scieee Open visual document viewer

Interplay between crystal-size and disorder effects in the high-energy optical response of photonic crystal slabs

Dorado, Luis A.; Depine, Ricardo A.; Lozano Barbero, Gabriel Sebastián; Míguez García, Hernán Ruy

Abstract

Experimental reflectance spectra have been obtained for colloidal crystals whose widths ranged from one to several sphere monolayers, and their features in the higher order band energy range have been reproduced theoretically. In order to fit the measured data, optical extinction has been introduced in the theoretical model, which accounts for structure imperfections and disorder, the main sources of losses in an actual measurement. A complex spectrum in the high frequency region is observed even for one ordered monolayer, being this peak structure gradually modified as more layers are piled up. This allowed us to identify which peaks are reminiscent of the optical reflectance features of a single close-packed layer and which are the result of building up a three dimensional periodicity. A clear correlation between the amount of extinction introduced in the fitting and the slab width has been found, which demonstrates that wider real crystals produce less diffusely scattered light. At the same time, we find that the optical response of thinner crystals is more robust against the introduction of extinction than that of thicker ones, for which the effect is dramatic.

Full text

In e play be ween c ys al-size and diso de e ec s in he high-ene gy op ical esponse o pho onic c ys al slabs Luis A. Do ado and Rica do A. Depine* G upo de Elec omagne ismo Aplicado, Depa amen o de Física, Facul ad de Ciencias Exac as y Na u ales, Uni e sidad de Buenos Ai es, Buenos Ai es C1428EGA, A gen ina Gab iel Lozano and He nán Míguez Ins i u o de Ciencia de Ma e iales de Se illa, Consejo Supe io de In es igaciones Cien í icas, Se illa 41092, Spain 共Recei ed 20 June 2007; e ised manusc ip ecei ed 13 Augus 2007; published 5 Decembe 2007兲 Expe imen al e lec ance spec a ha e been ob ained o colloidal c ys als whose wid hs anged om one o se e al sphe e monolaye s, and hei ea u es in he highe o de band ene gy ange ha e been ep oduced heo e ically. In o de o i he measu ed da a, op ical ex inc ion has been in oduced in he heo e ical model, which accoun s o s uc u e impe ec ions and diso de , he main sou ces o losses in an ac ual measu emen . A complex spec um in he high equency egion is obse ed e en o one o de ed monolaye , being his peak s uc u e g adually modi ied as mo e laye s a e piled up. This allowed us o iden i y which peaks a e eminis- cen o he op ical e lec ance ea u es o a single close-packed laye and which a e he esul o building up a h ee dimensional pe iodici y. A clea co ela ion be ween he amoun o ex inc ion in oduced in he i ing and he slab wid h has been ound, which demons a es ha wide eal c ys als p oduce less di usely sca e ed ligh . A he same ime, we ind ha he op ical esponse o hinne c ys als is mo e obus agains he in oduc ion o ex inc ion han ha o hicke ones, o which he e ec is d ama ic. DOI: 10.1103/PhysRe B.76.245103 PACS numbe 共s兲: 78.20.Bh, 78.40.