High-Q Optical Micro-cavity Resonators as High Sensitive Bio-chemical and Ultrasonic Sensors

Transcription

1 High-Q Optical Micro-cavity esoators as High Sesitive Bio-cheical ad Ultrasoic Sesors By Tao Lig A dissertatio subitted i partial fulfillet of the requireets for the degree of Doctor of Philosophy Electrical Egieerig i The Uiversity of Michiga Doctoral Coittee: Professor L. Jay Guo, Chair Associate Professor Xudog Fa Assistat Professor Tal Caro Assistat Professor Pei-Cheg Ku

2 Tao Lig All rights reserved

3 To My Faily ii

4 Acowledgeets First ad foreost I would lie to express y deepest gratitude to y advisor, Professor L. Jay Guo, for his supports ad guidace throughout y Ph.D. study, ad also the freedo that he gave to e durig y 5 years research life i the Uiversity of Michiga. His scietific ituitio ad experieced advice are crucial to overcoe probles i both theoretical ad experietal aspects. I have truly leared a lot of fro hi. I would also lie to tha y doctoral coittee ebers, Professor Xudog Fa, Professor Tal Caro ad Professor P.C Ku, for their useful coet ad advice to the wor i y thesis. Soe part of the bio-cheical sesig experiet is perfored with the help of Dr. Sheeree Majd ad Professor Michel Mayer i Bioedical Egieerig Departet. Also I will lie to say thas to Professor Xuedig Wag ad Dr. Zhixig Xie for useful discussio ad share the owledge of the photo-acoustic field. I would lie to tha all Guo group forer ad curret ebers. Forer ebers: Dr. Chug-Ye Chao traied e o the optical easureet before I joied Guo group. Dr. Jigsug Ki served as y etor i the clearoo durig y first year i Uiversity of Michiga ad trasferred useful fabricatio owledge to e. Dr. Dawe Li helped a lot i y first year s life i UM. Dr. Li-Jig Cheg, Dr. Myuggyu Kag, Dr. Pra Muherjee helped with all ids of fabricatio process probles related to dry etch, ao-iprit ad furace process. Dr. Philip Choi helped to solve various probles o iii

5 Naoex achie, Dr. Carlos Pia helped to prepare SSQ aterials for ipritig applicatios, Dr. Se Hyu A helped with soe flexible old preparatios ad soe useful discussio with Dr. Yi-Hao Che ad Dr. Tig Xu. All ids of helps fro curret ebers: Dr. Haofei Shi, Dr. Moo Kyu Kwa, Dr. Ji zhou, Yiuei Wu, Xig Tu, Hyoug Wo Baac, Sug-Liag Che, Alex Kapla, Hyusoo Ki, Brado Lucas, Jog G. O, Hui Joo Par, Ashwi Paday, Cheg Zhag, Peg Zhu, JaeYog Lee, Kyu-Tae Lee. Out of Guo group, I would lie to say this to all y frieds ad classates i the UM, Weiig Wag, Chea Xia, Meg Zhag, Guag Huag, Professor Wei Guo, Xiuqua Ma, Ju Yag, Professor Zetia Mi, etc. Also thas to all SSEL staff for helpig o various fabricatio probles. Fially, I would lie to give y sicere appreciatio to y parets, y wife ad y elder sister, without their cotiuous support, I ca ot go that far. iv

6 Table of Cotets Dedicatio...ii Acowledgeets... iii List of Figures... viii Abstract... xiii Chapter Itroductio.... Optical icrocavity based bio-cheical sesors.... Optical icrocavity based ultrasoic sesors...4. Dissertatio overview...5 Chapter Optical Microcavity esoators...8. Itroductio...8. Cavity paraeters..... Q factor..... Free spectru rage..... Fiesse F.... Optical ode i icrotube ad icrorig resoators....4 Mode field distributio....5 Loss i optical icro-cavity Surface scatterig loss adiatio loss Absorptio loss...8 Chapter Pris coupled icro-tubes as sesitive bio-cheical sesors.... Itroductio ad otivatio.... Experietal setup.... Bul refractive idex sesig..... Bul refractive idex sesig experiets... v

7 .. Siulatios of bul refractive idex sesig Discussios.... Surface sesig experiet..... esoat ode characterizatios..... Lipid oolayer detectio Lipid bi-layer detectio Self asseble layer by layer detectio Quatificatio of bio-fil layer bidig Study iteractio betwee the lipid ebrae ad proteis Coclusio...46 Chapter 4 Sesig properties of icro-tube resoator sesors Itroductio ad otivatio Field distributio i icro-tube resoator Bul refractive idex sesig sesitivity Surface sesig sesitivity Absorptio sesig sesitivity Sesitivity ehaceet usig a coupled cavity Coclusio...64 Chapter 5 High sesitive ultrasoic detectio usig polyer icrorigs Itroductio ad otivatio Fabricatio process of polyer icrorigs Noral old fabricatio process Siplified old fabricatio process Spectru ad SEM characterizatio Acoustic sesitivity Frequecy respose Device s perforace Iproveet Fabricatio process to iprove Q factor Loss characterizatio Acoustic sesitivity Coclusio...87 vi

8 Chapter 6 Ultra-high sesitive ultrasoic detectors usig ultra-high Q icrorigs Itroductio ad otivatio Devices siulatio ad fabricatio Acoustic sesitivity Agular respose Saller size polyer icro-rigs for high frequecy iagig Photoacoustic icroscopy usig polyer icrorigs Coclusio... Chapter 7 Coclusio Achieveet i high sesitivity bio-cheical sesors Achieveet i high sesitivity ultrasoic sesors Future wor Ultra-sall device Ultra-high Q factor device Acoustic detectors o fiber tip Micro-rig arrays for iagig applicatios... 9 Bibliography... vii

9 List of Figures Figure. Various optical icro-cavity biocheical sesors: a icrosphere[], b icrorig [8], c icrodis [], icro-tube [5]... Figure. a Fiber tip Fabry-Perot F-P cavity ultrasoic hydrophoe b O-chip polyer icrorig ultrasoic hydrophoe...5 Figure. a Part of St. Paul s Cathedral i Lodo, b ay optics poit of view of whisperig gallery ode...8 Figure. Scheatic of pris coupled icro-tube sesor syste... Figure. esoace curve shift due to the chage of liquid refractive idex i the icro-tube at icidet agle of 7.5 o... Figure. esoace wavelegth as a fuctio of the chage i liquid refractive idex i the icro-tube at a icidet agle of 7.5 o...4 Figure.4 a esoace curve shift due to the chage of liquid refractive idex i the icro-tube, b resoace wavelegth shift as a fuctio of the chage i liquid refractive idex at a icidet agle of ~5 degree....5 Figure.5 a adial electrical field itesity distributio for the resoace ode. b esoace wavelegth shift related to liquid s refractive idex chage for the resoace ode...8 Figure.6 a adial electrical field itesity distributio for resoace ode b esoace wavelegth shift related to liquid s refractive idex chage for resoace ode... Figure.7 The scheatic ray picture of resoace ode with light trasitted ito the ier boudary.... Figure.8 esoace cure shift due to chage of liquid idex i the icro-tube...4 Figure.9 adial electrical field itesity distributio for the resoace ode 6 E r viii

10 Figure. esoace curve shift due to the lipid oolayer bidig to the ier wall of icro-tube Figure. esoace curve shift due to lipid ebrae bidig to ier surface of icro-tube...7 Figure. esoace curve shift due to the electrostatic self assebly thi fil coatig at the ier wall of the silica icro-tube....9 Figure. Siulated resoace curve shift with the lipid ebrae thicess chage...4 Figure.4 esoace curve shift due to Aexi V bidig to POPC lipid bi-layer...4 Figure.5 esoace curve shift due to the atibiotic peptide alaethici iteract with lipid bilayer o the ier surface of the icro-tube...44 Figure.6 esoace wavelegth shift related to differet cocetratio of alaethici ijectio Figure 4. The oralized electric field itesity distributios with aziuthal uber M=7 ad radial order uber N=5 ad 7 deoted by E ad 7 E 7, respectively...48 Figure 4. Siulated bul refractive idex sesig sesitivity of differet radial order odes with the sae aziuthal uber M=7 by perturbatio ethod ad Mie scatterig ethod...5 Figure 4. Siulated surface sesig sesitivity of differet radial order odes with the sae aziuthal uber M=7 by perturbatio ethod ad Mie scatterig ethod. The bio-fil s refractive idex is assued to be Figure 4.4 Siulated absorptio sesig sesitivity of differet radial order odes with sae aziuthal uber M=7 by perturbatio ethod ad Mie scatterig ethod. The absorptio coefficiet of liquid is assued to be =.7c Figure 4.5 a a scheatic of ier coated silica icro-tube sesor, the cavity ca be decoposed ito two cavities: b a silica icro-tube cavity filled with low idex aterials ad c a water cylider cavity covered with low idex aterials...6 Figure 4.6 a Aticrossig behavior of the resoat wavelegths whe the refractive idex of liquid is aroud.5, b the Quality factor shows crossig behavior whe the refractive idex of liquid is aroud ix

11 Figure 4.7 Calculated field distributios i the cavity with differece refractive idex of liquid,, liquid refractive idex aroud.,,v liquid refractive idex aroud., V, V liquid refractive idex aroud.4..6 Figure 4.8 a sesitivity of two hybrid resoace ode chages with liquid refractive idex.b the sesitivity chages with the ier coated layer aterial refractive idex.with fixed thicess d=...64 Figure 5. Scheatic of fabricatio of deep old fro the shallow old. a A shallow old fabricated fro EBL ad IE. b The shallow old is cotacted oto the saple with 5K PMMA layer. c The ipritig process is perfored at high pressure ad high teperature. d The saple is separated fro the old ad the PMMA patters are created o the substrate. e PMMA residual layer is etched away by O plasa. f Metal as Ti/Ni is deposited ad lift-off process is perfored to trasfer the patter to etal fil. g The oxide layer is etched usig etal as. h The etal layer is reoved by the etal etchat Figure 5. Scheatic of siplified silico oxide old fabricatio process. a E-bea lithography o the 95 PMM layer o the silico substrate with silico oxide ad 5 Cr fil. b The saple is developed i MIBK:IPA=:. c The Cr layer is etched by usig PMMA as a etchig as, the the PMMA layer is reoved by hot acetoe. d The silico oxide layer is etched usig Cr as a etchig as, the reove the Cr layer by usig Cr etchat Figure 5. Trasissio spectru of the polyer icrorig...7 Figure 5.4 a SEM iage of the polyer icro-rig with =5, b the sidewall view of the polyer icrorig c SEM iage of the polyer icro-rig with soe holes o the top...7 Figure 5.5 The experietal setup to easure the oise-equivalet pressure ad sesitivity of a polyer icrorig resoator. the distace betwee the ultrasoud trasducer ad the resoator was Figure 5.6 a Optical trasissio spectru of a polyer icrorig resoator. The iput power was 4. W. b Sigle-shot acoustic wavefor easured by the resoator. The positive pea correspods to Pa. The optical probig wavelegth ad iput power were set to ad 5.5 W, respectively....7 Figure 5.7 a Acoustic sigal detected by a polyer icrorig resoator ad laser pulse profile detected by the photodetector. b Spectra of the sigals ad x

12 frequecy respose of the resoator. The detectio badwidth of the resoator was over 9 MHz at db Figure 5.8 Scheatic of typical silico old fabricatio process...76 Figure 5.9 Sidewall SEM iage of the polyer icro-rig fabricated fro the old: a without resist reflow process, b with resist reflow process, c with resist reflow ad theral oxidatio process Figure 5., High agificatio SEM picture of o the sidewall roughess...78 Figure 5..Trasissio spectru of polyer icro-rig fabricated fro the old a without resist reflow process, b with resist reflow process, c with resist reflow ad theral oxidatio process. All the blac dot curves are experietal data ad red lie curves are Loretz fittig cure...8 Figure 5..Trasissio spectra of polyer icro-rigs with differet iput power...8 Figure 5..Trasissio spectra of polyer icro-rigs with differet iput power...84 Figure 5.4 Trasissio spectru of polyer icro-rig ierged i DI water. b Sigle shot of acoustic wavefor easured by high Q polyer icro-rig..86 Figure 6. Siulated E filed itesity distributio of = polyer icro-rig at resoace wavelegth aroud 8 with botto ad top claddig are silica ad water....9 Figure 6. Sidewall view of the silico old with ew recipe...9 Figure 6. a SEM iage of the polyer icrorigs with =, b SEM iage of the sidewall of polyer icro-rig...9 Figure 6.4 The trasissio spectru of polyer icrorig with =...9 Figure 6.5 a Trasissio spectru of polyer icro-rig ierged i DI water. bsigle shot of acoustic wavefor easured by high Q polyer icro-rig...94 Figure 6.6 a Experietal setup for easure the agular respose of the polyer icro-rig, experietal data dot ad theoretical calculatio lie of agular respose of the polyer icro-rigs with D = 6 at MHz solid dot ad lie ad MHz epty dot ad dash lie b, ad with D = 4 at MHz solid dot ad lie ad 4 MHz epty dot ad dash lie c...96 xi

13 Figure 6.7 a The trasissio spectru of polyer icrorigs with = coarse sca, b the trasissio spectru of polyer icrorigs with = fie sca Figure 6.8 The trasissio spectru of polyer icrorigs with = coarse sca o 4 theral oxide wafer...99 Figure 6.9 The trasissio spectru of polyer icrorigs with = o substrate with 4 thic HSQ fil o 4 theral oxide wafer.... Figure 6. Siulated field distributio of =u polyer icrorig o theral oxide substrate i the air a ad i the water b... Figure 6. Siulated the field distributio of i water ad air of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 low idex buffer layer... Figure 6. Trasissio spectru of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 low idex buffer layer i the air a ad i the water b.... Figure 6. Trasissio spectru of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 low idex buffer layer i the water...4 Figure 6.4 left pael scheatic of a icrorig based PAM syste based o a icrorig resoator. iddle pael axiu aplitude projectio MAP iage of the USAF resolutio teplate group 7. right paels A-lie sigals alog the Z axis of the iages of the USAF resolutio teplate with a icrorig based PAM right upper pael ad covetioal PAM with Oda trasducer right lower pael. The iset at upper left shows a scaig electro icrograph of a polystyree icrorig with radius used i this experiet....6 Figure 6.5 MAPs o XY,XZ, YZ plaes of the ex vivo iages of the vasculature i a ouse bladder wall acquired with AOPAM upper row usig icrorig ad covetioal PAM usig Oda trasducer lower row....8 Figure 7. shows a silico slot waveguide hybrid with polyer aterial with radius =5...8 xii

14 Abstract High-Q Optical Micro-cavity esoators as High Sesitive Bio-cheical ad Ultrasoic Sesors by Tao Lig Chair: L. Jay Guo Optical icro-cavity resoators have quicly eerged i the past few years as a ew sesig platfor i a wide rage of applicatios, such as bio-cheical olecular detectio, eviroetal oitorig, acoustic ad electroagetic waves detectio. I this thesis, we will aily focus o developig high sesitivity silica icro-tube resoator bio-cheical sesors ad high sesitivity polyer icro-rig resoator acoustic sesors. I high sesitivity silica icro-tube resoator bio-cheical sesors part: We first deostrated a pris coupled silica icro-tube bio-cheical sesig platfor to overcoe the reliability proble i a fiber coupled thi wall silica icro-tube sesig platfor. I refractive idex sesig experiet, a uique resoace ode with sesitivity aroud 6/refractive idex uit IU has bee observed. Surface sesig experiets also have bee perfored i this platfor to detect lipid oolayer, lipid bilayer, electrostatic self asseble layer-by-layer as well as the iteractio betwee the lipid bilayer ad proteis. The a theoretical study o various sesig properties o the xiii

15 silica icro-tube based sesig platfor has bee realized. Furtherore, we have proposed a coupled cavity syste to further ehace the device s sesitivity above /IU. I high sesitivity polyer icro-rig resoator acoustic sesors part: We first preseted a siplified fabricatio process ad realized a polyer icrorig with a Q factor aroud 6. The fabricated device has bee used to detect acoustic wave with oise equivalet pressure NEP aroud Pa over -75MHz frequecy rag, which is coparable to state-of-art piezoelectric trasducer ad the device s frequecy respose also have bee characterized to be up to 9MHz. A ew fabricatio process cobied with resist reflow ad theral oxidatio process has bee used to iprove the Q factor up to 5 ad the device s NEP has bee tested to be aroud 88Pa over -75MHz rage. Further iprovig the device s Q factor has bee realized by shiftig the device s worig wavelegth to ear-visible wavelegth ad further reducig the device s sidewall roughess. A record ew high Q~4x 5 has bee easured ad the device s NEP as low as Pa has bee easured. Furtherore, a saller size polyer icrorig device has bee developed ad fabricated to realize larger agle bea forig applicatios. xiv

16 Chapter Itroductio Optical icro-cavity resoators have bee extesively studied fro later of cetury to ow i ay applicatios. I 969, Marcatili first proposed icro-rig resoators as itegrated optical wavelegth filters []. Microsphere [] ad icro-droplet [] resoators have bee ivestigated fro early of 98 s to realize lasig ad oliear optical process. Util 99s, chip based icro-cavity resoators have bee fabricated with the ature of seicoductor techology [4]. Those icro-cavity resoators have show several advatage properties, such as high Q factor, arrow resoace lie-width, copact size ad strog optical filed ehaceet iside cavities. With these uique characteristics, icro-cavity resoators have bee deostrated lots of applicatios i both optical passive ad active devices. Micro-cavity resoator based optical passive devices icluded: sigle resoator/high-order resoators filters [5,6], add/drop filters i wavelegth divisio ultiplexig WDM syste [7,8], tuable filters [9,], bio-cheical sesors[-5], echaical sesors [6] ad pressure sesors[7,8]. Micro-cavity resoator based optical active devices icluded: cotiuous wavelegth lasers [9,], aa lasers[,], optical switcher/odulators[-5], optical oliear oscillators[6,7,8] ad cavity based radiatio pressure effects [9] I this itroductio, we will aily itroduce the optical icro-cavity based bio-cheical sesors ad ultrasoic sesors.

