scieee Open visual document viewer

Enhanced backward wave propagation in evanescent waveguides loaded with split ring resonators

Carbonell, Jorge; Roglá, L. J.; Boria, Vicente E.; Marqués Sillero, Ricardo

Abstract

This paper reports on the design and experimental characterization of evanescent waveguides loaded with split ring resonators SRRs supporting enhanced backward wave propagation. With respect to previous negative refractive index devices based on metallic waveguides, the use of a direct coaxial probe excitation inside an evanescent waveguide drastically improves transmission and reflection characteristics of the periodic structure. Moreover, the addition of metallic windows interleaved between consecutive resonators offers an extra degree of freedom in order to control interresonator couplings, and hence, device parameters like overall transmission, reflection, and bandwidth of the backward wave pass band. A numerical and experimental analysis of different structures overcoming the high loss issue of previously published results is detailed. An experimental validation of the proposed enhancements is done for periodic structures, where no individual tuning of the resonator sections is performed. The proposed technology mainly offers the advantage of miniaturization, and it could be combined in a further step with standard design techniques to provide frequency selective devices.

Full text

J. Appl. Phys. 102, 044902 (2007); h ps://doi.o g/10.1063/1.2769141 102, 044902 © 2007 Ame ican Ins i u e o Physics. Enhanced backwa d wa e p opaga ion in e anescen wa eguides loaded wi h spli ing esona o s Ci e as: J. Appl. Phys. 102, 044902 (2007); h ps://doi.o g/10.1063/1.2769141 Submi ed: 18 Janua y 2007 . Accep ed: 28 June 2007 . Published Online: 20 Augus 2007 J. Ca bonell, L. J. Roglá, V. E. Bo ia, and R. Ma qués ARTICLES YOU MAY BE INTERESTED IN Elec ic- ield-coupled esona o s o nega i e pe mi i i y me ama e ials Applied Physics Le e s 88, 041109 (2006); h ps://doi.o g/10.1063/1.2166681 Spli ing esona o -based le -handed coplana wa eguide Applied Physics Le e s 83, 4652 (2003); h ps://doi.o g/10.1063/1.1631392 Enhanced backwa d wa e p opaga ion in e anescen wa eguides loaded wi h spli ing esona o s J. Ca bonell,a兲L. J. Roglá, and V. E. Bo iab兲 Ins i u o de Telecomunicaciones y Aplicaciones Mul imedia (iTEAM), Uni e sidad Poli écnica de Valencia, Camino de Ve a s/n, 46022 Valencia, Spain R. Ma quésc兲 Depa amen o de Elec ónica y Elec omagne ismo, Uni e sidad de Se illa, A . Reina Me cedes s/n, 41012 Se illa, Spain 共Recei ed 18 Janua y 2007; accep ed 28 June 2007; published online 20 Augus 2007兲 This pape epo s on he design and expe imen al cha ac e iza ion o e anescen wa eguides loaded wi h spli ing esona o s 共SRRs兲suppo ing enhanced backwa d wa e p opaga ion. Wi h espec o p e ious nega i e e ac i e index de ices based on me allic wa eguides, he use o a di ec coaxial p obe exci a ion inside an e anescen wa eguide d as ically imp o es ansmission and e lec ion cha ac e is ics o he pe iodic s uc u e. Mo eo e , he addi ion o me allic windows in e lea ed be ween consecu i e esona o s o e s an ex a deg ee o eedom in o de o con ol in e esona o couplings, and hence, de ice pa ame e s like o e all ansmission, e lec ion, and bandwid h o he backwa d wa e pass band. A nume ical and expe imen al analysis o di e en s uc u es o e coming he high loss issue o p e iously published esul s is de ailed. An expe imen al alida ion o he p oposed enhancemen s is done o pe iodic s uc u es, whe e no indi idual uning o he esona o sec ions is pe o med. The p oposed echnology mainly o e s he ad an age o minia u iza ion, and i could be combined in a u he s ep wi h s anda d design echniques o p o ide equency selec i e de ices. © 2007 Ame ican Ins i u e o Physics. 