd ed.(Bristol:IOP)Wiese,W.L.,Smith,M.W.,&Glennon,B.M.1966, Atomic Transition Probabilities,Vol.1(NSRDS-NBS 4;Washington:GPO)Williams,R.E.,et al.2000,AJ120,2735Zheng,W.et al.2004,ApJ,605,631Zuo,L.,&Phinney,E.S.1993,ApJ,418,28Inoue et al.:Upper limit of Lyman continuum escape 11Appendix A:A mean IGM opacity modelIn this appendix,we ﬁrst show that the number density distribution function of the Lyman limit systems (LLSs)reasonably agrees with an extension of the distribution function of the Lyman αforest,in contrast to previous works.That is,the intergalactic clouds may be described by a continuous distribution from low-to high-density through any redshift.Second,we present an analytic approximation of a mean opacity of the intergalactic medium (IGM)based on the distribution function.A mean IGM opacity at an observed wavelength λobs against a source with z =z S can be expressed as (e.g.,Møller &Jakobsen,1990;Madau,1995) τIGM λobs (z S ) = z S 0dz N HI ,up N HI ,lowdN HI ∂2N ∂z∂N HI =f (N HI )g (z ),(A.2)where f (N HI )∝N −βHI with the normalization as N HI ,up N HI ,low f (N HI )dN HI =1and g (z )=A (1+z )γ.We note here that the normalization A means the total number of the IGM cloud per unit redshift interval and is a function of N HI ,lowand N HI ,up .Observationally,the power-law index βmay be regarded as a constant equal to about 1.5for a wide range of column densities from log N HI /cm −2=12.5to 21.5,although it seems to depend on redshift and column density in detail (Sargent et al.,1989;Petitjean et al.,1993;Kim et al.,2002;P´e roux et al.,2003).For simplicity,we assume a single power-law index β=1.5for all redshifts and column densities.For the cloud number evolution with redshift,Madau (1995)adopted two diﬀerent evolutionary indices (γ)for the Lyman αforest (N HI <∼1017cm −2)and for the LLSs (N HI >∼1017cm −2),based on observations by Murdoch et al.(1986)and Sargent et al.(1989).However,the same index γfor these populations is compatible with recent ob-servations as shown in Fig.A.1.The LLS number evolution is reproduced from P´e roux et al.(2003)in this ﬁgure as the ﬁlled square points with error bars.The solid line is the number evolution of the Lyman αforest reported by Kim et al.(2002)and Weymann et al.(1998)multiplied by the reduction factor due to the lowest column den-sity considered;Kim et al.(2002)deal with Lyman αforest with log N HI /cm −2=13.644–17,whereas P´e roux et al.(2003)deal with the clouds with a column density larger than 1.6×1017cm −2,so that the multiplicative factor is (4.4×1013/1.6×1017)β−1≈0.017when β=1.5.There is a systematic disagreement along the vertical axis for high-z points between the solid line and the data,but the evolutionary slopes are consistent as noted by P´e roux et al.(2003).The high-z data are reproduced by the dashed line whose reduction factor is 0.01.On the other hand,the LLS number evolution reported by Sargent et al.(1989)and adopted by Madau (1995)is shown as the dotted line in Fig.A1and cannot reproduce the data points of P´e roux et al.(2003).Although the number evolution suggested by Stengler-Larrea et al.(1995)shows a better ﬁt (dash-dotted line),in this paper,we assume the case of the solid line.This means that we assume the same redshift evolution of the cloud number density over the entire column density range with β=1.5.Under a cloud distribution function with a single power-law index β(1<β<2)and the same evolutionary index γfor all range of the column density,we take a set of limiting column densities as σH N HI ,low ≪1and σH N HI ,up ≫1,for example,N HI ,low =1012cm −2and N HI ,up =1022cm −2.In this case,equation (A.1)can be approximated asτIGM λobs (z S ) ≈ z S 0Γ(2−β)A N β−1HI ,low (1+z )γσH (λobs /1+z )β−1dz ,(A.3)where Γ(2−β)is the usual Gamma function (Zuo &Phinney,1993).The factor,A N β−1HI ,low can be estimated fromthe observed number of clouds for a limited range of the column density with the power-law distribution.If the number of clouds with a column density between N l and N u (N HI ,low ≤N l <N u ≤N HI ,up )is denoted as A ′,wehave A N β−1HI ,low ≈A ′N β−1lwhen N HI ,up /N HI ,low ≫1,N u /N HI ,low ≫1,and β>1.According to Kim et al.(2002)and Weymann et al.(1998),we adopt (A ′,γ)=(6,2.5)for z >1.1and (34,0.2)for z ≤1.1against log N HI /cm −2=13.64–17(i.e.N l =4.4×1013cm −2).We will present an analytic approximation of equation (A.3),although we can integrate it numerically with a detailed function of the hydrogen cross section,σH .We adopt an approximated form of the cross section for the analytical formula.The point is that the cross section of the i -th line,σi (λ),can be neglected for a wavelength λout of a small range of |λ−λi |<∼b/c ,where λi is the central wavelength of the line,b is the Doppler parameter,and c is。