17 . Optical icrocavity based bio-cheical sesors ecetly, icro-cavity resoator based bio-cheical sesors have bee raised a lot of attetios because of their uique properties which ca reduce the device size by order of agitude, without sacrificig the iteractio legth by virtue of the high quality Q factor resoaces, thereby sigificatly reducig the aout of saple eeded for the detectio. The resoace effect also provides a effective log iteractio legth for the sesor to achieve sufficiet sesitivity. Also typical bio-sesig experiet requires that the devices ca hadle aqueous aalytes. Therefore fluidic hadlig capability is a idispesable part of the sesor platfor. High quality-factor icrosphere cavities usig Whisperig Gallery Mode WGM resoaces have bee deostrated to respod to a oolayer of protei adsorptio [], however itegratio with fluidic syste is very challegig ad typically requires fluidic chabers uch larger tha the active device eleet. Though icro-rig [,] ad icro-dis [] sesors ca be ass fabricated usig batch processig techiques, they suffer fro liited Q factors due to the surface roughess iduced scatterig loss.

18 a b c d Figure. Various optical icro-cavity biocheical sesors: a icrosphere [], b icrorig [8], c icrodis [], icro-tube [5] The silica icro-tube based resoator sesor is very attractive for biocheical sesig applicatio due to its ability to hadle aqueous aalytes ad its high Q factor []. The high-q ~ 6 is a result of the low scatterig loss due to very sooth surfaces ad log virtual iteractio legth. However there are also liitatios to the icro-tube resoators. Sice oly the ier surface ca be used as the sesor iterface, so the total evaescece optical field iteractig with the solutios is less tha half that of icro-rig [] resoators if the two devices have siilar field distributio for the guided ad evaescet waves, thereby resultig i a lower sesitivity. Moreover, the previously deostrated icro-tube resoator sesors required a tube thicess of less tha 5 icros [5] i order to obtai a refractive idex sesitivity of.6/iu. Sigificatly higher idex sesitivity ca be obtaied by further reducig the wall thicess to below sub-icro []. But such strategies cause the icro-tube to be very brittle ad difficult

19 to hadle i practical sesig applicatios. I this thesis, we will develop a ew strategy to solve this proble to realize reliable ad high sesitivity bio-cheical detectio.. Optical icrocavity based ultrasoic sesors ecetly, a optical cavity based ultrasoud detectio platfor has attracted icreasig attetios [4-7]. Copared with covetioal piezoelectric trasducers, the ew detectio platfor provides several advatages, such as preservig high sesitivity with reduced eleet sizes, high-frequecy ad widebad respose with siple fabricatio. I this ew optical cavity based detectio platfor, optically trasparet polyer aterial was used because of its high optical elastic coefficiet ad high deforability, which ca provide sesitive respose uder acoustic pressure. A polyer based fiber tip Fabry-Perot F-P cavity ultrasoic hydrophoe has bee deostrated with coparable sesitivity ad oise equivalet pressure NEP to curret piezoelectric PVDF ultrasoic sesig devices [6]. A iproved sesitivity has bee realized i a polyer based plae F-P device by icreasig the device s Q factor, which results fro a icreased reflectio of the ulti-layer stacs based -diesio photoic crystal irror [7]. Further iproveet of the sesitivity ca be achieved by further icreasig the reflectivity of the -D photoic crystal irror, but requires a sophisticated syste to precisely cotrol the thicess ad uifority of the deposited ultilayer fil. Except the photoic crystal based light cofiig echais, the total iteral reflectio TI echais has bee widely used to cofie the light i the cavity to achieve uch higher Q factor. High Q resoators usig the TI echais iclude icro-spheres [8,], 4

20 icro-diss [9,48], icro-rigs [4,4,4] ad icro-tubes [5,44]. By cobiig polyer aterial s high optical elastic coefficiet ad high deforability with the TI-based high Q factor icro-resoators, it is possible to deostrate high sesitivity ultrasouic detectors. a b Figure. a Fiber tip Fabry-Perot F-P cavity ultrasoic hydrophoe b O-chip polyer icrorig ultrasoic hydrophoe. Dissertatio overview Chapter provides a brief itroductio of optical icro-cavities ad aily focus o describig optical icro-cavity based bio-cheical sesor ad ultrasoic sesors. Chapter first presets the basic paraeters, such as Q factor, free spectru rage FS ad fieess. The the theory of optical icro-cavities is itroduced i cylidrical coordiator. Based o this theory, the field distributio i the icrorig ad icrotube are provided ad at last the optical loss is studied. 5

21 Chapter describes to use a pris couplig ethod to realize a reliable ad high sesitivity icro-tube sesig syste. I Bul sesig experiet, a uique high sesitivity 6/IU has bee observed ad cofired by siulatio. I surface sesig experiet, the detectio of lipid oolayer, lipid bilayer, self asseble layer-by-layer have bee realized, also the iteractio betwee the lipid ebrae ad proteis has bee studied. Chapter 4 shows various sesig echaiss of icro-tube sesors, icludes bul refractive idex sesig, surface sesig, absorptio sesig. Siulatio results of those sesig properties have bee well preseted. Fially a ew coupled cavity syste is itroduced to realize a ultrahigh sesitivity bio-cheical sesig platfor. Chapter 5 itroduces the ao-ipritig techique to fabricate chip-based polyer icro-rig as ultrasoic sesig eleets. First a siplified fabricatio process has bee created to fabricatio a polyer icro-rig with siilar Q factor as pervious reported. Secod, a iproved fabricated ethod has bee preseted to iprove the device s perforace by iprovig the sidewall roughess usig resist reflow ad theral oxidatio ethod. A low oise ad widebad ultrasoic detectio has bee deostrated by usig our polyer icro-rig devices. Chapter 6 presets a ew desig of polyer icro-rig worig at visible wavelegth rage to reduce aterial ad water absorptio loss. A ew process has bee created to further reduce the device s side wall roughess to realize a polyer icrorig with Q factor as high as 5x 5. Also the device s sesitivity has bee greatly iproved with the device s extreely high Q factor. Saller size polyer icro-rigs also have bee fabricated to realize high frequecy bea forig applicatio. 6

22 to be doe Chapter 7 presets all of the wo doe so far ad discusses the future experiets 7

23 Chapter Optical icrocavity resoators. Itroductio a b Figure. a Part of St. Paul s Cathedral i Lodo, b ay optics poit of view of whisperig gallery ode Circular optical icro-cavity resoators have bee studied over 4 years ad the eige odes of the circular cavities are called whisperig gallery odes. Acoustical whisperig gallery odes have bee studied fro early cetury by Lord ayleigh [45], i order to explai the curious property of the whisperig gallery at the doe of St Paul s Cathedral i Lodo Figure.a. This curious property ca be addressed lie that if two persos stad at opposite sides of the gallery, ad oe whispers aroud the wall of the doe, ad the whispers ca be heard clearly by aother perso. This effect does ot 8

24 wor if oe speas at ceter of the doe or orally to the edge of the doe. The aswer to this strage effect is that the soud bouces alog the curved wall of the gallery with very little loss, ad so it ca be heard at a very log distace. The reaso that it does ot wor if you spea at ceter or orally to the edge of the doe is that the icreased aplitude of the oise allows it to circulate the gallery ultiple ties, ad so the souds all get ixed up ad ca o loger be heard properly. It ca be viewed that there is a arrow regio ear the edge of the doe where the soud waves propagate with very low loss, ad this is ow as a whisperig gallery ode [45] i hoor of gallery where it was discovered. This pheoeo which was first observed by soud wave boudig aroud the edge of the wall, while sae picture ca also applied to the optical wave by assuig that the iside ad the outside of gallery regio are filled with high refractive idex aterial, low refractive idex optical trasparet aterial, respectively. By choosig properly excitatio ethod, the optical wave ca be coupled, geerated, cofied at the edge of the high refractive idex aterial by the total iteral reflectio. Figure.b shows the ray optical poit of view of whisperig gallery ode. I recetly, this pheoeo has bee foud i optical wave regio i icro circular cavity, such as icro-dis, icro-rig, icrosphere ad icro-tube. 9

25 . Cavity paraeters.. Q factor The ost coo ad iportat paraeter of the cavity is the quality factor Q factor, the defiitio of the quality factor for ay resoat eleets ca be siply fro oe equatio [46, 47]: Q U P Loss cavity - Where the is the resoat frequecy, U cavity is the store eergy i the resoator ad P Loss is the power dissipated by the cavity. Also we assue there is o source i the cavity. The P Loss ca be writte as P loss =-du c /dt by the defiitio. So equatio - ca be chage to: du c dt Q U c, ad the solutio to this tie-depedet differetial equatio is: ot / Q t / Q / U c t U c e U c e. So we ca defie the photo life tie i the cavity as: Q ph - By usig Fourier trasfer the cavity Q factor ca be easured through the spectru rather tha by usig rig dow tie: Q -

26 Where the is the resoat frequecy of the cavity, ad axiu of the resoat pea. is the full width half.. Free spectru rage The free spectru rage is defied as the frequecy spacig or wavelegth spacig betwee two adjacet odes. If we defie the two adjacet odes with resoat wavelegth at ad, the the Free spectru rage FS: FS= - L eff eff L g -4 Where eff is the effective idex of the resoat ode ad L is the cavity legth... Fiesse F [47]: The fiesse of a resoator is defied as the ratio of the FS to the resoat lie-width FS F -5 for a aziuthally syetric resoace ode, the F ca be calculated as F=Q/, where is the aziuthal uber of the resoace ode. A physical explaatio of fiesse is the roud trip uber of the photo i the cavity.

27 . Optical ode i icrotube ad icrorig resoators Optical odes i a dielectric icro-rig ca be obtaied by solvig the Maxwell equatio i cylidrical coordiator. Give the Maxwell equatio with o source: t t r B t r E,,, t r D t t r D t r H,,, t r B -6 Where,, t r E t r D r o ad,, t r H t r B are the liear costitutive relatios. By cosiderig that all filed H B D E,,, have the siusoidal tie depedet ter exp t i ad tie idepedet Maxwell equatio ca be expressed as followig: r B i r E r D r D i r H r B -7 For icro-rig case, we ca get the Helholtz equatio fro Maxwell equatio ad liear costitutive relatio:,, F z r F -8 where the F ca be either H or E. The detail of Helholtz equatio i cylidrical coordiates ca be writte as followig: ],, [ F z r z r r r r -9 For TE ode, F ca be H z ad for TM ode the F ca be E z. After separate variable we ca chage the equatio to differet two equatios:

28 , X r X r r X r r X eff - ] [ Z z dz Z d eff - Where F z =, z Z r X. Equatio - also ca be solved by usig separatio of variable ethod by writig:, r r X, ad the the equatio - ca be chage to two followig equatios: [ r r r r dr d eff - d d - The solutio to equatio - ca be siply writte as: e i, where is the a iteger uber, ad the equatio - s solutio is Bessel fuctio. For iteger uber Bessel equatio s solutio should be cobiatio of first order Bessel fuctio J ad secod order Bessel fuctio N. But whe r=, the secod order Bessel fuctio go to ifiity, so iside the icro-rig regio the solutio just ca be first order Bessel fuctio, ad outsider use cobiatio of first order ad secod order Bessel fuctio..4 Mode field distributio For siplicity, we oly cosider TM ode, The E z copoet ca be expressed as: r r r t i r B H t i r B N r B J t i r B J E z 4 ] exp[ ] ]exp[ ] exp[ -4

29 4 where is the aziuthal order uber. At the boudary, E z field eeds to satisfy the boudary coditio: r z r z z z r E r E E E -5 This leads to two equatios: B N B J J B -6 B N B J J B -7 At the boudary, E z field also eeds to satisfy the boudary coditio: r z r z z z r E r E E E -8 The we get two ore equatios ] [ ] [ ] [ 4 d B N d B J d H B -9 ] [ ] [ ] [ 4 d N B d J B d H B - Dividig Eq. -6 by Eq.-7 ad Eq. -9 by - we get: / / N J B B N J B B J J - / / N J B B N J B B H H - For equatio 9 we ca solve for B /B :

30 5 J H J H N H N H B B - Equatio 8 ad equatio 9 or will be used to fid the resoace wavelegth with differet aziuthal ubers i the cavity. We defie: C B B, the B =CB A B J N CJ B B 4 A B H N CJ B B Fially the wave equatio ca be writte as: r r r t i r H A B t i r N r CJ B t i r J A B E z ]} exp[ { ]} ]exp[ {[ ]} exp[ { -4 Where: J N CJ A, H N CJ A, ad the paraeter B ca be obtaied fro the oralizatio of the E field..5 Loss i optical icro-cavity I optical icrocavity, due to presece of loss echaiss, such as radiatio loss cofieet loss, aterial absorptio loss ad surface scatterig loss, eergy will be dissipated outside of the cavity. The extet to describe how good the resoat syste to

31 store eergy is cooly expressed by the Quality-factor, which is defied by the eergy storage tie oralized with respect to the period of oscillatio U P Q cavity Loss Where the is the resoat frequecy, U cavity is the store eergy i the resoator ad P Loss is the power dissipated by the cavity. The over all Q factor ca be describe as followig expressio: Q Q Q Q Q -5 total absorptio. radiatio scatterig couplig Where Q, Q absorptio, Q radiatio, Q scatterig, Qcouplig total are total Q, absorptio loss related Q, radiatio loss related Q, scatterig loss related Q ad couplig loss related Q,respectively. I here we will aily focus o discussig surface scatterig loss ad aterial absorptio loss..5. Surface scatterig loss I optical waveguides, whe the light propagate alog the waveguide, rado boudary iperfectios or surface roughess will scatter the eergy out of the waveguide to for the scatterig loss. The scatterig loss is particularly iportat whe the waveguide is bedig, because ceter of optical filed is shiftig away fro the ceter of the waveguide ad which will icrease the field itesity at the boudary of the waveguide where surface iperfectio is located. Thus, the bedig waveguide will suffer higher surface scatterig loss copared to the straight waveguide with sae diesio. The roughess at sidewall is orally geerated fro photo-lithography step, 6