关DOI: 10.1063/1.2769141兴 I. INTRODUCTION In he las ew yea s, nega i e e ac i e index ma e ials ha e opened mo e possibili ies o he design o mic owa e de ices. Gene ally included in a b oade ca ego y o “me ama e ials,” hese a i icial composi es owe hei p op- e ies o subwa eleng h de ails o s uc u e a he han o hei chemical composi ion. Possible applica ions o his ype o de ices a e based on he use o hei unique ea u es, such as backwa d wa e p opaga ion,1nega i e e ac i e index,2 o e anescen wa e egene a ion.3E en i o da e he opic is s ill an eme ging esea ch line, one can ci e, among o he s, p omising po en ial applica ions such as imaging below he di ac ion limi wi h he so-called “supe - esolu ion,”4 cloaking,5,6de ice minia u iza ion,7o , e y gene ally speak- ing, any unusual p opaga ion e ec no achie able wi h na u- al “ma e ials.” In his con ex , a numbe o wa eguide-based backwa d p opaga ion de ices ha e been p oposed ecen ly, s a ing om he o iginal p oposal in Re . 8. In ha pape , le - handed o backwa d wa e p opaga ion was ob ained o a s uc u e composed o an a ay o spli ing esona o s 共SRRs兲9inside an e anescen wa eguide. This composi e s uc u e possesses a le -handed ansmission band sligh ly abo e he esonan equency o he esona o s, which is gen- e a ed h ough he combina ion o a nega i e e ec i e pe - mi i i y syn hesized by he e anescen wa eguide enclo- su e, and a nega i e e ec i e pe meabili y syn hesized h ough he a ay o SRRs. A e his pionee ing pape , a numbe o s udies ollowed using wa eguide s uc u es as he basis o hei design.10–14 Ye , he common d awback o poo ansmission pe o mance, which could only be pa - ially alle ia ed h ough he use o long ape ing sec ions, 共Re . 15兲, was acknowledged. Two enhancemen s we e heo- e ically in oduced in, Re . 16, which is a ecen con ibu- ion aiming a imp o ing he poo ansmission le els p e i- ously epo ed. These can be ci ed as he use o a coaxial exci a ion o eed he wa eguide s uc u e and he use o induc i e windows in e lea ed be ween he SRR a ay. This echnology makes he ansmission s uc u e e y compac in size wi h espec o o he con en ional wa eguide con igu- a ions. The p esen pape analyzes bo h ypes o enhance- men s no only in e ms o ansmission, bu also in e ms o ma ching, in e esona o coupling, loca ion, and wid h o he pass band and loss con ibu ions, hus, pa ing he way o hei u u e use owa d he applica ions. Fu he mo e, expe i- men al esul s con i ming he heo e ical wo k ha e been ob- ained. In his pape , Sec. II e iews he p oposed s uc u es, and he causes o pe o mance imp o emen associa ed wi h he wa eguide heigh educ ion, wi h he use o he coaxial eed- ing p obe and wi h he use o he induc i e windows. Sec ion III desc ibes he design and manu ac u ing o a p o o ype and he expe imen al alida ion o he nume ical p edic ions. Also, a measu ed esul gi ing e idence o backwa d wa e p opaga ion is p o ided, oge he wi h a sho discussion on a兲Elec onic mail: [email p o ec ed].es b兲Elec onic mail: [email p o ec ed].es c兲Elec onic mail: [email p o ec ed] JOURNAL OF APPLIED PHYSICS 102, 044902 共2007兲 0021-8979/2007/102共4兲/044902/8/$23.00 © 2007 Ame ican Ins i u e o Physics102, 044902-1 he main ea u es associa ed wi h he p o o ypes. Finally, Sec. IV summa izes he conclusions o his wo k. II. TOPOLOGIES FOR IMPROVING THE TRANSMISSION RESPONSE OF AN EVANESCENT WAVEGUIDE LOADED WITH AN SRR ARRAY All subsequen designs a e based on a common SRR geome y. SRRs a e all o he b oadside-coupled ype: one spli ing on each side o he subs a e and symme ically o ien ed wi h espec o hei axis.17 B oadside-coupled eso- na o s do no p esen bianiso opy and ha e a educed elec- ical size as compa ed o he pa icles p oposed by Pend y in Re . 