⫺q, 42.70.Qs, 41.20.Jb Imp o emen s in he p ocesses o ab ica ion o sel - assembled h ee dimensional 共3D兲pho onic c ys als 共PhCs兲, which a e ma e ials wi h a spa ially pe iodic dielec ic unc- ion in all h ee dimensions, ha e made possible he obse - a ion o high quali y op ical spec a in highe o de band equencies.1–3Also, he appea ance o localized modes in- side band gaps in he pho onic band s uc u e due o he addi ion o local de ec s is a well-known e ec , which is analogous o he doping o semiconduc o s.4Poin , line, o plane de ec s a e c ea ed wi hin he c ys al by locally adding o emo ing ma e ial in a con olled manne ,5,6so hey a e di e en om in insic de ec s, which a e unin en ional dis- up ions o he spa ial pe iodici y o he dielec ic unc ion ha a ise du ing he ab ica ion p ocess. Despi e new ab i- ca ion echniques ha ha e alle ia ed he e ec o diso de in he so-called low-ene gy ange, whe e he la ice cons an is less han he wa eleng h o ligh , e y ecen ly, i has been demons a ed ha ex inc ion due o in insic de ec s in 3D PhCs s ongly a ec s he shape o he expe imen al spec a in he high-ene gy ange, whe e he la ice cons an is g ea e han he wa eleng h.7Also, calcula ions o he pho onic band s uc u e, pe o med conside ing ex inc ion, ha e shown a clea co ela ion be ween he beha io o he imagina y pa o he wa e ec o and he op ical ea u es obse ed in he spec a. This means ha diso de , while ha ing a small in- luence on he measu ed spec a in he low-ene gy ange, ampli ies i s e ec s a highe ene gies and, consequen ly, u - he imp o emen o he ab ica ion echniques is equi ed o achie e op ical quali y a hose pho on ene gies.8 In o de o pe o m a alid compa ison be ween he op i- cal spec a o ini e c ys al slabs and he pho onic band s uc- u e o an in ini e c ys al, PhC slabs o se e al laye s we e analyzed in a p e ious wo k; an excellen ag eemen be ween he expe imen al and he calcula ed op ical esponse a high ene gies has been ound. He e, we p esen an analysis o he beha io o he op ical spec a ea u es in he high-ene gy ange as he wid h o he c ys al slab is g adually inc eased. We compa e he expe imen al and heo e ical e olu ions o he specula e lec ance spec a as we inc ease he c ys al hickness om 1 o 18 sphe e laye s. We ound complex spec a in he high equency egion e en o he ini ial monolaye s uc u e,9being his peak s uc u e g adually modi ied as mo e laye s a e piled up. This allowed us o iden i y which peaks a e eminiscen o he op ical e lec- ance ea u es o a single close-packed laye and which a e he esul o building up a h ee dimensional pe iodici y. A he same ime, we ind a clea co ela ion be ween he c ys al size and he amoun o ex inc ion needed o a ain he bes i . The in e play o c ys al size and diso de e ec s on he high- ene gy op ical esponse is ho oughly analyzed, inding a much mo e d ama ic e ec o he o me on he e lec ance spec a o la ge c ys als. Pho onic c ys al ilms we e made by e apo a ion induced sel -assembly10 on glass subs a es o polys y ene sphe e 共IKERLAT, a e age diame e o 700 nm, polydispe si y be- low 3%, densi y ␳ =1.1 g/cm3, e ac i e index n=1.59兲sus- pended in wa e wi h pa icle olume ac ion anging om 0.05% o 0.20% and e apo a ed a empe a u es anging om 30 o 60 °C. Se e al independen measu emen s o he high-ene gy op ical esponse o simila s uc u es can be ound in he li e a u e.1,11,12 Thus, he echnique is su i- cien ly ma u e and eliable o p o ide ep oducible esul s in di e en labo a o ies. The ypical c ys al g ow h di ec ion is he 关111兴, and we will ocus on he op ical p ope ies when ligh impinges in ha pa icula di ec ion since hey a e bes known.13 Thus, in his case, a monolaye consis s o a ian- gula la ice o close-packed sphe es and he slab wid h is inc eased by piling monolaye