32 etal lift-off step or dry etchig step durig the device fabricatio ad it shows a rado fashio. Differet approaches have bee developed to estiate the scatterig loss i optical waveguide, such as a volue curret ethod [48], a perturbatio approach[49]. Oe of the siplest ways to describe the scatterig loss i straight ad bedig waveguide ca be expressed as followig equatios [5,5] : y, straight scatt, straight eff cl Es -6 y, bedig, bedig eff cl E E scatt -7 Where is the variace of the surface roughess at the iterface, eff is the effective refractive idex of the guided ode i the waveguide ad cl is the refractive idex of the claddig.5. adiatio loss The optical wave withi a straight waveguide ca be well cofied by total iteral reflectio at the dielectric ad air boudary ad guided i the waveguide without loss. However, it is ow that total iteral reflectio at a curved iterface is icoplete, ad leads to a trasitted wave, which causes loss of optical eergy. This loss echais is called optical tuelig loss [5] or bedig loss. This optical tuelig pheoeo ca be uderstood by drawig a aalog to the quatu echaical treatet of a -D particle i a potetial well which gives the physical eaig of the bedig loss. But a approxiatio expressio of the bedig loss will give us ore uderstadig how the bedig loss chages with waveguide geoetrics. The bedig loss ca be expressed 7

33 as[5]: radiatio co s cl y, curved curved sw e exp{[ W s curved s s l s curved curved ] } s c -8 Where curved is the propagatio costat i curved waveguide, ad, y, curved, c are wave vector i vacuu, y i curved waveguide ad bedig radius of the waveguide, respectively. Fro this expressio, we ca clearly see that the bedig loss is expoetial icreasig whe the bedig radius is reduced. So by choosig a proper bedig radius, we ca greatly reduce the bedig loss i our icro-rig resoator desig..5. Absorptio loss Silica was used as the aterial for fabricatig the capillary tube which was used as the icrocavity resoator i our bio-cheical experiet. Silica has low absorptio losses over very board spectru rage. The iiu loss aroud.db/ located aroud.55 μ wavelegth rage, for which it has becoe the operatig wavelegth for fiber-optic telecouicatios. With such low absorptio value, the absorptio liited Q factor ca be calculated to be Q eff x, but oral bio-cheical experiet requires silica icrocavity resoator surrouded by water which have very high absorptio coefficiet at.55 wavelegth rage, so a silica icro-tube resoator will show a Q factor aroud 4 to 6 whe it was used as a bio-cheical sesig eleet. Polyer aterial also has bee used as optical waveguide aterial for very log tie, because it is easy to be prepared, ca be dope with all ids of fuctioal aterials ad 8

34 ca be fabricated o various substrates. But polyer aterials have show relative higher absorptio loss aroud few db/c copared to silica or silico at telecouicatio wavelegth rage. This is aily due to carbo-hydroge bods haroic absorptio at ear-i rage. Oe possible ethod to reduce the absorptio loss is to replace the H ato i the polyer by F ato. Aother possible ethod to reduce absorptio loss is to ove device s worig wavelegth fro ear I to visible or ear visible wavelegth rage where polyer aterials have show uch lower loss [54], so it has bee applied to fabricate polyer optical fiber i short distace optical couicatio. ecetly PMMA has bee used to fabricate optical icrocavity to shows the aterial absorptio los as low as.44 db/c, yieldig Qabs. 8 [55 ]. 9

35 Chapter Pris coupled icro-tubes as sesitive bio-cheical sesors. Itroductio ad otivatio Silica capillary tube resoator sesor has attracted a lot of iterests o biocheical sesig applicatios ot oly due to its ability to hadle aqueous aalytes, but also its high Q factor., The high-q varies fro 4 to 6 is a result of the low scatterig loss due to its sooth surfaces which is beefited fro its high teperature fabricatio process. However there are also liitatios to the icro-tube resoators. Sice oly the ier surface ca be used as the sesor iterface, so the total evaescece optical field iteractig with the solutios is less tha half that of icro-rig resoators if the two devices have siilar field distributio for the guided ad evaescet waves, thereby resultig i a lower sesitivity. Moreover, the previously deostrated icro-tube resoator sesors required a tube thicess of less tha 5 icros i order to obtai a refractive idex sesitivity of.6/iu [5]. Sigificatly higher idex sesitivity ca be obtaied by further reducig the wall thicess to below sub-icro [,56],. But such

36 strategies cause the icro-tube to be very brittle ad difficult to hadle i practical sesig applicatios. I this chapter, we will propose ad deostrate to use the pris coupled thic wall above icro-tube sesor syste ot oly to overcoe the reliability proble, but also to realize high sesitivity detectio. Bul refractive idex sesig ad surface sesig experiets have bee deostrated ad experietal results atch well with the siulatio results. Also this ew platfor has bee used to realize sesitivity surface sesig experiet by detectig lipid oolayer, lipid billayer ad electrostatic self asseble layer by layer bidig to the ier surface of the icrotube. Moreover the iteractios betwee the lipid ad protei have bee studied.. Experietal setup Figure. Scheatic of pris coupled icro-tube sesor syste

37 Figure. shows the scheatic of the experietal setup. Light icidet oto a SF-glass pris was evaescetly coupled ito a fused silica icro-tube through a cotrolled air gap. The fused-silica icro-tube was purchased fro Polyicro Ic. ad has a ier radius of = 4 µ ad a outer radius of = 66 µ. I our experiet, the silica icro-tube was first coected to two pieces of plastic tube at each ed. The aalyte solutio flowig through the silica tube was cotrolled by a peristaltic pup. Next, the silica tube was positioed perpedicular to the icidet light plae ad outed at the ceter of the pris. The couplig gap betwee the icro-tube ad the pris was cotrolled by a deposited Al fil with 4-5 thicess o oe side of the silica icro-tube. The silica icro-tube was the boded to the pris ad outed o a rotatio stage. The output fro a wavelegth tuable diode laser aroud 55 wavelegth Satec TSL- was colliated by a GIN les colliator ad icidet oto the pris with a spot size of ~ 5u. The icidet agle was tued by rotatig the rotatio stage so as to satisfy the phase atchig coditio to the differet order resoace odes i the icro-tube resoator. The light reflected fro the pris, which was coupled out of the resoator, passed through a aperture ad the a polarizer, ad fially was focused oto a photo-detector. The polarizer was used to select the TE ad TM polarizatio i the output fro the resoator. I our experiet we focused o the easureet of the TE polarizatio, because the TE ode was observed to posses higher Q factors tha did the TM odes.

38 . Bul refractive idex sesig.. Bul refractive idex sesig experiets -6 fluidic refractive idex..6x fluidic refractive idex.4.x -6 fluidic refractive idex.8 Itesitya.u..8x -6.4x -6.x -6.6x -6.x Wavelegth Figure. esoace curve shift due to the chage of liquid refractive idex i the icro-tube at icidet agle of 7.5 o The idex sesig experiets were perfored by chagig the refractive idex of the liquid that flows through the icro-tube, ad siultaeously oitorig the correspodig resoace wavelegth shifts. The refractive idices of the liquid were cotrolled by preixig a very sall aout of isopropyl alcohol IPA with deioized DI water. The refractive idex of the solutio ca be estiated based o the dielectric costats of both liquids ad the olar ratio of each copoet [57]: - IPA water where is the olar ratio of IPA i the solutio. The refractive idices of water ad IPA are tae to be. ad.78 aroud the 55 wavelegth rage

39 esoace wavelgth shift p The first pea The secod pea The third pea The fourth pea efractive idex chage Figure. esoace wavelegth as a fuctio of the chage i liquid refractive idex i the icro-tube at a icidet agle of 7.5 o The bea illuiated o the pris is a Gaussia bea that covers a rage of spatial wave vector copoets. Therefore, the light ca be coupled ito several resoace odes withi a certai wavelegth rage that satisfies the phase atchig coditio. Fig.. shows three resoace spectra easured fro the reflected bea for liquids with three differet refractive idices filled i the icro-tube. The refractive idex differece betwee each spectru was ~ 4 x -4, ad the icidet agle to the pris was ~ 7.5 o. Four resoat odes are show i each spectru. These four resoat odes shift differetly for the sae fluidic idex chages of 4x -4 : the resoace wavelegth shift is ~ 9p for the first pea, ~ 8p for the secod pea, ~ p for the third ad ~ 54p for the fourth. Fig.. shows the resoat wavelegth shifts as a fuctio of the refractive idex chage. By liearly fittig the curve i Fig.., the device sesitivity, defied as the resoace shift per uit refractive idex uit IU for the four resoace odes are ca /IU, 7 /IU, 74.7 /IU ad 5 /IU, respectively. 4

40 The sesitivity of these resoace peas is 4 ties higher tha the sesitivity reported by White et al., which was about.6/iu. Itesitya.u. 5.x -6 fluidic refractive idex. 5.x -6 fluidic refractive idex.5 fluidic refractive idex. 4.8x -6 fluidic refractive idex.5 4.6x x -6 4.x -6 4.x -6.8x -6 a.6x Wavelegth Wavelegth shiftp The first Pea The secod pea The third pea b 5 5 efractive ide chage x -5 Figure.4 a esoace curve shift due to the chage of liquid refractive idex i the icro-tube, b resoace wavelegth shift as a fuctio of the chage i liquid refractive idex at a icidet agle of ~5 degree. Further decreasig the icidet agle to ~ 5 o allows the excitatio of uch higher order resoace odes i the icro-tube resoator. Figure.4 a shows four resoace spectra easured for differet liquid idices at a icidet agle of ~ 5 o. Agai three resoace odes show differet shifts. For the sae idex chage of 5-5, the resoace shifts are p for the first ode, 8p for the secod ad p for the third. Figure.4b shows the resoat wavelegth shift as a fuctio of the refractive idex chage. The sesitivity for the three resoace odes are 6 /IU, 6 /IU, ad 8 /IU, respectively. The sesitivity correspodig to the first ad secod ode is uch higher tha what ca be achieved with the typical evaescece wave sesors... Siulatios of bul refractive idex sesig We aticipate that the very high sesitivity to the refractive idex chage i the silica icro-tube is due to the large optical field peetratig ito the fluidic core regio. I order to uderstad the physics behid the observed high sesitivity, we ivestigated the optical field distributio i the icro-tube usig cylidrical coordiates z, r ad. The 5

41 6 tie idepedet field distributio for the resoace ode ca be separated ito a radial-depedet ode copoet ad a aziuthal-depedet phase ter, exp i r Z, where Z is the aplitude of the axial agetic TE or electrical TM odal field, is the aziuthal quatizatio uber. For the TE ode i the icro-tube we oly eed to cosider three copoets: H z, E r ad E. The radial depedet field H z ca be expressed by Bessel fuctio i the followig [58]: r r r r DH r CN r BJ r AJ H Z ] [ - where, ad are the refractive idices of the liquid core regio, the silica icro-tube regio ad the outside air regio, is the wave-vector i the vacuu; r J, r N ad r H are respectively the th order cylidrical Bessel fuctio, Neua fuctio ad Hael fuctio of the first id. By atchig the boudary coditios at the liquid/silica ad silica/air iterfaces, we obtai the eige fuctio equatio, which ca be used to calculate the resoace wavelegth. The TE ode resoace with aziuthal uber is deteried by: / / N C J B N C J B J J - where J H J H N H N H C B

42 For a give uber, there are a series of satisfyig equatios, which are referred to as the th =,,,... order resoace ode. The radial ad tagetial copoets of the electrical field i the TE ode ca be expressed i ters of H z as: E r H z -4 o r r ad E i o r H r z -5 The agetic field distributio i the icro-tube resoator at the resoace wavelegth ca be obtaied fro Eq. - ad the electrical field distributio ca be obtaied fro equatios -4 ad -5. I the siulatio we chose the followig paraeters for the icro-tube ad its eviroet: ier radius = 4 µ, outer radius = 66 µ, ad the refractive idices of the silica icrotube =.45, the liquid water i the silica tube =., ad the air outside the icro-tube =.. The electrical field is resposible for the resoace wavelegth shift due to the refractive idex chage i the liquid. Therefore, our study will aily focus o the electrical field distributio i the icro-tube. I the icro-tube resoator the radial electrical field of TE ode is uch larger tha the agular electrical field, thus it will suffice to plot the radial electrical field E r distributio i the silica icro-tube. Figure.5a shows the electrical 5 field distributio of the TE ode i the silica icro-tube resoator with a resoace 74 wavelegth aroud.55 µ, where the subscript, 74, stads for aziuthal ode uber ad the superscript, 5, stads for radial ode uber. As show i Fig..5a, the optical field is predoiatly guided i the high idex regio silica tube wall ad decay 7

43 expoetially i the low refractive idex regios liquid core ad air regios. Such a ode is cofied i the tube wall by the total iteral reflectios fro the silica/air ad 5 silica/water iterfaces. Thus, refractive idex chage trasduced by thete ode is via 74 a typical evaescece wave sesig echais. Itesitya.u Liquid Silica-tube Wall Air a adius distaceu esoace pea shiftp efractive idex chage b Figure.5 a adial electrical field itesity distributio for the resoace ode. b esoace wavelegth shift related to liquid s refractive idex chage for the resoace ode We siulated the resoace wavelegth shift related to the refractive idex chage of liquid by usig the Mie theory. I Mie theory calculatio [59], the optical field outside the tube wall is replaced by: H z H H -6 ic. scat. z z ic. where Hz J r represets the icidet wave, ad H scat. z DH r represets the scattered wave. The coefficiet of the scattered wave D ca be calculated by atchig H z ad E at the boudaries r = ad r 8

44 =. The resoace curve ca be obtaied by plottig the coefficiet D as a fuctio of the wavelegth by choosig a specific aziuthal ode uber. Whe chagig the refractive idex of the liquid by x -4, the resoace pea gradually oves to the loger wavelegth. The resoace pea shift as a fuctio of the refractive idex chage is plotted i Fig..5 b. A liear fit of the calculated data poits gives the sesitivity of this ode, which is ~ 7/IU. This uber is very close to the experietal sesitivity obtaied at the icidet agle of ~ 7.5 degree. Geerally, for such resoace ode havig a evaescece wave i the liquid core regio, the sesitivity orally is lower tha /IU. Usig a siilar approach, we ivestigated resoace odes havig higher radial 7 ode ubers. The radial electrical field distributio of the TE ode i the silica icro-tube resoator is plotted i Fig..6 a. The radial electrical field still decays expoetially i the air regio, however the electrical field distributio i the liquid regio is drastically differet fro that of the typical evaescece wave resoace ode 5 e.g. TE discussed above. Not oly does the radial electrical field for the first pea 74 copletely ove ito the liquid regio, but also it reaches a agitude that is sigificatly higher tha that i the silica tube walls. The optical field i the liquid regio is o loger a evaescet field, but rather accout for about 6% of the total field. Such a resoace ode with strog electrical field i the liquid regio is very sesitive to the liquid s refractive idex chage, aig it a excellet choice for idex sesig. We further siulated the resoace pea shift resultig fro the idex chage of liquid by usig the Mie scatterig ethod. The results are show i Fig..6 b. As ca be see i 7 9

45 the figure, the idex sesitivity of this ode is aroud 56/IU, which is very close to the 6/IU that we observed i the experiet. 5 5 Itesitya.u. 5 5 Liquid a Silica-tube Wall Air esoace shiftp 5 5 b adius distaceu efractive idex chage Figure.6 a adial electrical field itesity distributio for resoace ode b esoace wavelegth shift related to liquid s refractive idex chage for resoace ode.. Discussios I order to uderstad the origi of resoace odes which have the sigificat field peetratio ito the liquid regio, we preseted a ray optics picture o this id of ode. This id of resoace odes exist as if the rays are bouced by the liquid/silica iterface ad cofied i the liquid regio. Fro the ray optics poit of view, whe the light propagatig i the liquid regio is icidet oto the liquid/silica iterface r =, the light will be partially reflected tered ray ad partially trasitted at this ier boudary; ad the trasitted rays cotiue to propagate withi the silica tube wall util they hit the silica/air iterface r =, where they udergo total iteral reflectio at this outer boudary tered ray. Whe these two rays eet agai at the liquid/silica

46 iterface at the sae locatio ad are i phase, they will iterfere costructively [6]. Furtherore, if the two rays for closed loop withi the tube circuferece, a resoace ode is created. The scheatic ray picture of this ode is show i Figure.7. This type of odes ca exist oly uder certai coditios related to the geoetry of the icro-tube ad the refractive idices of the three regios. Let us assue that the refractive idex of the liquid is, icro-tube ad outer edia. We first defie the icidet ad reflectio agles at the ier boudary r = as ad that at the outer boudary r = as. I order to have the light trasitted through the ier boudary still well cofied at the outer boudary by the total iteral reflectio, the agles ad should satisfy the followig three coditios: si si, si ad si -7 These coditios will lead to aother coditio o the diesios of the icro-tube:. Therefore if the liquid iside the tube is water ad outside edia is air that correspods to our experiet, the requireet o the ratio of the outer ad ier radius of the icro-tube is.. This coditio is satisfied for the silica icro-tube used i our experiet / = 66/4 =.4 <..