9. Squa e-shaped de ices we e selec ed since pe o - mance was much be e han o ing-shaped ones 共be e ansmission and e lec ion cha ac e is ics兲, especially o he expe imen al esul s. This choice was jus i ied because o he coupling imp o emen be ween he i s esona o and he coaxial p obe and also be ween neighbo ing esona o s. Mo eo e , he ab ica ion p ocess p oduced mo e accu a e samples in he case o he squa e-shaped esona o s. The la e al dimension o his “pa icle” is l=3.2 mm and i has me allic ack wid hs o c=0.5 mm. The ing gap o sli is 0.5 mm wide. These dimensions, pa e ned on a low pe mi - i i y subs a e 共␧ =2.2兲wi h a hickness o 0.508 mm, gi e an app oxima e esonan equency o =8.5 GHz. The pe- iodic a ay has a pe iod o p=4 mm. Fo he simula ions, i losses a e included o he SRRs, ini e conduc i i y o cop- pe 共 ␴ Cu=5.8⫻107S/m兲and dielec ic loss angen o he subs a e 共 an d=0.0009兲a e aken. A. In luence o a educed heigh wa eguide S a ing om he s uc u e p oposed in Re . 12, he au- ho s ha e i s analyzed nume ically he double in luence o he use o educed heigh wa eguides ins ead o a ull heigh 共o s anda d兲wa eguide. Nume ical simula ions h oughou his wo k use he comme cial so wa e simula o Anso HFSS. S anda d wa eguide dimensions o X-band ope a ion co espond o hose o a WR-90 共22.86 mm⫻10.16 mm兲, whe eas he educed heigh wa eguide has a e ical dimen- sion sligh ly bigge han he SRR dimension 共4.1 mm兲.An analysis in e ms o Spa ame e s shows ha he use o a mo e-con ined shielding s uc u e a ound he esona o s in- c eases no only he amoun o powe being in oduced in- side he pe iodic s uc u e, bu also he in e esona o cou- pling. This is di ec ly obse able in Fig. 1, since ansmission and e lec ion le els a e bo h imp o ed. A ansmi ed back- wa d wa e band is obse able close o he esonan e- quency o he SRRs. The numbe o esonan equencies in his ansmi ed band equals heo e ically he numbe o eso- na o s in he ini e-leng h s uc u e 共4, in his case兲, howe e , he e ec o ma e ial losses in he dielec ic and conduc o s, combined wi h a high misma ch 共 ep esen ed by S11兲, e- duces he numbe o esonan peaks o dips. Mo eo e , e- ducing he heigh o he wa eguide also inc eases in e eso- na o couplings by widening he bandwid h o he backwa d wa e band. The inc eased in e ac ion be ween he esona o s shi s upwa ds he indi idual esonan equencies obse - able in he ansmi ed ange. In bo h cases, eeding o he simula ed s uc u es is done di ec ly h ough he de ini ion o access po s a bo h ends, wi h hei shapes co esponding o he wa eguide ans e se dimensions. The exci a ion o he esona o s inside he wa eguide can be in e p e ed h ough he o iginal ac ional a ea con- cep in oduced by Pend y o he SRRs in Re . 9共a ea o he SRR o e a ea o he uni cell兲. I was shown in Re . 18 ha he coupling o he ene gy ca ied ou inside he wa eguide by he dominan TE10 mode o he esona o is based on his ac ional a ea concep . Since he po ion o he educed heigh wa eguide occupied by he SRR is bigge , much mo e magne ic- ield lines will exci e he esona o s. I is ob ious ha coupling be ween consecu i e esona- o s also depends on he sepa a ion be ween hem. The e is a signi ican in e ac ion among neighbo ing SRRs, bu his only holds o small sepa a ion dis ances. Hence, i is he collec i e esponse o he a ay ha in luences he bandwid h o he de ice. Using a educed heigh wa eguide s eng hens he in e ac ion be ween consecu i e esona o s and he ma ching o exci a ion o he pe iodic s uc u e, yielding an inc eased bandwid h. B. Coaxial p obe exci a ion The use o a educed heigh wa eguide, as men ioned in he p e ious pa ag aph, p ecludes he use o a s anda d wa eguide ansi ion in o de o eed he pe iodic s uc u e, hus leading o he adop ion o an op imized exci a ion FIG. 1. 