s up in he sequence PHYSICAL REVIEW B 76, 245103 共2007兲 1098-0121/2007/76共24兲/245103共4兲©2007 The Ame ican Physical Socie y245103-1 ABCABC¯. As i is usual, he pho on ene gy is exp essed in educed uni s a/␭, whe e ais he la ice cons an o he classical cubic cell, whose alue is a=冑2 ␾ in an cc close- packed s uc u e, ␾ being he sphe e diame e and ␭ he wa eleng h o he inciden ligh . Re lec ance measu emen s we e pe o med using a Fou- ie ans o m in a ed spec opho ome e 共BRUKER IFS- 66兲a ached o a mic oscope. A 4⫻objec i e wi h a nume i- cal ape u e o 0.1 共ligh cone angle o ±5.7°兲was used o i adia e he la ices and o collec he e lec ed ligh a qua- sino mal incidence wi h espec o i s su ace. A spa ial il e was used o selec i ely de ec ligh om squa e spo s o 1000⫻1000 ␮ m2. Simula ed e lec ance spec a we e calcula ed by means o he ec o Ko inga-Kohn-Ros oke me hod.14,15 To ensu e con e gence in he high-ene gy ange, we used 41 wo di- mensional 共2D兲 ecip ocal la ice ec o s in he plane wa e expansions and sphe ical wa es wi h angula momen um up o lmax=9. These alues a e he minimum ones ha p o ide good con e gence in ha ange. Lowe alues o lmax and less ecip ocal la ice ec o s s ill gi e a co ec esul o lowe -ene gy bands, bu i is no su icien o p o ide us ul in o ma ion on he egion we a e in e es ed in. Measu e- men s we e pe o med wi h he PhC slab deposi ed on a glass subs a e o e ac i e index 1.53, so his subs a e is also included in he heo e ical model. The incoming me- dium is ai 共 e ac i e index=1兲. The sphe es a e embedded in ai and ha e a complex dielec ic cons an ␧s=2.5+i␧i, which co esponds o la ex sphe es 共 e ac i e index=冑2.5 ⬵1.58兲. An imagina y pa ␧iis added o he dielec ic con- s an o he sphe es in o de o ake in o accoun all possible sou ces o losses in ac ual measu emen s. Then, he alue o ␧iis chosen o i he expe imen al da a. Ene gy losses due o powe dissipa ion a e negligible in la ex sphe es, so s uc u e impe ec ions and diso de a e e- sponsible o he expe imen al de ia ion o he spec um shape wi h espec o he heo e ical op ical esponse wi hou ex inc ion.7In he high-ene gy ange, he a e age size o hese s uc u al impe ec ions, measu ed in wa eleng hs, be- comes mo e impo an han ha in he low-ene gy bands. Fo his eason, hese in insic de ec s16 cause ligh o be sca - e ed di usely, emo ing pa o he ene gy om he specu- la ly e lec ed o ballis ically ansmi ed beam. In his wo k, we will ocus on he analysis o specula e lec ance esul s, since hey can be eadily a ained by mic ospec oscopy om single domains o he c ys alline ilm using a low nume ical ape u e objec i e, ensu ing he hickness uni o mi y o he es ed spo and he no mal incidence o he beam. This is a much mo e complica ed ask o be pe o med in ansmission mode, since he inciden beam is ypically ocused wi h a la ge nume ical ape u e objec i e. In Figs. 1and 2, he measu ed and calcula ed specula e lec ances 共o R0,0兲can be seen as a unc ion o he la ice cons an a/␭ o di e en c ys al slab wid hs. Ve y good ag eemen be ween he expe imen al and heo e ical e lec- ance spec a has been ob ained by in oducing ex inc ion in he heo e ical model, as we ha e demons a ed o much hicke c ys als be o e. Figu e 1shows he expe imen al and heo e ical specula e lec i i y spec a o PhC slabs o a ew laye s 共N=numbe o laye s兲. The alue o ␧iin he calcula ed spec a was chosen o ob ain he bes i . No e ha all oscilla ions p esen a simila heigh wi h espec o he backg ound in bo