47 Figure.7 The scheatic ray picture of resoace ode with light trasitted ito the ier boudary. The detectio liit of the icro-resoator-based refractive idex sesig device is directly related to the Q factor of the resoator ad the sesitivity of the resoace ode discussed above. I our experiet we ca resolve the oe twety-fifth of the resoace lie width chage, give that resoace ode with the high field iside the liquid regio has a Q factor of ~ x 4. The liit of the sallest detectable liquid idex chage is ~ 5 x -6 refractive idex uits. The Q factor of the icro-tube is deteried by the total loss of resoace ode i the resoator, which icludes the radiatio loss, the absorptio loss ad the scatterig loss. Thus, the resoator s overall Q factor ca be expressed as: Q -8 Q Q Q radiatio absorptio scatterig Because the icro-tube fors a curved waveguide, eve for the special ode described above that has high field itesity i the liquid regio, the optical wave trasitted ito the liquid regio still ca retur to the sae loop ad well cofied by the total iteral

48 reflectio at outer boudary; thus, the radiatio loss i this id of resoace ode is still very low. The scatterig loss also ca be reduced by choosig the ode with low field itesity at the ier boudary iterface. The Mie scatterig ethod revealed that the itrisic Q value of this id of ode is aroud 7 without cosiderig the water absorptios loss. Whe the loss due to high water absorptio aroud.55 µ wavelegth is tae ito accout i the Mie Scatterig calculatio, the Q factor is reduced to about 5 x 4, which is cosistet with experietal result of Q ~ x 4. Therefore we believe that the water absorptio loss liited Q is the doiat ter i deteriig the overall Q factor of the resoace ode. Based o these cosideratios, we ca ifer that the refractive idex detectio liit of the icro-tube resoator sesor ca be greatly icreased by ovig the worig wavelegth fro ear the I to the visible rage. Usig the Mie theory calculatio we predict the itrisic Q of the resoace ode with high field iside the liquid regio is above 5x 6 i the visible rage. If i experiet we ca achieve a Q factor i visible rage aroud 6, the sallest detectio liit ca reach ~ 5 x -8 refractive idex uits, resultig i the icro-tube resoator beig oe of the best idex sesig devices.. Surface sesig experiet.. esoat ode characterizatios Before we start to do the surface sesig experiet, we eed pic up a resoace ode which will be highly sesitive to the bio-fil absorbed to the ier surface of the icro-tube. To test the sesitivity of the resoace ode i icro-tubes, a idex sesig experiet is first perfored by chagig the refractive idex of the liquid that flows

49 through the icro-tube, ad siultaeously oitorig the correspodig resoace wavelegth shifts. The refractive idices of the liquid were cotrolled by preixig a very sall aout of isopropyl alcohol IPA with deioized DI water. The refractive idex of the solutio ca be estiated based o the dielectric costats of both liquids ad the olar ratio of each copoet: Figure..8 shows two resoace spectra easured fro the reflected bea for liquids with two differet refractive idices filled i the icro-tube. I the idex sesig experiet, oe liquid is the DI water ad aother is DI water ixed with IPA with volue ratio :.. The refractive idex differece betwee each spectru was ~ x -4, ad the icidet agle to the pris was ~ 6 o. Oe resoat ode is oitored i each spectru ad the resoace wavelegth shift is aroud 8p for idex chage x -4. The sesitivity is estiated by /=9/IU..7 liquid with idex. liquid with idex..65 Itesitya.u W avelegth Figure.8 esoace cure shift due to chage of liquid idex i the icro-tube 6 Figure.9 shows the electrical field distributio of TE 77 ode i the silica icro-tube resoator where the aziuthal ode uber is 77, ad the radial ode uber is 6. As show i Figure.9, the optical field is predoiatly guided i the 4

50 high idex regio silica tube wall ad decay expoetially i the low refractive idex regios liquid core ad air regios. Ulie the special ode we exploited for bul refractive idex sesig that has ai optical field residig i the liquid core regio, i the experiet o detectig lipid ebrae bidig o the tube wall, we prefer the resoace ode represeted i Fig.9 sice it offers a very large electrical field right at the ier surface of the tube wall, allowig axiu iteractio with the thi lipid ebrae layer. Such a ode is cofied i the tube wall by the total iteral reflectios fro the silica/air ad silica/water iterfaces...6 Liquid regio Silica regio Air regio Itesitya.u adial distacemicro Figure.9 adial electrical field itesity distributios for the resoace ode E.. Lipid oolayer detectio 6 r 77 The icro-tube resoator sesor syste is used to perfor first surface sesig experiet by detectig the foratio of a lipid oolayer o the ier wall of the icro-tube. It starts with a piraha solutio H SO 4 :H O =: cleaed icro-tube treated with a silae to reder its ier surface to hydrophobic. After surface treatet, 5

51 the icro-tube is filled with DI water ad the resoat curve is recorded as a baselie, the -palitoyl--oleoyl-s-glycero--phosphocholie POPC liposoe is ijected ito the icro-tube. It is ow that POPC lipid will for a oolayer to such a surface by orietig their hydrophobic tails toward the hydrophobic wall [6]. The POPC liposoe is stayig i the icro-tube for is to ae sure that it fors a good oolayer coverage o the ier wall of the icro-tube, after that we wash those ubouded liposoe out with DI water. All the spectra are recorded i the figure.. The resoace et shift i this case is ~p, which is the cotributio fro the oolayer coatig o the ier wall of the icro-tube. Figure. esoace curve shift due to the lipid oolayer bidig to the ier wall of icro-tube... Lipid bi-layer detectio The secod experiet to characterize our icro-tube resoator sesor syste is to detect the lipid bi-layer absorbed o to the ier wall of the icro-tube by usig the 6

52 resoat ode with siilar sesitivity i our pervious experiet. The resoat spectra of the process are show i the figure.. First the piraha cleaed icro-tube is filled with the buffer solutio Tris as to establish the baselie, the the POPC liposoe is ijected ito the silica icro-tube, after iutes the resoat spectru was easured ad showed a red shift of ~74 p. This shift is due to the lipid ebrae bidig to the ier surface of the icro-tube ad the uboud liposoe reaied i the solutio. After that the Tris buffer is used to wash away the uboud liposoe i the icro-tube ad replace it with Tris buffer ad the resoat curve is show a bac shift aroud p. The et shift is about 44p which represets the lipid ebrae absorbed oto the ier surface of the icro-tube. The resoat wavelegth shift fro the lipid bi-layer is alost double of the shift fro the lipid oolayer, which is quite reasoable. after saturatio wash with Tris buffer just start with Tris buffer filled with liposoe ad go to saturatio.6 Itesitya.u wavelegth Figure. resoace curve shift due to lipid ebrae bidig to ier surface of icro-tube 7

53 .. Self asseble layer by layer detectio To further ivestigate the capability ad the repeatability of the icro-tube resoator sesor syste i detectig the adsorptio of a ultrathi fil, we perfored aother experiet by ivestigatig a layer-by-layer polyer fil foratio o the ier wall of silica icro-tube. This experiet is doe by usig the electrostatic self-assebly of aoeter thic polyelectrolyte fils. Electrostatic self-assebly relies o the electrostatic iteractios betwee a adjacet polycatio ad polyaio layers. The substrate is preprocessed to have charges e.g., the oxide surface has a egative charge i dilute acid solutio. I our experiet, we used poly-diallyldiethylaoiu chloride PDDA as the polycatio ad polydye s-9 as the polyaio. To start, the icro-tube is cleaed with piraha solutio for iutes, which helps to for O - H + groups o tube surface where H + ios ca be readily exchaged with other catios. The catioic solutio PDDA solutio is flow through the tube to allow the egatively charged silica wall surface to be copletely covered with a layer of positively charged PDDA olecules. After a DI water rise, ioic solutio poly dye S-9 flows through the tube ad for a egatively charged oolayer that bod to the pervious catioic layer by electrostatic iteractio [6]. This sequece ca be repeated to for a ulti-layer coatig at the ier tube wall. The experiet results of coatig 7 alteratig oo-layers are show i Fig... The resoace shift iduced by PDDA oolayer is ~ 9p, while that of poly dye s-9 oolayer is ~ 4p. This reaso for this differece is that the refractive idex ad thicess of poly dye s-9 oolayer is uch larger ad thicer tha that of PDDA. 8

54 Figure. esoace curve shift due to the electrostatic self assebly thi fil coatig at the ier wall of the silica icro-tube...4 Quatificatio of bio-fil layer bidig We siulated the resoace wavelegth shift related to the refractive idex chage of liquid ad due to the adsorptio of a orgaic layer o the ier tube wall by usig the Mie theory. By chagig the refractive idex of the liquid with x -4, the resoace pea gradually oves to loger wavelegth. By fittig the calculated data poits, we ca get this ode s sesitivity for refractive idex as ~ 7/IU. This uber is very close to the experietal sesitivity obtaied at idex sesig experiet which is aroud.9/iu, which also supports our iterpretatio of the optical field distributio for such a resoace ode Usig the Mie scatterig ethod, we ivestigate the resoace wavelegth shift due to the lipid ebrae absorbed oto the ier wall of the icro-tube. We eep the 9

55 4 icro-tube s geoetry the sae lie before ad we cosider the lipid bilayer s refractive idex aroud.46. Whe the ier surface of the tube is coated with lipid ebrae, for TE ode the agetic field distributio ca be described i the followig for: ] [ ] [ d d d d d d H B N B J B N B J B J B H z -9 where,, ad 4 are the refractive idices of the liquid core regio, lipid ebrae regio, the silica icro-tube regio ad the outside air regio. After atchig the boudary coditios, the resoace coditio ca be give by the followig equatio: / / / / d N d J B B d N d J B B d N d J B B d N d J B B - Where d J d J d J d J d N d J d N d J B B ad d J d H d J d H d N d H d N d H B B Figure. shows the resoace curves shift to loger wavelegth whe the thicess of the thi fil layer absorbed o the icro-tube is icreased icreetally by. This is because the resoace ode s effective refractive idex i the cavity is icreased with the icrease i adsorbed layer thicess. The the calculated resoace wavelegth shift is plotted with respect to the chage of the thicess of the lipid ebrae, we fid that the experietally observed 44p shift correspods to a thicess of the lipid ebrae

56 thicess aroud 4-5 with the typical lipid ebrae with the refractive idex aroud.46. This result is cosistet with a bilayer lipid ebrae thicess reported i the literature.. Itesity a.u o absorbed layer = Wavelegth Figure. Siulated resoace curve shift with the lipid ebrae thicess chage The siulatio fro the layer by layer coatig also shows that the PDDA with the refractive idex aroud.46 ad poly dye s-9 s refractive idex aroud.6, ad both of the oolayer s thicess is aroud, which is cosistet to the reported by other paper...5 Study iteractio betwee the lipid ebrae ad proteis First we have studied the iteractio betwee the lipid bilayers ad Aexi V proteis: Aexi V is a eber of a large faily of proteis, aexis, which bid to egatively-charged lipids i a Ca + -depedet aer [6]. This protei is ow to disrupt the blood coagulatio by forig -diesioal crystals o the surface of cell ebraes. Figure.4 shows the experietal results of aexi V iteracted with POPC lipid bi-layer. The solid lie curve correspods to the baselie as the icro-tube 4

57 filled with 8M Cacl buffer solutio; the dash lie curve a et shift of 4p is a result of cotiuous flow of the aexi V protei solutio with cocetratio 6.67g/l for iutes followed by washig with 8M Cacl solutio; ad the dash dot lie curve a et shift of 8p was obtaied after cotiuous flow of the aexi V protei solutio for 4 iutes followed by 8M Cacl washig. The resoace shift is relatively sall as copared with the shift we observed i the lipid bilyer experiet. We believe this is because the Aexi V protei does ot have full coverage whe it bids to the surface of the lipid ebrae [64]. Figure.4 esoace curve shift due to Aexi V bidig to POPC lipid bi-layer Secod we have studied the iteractio betwee the lipid bilayers ad Alaethici proteis: Alaethici is a -aio-acid peptide fro the fugus Trichodera viride ad it is well ow that alaethici acts by creatig pores i the cell ebrae to for the 4

58 io chael i the lipid ebrae [65]. The first experiet we have doe is to show the resoat curve shift with tie at fixed alaethici cocetratio. First we tae the sigal fro the silica icro-tube fillig with DI water as a baselie, the the icro-tube is filled with POPC liposoe for 5 is to ae sure that the lipid bi-layer ebrae is well fored o the ier wall of the icro-tube, after that the u-bouded POPC liposoe is washed out with the DI water for 5 is, we ca clearly observe the resoat wavelegth red shift due to the POPC lipid ebrae bidig to the ier wall of the icro-tube. The the atibiotic peptide alaethici with cocetratio 5g/l is ijected ito the icro-tube ad we collect the optical sigal at differet ties after the ijectio. We ca fid that with the tie icreased the resoat wavelegth shift to short wavelegth. I the first is the shift is large, after that the shift is gettig saller ad saller, which is because the cocetratio of the alaethici get reduced after it stay i the silica icro-tube for soe tie. After 4 is the resoace blue shiftig is gettig saturated ad the total shift is about 9p i Figure.5. This blue shift which ca be explaied that the refractive idex of the whole fil absorbed o the ier wall is decreased or the fil thicess of the lipid ebrae is reduced. 4

59 Figure.5 esoace curve shift due to the atibiotic peptide alaethici iteract with lipid bilayer o the ier surface of the icro-tube Aother experiet we have doe is that we start with lower cocetratio of alaethici ad we gradually icrease the alaethici cocetratio fro 5g/l to g/l, 5g/l, g/l, g/l ad 5g/l, the we oitored the resoace wavelegth shift after the icro-tube is filled by with differece alaethici cocetratio solutios for is. At this experiet we also choose the resoace ode which has the high electrical field at the ier boudary by testig its bul idex sesitivity which is aroud /IU ad the Q factor of this resoace ode is aroud.6x 4. We ca clearly see fro the figure.6 that whe the alaethici cocetratio chages fro 5g/l to g/l, the resoace wavelegth is cotiuously shiftig to the loger wavelegth ad the shift is quiet liear. Fro g/l to 5g/l the resoace wavelegth is shiftig to shirt wavelegth ad the shift is quiet draatically. Those ca be explaied that at low cocetratio lower tha 44