共Colo online兲Re lec ion and ansmission coe icien s o a ou spli ing esona o backwa d wa e s uc u e in WR-90 including induc i e windows. Inse s show s anda d heigh and educed heigh wa eguide p o o- ypes. These nume ical esul s om ull-wa e simula ions include me allic and dielec ic losses in he esona o s. 044902-2 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲 sou ce o he pe iodic SRR a ay. Coaxial exci a ion o he de ice was hen selec ed as an imp o ed means o inc easing a ansmi ed signal ac oss an e anescen wa eguide. A his s age, “e anescence” can be p oduced wi h pe iodic induc- i e windows, as shown in he inse s o Fig. 1, o by di ec ly using a wa eguide wi h educed ans e se dimensions 共wid h and heigh 兲, which is he s a ing poin o his subsec- ion. The desc ip ion o he analyzed s uc u e can be ound in Re . 16, as a wa eguide app oach d awn om he plana p oblem s udied in Re . 18. Le us ecall ha he uni cell size in all he pe iodic s uc u es employed in he p esen pape is 4 mm 共in he p opaga ion di ec ion兲, and he guided wa eleng h a 8.5 GHz 共Xband兲is ␭g=55.5 mm in he WR-90. Roughly, he pe iodic s uc u es ha e a ela i e uni cell size o ␭g/14 as compa ed o he s anda d wa eguide case. Figu e 2illus a es he simula ed magne ic ield Hgen- e a ed in he longi udinal plane o a p o o ype s uc u e wi h 3 SRRs. This p o o ype includes a sho -ci cui ed coaxial p obe exci a ion eeding a SRR a ay in an e anescen wa e- guide. The e y s ong magne ic ield gene a ed is op imally o ien ed since i is pa allel o he axis o he SRRs in he a ay plane. This p o ides enhanced coupling be ween he p obe and he esona o s, which causes he educ ion o he e lec ion coe icien , and hence he imp o ed ma ching o he s uc u e. In his con igu a ion di ec p obe- o-p obe coupling may a ec he inal ansmission h ough he de ice. Fo hese pu poses, he s uc u e wi hou he me allic esona o s, al- hough including he dielec ic subs a e, has been simula ed. Resul s 共no shown兲demons a e ha o a p obe- o-p obe dis ance o 7.2 mm 共co esponding o a wo-SRR de ice兲, he ansmi ed signal is no signi ican and is below −50 dB. The e y e anescen wa eguide enclosu e p ecludes he di- ec ansmission in he absence o he SRRs. On op o he in luence o he educed heigh wa eguide, he ma ching p oblem is clea ly dependen on he dis ance om he p obes o he i s o las esona o s o he a ay. This can be no iced in Fig. 3, which shows nume ically com- pu ed Spa ame e s o a es s uc u e wi h ou SRRs. Ma- e ial losses a e no included in hese simula ions. Fo h ee alues o he dis ance om he i s and las esona o o he coaxial p obe 共d→0, d=0.4, and d=1 mm兲, he plo s show ha enhanced ma ching, wi h a educed e- lec ion le el, is achie ed when he p obe is close o he esona o s. This is appa en in he ansmission cu e whe e, o d=1 mm, e y selec i e esonan peaks a e de ined in he ansmission band. S11 shows ha only a hese indi idual equencies is he s uc u e comple ely ma ched. Again, he numbe o hese peaks co esponds o he numbe o esona- o s in he pe iodic s uc u e. As he coaxial p obe is close o he dielec ic subs a e wi h he p in ed SRR acks, and in he limi case bo h a e “ angen ” wi h d→0mm共Ohmic con ac be ween he p obe and he SRR ings, and hence, sho ci cui s a e a oided because o he ci cula shape o he coaxial p obe, see he inse in Fig. 3兲, he selec i i y o hese esonan peaks is educed wi h an inc eased o e all ansmis- sion le el. The e lec ion coe icien also shows an imp o ed beha io wi h educed le el as compa ed o he p e ious cases. Coupling o he ene gy o he de ice is enhanced since he elec omagne ic 共EM兲 ields a e e anescen inside he wa eguide o educed dimensions, and any small sepa a ion would imply a coupling loss. Only when hey a e e y close o he p obes, he i s and las esona o s s ongly in e ac wi h hem. The coaxial p obes locally gene a e high EM FIG. 2. 共Colo online兲Magne ic ield Hgene a ed om he sho -ci cui ed coaxial p obe. No e he ield o ien a ion is aligned wi h he spli ing eso- na o axis. FIG. 3. 共Colo online兲Nume ical esul s o di e en alues o he dis ance om he coaxial p obes o he i s and las esona o s o he a ay 共calcu- la ed ansmission and e lec ion coe icien s o a ou SRR s uc u e兲.In- se s show la e al and op zoomed iews o he design wi h indica ions on he dis ance om he p obe o he i s esona o . 