h he expe imen al 共black solid lines兲and he simula ed 共 ed dashed lines兲spec a, which was ac ually he c i e ion ollowed o choose he bes i . Di e ences in he baseline a e due o he ac ha only one in e ace o he glass subs a e is conside ed o he calcula ions in o de o a oid he p esence o e y sho equency lobes in he spec- a. These a ise om he in e e ence be ween beams e- lec ed a he uppe and lowe aces o he hick subs a e 共1mm兲unde conside a ion. Also, some dispe sion in he eal pa o he dielec ic cons an was added, which accoun s o a small e ac i e index inc ease o ma e ials as we mo e owa d highe ene gies. In Fig. 1共a兲, wo main peaks o he monolaye 共N=1兲can FIG. 1. 共Colo online兲Measu ed 共solid lines兲and calcula ed 共dashed lines兲specula e lec ance spec a o glass-suppo ed slabs o Nlaye s composed o sphe es o dielec ic cons an ␧s=2.5+i␧i in ai . 共a兲N=1, ␧i=0.30. 共b兲N=2, ␧i=0.25. 共c兲N=3, ␧i=0.15. 共d兲 N=4, ␧i=0.13. FIG. 2. 共Colo online兲Measu ed 共solid lines兲and calcula ed 共dashed lines兲specula e lec ance spec a o glass-suppo ed slabs o Nlaye s composed o sphe es o dielec ic cons an ␧s=2.5+i␧i in ai . 共a兲N=6, ␧i=0.08. 共b兲N=13, ␧i=0.06. 共c兲N=18, ␧i=0.06. DORADO e al. PHYSICAL REVIEW B 76, 245103 共2007兲 245103-2 be seen: one nea a/␭=0.55 and he o he a a/␭=1.0. These peaks mo e g adually owa d highe ene gies as he slab wid h is inc eased, as Figs. 1共a兲–1共c兲show. Also, seconda y peaks ela ed o he Fab y-Pe o oscilla ions appea be ween hese main peaks o slabs o N=2, 3, and 4 laye s. Since all ou measu emen s we e aken a om he edges o he c ys- al, in which di e en ypes o s acking sequences ha e been epo ed,17 he ai ness o he i ings we a ain conside ing only he ABCABC¯sequence indica es ha an cc la ice is buil as mo e monolaye s a e piled up. In Fig. 2, he e lec i i y spec a o PhC slabs o N=6, 13, and 18 laye s can be app ecia ed. The main peak in he low-ene gy band ends o i s inal posi ion, which is a a/␭ =0.61, as can be p edic ed om he pho onic band s uc u e o he sphe e cc la ice unde conside a ion, because he e is a pseudoband gap in he ⌫Ldi ec ion cen e ed a his pa - icula ene gy.7The i s peak in he high-ene gy egion has mo ed o a/␭=1.07, and a second double degene a ed peak nea a/␭=1.2 is p esen . I is in e es ing o no e ha his second peak becomes double degene a ed in slabs wi h a leas six laye s. In o he wo ds, we mus comple e a leas wo ABC sequences in he slab in o de o ob ain he peak a a/␭=1.2. Figu e 2also shows ha his peak is di icul o see in he measu ed da a o N=6 and 13 laye s, bu Fig. 2共c兲 o N=18 laye s con i ms i s exis ence. The e is a second peak s uc u e in he ange o 1.5⬍a/␭⬍1.8, which is composed o wo peaks o simila ampli udes. They can be obse ed in Figs. 1共c兲,1共d兲, and 2, so he numbe o laye s needed mus be Nⱖ3, which co esponds o one o mo e comple e se- quences o ABC. The g adual appea ance o hese high- ene gy peaks as he numbe o sphe e laye s is inc eased can be p edic ed om he pho onic band s uc u e calcula ed wi h ex inc ion.7Howe e , o a/␭⬎1.63, di ac ed spo s emo e pa o he ene gy om he specula ly e lec ed beam. Then, abo e ha ene gy, bo h diso de and di ac ion con ibu e o he educ ion o in ensi y o he specula ly e- lec ed ligh . The a io be ween he powe ca ied by each di ac ed mode and he inciden powe is gi en by a e lec ion coe i- cien Rp,q, whe e he pai o in ege s 共p,q兲indica e he o de o di ac ion, and 共p,q兲=共0,0兲co esponds o he specula ly e lec ed ligh , he quan i y ha has been