60 g/l the alaethici is just bidig to the surface to the lipid ebrae which will icrease the thicess of the fil, so the resoace wavelegth will shift to loger wavelegth. But at the high cocetratio higher tha g/l, alaethici will aggregated together to for a pore o the lipid ebrae ad those holes ca be filled by lower idex edia water, so the total effective idex of the fil will reduced ad which will cotribute to the blue shift of the resoace wavelegth. If we copare those two experiet results, it is iterestig to fid that whe the alaethici cocetratio is chage fro low to high the total shift alost be zero, but with high cocetratio we ca see that the shift is egative, which eas that those alaethicis which is bidig to the surface of lipid ebrae at lower cocetratio still will stay o the lipid ebrae ad which will ot have the cotributio to for pores i the lipid ebrae. Ad those bidig o the surface alaethici will ot effect those oe who iserted ito the lipid ebrae to for pores, so the total blue shift to with this two case are all aroud p. Figure.6 esoace wavelegth shift related to differet cocetratio of alaethici ijectio. 45

61 .4 Coclusio We have successfully deostrated usig pris coupled icro-tubes as ew bio-cheical sesig platfor. I bul refractive idex sesig experiets, we have observed a uique resoace ode with very high sesitivity aroud 6/IU which is very good for bul idex sesig applicatio. I surface sesig experiets, we have successfully deostrated to detect lipid oo-layer, lipid bi-layer ad electrostatic self asseble layer by layer bidig to ier surface of silica icro-tubes ad the siulatio results atch well with those experiets results. Furtherore, the iteractios betwee the lipid bi-layer to soe proteis have bee studied i our sesor syste. 46

62 Chapter 4 Sesig properties of icro-tube resoator sesors 4. Itroductio ad otivatio I the silica icrotube resoator sesor syste we observed two ids of sesig regies i our previous experietal studies. The first is the typical evaescet sesig regie i which the optical wave is cofied betwee the ier ad outer boudaries of the icrotube wall by total iteral reflectios; oly the evaescet wave extedig ito the icrotube liquid core will iteract with the liquid or gas to realize the sesig fuctio. The secod is a oevaescet sesig regie i which the optical field is cofied by total iteral reflectio at the outer boudary, but the optical wave is partially trasitted ad partially reflected at the ier boudary. The costructive iterferece betwee the two waves establishes a uch stroger optical field iside the icrotube, which also peetrates deeper ito the liquid core regio ad greatly ehaces the sesitivity due to the sigificatly iproved field overlap with the aalytes i the liquid. The purpose of this chapter is to provide detailed theoretic alaalysis of the two sesig odes ad ivestigate their advatages ad liitatios. I this chapter, we first plotted the field distributio of the correspodig WGMs with differet radial orders. 47

63 Based o the field distributio, the icrotube s various sesig properties were ivestigated i detail by the perturbatio ethod for both evaescet ad oevaescet regies, which iclude bul refractive idex, surface, ad absorptio sesig properties. Also we proposed a ultra-high sesitivity sesor by usig a coupled cavity platfor.. 4. Field distributio i icro-tube resoator Figure 4. The oralized electric field itesity distributios with aziuthal uber 5 7 M=7 ad radial order uber N=5 ad 7 deoted by E 7 ad E 7, respectively The oralized electric field itesity distributios fro the equatio i chapter with aziuthal uber M=7 ad radial order uber N=5 ad 7 deoted by 5 E 7 ad 7 E 7, respectively are show i Fig. 4.. I the siulatio we chose the sae 48

64 icrotube diesios reported i our previous experiet: ier radius =4, outer radius =66, liquid refractive idex =., silica refractive idex =.45, ad air refractive idex =.. We ca clearly see that with the icrease of the radial order uber the optical field is peetrated deeper ito the liquid regio. We foud that whe the radial order uber is equal to or saller tha 5 the first pea of the electric field itesity coutig fro the ceter of the tube ad the rest of the electric field itesity peas reai i the high idex regio silica regio. This eas that the optical field is well guided withi the silica wall ad the wave is cofied by the total iteral reflectios at both the ier ad outer wall boudaries with oly a sall evaescet field peetratig ito the liquid regio. Such odes will be referred to as evaescet sesig odes for sesig applicatios. However, if the radial uber is larger tha 5 e.g., ode 7 E 7, the first pea of the electric field oves copletely ito the liquid regio because the optical wave is partially trasitted ad partially reflected at 4. Bul refractive idex sesig sesitivity After the electrical filed itesity distributio i the icro-tube resoators are obtaied, the sesitivity of the differet resoace odes ca be calculated by the perturbatio theory. The bul refractive idex sesitivity S, defied as the ratio of the resoace wavelegth shift to the bul idex chage, ca be described by the followig equatio [66]: 49

65 S bul V V E z E dr z dr 4- where is the resoace wavelegth, V is the whole space ad V is the space fro the origi to the ier boudary. Figure 4. Siulated bul refractive idex sesig sesitivity of differet radial order odes with the sae aziuthal uber M=7 by perturbatio ethod ad Mie scatterig ethod Usig the electric field distributios obtaied for resoace odes of the differet radial uber N but the sae aziuthal uber M=7, the bul refractive idex sesitivity ca be calculated usig Eq. 4- ad plotted i Fig. 4.. The result fro the perturbatio ethod blac curve i Figure 4. atches well with that fro Mie scatterig ethod red curve which is very accurate to predict the resoace wavelegth i the icro-cavity. Clearly the bul refractive idex sesig sesitivity icreases 5

66 draatically as the radial uber icreases fro N= to N=7. This is because with higher radial order uber the optical field peetrates deeper ito the liquid regio. As stated earlier, whe the radial uber N is saller tha 5, the optical wave is still well cofied by the total iteral reflectio at the ier boudary ad oly wea evaescet field resides i the liquid regio ad we call this regie with N < 5 as evaescece sesig regie. Whe the radial uber N is larger tha 5, the first E-field pea of the resoace odes oves ito the liquid regio ad we call the regie with N > 5 as o-evaescece sesig regie. For bul refractive idex sesig the o-evaescece regie is preferred due to icreased field overlap with the liquid aalyte. The sesitivity ca be ehaced by a factor of 6 if we copare the axia sesitivity i the o-evaescece regie e.g. N=7 to that i the evaescece regie e.g. N=5. A iterestig fidig fro this calculatio is that the sesitivity i the o-evaescece regio for N>7 oscillates betwee /IU to 6/IU with the icrease i the radial order uber. Therefore it is iportat to choose the proper resoace order whe perforig the bul idex sesig experiet. This ca be doe, e.g. by tuig the icidet agle, i the experietal settig. The oscillatory behavior is soewhat cotradictory to our expectatio that as the radial order icreases the optical field peetrates deeper ito the fluid regie ad the sesitivity would cotiue to icrease rather tha oscillates. We believe this behavior is due to a itricate iterferece effect betwee the two optical rays. 5

67 4.4 Surface sesig sesitivity I a typical surface sesig experiet, the bio-olecules adsorbed oto the surface of the ier wall of a icro-tube will cause a resoace shift, ad oe ca defie the surface sesig sesitivity as the ratio of the resoace wavelegth shift to the thicess of the bioolecular fil. Additio of the bioolecular fil to the ier surface of the tube will chage the wave-vector of a supported ode fro to ad the electrical field fro E to E. Both the uperturbed E ad perturbed E electrical fields satisfy the wave equatio: E 4- E E 4- E where ad are the dielectric costat distributios of the icro-tube syste before ad after the bio-olecular adsorptio. Multiplyig Eq. 4- by E ad itegratig over the etire space gives: * * * E Edr E Edr E fil v liquid v E dr 4-4 where fil ad liquid, are the dielectric costats of the bio-olecular fil ad the liquid, respectively; is the total space of the icro-tube ad is the space occupied by the adsorbed bio-fil o the icro-tube. By usig a equality [67]: 5

68 v E * Edr E Edr E Edr * v v the followig equatio ca be derived: * E Edr E fil liquid Edr v 4-5 Assuig that the adsorbed bio-olecular layer thicess T is very thi, the electric field withi the fil ca be tae to be a costat, ad the overall chage of the field distributio i the resoator is very sall, so E ca be approxiated by E i the perturbatio calculatio. We ca the obtai the followig result: E * v fil liquid E E dr E dr E fil liquid T E E dr v 4-6 ad the sesitivity i surface sesig ca be the expressed as: S surface E T v fil E E dr liquid 4-7 Fro Eq. 4-7, we ca see that the sesitivity ca be ehaced by icreasig the ratio of the field itesity at the tube ier boudary to the total optical field, reducig the tube thicess, or elargig the ier radius of the icro-tube. The first coditio ca be achieved by excitig higher order evaescet odes. educig the tube thicess also 5

69 decreases the optical field iside the tube wall, which has a siilar effect to the first coditio. Icreasig the ier radius creates the evaescece field at the ier surface [68], which leads to iproved sesitivity. Calculated surface sesig sesitivity by perturbatio ethod for differet radial ode ubers with the sae aziuthal uber M=7 atches well with the results fro the Mie scatterig ethod ad the results are plotted i Figure 4.. I the evaescet sesig regie, the sesitivity icreases with radial uber because of the icrease of the electrical field at the ier boudary. I the o-evaescet sesig regie, the oscillatory behavior as a fuctio of the ode ubers has bee agai observed but with opposite tred to the case of bul idex sesig sesitivity. Figure 4. Siulated surface sesig sesitivity of differet radial order odes with the sae aziuthal uber M=7 by perturbatio ethod ad Mie scatterig ethod. The bio-fil s refractive idex is assued to be.46 54

70 4.5 Absorptio sesig sesitivity Detectio of the presece of certai aalytes i the fluid ca also be accoplished if the aalytes exhibit absorptio i the wavelegth rage of iterest. The sesitivity ca be ehaced by usig the icro-resoator structure due to the icreased iteractio legth betwee the light propagatig i the resoator ad the aalytes. For siplicity, we cosider the case of aalytes dissolved i the bul liquid flowig through the icro-tube. Light absorptio ca be icluded i our odel by taig ito accout the iagiary part of the refractive idices. The presece of the light absorbig aalytes chages the iagiary part of refractive idex of the liquid fro to, which will cause the iagiary part of the resoace wave vector chage fro to ad the electrical field of the resoace ode chages fro E to E. The uperturbed ad the perturbed electrical field still satisfy the wave vector equatio: E i E 4-8 E i E 4-9 where the ad are the dielectric costat distributios for icro-tube filled with liquid with optical absorptio coefficiet ad, respectively. I the liquid regio the dielectric costats are i ad i. I real sesig experiet ad are uch saller tha, so the resoace wavelegth ad the field distributio will ostly reai uchaged. Oly the resoace ode s Q factor will 55

71 chage with the optical absorptio coefficiet. Fro the equatios 4-8 ad 4-9, we ca give aother ew equatio: i E E dr i E E dr 4- By coparig the real ad the iagery part of equatio 4- ad we assue the Q factor of the icro-tube resoator is predoiatly deteried by the absorptio loss, we ca fid two equatios: E E dr v E E dr v 4-4- The quality factor of the icro-tube resoator ca be defied as flowig: Q i, i where i=,., We defie a iteractio factor of the resoace ode v v E E E E dr dr, the we ca get: 4 Q 4-56

72 We defie the sesitivity of the absorptio sesig as the ratio of the Q factor chage with the chage of the iagiary part of the refractive idex of the liquid chage: S abs Q 4 Q 4-4 It is iterestig to show that the absorptio sesitivity is icreased with decreased, which also eas that the sesitivity is icreased with icreasig Q factor, cosistet to oe s expectatio. The siulated absorptio sesig sesitivity with differet radial uber sae aziuthal uber M=7 of the resoace odes with absorptio coefficiet =.7c - is show i the figure 4.4. The blac curve is siulated by perturbatio ethod ad the red curve is siulated by Mie scatterig ethod. We ca clearly see that the sesitivity for radial uber N<6 liearly decreases with the icreasig radial uber. This is because odes with higher radial ode uber have ore overlap with the absorbig edia ad therefore lower Q factor. The perturbatio ethod agai atches very well with Mie scatterig ethod whe N is saller tha 8, after that the Mie scatterig shows saller sesitivity coparig with the perturbatio ethod. We thi this is because that at large radial uber the Q factor is ot oly doiated by the absorptio loss, but also the radiatio loss due to bedig also becoes iportat. 57

73 Figure 4.4 Siulated absorptio sesig sesitivity of differet radial order odes with sae aziuthal uber M=7 by perturbatio ethod ad Mie scatterig ethod. The absorptio coefficiet of liquid is assued to be =.7c Sesitivity ehaceet usig a coupled cavity As we ca fid fro sectio 4. that the icro-tube resoator sesor s sesitivity ca ot be further iproved by usig uch higher order ode. Ad the sesitivity starts to oscillate betwee /IU ad 6/IU betwee differet higher-order-odes. Ca we still further iprove the device s sesitivity? Couple of ethods have bee proposed ad realized i cavity based icrocavity sesor syste: Professor Arold s group has realized a idea to iprove the device s sesitivity by coatig the surface of the resoator with a relatively high refractive idex ao-layer. The high refractive idex layer teds to draw light i ad pulls the eergy fro the sphere iterior ad brigs it closer to the surface. As a result, the evaescet field of the coated icrosphere would exted deeper ito the outside ediu cotaiig the olecules to be detected. They 58

74 experietally reported the sesitivity ehaceet aroud 7 ties, ad theoretically the sesitivity ca be ehaced up to 5 ties [69]. M. Suetsy [7], et al reported to usig a curable low idex polyer to fix the positio betwee the capillary tube ad the couplig fiber after it solidified. Meawhile, the capillary tube s wall ca be thied dow to few hudred aoeters to push ore optical field ito the liquid regio to realize uch higher sesitivity. The sesitivity has bee deostrated as high as 8/IU i the bul refractive idex sesig experiet. But there still has sigificat part of the optical field i the outside polyer regio, so the sesitivity ca ot be further iproved up to /IU. I here, we proposed a ew id of resoace ode as a sesig ode i low idex aterial ier coated icro-tube resoator sesors. Such a id of resoat ode ca have a extreely high optical field i the liquid regio, which could lead to extreely high optical sesitivity up to /IU. The proposed ier-coated silica icro-tube resoator sesor is show i the Figure 4.5a, where ad are the ier ad outer ier radii of the silica icro-tube ad - is the thicess of the ier coated low refractive idex layer. Ad we treat it as four-layer syste i cylidrical coordiate i which the first layer is liquid core with refractive idex., water, the secod layer is coated low refractive idex layer with refractive idex.75, Teflo AF 4, the third layer is silica tube layer with refractive idex.45, Silica ad the fourth layer is the air with refractive idex 4, air. The radial field distributio i the cavity ca be described i the followig: E r [ BJ [ B4J BJ r r r BN r] r r B N r] r B 6 H r 5 4 r

75 Where J, N ad H fuctio of the first id, respectively. are the th Bessel fuctio, Neua fuctio ad Hael b a 4 c 4 Figure 4.5 a a scheatic of ier coated silica icro-tube sesor, the cavity ca be decoposed ito two cavities: b a silica icro-tube cavity filled with low idex aterials ad c a water cylider cavity covered with low idex aterials Such a cavity also ca be decoposed ito two idividual cavities which are show i the Fig. 4.5b which is a silica icro-tube filled with low idex aterial ad Fig. 4.5c which is a liquid icro-cylider with a low refractive idex edia outside. Whe we put those two resoator cavity together, those two sets of resoace i those two cavities ca be coupled to each other ad uderstood i ters of x Hailtoia atrix [7]: H E W V E Where E ad E are the coplex eergies of the ucoupled syste, W ad V are the couplig costats betwee two differet states. For wea 6