044902-3 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲 ields due o he sho ci cui s wi h he wa eguide enclosu e. The s ong cu en gene a ed locally by he sho ci cui is, hence, con ibu ing o he SRR exci a ion 共which is essen- ially magne ic兲. I is also possible o obse e in Fig. 3 ha he ou o band ejec ion le el is educed wi h a smoo he all o he ansmission cu e a high equencies. Ne e heless, o e all he ansmi ed bandwid h is in p ac ice no depen- den on his pa ame e . The ansmission and e lec ion cha ac e is ics o di e - en leng h s uc u es ha e been simula ed in he case whe e he p obes a e quasi in con ac 共d→0 case兲wi h he i s and las esona o s o he a ay 共see Fig. 4兲. Simula ions do in- clude losses in he conduc o s and he dielec ic subs a e. T ansmission cu es show a ela i ely high le el band cen- e ed close o he esonance equency o he SRRs wi h low losses. These losses sligh ly inc ease wi h he numbe o esona o s bu a e smalle han 4 dB e en o he case o he 6 SRRs s uc u e. An es ima e o he a enua ion cons an e alua ed om he maximum ansmi ed le el in he pass band o he di e en leng h s uc u es gi es an app oxima e alue o ␣ =1.06 dB/cm. This alue clea ly imp o es he pe o mance o he esul s p e iously ob ained by he au ho s in Re . 12, which was a ound ␣ =4 dB/cm, and is compa- able o he esul s o Re . 10, claiming a=0.8 dB/cm. I is impo an o no e ha o ob aining such low alues ex e nal wa eguide une s we e employed in his las wo k o ma ch he e anescen s uc u e. Also, an e anescen wa eguide o a much la ge c oss sec ion 共12 mm⫻12 mm agains 4.1 mm⫻4.1 mm in his case兲was used, hence, his imple- men a ion p o ides mo e compac ness 共ope a ion equencies a e in bo h cases close o 8 GHz兲. A lowe e anescence o he wa eguide in Re . 10 lowe s he a enua ion cons an , since in e esona o coupling is based on p oximi y and e a- nescen ields. The bandwid h o he ansmi ed band e- mains cons an wi h he leng h a ia ions, whe eas ou o band le els dec ease as he de ices a e longe , and hence, a e mo e “selec i e.” F ac ional bandwid h can be es ima ed a ound 5.7%共0.5 GHz bandwid h wi h 8.5 GHz cen al e- quency兲. These esul s a e also in line wi h p e iously pub- lished esul s o de ices based on spli ing esona o s 共in- insically na ow band s uc u es兲, wi h, o example, 5% bandwid h in Re . 10. Main conclusion is, he e o e, ha he good ma ching p o ided by he coaxial p obe eeding pe - mi s a he high ansmission le els e en o pe iodic s uc- u es su icien ly long o p o ide a good selec i i y. C. Induc i e windows The use o induc i e windows o idges in be ween he spli ing esona o s is a means o con olling he ene gy coupling be ween consecu i e esona o s, as i was s a ed in Sec. II A. The e o e, i is possible o include hem in he p o o ype oge he wi h he coaxial eeding p obe. This o e s an ex a deg ee o eedom in o de o de ine, wi hin a ce ain ange o alues, he e lec ion cha ac e is ics o he de ice as well as he bandwid h o he backwa d wa e pass band. The window canno ha e a la ge ape u e han he e anescen wa eguide wid h. In his case, he e would be no window and i he window we e e y igh 共small ape u e兲 he ene gy low would be supp essed. Fo compa ison pu poses, he pe- iodic a ay o SRRs o he p e ious sec ion has been eused bu including in e lea ed me allic windows inside he e a- nescen wa eguide. Fo a wo SRR de ice, wi h one window in e lea ed be ween he esona o s, Fig. 5shows he nume ical esul s in e ms o Spa ame e s and bandwid h as a unc ion o he window ape u e. I is obse able ha he e is an op imum window size ha minimizes e lec ion losses gi ing an im- p o ed ma ching. This op imum alue o he window ape - u e is sligh ly highe han 2.9 mm. Wi h espec o he ans- mission coe icien , S21 has a la cha ac e is ic wi h a high and qui e cons an le el in a la ge ange o window sizes. Fo FIG. 4. 