measu ed o his wo k. The onse o di ac ion in ai a a/␭=1.63 can be p edic ed by equi ing he conse a ion o he angen ial componen o he wa e ec o s o a 2D iangula la ice, ha is, a ␭艌冑2 n冑p2+共2q+p兲2 3,共1兲 whe e nis he e ac i e index o he di ac ion medium 共n=1 o ai 兲. This same exp ession explains he appea - ance o di ac ed modes only wi hin he glass subs a e 共n=1.53兲and no in he incoming medium a a/␭=1.07 o he suppo ed la ice.7In Fig. 3, we compa e he calcula ed spec um o R0,0 and ha o 兺共p,q兲⫽共0,0兲Rp,q共 he sum o he in ensi ies o all di ac ed modes in he homogeneous inci- den medium o e ac i e index n=1兲 o he slab composed o 18 laye s. The o al e lec i i y R=兺共p,q兲Rp,qis also plo ed in Fig. 3共a兲, whe e he in ensi y educ ion o R0,0 compa ed o Rcan be app ecia ed o pho on ene gies a/␭⬎1.63. The impac o diso de in he op ical esponse o he glass-suppo ed PhC slab can be obse ed in Fig. 4, whe e he specula e lec ance spec a o N=1 and 5 laye s is shown o di e en alues o he imagina y dielec ic con- s an ␧i. The a enua ion o he e lec ance alues compa ed o he case ␧i=0.0001, which co esponds o a slab wi h al- mos no diso de , can clea ly be seen in bo h cases. Howe e , his e ec is much mo e d ama ic in he case o he i e laye la ices, whose maximum a a/␭=1.15 is dec eased by a ac- o abo e 10, while o he monolaye , he co esponding maximum a a/␭=1.0 is educed by ac o smalle han 2. This is a heo e ical con i ma ion ha he hinne he c ys als, FIG. 3. 共Colo online兲共a兲Calcula ed specula 共solid line兲and o al 共dashed line兲 e lec ance spec a, R0,0 and R, o a glass- suppo ed slab o 18 laye s and sphe es o dielec ic cons an ␧s =2.4964+0.06iin ai . 共b兲Re lec ed di ac ed ligh , 兺共p,q兲⫽共0,0兲Rp,q, spec a. The e ical dashed line indica es he onse o di ac ion. FIG. 4. 共Colo online兲Calcula ed specula e lec ance spec a o 共a兲a glass-suppo ed monolaye 共N=1兲and 共b兲a glass- suppo ed slab o N=5 laye s, bo h wi h sphe es o dielec ic con- s an ␧s=2.5+i␧iin ai o di e en alues o ␧i. INTERPLAY BETWEEN CRYSTAL-SIZE AND DISORDER…PHYSICAL REVIEW B 76, 245103 共2007兲 245103-3 he mo e obus i s op ical esponse agains he in oduc ion o ex inc ion. In he expe imen al spec a o Figs. 1and 2, we can ap- p ecia e a g adual inc ease o he e lec ance ampli udes as he numbe o laye s is inc eased. Also, he imagina y pa ␧i o he dielec ic cons an o he sphe es dec eases as he numbe o laye s Nis inc eased in he calcula ed spec a. This endency is plo ed in Fig. 5, whe e ␧ican be seen as a unc ion o N. Hence, he e exis s a clea co ela ion be ween he ex inc ion in oduced and he wid h o he slab, which indica es ha di usely sca e ed ligh inc eases o hinne c ys als. In he case o c ys als o a ew laye s, besides he e ec o diso de , in e ac ions wi h he glass subs a e p o- duce an addi ional a enua ion o he e lec i i y alues. The e o e, he inc ease o ␧ias he c ys al wid h is dec eased means ha mo e di usely sca e ed ligh is p oduced, he combina ion o diso de and glass-c ys al in e ac ion being esponsible o his e ec . Ne e heless, o c ys als wi h six o mo e laye s, ␧ibecomes p ac ically a cons an , as Fig. 5 shows, which means ha any su ace e ec and glass-c ys al in e ac ion a e negligible, he alue o ␧ibeing a measu e o he deg ee o diso de in hese cases. Thus, we ound a me hod o quan i ying diso de in a eal c ys al because we can use a single pa ame e such as ␧i, which is de e mined by i ing he heo e ical model o he expe imen al da a, as a measu e o he s uc u al quali