76 couplig VW I E I E, there is a crossig for the real part of the eergy ad ati-crossig of the iagiary part. For a strog couplig coditio, VW I E I E ad there is a ati-crossig of the real part of the eergy ad a crossig of the iagiary part. 57 a b esoace Q factorx efractive idex of Liquid efractive idex of liquid Figure 4.6a Ati-crossig behavior of the resoat wavelegths whe the refractive idex of liquid is aroud., b the Quality factor shows crossig behavior whe the refractive idex of liquid is aroud. Figure 4.6 a shows a pair of coupled ode s resoace wavelegth chages with the refractive idex of the liquid core. A clear ati-crossig [7, 7, 74] behavior was show i the figure 4.6a, where two resoace wavelegths coe close but the repel each other at liquid refractive idex aroud., ad resoace Q factor presets a crossig i figure 4.6b at the sae liquid refractive idex positio. These pheoea idicate that those two odes are strog coupled. No Q factor ehaceet was observed aroud ati-crossig regio, this aily is due to the Q factor i the cavity is doiated by water absorptio loss. 6

77 Itesitya.u A B C A: Liquid regio B: Polyer regio C: Silica regio D: Air regio D Itesitya.u A B C D adial distace adial distace Itesitya.u.. A B C D Itesitya.u.. A B C D V Itesitya.u adial distace V A B C D Itesitya.u adial distace. A B C VI D adial distace adial distace Figure 4.7 Calculated field distributios i the cavity with differece refractive idex of liquid,, liquid refractive idex aroud.5,,v liquid refractive idex aroud., V, V liquid refractive idex aroud.8. The device s strog couplig pheoea have bee studied by plottig the device s spatial field distributio with differet refractive idex of the liquid core. Figure 4.7 shows the calculated field distributios i the cavity with differece refractive idex of liquid:, belog to the pair of resoace ode with liquid refractive idex =.5,,V belog to the pair of resoace ode with liquid refractive idex =., V, V belog to the pair of resoace ode with liquid refractive idex =.8. Figure 4.7 I, III, V belog to the lower brach of the resoace ode i figure 4.6a ad Figure4.7 6

78 II, IV, VI belog to the up brach of the resoace ode i figure 4.6a. We ca clearly see that the ode i I or VI is aily a liquid-lie ode alost all the optical field is cofied i the liquid regio, oly very sall part of the field is i the silica tube regio ad the ode i II or V is aily a silica tube ode, ost of the field is cofied i the silica regio ad oly part is i the liquid regio. Mode III ad IV are the hybridized ode, which are the super-positios of the ode I or VI ad II or V far away fro the ati-crossig regio. Whe we sweep the liquid refractive idex cross the ati-crossig the regio, we ca see that the liquid ode start to chage fro oe brach to aother brach. This is a clear evidece of the strog couplig that the resoace ode exchages their ode patter as well as the eergy while passig through the ati-crossig [75] regio. The device s refractive idex sesig sesitivities have bee obtaied by taig the differetial of the curves i the figure 4.6 a ad the sesitivities of those paired odes chagig with refractive idex of liquid core have bee show i the figure 4.8a. The couplig ature aes those two odes sesig sesitivity to have a opposite tred to the chagig of the refractive idex of liquid. Ad the sesitivity ca achieve as high as 967/IU. Figure 4.8b shows that the device s sesitivity ca be icreased above /IU by reducig the ier coated aterial s refractive idex which is aily due to chage the couplig stregth betwee those two cavity odes. 6

79 Sesitivity/IU a efractive idex of liquid Sesitivity/IU b Ier coated layer refractive idex Figure 4.8 a sesitivity of two hybrid resoace ode chages with liquid refractive idex.b the sesitivity chages with the ier coated layer aterial refractive idex.with fixed thicess d= 4.7 Coclusio We provide a theoretical study o various sesig properties o silica icro-tube resoator sesors. We foud that the bul refractive idex sesig sesitivity icreases with the radial order uber i the evaescet sesig regie ad oscillates i the o-evaescet sesig regie. The o-evaescet ode is particularly suitable for bul refractive idex sesig ad the sesitivity ca achieve 6/IU i our structure. The evaescet ode, havig a high electric field agitude at the ier boudary, is preferred i the surface sesig experiet with sesitivity as high as p/ ad high Q resoace ode is desirable for absorptio-based sesig. Coupled cavity platfor was proposed to realize ultra-high >/IU sesitivity sesig. 64

80 Chapter 5 High sesitive ultrasoic detectio usig polyer icrorigs 5. Itroductio ad otivatio I ultrasoud pulse-echo iagig ad photoacoustic also called optoacoustic iagig, high resolutio is achieved usig high-frequecy ultrasoud trasducers. High-frequecy above MHz ultrasoud iagig has bee applied to itravascular iagig, edosoography, sall aial iagig, si iagig ad ophthalology, ad bioedical applicatios of high-frequecy photoacoustic iagig iclude icrovasculature visualizatio, fuctioal iagig, ad itravascular iagig. I these refereced wors, piezoelectric trasducers were used. Sall eleet size ad spacig ad large eleet cout greatly liit the feasibility of realizig high-frequecy two-diesioal D arrays based o piezoelectric trasducers. However, to achieve high frae rate i three-diesioal D ultrasoud ad photoacoustic iagig, D arrays are required. Oe way to avoid the difficulties iheretly associated with piezoelectric trasducers is to detect ad geerate ultrasoud optically. Optical detectio of ultrasoud has bee studied for decades. Its advatages iclude iuity agaist electroagetic iterferece at optical eds, easier realizatio of large ad dese arrays 65

81 with eleet sizes of μ, high acoustic badwidth, ad depedecy of sigal-to-oise ratio SN o optical probig power istead of detector size. ecetly, a optical cavity based ultrasoud detectio platfor has attracted icreasig attetios. Copared with covetioal piezoelectric trasducers, the ew detectio platfor provides several advatages, such as preservig high sesitivity with reduced eleet sizes, high-frequecy ad widebad respose with siple fabricatio. I this ew optical cavity based detectio platfor, optically trasparet polyer aterial was used because of its high optical elastic coefficiet ad high deforability, which ca provide sesitive respose uder acoustic pressure. High Q resoators usig the TI echais iclude icro-spheres, icro-diss, icro-rigs ad icro-tubes. By cobiig polyer aterial s high optical elastic coefficiet ad high deforability with the TI-based high Q factor icro-resoators, it is possible to achieve high sesitivity acoustic detector by usig polyer icro-resoator. 5. Fabricatio process of polyer icrorigs 5.. Noral old fabricatio process Previous our group has deostrated to use Nao-ipritig techique to fabricate polyer icro-rig resoator [76]. The fabricatio process starts with creatig a aster old. First, e-bea lithography is used to create icro-rig coupled with straight waveguide patters i the PMMA layer ad the the patters are trasferred to the SiO layer by usig reactive io etch IE. After IE, the PMMA residual layer is reoved by hot acetoe. This aster old is called shallow old. The the shallow old is treated with surfactat to reduce the surface eergy. The the shallow old is used to 66

82 fabricate a deep old which is used to directly fabricate the polyer icro-rig by ao-ipritig techique. The shallow old is used to create patters i a layer of 5K PMMA o the silico substrate with theral oxide usig ipritig techique. After that the residual layer is etched away by O plasa. The a lift-off process is perfored to trasfer the patters to the etal fil Ni or Cr. With this etal as a etch as, we ca etch deep ad vertical patters by usig IE. After IE, the etal as is reoved by etal etchat. The whole fabricatio process is show i the Figure 5. [77] Figure 5. Scheatic of fabricatio of deep old fro the shallow old. a A shallow old fabricated fro EBL ad IE. b The shallow old is cotacted oto the saple with 5K PMMA layer. c The ipritig process is perfored at high pressure ad high teperature. d The saple is separated fro the old ad the PMMA patters are created o the substrate. e PMMA residual layer is etched away by O plasa. f Metal 67

83 as Ti/Ni is deposited ad lift-off process is perfored to trasfer the patter to etal fil. g The oxide layer is etched usig etal as. h The etal layer is reoved by the etal etchat. 5.. Siplified old fabricatio process Electro bea lithography a Electro-bea resist PMMA Cr fil Silico oxide layer Silico substrate b PMMA Developet c Cr dry etch, the reove PMMA d Silico oxide etch, the reove Cr fil Figure 5. scheatic of siplified silico oxide old fabricatio process. a E-bea lithography o the 95 PMM layer o the silico substrate with silico oxide ad 5 Cr fil. b The saple is developed i MIBK:IPA=:. c The Cr layer is etched by usig PMMA as a etchig as, the the PMMA layer is reoved by hot acetoe. d The silico oxide layer is etched usig Cr as a etchig as, the reove the Cr layer by usig Cr etchat. Our previous fabricatio process eed fabricate a shallow old first ad the create the deep old usig the shallow old. There are four etchig steps i the whole old 68

84 fabricatio process, which could icrease the surface roughess with too uch etchig steps. I order to siplify the fabricate process ad well cotrol the surface roughess, we developed a ew fabricatio process start with e-bea lithography o 5 thic 95 PMMA A4 o silico substrate with 5 Cr layer o theral oxide, after that the exposed saple is developed i developet with MIBK:IPA=: for i. ad rise i IPA for aother i. The the Cr layer is etched away with Cl gas by usig the PMMA layer as a etch as, after that the PMMA layer is reoved i hot acetoe. The ext step is to etchig the silico oxide layer by usig Cr as a etch as ad the the Cr layer is reoved by usig Cr etchat. The whole process flow is showig o the figure 5.. After the deep old fabricatio, the polyer icro-rig coupled with straight waveguide patters is created usig ao-ipritig lithography ad the followed by O plasa to reove the residual layer. 69

85 5. Spectru ad SEM characterizatio Figure 5. Trasissio spectru of the polyer icrorig The fabricated polyer icrorigs were characterized by usig a tuable laser at.55 wavelegth rage. The easured trasissio spectru of polyer icrorig is show i the Figure 5. ad the Q factor of the resoace pea is aroud 6. The SEM iage of the polyer icrirg is showig i the Figure 5.4a, ad the sidewall view of the polyer icrorig is showig i Figure 4.4b. We ca clearly see the roughess o the sidewall which is created durig the reactive io etchig process. Figure 5.4 c shows soe holes o the top part of the polyer icrorig. This ay be geerated durig the IE process, soe of the etal particles are re-deposited i the trech ad the those 7

86 particles ca act as a as to for soe pillars i the trech. This proble ca be solved by replacig the etal as with resist as. a b c Figure 5.4 a SEM iage of the polyer icro-rig with =5, b the sidewall view of the polyer icrorig c SEM iage of the polyer icro-rig with soe holes o the top 7

87 5. Acoustic sesitivity Ultrasoud Trasducer Oscilloscope Tuable Laser Polarizatio Cotroller Water Photo-detector Figure 5.5 The experietal setup to easure the oise-equivalet pressure ad sesitivity of a polyer icrorig resoator. the distace betwee the ultrasoud trasducer ad the resoator was.5. The setup show i Figure 5.5 was used to easure the NEP ad sesitivity of a polyer icro-rig ultrasoud detectio device. A cotiuous-wave tuable laser source HP 868F, Agilet Techologies, Sata Clara, CA was coected to the device s iput fiber, ad the output fiber was coected to a photodetector 8-FC, New Focus, Sa Jose, CA, whose DC output has a gai of V/A ad AC output has a gai of 4* 4 V/A ad a electrical badwidth of 5Hz 5MHz. These photodetector outputs were coected to a digital oscilloscope WaveSurfer 4, LeCroy, Chestut idge, NY for data collectio. Usig the DC output, we easured the device s trasissio spectru, of which oe saple curve is show i Figure 5.6a, with a iput power of 4. W. The Q-factor was estiated to be 6,. Accordig to the off-resoace light trasissio, 9% of the probig light was collected by the photodetector. A calibrated -MHz ufocused trasducer V6, Paaetrics NDT, Waltha, MA with a.8 diaeter was used to isoify the rig. It outputs a pea pressure of Pa aroud its 7

88 surface whe drive by a pea-to-pea -V oe-cycle MHz siusoidal wave. The acoustic couplig ediu was deioized water. The optical probig wavelegth ad iput power were set to ad 5.5 W, respectively. To easure the ultrasoud sigals, the AC output of the photodetector was used. Figure 5.6 b shows a recorded sigal trace. The rigig followig the ai sigal was due to reflectios withi the silico substrate, which ca be reoved by chagig substrate aterials ad/or structures. Sice Pa leaded to a output voltage of V, the correspodig shift of trasissio spectru was 6 p accordig to its slope, ad the sesitivity of the polyer icrorig ultrasoud detectio device was V/Pa. The root-ea-squared oise levels were.5,.,.5, ad. V over 5, 5 ad 75, MHz, respectively, ad the correspodig NEPs, a easure of the iiu detectable pressure of the device, were.4,.,.pa..4 a.4 b Voltage V... Voltage V Wavelegth.9.. Tie s Figure 5.6 a Optical trasissio spectru of a polyer icrorig resoator. The iput power was 4. W. b Sigle-shot acoustic wavefor easured by the resoator. The positive pea correspods to Pa. The optical probig wavelegth ad iput power were set to ad 5.5 W, respectively. Copared to the previous best result, 4. Pa over 5 75 MHz[78], we have iproved NEP by ore tha oe order of agitude. The sae level of NEP was 7

89 achieved usig a Fabry Pérot OUT with a detectio badwidth of MHz ad a detector diaeter of 5.As a referece, a 75-µ piezoelectric PVDF trasducer HPM75/, Precisio Acoustics, Dorchester, Dorset, UK; has a NEP of 6 Pa [=6 μv/ V/Pa] over a -MHz badwidth eve if oise oly coes fro a dedicated preaplifier HP, Precisio Acoustics. I additio, polyer icrorig ultrasoud detectio devices NEP ca be further reduced by couplig ore detectio light i or icreasig their Q-factor 5.4 Frequecy respose A widebad optoacoustic source was realized to easure the detectio badwidth of a polyer icrorig ultrasoud detectio device. A - thic chroiu fil was deposited oto a glass substrate. Illuiatig such a fil with a wide-spot aosecod laser pulse geerates a plaar optoacoustic wave with a teporal profile duplicatig the excitig pulse shape after vertical oralizatio [79, 8]. I a short propagatio distace, the wave ca aitai its teporal shape ad therefore be used as a acoustic source with a ow spectru. We put the chroiu fil 54 µ fro a polyer icrorig ultrasoud detectio device ad illuiated it usig a 5 pulsed frequecy-doubled Nd-YAG laser Surelite I-, Cotiuu, Sata Clara, CA with a spot size of 4.5 i diaeter. The acoustic couplig ediu was deioized water. A acoustic sigal detected by the device ad a laser pulse profile detected by the sae photodetector are show i Fig. 5.7a. Their close teporal duratios suggest a short ipulse respose of the device. Figure 5.7b shows the sigals spectra together with the 74

90 estiated frequecy respose of the device, which was obtaied by taig the differece of these two spectra to equalize the effects of fiite laser pulse duratio ad photodetector badwidth. The detectio badwidth of the P polyer icrorig ultrasoud detectio device was over 9 MHz at db..5 N o r a l i z e d A p l i t u d e.5 a Tie s Acoustic sigal Laser pulse M a g i t u d e d B - b - Acoustic sigal Laser pulse Frequecy respose of POUD Frequecy MHz Figure 5.7 a Acoustic sigal detected by a polyer icrorig resoator ad laser pulse profile detected by the photodetector. b Spectra of the sigals ad frequecy respose of the resoator. The detectio badwidth of the resoator was over 9 MHz at db. 75

91 5.5 Device s perforace Iproveet 5.5. Fabricatio process to iprove Q factor Electro bea lithography a Electro-bea resist PMMA Silico o xide layer Silico substrate b PMMA Developet c Silico Oxide Etchig d Silico Etchig Figure 5.8 Scheatic of typical silico old fabricatio process I order to fabricate high Q factor polyer icro-rigs usig ao-iprit techique, a old with sooth sidewalls is eeded. May aterials such as glass, polyer, etal, ad silico have bee exaied for old applicatios. Aog these aterials, silico is the ost attractive aterial cadidate due to the excellet processibility of silico ad a wide rage of fabricatio processes developed by the itegrated circuit IC idustry. Especially dry etchig of silico has bee show to achieve sooth ad vertical sidewalls [8], which is ideal for our applicatio. Our silico old is fabricated usig electro bea lithography followed by reactive io etch IE. Two iportat odificatio steps are used to further sooth the silico old sidewalls. The fabricatio starts with a silico 76