共Colo online兲T ansmission and e lec ion cha ac e is ics o di e - en leng h pe iodic s uc u es. FIG. 5. 共Colo online兲T ansmission and e lec ion coe icien s a he cen al equency o he pass band and −3 dB bandwid h as a unc ion o he in e - lea ed window ape u e wid h o a wo SRR de ice. 044902-4 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲 low alues o he window wid h, he S21 esponse is de- g aded wi h lowe ansmission. In e ms o bandwid h, i is easily deduced ha he bandwid h o he de ice inc eases oge he wi h he ape u e wid h. This bandwid h a ia ion is, hence, associa ed wi h he con ol o he in e esona o cou- plings, jus as in a classical design ule in il e heo y. Figu e 6shows he ansmission and e lec ion cha ac- e is ics o se e al s uc u es wi h di e en numbe o eso- na o s. The me allic idges employed ha e a hickness o 1 mm and a la e al sepa a ion o 2.7 mm; hese a e common alues o all cu es since he analysis is based on ini e-size pe iodic s uc u es. I is possible o obse e ha , by includ- ing he induc i e windows, he wo SRRs cu e shows a e y la ansmission band wi h a educed in-band e lec ion loss. Gene ally speaking o he whole se o cu es, bandwid h is educed wi h espec o he cu es in Fig. 4, and a ce ain con ol o he cha ac e is ics o he pass band can be achie ed h ough he adop ion o a speci ic window ape u e alue. The op imum alues, maximizing ansmission and minimizing e lec ion, o he window ape u es depend on he numbe o esona o s o he s uc u e. Ob aining hese op imum alues would equi e employing, o example, il e design echniques, which is no he pu pose o his wo k. III. EXPERIMENTAL VALIDATION AND DISCUSSION A p o o ype has been ab ica ed in o de o alida e he expec ed pe o mance enhancemen o he p oposed ea- u es, see Fig. 7. The e anescen wa eguide has ans e se dimensions o 4.1 mm⫻4.1 mm, and a segmen ed ap- p oach has been used in o de o es di e en leng h de ices, as in Re . 12. Two segmen s o a ull wid h e anescen wa e- guide and an in e lea ed window sec ion a e isible on he op pic u e o Fig. 7. P o o ypes ha e been ab ica ed includ- ing ei he wo o h ee esona o s in he p opaga ion di ec- ion. Also, wo coaxial p obe ansi ions ha e been imple- men ed s a ing om s anda d coaxial SMA connec o s inse ed in o he wa eguide body e mina ions. These con- nec o s ha e a long cen al conduc o in o de o assu e he con ac o he p obe wi h he op o he e anescen wa e- guide p oducing a sho . An in e lea ed window wi h an ap- e u e wid h o 2.7 mm has been inse ed be ween he eso- na o s, a alue close o he S11 op imum poin o he wo SRR s uc u e. This choice is jus i ied in o de o euse he same window elemen s o measu e a h ee spli ing esona- o line wi h mode a ely good ansmission and e lec ion esponse. All he samples ha e been mechanically milled on a dielec ic subs a e o ype Nel ec NY9220 om Nelco, wi h he p e iously men ioned elec ical and geome ical cha ac e is ics. An HP 8510C Vec o Ne wo k Analyze has been used, wi h a s anda d coaxial h ough- e lec -line 共TRL兲calib a- ion, o measu e he Spa ame e s in he magni ude and phase o he p o o ypes. FIG. 6. 共Colo online兲T ansmission and e lec ion cha ac e is ics o di e - en leng h pe iodic s uc u es including in e lea ed induc i e windows. FIG. 7. 共Colo online兲Segmen ed p o o ype composed o coaxial- o- e anescen wa eguide ansi ions, e anescen wa eguide po ions and squa e spli ing esona o a ays 共 op兲; compac wo SRR p o o ype wi h o al leng h o 12.6 mm 共bo om兲. 044902-5 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲 Figu e 8shows a compa ison o he nume ical and mea- su ed esul s o a wo SRRs de ice in e ms o ansmission and e lec ion coe icien s. Bo h pa ame e s show a good ag eemen be ween he expec ed and he measu ed cha ac e - is ics, especially o he cen al equency and he bandwid h o he pass band. The la e is limi ed due o he in e lea ed window e ec and his ac esul s in he p esence o