y o he PhC slab. A simila me hod has al eady been p oposed o he same ype o PhC ilms bu based on he i ing o he low-ene gy op ical e- sponse o he la ices.18 In e es ingly, he same end o he ex inc ion needed o i he expe imen al cu e e sus he numbe o close-packed laye s in he ilm is ound. In conclusion, we ha e ealized an in eg al app oach o he op ical esponse o 3D PhC slabs as a unc ion o he slab wid h in he highe o de band ene gy ange. By analyzing he g adual modi ica ion o ha esponse om one o se e al monolaye s, we could iden i y which peaks a e eminiscen o he op ical e lec ance ea u es o a single close-packed laye and which a e he esul o building up a h ee dimen- sional pe iodici y. Fu he mo e, we ha e ound a clea co e- la ion be ween he ex inc ion in oduced in ou heo e ical model and he wid h o he eal slab, which indica es ha wide c ys al slabs p oduce less di usely sca e ed ligh and p o ides a me hod o quan i ying diso de in hese cases. This wo k has been ealized in he amewo k o a join Spanish-A gen inian coope a ion p ojec CSIC-CONICET 共G an No. 2005AR0070兲. R.A.D. and L.A.D. acknowledge suppo om Consejo Nacional de In es igaciones Cien í i- cas y Técnicas 共CONICET兲, Uni e sidad de Buenos Ai es 共UBA兲, and Agencia Nacional de P omoción Cien í ica y Tecnológica 共ANPCYT-BID 802/OC-AR03-14099兲. H.M. is g a e ul o inancial suppo om he Ramón A eces Foun- da ion and he Spanish Minis y o Science and Educa ion unde G an No. MAT2005-03028. *[email p o ec ed] 1H. Míguez, V. Ki ae , and G. Ozin, Appl. Phys. Le . 84, 1239 共2004兲. 2F. Ga cía-San ama ía, J. F. Galis eo-López, P. V. B aun, and C. López, Phys. Re . B 71, 195112 共2005兲. 3K. Wos yn, Y. Zhao, B. Yee, K. Clays, A. Pe soons, G. de Scha- e zen, and L. Hellemans, J. Chem. Phys. 118, 10752 共2003兲. 4E. Yablono i ch, T. J. Gmi e , R. D. Meade, A. M. Rappe, K. D. B omme , and J. D. Joannopoulos, Phys. Re . Le . 67, 3380 共1991兲. 5P. V. B aun, S. A. Rinne, and F. Ga cia-San ama ia, Ad . Ma e . 共Weinheim, Ge .兲18, 2665 共2006兲. 6A. A senaul , F. Fleischhake , G. on F eymann, V. Ki ae , H. Míguez, A. Mihi, N. Te eaul , E. Vek is, I. Manne s, S. Ai chi- son, D. Pe o ic, and G. A. Ozin, Ad . Ma e . 共Weinheim, Ge .兲 18, 2779 共2006兲. 7L. A. Do ado, R. A. Depine, and H. Míguez, Phys. Re . B 75, 241101共R兲共2007兲. 8Z. Y. Li and Z. Q. Zhang, Phys. Re . B 62, 1516 共2000兲. 9S. Yano, Y. Segawa, J. S. Bae, K. Mizuno, S. Yamaguchi, and K. Oh aka, Phys. Re . B 66, 075119 共2002兲. 10 J. F. Be one, P. Jiang, K. S. Hwang, D. M. Mi leman, and V. L. Col in, Phys. Re . Le . 83, 300 共1999兲. 11 J. F. Galis eo-López and C. López, Phys. Re . B 70, 035108 共2004兲. 12 S. Wong, V. Ki ae , and G. A. Ozin, J. Am. Chem. Soc. 125, 15589 共2003兲. 13 P. Jiang, J. F. Be one, K. S. Hwang, and V. L. Col in, Chem. Ma e . 11, 2132 共1999兲. 14 N. S e anou, V. Yannopapas, and A. Modinos, Compu . Phys. Commun. 113, 4977 共1998兲;132, 189 共2000兲. 15 N. S e anou, V. Ka a hanos, and A. Modinos, J. Phys.: Condens. Ma e 4, 7389 共1992兲. 16 E. Lido ikis, M. M. Sigalas, E. N. Economou, and C. M. Souk- oulis, Phys. Re . Le . 81, 1405 共1998兲. 17 X. Checou y, S. Enoch, C. López, and A. Blanco, Appl. Phys. Le . 90, 161131 共2007兲. 18 J. F. Galis eo-López, M. Galli, M. Pa ini, A. Bales e i, L. C. And eani, and C. López, Phys. Re . B 73, 125103 共2006兲. FIG. 5. Imagina y pa o he dielec ic cons an o he sphe es as a unc ion o he numbe o laye s o he PhC slab. DORADO e al. PHYSICAL REVIEW B 76, 245103 共2007兲 245103-4