92 wafer with a iitial 4 theral oxide layer. Electro bea lithography is used to create a patter with a icro-rig ad a straight bus waveguide o a 8 thic positive electro-bea resist 95 PMMA. The patters are the trasferred to a silico oxide layer usig IE with PMMA as the etch as. After this step the PMMA as is reoved i hot acetoe. The icro-rig patter is further trasferred oto the silico wafer by deep silico etchig process with the silico oxide layer as etch as. After deep silico etchig, the silico oxide asig layer is reoved by buffered hydrofluoric acid BHF, which copletes the aster old fabricatio the scheatic of the fabricatio process is showig i Figure 5.8. Two iportat steps have bee added to i this silico old fabricatio process for gettig sooth sidewalls. a c b Figure 5.9 Sidewall SEM iage of the polyer icro-rig fabricated fro the old: a without resist reflow process, b with resist reflow process, c with resist reflow ad theral oxidatio process. The first iportat step is the PMMA resist reflow, which is applied before the silico oxide etch step. By choosig a suitable teperature ad tie duratio, this reflow process ca greatly reduce iperfectios i the PMMA resist patters ad harde the 77

93 edge of the PMMA resist. Too low of a teperature will ot cause the resist to reflow, while too high of a teperature will cause deforatio i the couplig gap regio, which could sigificatly icrease the optical loss. After a uber of experiets, the appropriate PMMA resist patters reflowig teperature ad tie duratio was deteried to be 5 o C for 9 secods. Figure 5.9 show the SEM pictures of iprited polystyree icro-rig waveguide sidewalls by usig the silico olds with 5.9b ad without 5.9a PMMA resist reflow process, respectively. We ca clearly see fro Figure 5.9a that there is huge roughess o the sidewall of the polyer icro-rig, which is thought to be due to the daage caused by IE o the edge of the PMMA patter durig etchig of SiO, which evetually gets trasferred oto the silico old. However, i Figure 5.9b, the sidewall of the polyer icro-rig fabricated fro the old with PMMA reflowig process has relatively sall vertical roughess, which eas that the edges of the PMMA are well protected durig the IE process. Figure 5., High agificatio SEM picture of the sidewall roughess The secod iportat step is the theral oxidatio followed with BHF etchig step, 78

94 which was used to sooth out the roughess after deep silico etchig. The oxidatio step is perfored i a high teperature furace that grows aroud theral oxide o the silico surface. As the oxidatio cosues about 44% of Si, the rough Si surface layer is coverted to SiO i this step ad reoved by BHF etchig after the oxidatio step. Figure 5.9c shows the sidewall of the iprited polyer icro-rig usig the silico old after theral oxidatio ad BHF etchig step. We see a ajor iproveet i the sidewall roughess copared to Figure 5.9b. Most of the sharp ripple-lie roughess ad sall holes o the sidewall see i Fig. 5.9b have disappeared i Figure 5.9c ad figure Loss characterizatio The Q factors of waveguide-coupled icro-rigs ca be classified as two copoets: Q itrisic ad Q couple, where the forer is the Q of a isolated resoator ad the latter taes ito accout of the couplig loss to the bus waveguide. Q itrisic is priarily liited by the optical losses i a isolated resoator, ad ca be attributed to these loss echaiss: radiatio loss, surface scatterig loss, aterial absorptio loss. Therefore the icro-resoator s overall Q factor ca be expressed as: Q total, 4- Q Q Q Q Q Q itrisic couple rad scatt abs couple where Q rad, Q scatt, ad Q abs are radiatio loss-related Q, surface scatterig loss-related Q, ad absorptio loss-related Q, respectively. The total loss-related Q ad couplig related Q ca be extracted fro the fitted trasissio spectru[8], the radiatio related Q ca be obtaied fro the siulatio, the absorptio loss-related Q ca be easured fro a 79

95 theral bi-stability effect will be described i ore detail below, ad fially the scatterig related Q ca be extracted fro equatio 4-. Figure 5.a ad 5.b show the trasissio spectru of polyer icro-rig iprited fro the silico old without ad with resist reflowig process, respectively. Figure 5. a shows the total Q of aroud 4, ad Figure 5.b shows the total Q of aroud 4. The aplitude atteuatio factor ca be obtaied fro icro-rig trasissio equatio, which ca be expressed as: T a a a cos a cos, 4- where is aplitude self-couplig coefficiet, a is the aplitude atteuatio factor, ad is the roud trip phase. The aplitude atteuatio is due to the various optical losses i the icro-rig, ad therefore ca be used to obtai the Q itrisic. The aplitude atteuatio factor a is related to the itrisic Q fro the followig equatio[8] : Q it 4 eff risic 4- l a where is the radius of the icro-rig, eff is the effective refractive idex of the resoace ode ad is the resoace wavelegth. 8

96 . Noralized itesitya.u. Noralized itesitya.u a Wavelegth b Noralized Itesitya.u c Wavelegth Wavelegth Figure 5..Trasissio spectru of polyer icro-rig fabricated fro the old a without resist reflow process, b with resist reflow process, c with resist reflow ad theral oxidatio process. All the blac dot curves are experietal data ad red lie curves are Loretz fittig cure By fittig with equatio 4-, we ca get the aplitude self-couplig coefficiet =.99 ad the aplitude atteuatio factor a =.95 for the icro-rig device fabricated usig the old without resist reflow process. The calculated itrisic Q accordig to Eq 4- is aroud. 4, correspodig to a propagatio loss of aroud.7 db/c. For the device created fro the old with resist reflow process, we fid that the aplitude self-couplig coefficiet =.996 ad the aplitude atteuatio factor a =.975. The 8

97 calculated itrisic Q is aroud.5 4, correspodig to the propagatio loss of 7.7 db/c. Therefore the resist reflow process ca reduce the optical propagatio loss by a factor of. Figure 5.c shows the trasissio spectru of the waveguide coupled icro-rig devices fabricated fro the old ade with the process icludig the resist reflow ad theral oxidatio steps. The total Q is fitted to be aroud. 5, ad the aplitude self-couplig co-efficiet = ad the aplitude atteuatio factor a =.994. The itrisic Q is calculated ~.5 5, which represets propagatio loss of.8 db/c. Theral oxidatio step further helps with reducig surface roughess leadig to a total reductio of propagatio loss by ore tha oe order of agitude. COMSOL ulti-physics software was used to siulate the radiatio loss of the polyer icro-rig devices [84]. By taig the advatage of the axial-syetry of the resoace odes i the icro-rig, the -diesioal eigevalue proble ca be trasfored to a equivalet -diesioal proble. The exact diesios of icro-rigs ad the refractive idex of polyer, which are used i the siulatio, are obtaied fro scaig electro icroscope SEM ad spectroscopic ellipsoetry. By solvig the eigevalue proble i COMSOL, we obtaied the eige frequecy for the trasverse electrical ode s i the icro-rig with real part equal to ad iagery part equal to Hz. Therefore the radiatio loss-related Q = e/i [85] is

98 with iput power w with iput power 5w with iput power w with iput power w with iput power w Wavelegth Figure 5..Trasissio spectra of polyer icro-rigs with differet iput power To deterie the absorptio loss-related Q factor i the polyer icro-rig, we adopted a ethod based o the theral-istability pheoeo of the polyer resoators. We easured the trasissio spectru for differet iput powers, ad the results are show i Figure 5.. Whe icreasig the iput power, we foud that the trasissio spectra exhibit two otable chages: resoace pea shift to a shorter wavelegth, which is aily due to the egative value of the opto-theral coefficiet of the polyer, ad distortio of resoace lie shape, which is due to the aterial absorptio iduced theral bi-stability effect [86]. The absorptio loss-related Q ca be extracted fro the liear relatio betwee the iteral cavity eergy ad absorbed power by followig the ethod described i referece 86. Assuig steady-state coditio, the absorbed power i the cavity ca be expressed P abs =T/ th, where T is the teperature chage fro the cavity, ad effective theral resistace th, which icludes the theral resistace of heat si fro the polyer cavity to abiet ad theral resistace of 8

99 cavity itself, was odeled usig COMSOL ulti-physics to be 9.5 W/K. The absorbed power the ca be expressed as: P abs d dt th, 4-4 where is the refractive idex of the polyer, is the resoace wavelegth, d/dt is the thero-optical coefficiet, is the theral expasio coefficiet ad is the resoace wavelegth shift. Figure 5. shows the power depedece of absorptio effect of the polyer icro-rig. The itra-cavity eergy is calculated accordig to the referece 6. The liear absorptio coefficiet ca be extracted to be Hz, which ca be used to calculate the absorptio loss-related Q =. 5 [87]. This correspods to a aterial absorptio loss of.db/c for polystyree used for our devices, which is cosistet to the published data [88]. 5 Absorbed powerw Itra-cavity eergyfj Figure 5..Trasissio spectra of polyer icro-rigs with differet iput power Fially, the surface scatterig loss-related Q ca be calculated fro the equatio 4-84

100 ad is ~. 6, ad the surface scatterig loss ca be show as low as. db/c. Fro this aalysis, we ca ow that the doiated loss i polyer icro-rigs at.55 wavelegth rage is the aterial absorptio loss ~.db/c, which is attributed to the carbo-hydroge bods haroic absorptio at ear-i rage. Such absorptio loss ca be iiized by replacig H with F atos. Fortuately, at visible wavelegth rage, polyer aterials ca have the absorptio loss as low as.4db/c [89]. Therefore, we believe that our polyer icro-rig s total Q ca be greatly icreased by ovig the worig wavelegth fro NI to the visible rage Acoustic sesitivity A MHz ufocused trasducer V6, Paaetrics NDT, Waltha, MA was used to characterize the ultrasoic sesitivity of the high Q polyer icro-rig. Durig the experiet, the icro-rig device is ierged i de-ioized DI water used as a couplig edia for acoustic wave. The device s trasissio spectru Figure 5.4a was easured i DI water with a Q factor of 4 ad couplig cotrast of 9% at resoace wavelegth. The itrisic Q of the polyer icro-rig i water is fitted to be 4 4, which is early 4 ties lower tha that i air. This reductio of Q is aily due to the icrease i bedig loss due to reduced refractive idex cotrast betwee the waveguide core ad the water claddig ad absorptio loss by the surroudig DI water. The trasducer is drive by a V pea-to-pea oe cycle MHz siusoidal wave, outputtig a pea pressure of Pa, which was calibrated by a coercial hydrophoe. The laser iput was set at wavelegth ad the power at 5 W. Whe the acoustic pressure pulse is hittig o the polyer icro-rig, it odulates the resoace 85

101 wavelegth, thereby the output power at fixed probig wavelegth. Figure 4.4b shows the recorded sigal trace fro a sigle-shot acoustic wave. The device produces a output of 89 V with a iput of Pa acoustic pressure, which eas the device s acoustic sesitivity is aroud 6. V/Pa. This is. ties larger tha what we recorded before usig a device with Q of ~6 ad the sae easureet syste, icludig the laser, photodetecor, ad oscilloscope. The root-ea-square oise levels were.8,.7, ad. V for 5, 5, ad 75 MHz badwidths, respectively. Thus, the correspodig NEPs are 5, 74, ad 88 Pa for 5, 5, 75 MHz badwidths, respectively. Therefore we have iproved the NEPs early three ties as copared with our previous best result of Pa, which represets the highest sesitivity ultrasoud trasducer of siilar physical size.. Noralized itesitya.u Wavelegth a b Figure 5.4 Trasissio spectru of polyer icro-rig ierged i DI water. b Sigle shot of acoustic wavefor easured by high Q polyer icro-rig 86

102 5.6 Coclusio A siplified fabricatio process has bee deostrated to fabricate polyer icro-rig with ore cotrollable couplig gap ad roughess. This ewly fabricated device has show the Q factor aroud 6. By usig it as a ultrasoud detector, a oise equivalet pressure of.pa over -75MHz ad a detectio badwidth of over 9MHz at -db were easured. By further iprovig the fabricatio process usig resist reflow ad theral oxidatio ethods, we have successfully fabricated a silico old with sooth sidewalls. The iprited devices ca achieve a itrisic Q factor as high as.5 5 i air, which correspods to optical propagatio loss of.8 db/c. Due to icreased Q factor of the polyer resoator i water, we have obtaied a oise equivalet pressure of 88 Pa over -75MHz which has iproved aroud ties. Further reductio of NEPs is possible by aig uch higher Q polyer icro-rigs.. 87

103 Chapter 6 Ultra-high sesitive ultrasoic detectors usig ultra-high Q icrorig 6. Itroductio ad otivatio Further iprovig our polyer icrorig detector s sesitivity to lower dow device s detectio liit as low as tes of Pascal will greatly beefit to i vivo photoacoustic iagig where a low laser eergy is required for safety reasos [9]. I order to realize uch higher sesitivity, a polyer icrorig detector with ultra-high Q factor is desired. But polyer aterials high absorptio loss at ear-i rage, which is attributed by the carbo-hydroge bods haroic absorptio, will prevet us to achieve ultra-high Q factor. Oe possible ethod to iiize polyer aterial s absorptio loss is to replace hydroge ato with fluorie atos [9]. But by dopig fluorie ato i the polyer, polyer aterial s refractive idex will get decreased which will liit its applicatio to for a copact device. Fortuately, at visible or ear visible wavelegth rage, polyer aterial ca have the very low absorptio loss. Polyer aterial, such as PMMA, has bee used to fabricate polyer optical fiber worig aroud 8 wavelegth because of its low loss. Therefore, we believe that our polyer icro-rig s 88

104 total Q ca be greatly icreased by ovig the worig wavelegth fro NI to the visible or ear visible wavelegth rage. Aother advatage to ove the worig wavelegth to ear visible is that it ca allow uch sall device size to realize siilar Q factor copared to ear I wavelegth rage, which is good for high frequecy ultrasoud detectio. To detect high-frequecy ultrasoud wave, a detector with both widebad respose ad sall eleet size is eeded. Sall device size iiizes the spatial averagig effect for high-frequecy waves, which is essetial for high-resolutio iagig. For exaple, phased-array iagig systes worig at a ceter frequecy of MHz require / eleet size ad spacig o the order of 5, where is the acoustic wavelegth. Aother exaple is that the sall device size for toographic iagig provides high resolutio ad high cotrast over a large iagig regio [9, 9]. Although the piezoelectric aterial polyviylidee fluoride PVDF based eedle hydrophoes ca reach the requireet of wide badwidth ad sall eleet size e.g., 4 HPM4/, Precisio Acoustics, Dorchester, Dorset, UK, the device lacs sufficiet sesitivity: the oise-equivalet pressure NEP is relatively high ~ Pa, ad liits the iagig depth. Moreover, arrays with sall eleet size ad spacig ad large eleet cout, required for real-tie iagig, are very difficult to realize usig piezoelectric trasducers because of the icreased oise level, coplexity of electrical itercoects ad fabricatio challeges. Optical icrorig detectio of ultrasoud could potetially address the above issues. It ca achieve a low NEP with relative sall eleet size ad widebad respose, ad would be easier to create dese arrays with sall eleet size. 89

105 6. Devices siulatio ad fabricatio Figure 6. shows the E field itesity of the polyer icrorig with = at resoace wavelegth aroud 8.The polyer icrorigs with refractive idex aroud.585 with width ad height are all equal to ad it is sittig o the silica substrate with refractive idex =.45 with water as top claddig. Fro the figure, we still ca see that soe of the field is leaig dow to the substrate regio. The siulated radiatio loss liited Q ca be above 7, which is uch higher tha oral absorptio loss liited Q. so we thi the desig the structure is good eough to realize higher Q factor. Figure 6. Siulated E filed itesity distributio of = polyer icro-rig at resoace wavelegth aroud 8 with botto ad top claddig are silica ad water. 9