only one esonance in he S11 cha ac e is ic. Table Isumma izes he nume ical and expe imen al esul s ob ained a he cen- al equency o he pass band. A e y high ansmission has been measu ed, combined wi h a mode a e selec i i y o he pass band and a e y low e lec ion loss, hus alle ia ing he ma ching p oblems o p e ious implemen a ions. Nume ical esul s conside losses in he spli ing esona o s.19 In o de o p o e he le -handed o backwa d wa e na- u e o his pass band, basically ela ed o he spli ing eso- na o beha io and he e anescen shielding, he measu ed phase di e ence be ween wo di e en leng h ansmission lines 共l1and l2, wi h l2⬎l1兲is displayed in Fig. 9. I is pos- sible o analyze he sign o he e ac i e index associa ed wi h a p opaga ing s uc u e wi h espec o he phase di e - ence sign, analyzing i as aken om Re . 20, which assumes he enginee ing con en ion ej ␻ o he a ia ion o he elec- omagne ic wa e ⌬ ␾ = ␾ 2− ␾ 1=− ␤ l2+ ␤ l1 =− ␤ 共l2−l1兲 =− ␻ n共 ␻ 兲 c共l2−l1兲.共1兲 Then, i he phase di e ence is posi i e his would indica e he p esence o a nega i e e ac i e index o he p opaga- ion o he EM wa e. This is he case o he equency ange in Fig. 9, co esponding o he pass band o Fig. 8, i.e., 8.1–9 GHz, which is in u n close o he esonance equency o he esona o s. Abo e his equency ange, o he ejec ed band whe e he e is no ansmission, bo h phase cu es ha e equal alues, hence p oducing a null phase di e ence wi h he con- sequence o p ohibi ed p opaga ion. Abo e app oxima ely 10 GHz, he phase di e ence would u n nega i e, hen indica - ing a posi i e e ac i e index o a p opaga ing band in an uppe equency ange. Figu e 10 shows he measu ed cha ac e is ics 共S11 and S21兲 o a wo SRR de ice including o no he in e lea ed window. In his compa ison, he wid h o he e anescen wa eguide is 4.1 mm and he window employed has an ap- e u e o 2.7 mm. While he a ia ion in he ansmission le el in he pass band is small 共a ound 1.4 dB in he middle o he pass band兲, he in luence o he me allic window is g ea in e ms o he e lec ion coe icien and impo an a ia ions in he S11 cha ac e is ic can be achie ed. Fo in- s ance, since he e is a s ong in e ac ion be ween bo h eso- na o s in he absence o he induc i e window, wo eso- nances a e isible in he S11 cu e. In oducing he in e lea ed window limi s his in e ac ion and de-couples bo h esona o s, gi ing jus one esonance in he S11 cu e. In addi ion, pass band selec i i y is inc eased wi h he use o he in e lea ed window wi h a sha pe all a bo h sides o he pass band. FIG. 8. 共Colo online兲Re lec ion S11 and ansmission S21 coe icien s o a wo SRR p o o ype, compa ison be ween measu ed and nume ical esul s. TABLE I. Spa ame e compa ison o wo and h ee SRR de ices. S uc u e S-pa ame e Type Value 共dB兲 F equency 共GHz兲 2 SRRs S21 Measu ed −1.71 8.48 Simula ed −0.85 8.48 S11 Measu ed −44.65 8.42 Simula ed −18.64 8.49 3 SRRs S21 Measu ed −3.90 8.70 Simula ed −1.62 8.73 S11 Measu ed −17.79 8.73 Simula ed −19.70 8.73 FIG. 9. 共Colo online兲Unw apped ansmission S21 phase measu ed o wo and h ee SRR pe iodic s uc u es including in e lea ed windows. 044902-6 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲 A h ee SRR s uc u e has also been measu ed in e ms o he Spa ame e s, wi h he esul s ob ained shown in Fig. 11. The ag eemen be ween he p edic ed alues o ull-wa e simula ions and he measu ed da a is easonably good. The window size emains cons an wi h espec o he wo SRR p o o ype, and hence, i is no an op imum alue in o de o ma ch he de ice. Ne e heless, o his pe iodic s uc u e s onge selec i i y has been ob ained 共wi h lowe ou -o - band ansmission le els兲and s ill a good ansmission has been measu ed. Resul s a e also included in Table I.I may be use ul o he applica ions ha a low numbe o esona o s allow o a sha p all o he pass-band edges. In con as , and as p edic ed in he simula ions, losses sligh ly inc ease wi h he numbe o esona o s in he lines, bu a e s ill accep able as compa ed o p e ious implemen a ions o wa eguide- based me ama e ial lines. Disc epancy be ween heo e ical and expe imen al alues may be explained because o ab i- ca ion ole ances, since SRRs a e qui e sensible o he me- chanical milling ole ances. IV. CONCLUSIONS In his wo k, he e y high losses associa ed wi h e- cen ly epo ed wa eguide-based me ama e ial ansmission lines a e pa ially o e come wi h he use o a di ec coaxial exci a ion and he use o in e lea ed induc i e windows in- side a backwa d wa e e anescen wa eguide loaded wi h SRRs. These wo ea u es imp o e he ansmission, e lec- ion, and bandwid h cha ac e is ics o p e iously published p o o ypes and p o ide a means o con olling he basic cha - ac e is ics o backwa d wa e p opaga ion in he pe iodic s uc u es. Measu ed ansmi ed le els may allow he u u e use o his echnology owa d he applica ions. Also, his s udy opens he possibili y o use s anda d syn hesis ech- niques in o de o mo e om pe iodic s uc u es o egula il e designs, based on backwa d wa e p opaga ion. The main ad an age o his ype o po en ial echnology is he e y compac size o he analyzed p o o ypes. ACKNOWLEDGMENTS This wo k was suppo ed by he Minis e io de Educación y Ciencia, Spanish Go e nmen , unde a Ramón y Cajal Con ac and unde he Resea ch P ojec s TEC-2004-04313- C02-01 and TEC-2004-04249.C02-02 and by Gene ali a Va- lenciana unde p ojec s ACOMP06/013 and GVA/2007/215. The au ho s a e indeb ed o Be na Be na do o he ab ica- ion o he SRR samples, and o Jose Vicen e Cas elló o he p epa a ion o he SMA connec o s o he ansi ions. 1V. G. Veselago, So . Phys. Usp. 10,509共1968兲. 2R. A. Shelby, D. R. Smi h, and S. Schul z, Science 292,77共2001兲. 3A. G bic and G. V. Ele he iades, IEEE T ans. Mic ow. Theo y Tech. 51, 2297 共2003兲. 4J. B. Pend y, Phys. Re . Le . 85, 3966 共2000兲. 5A. Alu and N. Enghe a, Phys. Re . E 72, 016623 共2005兲. 6D. Schu ig, J. J. Mock, B. J. Jus ice, S. A. Cumme , J. B. Pend y, A. F. S a , and D. R. Smi h, Science 314, 977 共2006兲. 7J. Ga cia-Ga cia, J. Bonache, I. Gil, F. Ma in, Md. C. Velazquez- Ahumada, and J. Ma el, IEEE T ans. Mic ow. Theo y Tech. 54,2628 共2006兲. FIG. 10. 共Colo online兲Measu ed ansmission 共 op兲and e lec ion 共bo om兲 coe icien s o a wo SRR s uc u e including o no an in e lea ed induc- i e window. FIG. 11. 共Colo online兲Re lec ion S11 and ansmission S21 coe icien s o a h ee SRR p o o ype, compa ison be ween measu ed and nume ical esul s. 044902-7 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲 8R. Ma qués, J. Ma el, F. Mesa, and F. Medina, Phys. Re . Le . 89, 183901 共2002兲. 9J. B. Pend y, A. J. Holden, D. J. Robbins, and W. J. S ewa , IEEE T ans. Mic ow. Theo y Tech. 47, 2075 共1999兲. 10S. H aba , J. Ba olic, and Z. Sipus, IEEE T ans. An ennas P opag. 53,110 共2005兲. 11I. A. Esh ah, A. Ahmed Kishk, A. B. Yako le , and A. W. Glisson, IEEE T ans. Mic ow. Theo y Tech. 53, 3298 共2005兲. 12J. Ca bonell, L. J. Roglá, V. E. Bo ia, and D. Lippens, IEEE T ans. Mi- c ow. Theo y Tech. 54,1527共2006兲. 13C. A. Fe nandes, R. Ma qués, and M. Sil ei inha P oc. Eu opean Mic ow. Associa ion 2,66共2006兲. 14R. Ma qués, J. Ma el, F. Mesa, and F. Medina, Mic owa e Op . Technol. Le . 35, 405 共2002兲. 15T. Decoopman, O. Vanbésien, and D. Lippens, IEEE Mic ow. Wi el. Com- pon. Le . 14, 507 共2004兲. 16J. Ca bonell, L. J. Roglá, V. E. Bo ia, and R. Ma qués, P oceedings o he Eu opean Mic owa e Con e ence 共IEEE, New Yo k, 2006兲. 17R. Ma qués, F. Medina, and R. Ra ii-El-Id issi, Phys. Re . B 65, 144440 共2002兲. 18J. Ca bonell, L. J. Roglá, and V. E. Bo ia P oc. Eu opean Mic ow. Assoc. 2,52共2006兲. 19M. J. F ei e, R. Ma qués, F. Medina, M. A. G. Laso, and F. Ma in, Appl. Phys. Le . 85,4439共2004兲. 20O. F. Siddiqui, M. Mojahedi, and G. V. Ele he iades, IEEE T ans. An en- nas P opag. 51,2619共2003兲. 044902-8 Ca bonell e al. J. Appl. Phys. 102, 044902 共2007兲