106 By switchig worig wavelegth to a short wavelegth, oe potetial proble is that it will icrease the surface scatterig. This is because that the scatterig is 4 proportioal to. By switchig the worig wavelegth fro.55 to.8u, it could potetially icrease the surface scatterig by 6 ties. So further iprovig the device s sidewall roughess is desired. Also pervious we have deostrated to use E-bea resist as a etch as to etch silico oxide, the use silico oxide as a etch as to etch Silico. The proble is that the silico oxide etchig recipe requirig higher plate bias above 5W which will daage the resist as durig the etchig process. Evetually, it will cause the roughess o the silico old. I order to solve this proble, we develop a ew etchig process by oly usig e-bea resists as a etch as. Figure 6. sidewall view of the silico old with ew recipe First resist reflow [86] process has bee applied to harde the edge of the resist patter to prevet potetial daage durig the reactive io etch process. By choosig suitable teperature ad tie duratio, this reflow process ca greatly reduce iperfectios i the PMMA resist patters ad harde the edge of the resist. Optial teperature ad heatig 9

107 tie were foud to be 5 o C ad 9 secods, respectively. Next the resist patters were trasferred ito the silico layer usig SF 6 gas ad C 4 F 8 gas based iductively coupled plasa ICP reactive-io etch IE o a Surface Techology Systes STS deep silico etcher syste. A low plate bias was used i the etchig recipe to reduce the daage to the reflowed PMMA as ad the SF 6 gas ad C 4 F 8 gas flow were optiized to iiize the sidewall roughess [94]. Figure 6. shows the etched silico trech with the resist reaied o top of the silico trech. After IE, the rest of PMMA as is reoved i hot acetoe. A cobiatio of resist reflow ad odified Bosch process for Si etchig are the eys to produce silico aster with sooth sidewalls a b Figure 6. a SEM iage of the polyer icrorigs with =, b SEM iage of the sidewall of polyer icro-rig The detail of usig the aoiprit process to create polystyree icro-rigs ca be foud i referece. Figure 6.a shows the top view of a scaig electro icroscopy SEM picture of a polystyree icro-rig resoator with a diaeter D = 6 coupled to a straight waveguide. Figure 6.b shows the SEM picture of the sidewall of the iprited polyer icro-rig. The soothess is draatically iproved as copared with our previous results. The device s trasissio spectru was easured 9

108 usig a tuable laser New Focus TLB-6 with tuig rage fro A sigle ode fiber Nufer 78-HP ad a covetioal ultiode fiber were aliged to the iput ad output waveguide, respectively. The polarizatio of the iput light was cotrolled by a fiber based optical polarizatio cotroller; ad for cosistecy the iput light was fixed as TE polarizatio. The output light sigal fro the ultiode fiber was easured by a photo-detector. The easured trasissio spectru i de-ioized water i Fig.6.4 shows a sharp resoace dip with a Q factor of 4 5. The device s itrisic loss of. db/c is extracted fro the trasissio spectru, which iplies a itrisic cavity Q factor of The bedig ad leaage loss liited Q is obtaied to be 7 ad the aterial absorptio liited Q ca be also as high as 7, so we believe that the Q factor of our curret 6 polyer icro-rigs is still liited by the surface scatterig loss, which ca be further iproved i the future.. Itesitya.u Wavelegth Figure 6.4 The trasissio spectru of polyer icrorig with = 9

109 6. Acoustic sesitivity..8 Itesitya.u X Voltage V a -.8 b Wavelegth Tie s Figure 6.5 a Trasissio spectru of polyer icro-rig ierged i DI water. bsigle shot of acoustic wavefor easured by high Q polyer icro-rig The experietal setup used to characterize the device s acoustic sesitivity is the sae as i the last chapter. The icro-rig resoator detector was iersed i de-ioized water which served as top claddig of the icro-rig ad couplig ediu for acoustic wave. The output wavelegth of light fro the tuable laser was set to where the resoace trasissio curve has the highest slope i figure 6.5a. The laser power fro the iput fiber was about W ad a ~ W laser power was coupled ito the iput waveguide. The output power was collected by a ultiode fiber which is coected a low-oise photodetector New Focus, 8-FC. The photo-detector output is coected to a oscilloscope to collect the data. A MHz ufocused trasducer V6, Paaetrics NDT, Waltha, MA, drive by a 5 V pea-to-pea oe cycle MHz siusoidal wave with a output pea pressure of 5 Pa, was used to isoify the 94

110 icro-rig. Whe the acoustic pressure pulse hits o the polyer icro-rig, it odulates the resoace wavelegth ad thereby the output power at fixed probig wavelegth. Figure 6.5b shows the recorded sigal trace fro a sigle-shot acoustic wave. The device produces a output of V with a iput of 5 Pa pea acoustic pressure, which eas the device s acoustic sesitivity is aroud 66.7 V/Pa. The root-ea-square oise levels were.7,. ad.4 V over -5, -5 ad -75 MHz badwidth, respectively, leadig to NEPs of.5, 5. ad.4 Pa at the correspodig badwidth. Copared to our published best results, 88 Pa over -75 MHz, we have further iproved the NEPs by over 4 folds. This result is 6 ties better tha the best Fabry-Perot cavity based optical ultrasoud detector [95] ad ties better tha the siilar size piezoelectric polyviylidee fluoride PVDF trasducer HPM75/, Precisio Acoustics, Dorchester, Dorset, UK. Sice our siulatio results show that the device s itrisic Q ca be as high as 7, we aticipate that the device s NEP ca reach sigle digit Pa by further iprovig the fabricatio ad icreasig the device s Q factor. Detectors with such ultra-low NEPs will directly beefit the ultrasoud ad photoacoustic iagigto sigificatly icrease the iagig depth. It also pushes the photoacoustic iagig towards cliical applicatios o huas because uch lower laser fluece is required. 95

111 6.5 Agular respose Noralized Aplitude a 5 blac bead i gel Nd:YAG pulse laser Oscilloscope y Tuable Laser Photo-detector Polarizatio xy ovig stage x Cotroller Wate r.. b Agular respose of rig D=6 c Agular respose of rig D= MHz MHz Noralized Aplitude MHz 4MHz Agledegree Agle degree Figure 6.6 a Experietal setup for easure the agular respose of the polyer icro-rig, experietal data dot ad theoretical calculatio lie of agular respose of the polyer icro-rigs with D = 6 at MHz solid dot ad lie ad MHz epty dot ad dash lie b, ad with D = 4 at MHz solid dot ad lie ad 4 MHz epty dot ad dash lie c The agle-depedet sesitivity of the polyer icrorig was characterized usig a photoacoustic ethod. Fig. 6.6 a shows the scheatic of the experietal setup. A 5 polystyree blac bead ebedded i the gel was illuiated by a pulsed, frequecy-doubled Nd:YAG laser at 5 wavelegth with pulse duratio of 6 s Surelite I-, Cotiuu, Sata Clara, CA. The eergy fro the pulsed laser was efficietly absorbed by the blac bead, geeratig a spherical acoustic wave by the optoacoustic effect. The photoacoustic wave was the detected by the polyer icrorig detector at a distace of about fro the bead. The detectors, with diaeters of 6 96

112 ad 4 detail will show i ext sectio, are liearly scaed to receive the photoacoustic sigals at differet agles. Badpass filters cetered at differet frequecies are applied to the recorded sigals at differet agles to extract the sigal levels at those frequecies. The theoretical agular respose ca be described by cosiderig a rig trasducer: D J asi [96], where is the wave-vector of the icidet acoustic wave, a the radius of the rig trasducer ad the icidet agle of the acoustic wave. Fig. 6.6 b ad c show the theoretical ad experietal agular resposes of the polyer icrorig with a diaeter D = 6 ad 4, respectively. The theoretical curves fit well to our experietal data. Followig siilar calibratio used for 6 icro-rigs, the Q factor of 7 4 with water claddig ad a NEP of aroud. Pa for 4 icrorigs were obtaied. The worse NEP is because the bedig-iduced loss becoes doiat for 4 case. For bea-forig applicatios, the detector s agular respose should have -6 db beawidth of 4 o. Uder this criterio, the icrorig detectors with D = 6 ad 4 ca be used as a iagig eleet for acoustic cetral frequecy of ad MHz, respectively. Copared to our previous polyer icrorig hydrophoe with D =, worig aroud MHz rage, the ew detectors have doubled ad tripled the frequecy rage with better or siilar sesitivity, correspodig to resolutios of 75 ad 5, respectively. The sall size device will also beefit the toographic iagig, eablig a uiforly high resolutio ad high cotrast over a large regio of iterest. The details of high-resolutio photoacoustic iagig applicatios usig the ultra-low oise sall size icrorig detectors ca be foud i our recet wor. 97

113 6.6 Saller size polyer icro-rigs for high frequecy iagig Itesitya.u a Itesitya.u b Wavelegth Wavelegth Figure 6.7 a The trasissio spectru of polyer icrorigs with = coarse sca, b the trasissio spectru of polyer icrorigs with = fie sca. As we have show i the pervious sectio that saller size icro-rig detectors are good for high frequecy ultrasoic iagig applicatios. This is because that sall size device has larger agular respose. By usig sae fabricatio techiques that have used i the previous sectio, we have fabricated = polyer icrorigs with reasoable Q factor ad sesitivity. Figure 6.7a shows the trasissio spectru of the =u polyer icrorig at coarse scaig ode. Ad figure 6.7b shows the trasissio spectru of oe resoace ode at fie scaig ode. We ca fid that the device s Q factor is aroud 7x 4 i the water. The reductio of the Q is aily due to the icreasig of the bedig loss at saller device size. But with curret Q factor, it is already good 98

114 eough to realize sesitivity high frequecy iage with detector NEP aroud.pa ad larger agular respose up to MHz frequecy..4 Itesitya.u Wavelegth Figure 6.8 The trasissio spectru of polyer icrorigs with = coarse sca o 4 theral oxide wafer I order to further icrease the acoustic iage resolutio, a icrorig acoustic detector with a radius as sall as has bee developed, but with its copact size, it will suffer uch larger leaage loss copared to large size device. The trasissio spectru of the iprited polyer icrorig with device s = has bee easured ad show i the figure 6.8. We ca clearly see that after reducig the device s size, the free spectru rage of the resoator is icreased ad oly oe resoace ode is showig up withi wavelegth bad-width. Ad the Q factor is ideed uch worse tha the device with radius =u ad it is aroud 4-5x i the water which traslated to optical loss aroud 9dB/c. With this Q factor 4-5x, we ca estiate the device s pressure detectio liit is aroud.5pa which is oe order agitude higher tha the device with radius =u. This will liit the device s applicatio i high sesitivity iagig applicatios. 99

115 Itesitya.u..8x -5.6x -5.4x -5.x -5.x -5 8.x -6 6.x -6 4.x -6.x Wavelegth Figure 6.9 The trasissio spectru of polyer icrorigs with = o substrate with 4 thic HSQ fil o 4 theral oxide wafer. Oe way to iprove the device s Q is to use a uch higher refractive idex polyer aterial as the waveguide s core aterial, such as Polychloro-p-xylee =.69, Polyviyl pheyl sulfide =.6568, PolyN-viyl carbazole =.686. But the proble is that those aterials orally have very high glass trasitio teperatures above 8 o C, which will prevet us to use the theral ao-iprit process to fabricate cirorig devices. Also it is really hard to for a relative thic fil above 5 by oral spi coatig ethod because of its low dissolutio rate i the solvet. Therefore a covetioal waveguide fabricatio process, such as a photolithography cobied with a dry etchig process, will ot fit for fabricatig those high refractive idex polyer waveguides. Aother way to icrease the device s Q factor is to reduce the refractive idex of the botto substrate. A couple of low refractive idex aterials ca be cosidered as the cadidates, such as HSQ =.4, Cytop =.4, SSQ =.4. After

116 a few tests, we choose HSQ aterial as the low idex aterial i our experiet. This is because the aterial s property is very stable after it is theral cured. The trasissio spectru of =u polyer icrorig fabricated o substrate with 4 HSQ fil o 4 theral oxide is show i the figure 6.9, we ca see that resoace dip s liewidth is uch arrower tha liewidth of the resoace dip i figure 6.8 ad the Q factor of the device i the water ca reach x 4 which is 5 tie higher tha the oe without 4 HSQ fil i the botto substrate. Air Water Theral oxide Theral oxide oxide a b Figure 6. Siulated field distributio of =u polyer icrorig o theral oxide substrate i the air a ad i the water b Aother iterestig way to reduce the device s bedig loss is to hybrid the device with a etal. But directly fabricate the waveguide o the etal will cause too uch loss to guide ode i the waveguide, oe way to reduce the loss fro the etal is to itroduce a low idex thi buffer layer betwee the waveguide aterial ad etal. Figure 6. shows the siulated the field distributio i the air a ad water b of the =u polyer icrorig o theral oxide substrate. We ca clearly see fro figure 6.a

117 that there is a large out of optical field is leaig dow to the substrate which is aily due to the low idex cotrast betwee the polyer ad theral oxide. After the device is erged i the water, we ca fid fro figure 6.b that the leaage to the substrate is gettig icreased ad also the leaage to the water also starts to show up. Ad the siulated Q i the air ad water are x 4, 4x, respectively. Figure 6. shows the siulated results of the field distributio i water ad air of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 low idex buffer layer. As we have aticipated that the etal fil really ca help to reduce the leaage field to the botto substrate, but it could icrease the etal related absorptio loss due to the presets of the etal fil. The siulatio shows that the after addig etal ad low idex buffer layer the Qs have bee i iproved tox 5 ad x 4 i the air ad water, respectively. Air Water Metal layer Buffer layer Silico a Silico b Figure 6. Siulated the field distributio of i water ad air of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 HSQ buffer layer

118 Itesitya.u... Itesitya.u Wavelegth a Wavelegth b Figure 6. Trasissio spectru of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 HQS buffer layer i the air a ad i the water b. The trasissio spectru Figure 6. of f =u polyer icrorig o the silico substrate with a etal fil ad a 4 low idex buffer layer has bee easured i the air ad i the water, respectively. I the air the device s Q factor is aroud x 5 ad i the water the Q factor is aroud x 4, which atched well with the siulatio results fro the COMSOl. This etal hybrid with low idex aterials substrate has also bee characterized to act as a substrate for ultrahigh Q device which has bee fabricated o our previous sectios i this chapter. Experietal easured device s Q i water is aroud x 5 Figure 6. which is very close to the result we got fro the device o 4 theral oxide substrate. This tells us that etal hybrid with low idex polyer substrate ca have the sae or better cofieet property copared to 4 theral oxide substrate. Aother iterestig thig is that it ca be fabricated o various substrates, such as flexible plastic fils ad etal blocs. This aes it easy to fabricate low optical loss devices o flexible or etal substrates.

119 4.x -6 4.x -6 Itesitya.u..8x -6.6x -6.4x -6.x -6.x -6.8x Wavelegth Figure 6. Trasissio spectru of =u polyer icrorig o the silico substrate coated with a etal fil ad a 4 low idex buffer layer i the water 6.7 Photoacoustic icroscopy usig polyer icrorigs With our curret polyer icro-rig resoators detector s very widebad respose ad high sesitivity, we hope that icro-rigs ca provide high axial resolutio i photoacoustic icroscopy. I the experiet, a thi fil optical resoator is used to replace the piezoelectric acoustic receivers. The iheret broad bad frequecy respose of the optical thi fil structure to acoustic waves, cobied with a focused optical excitatio bea ca produce both axial resolutio ad lateral resolutio coparable to that achieved i optical icroscopy. The left pael i Figure 6.4 presets the scheatic of photoacoustic icroscopy PAM setup based o the highly sesitive broad badwidth icrorig resoator. The resoator has rig-shaped for coupled with a straight waveguide servig as optical iput 4

120 ad output. Such a icrorig detector has bee show to have a alost flat bad respose up to ~ MHz. The icrorig resoator fabricated o a silico chip was covered by a Mylar protective layer to scree out the excitig laser bea fro a Nd:YAG laser Spot---5, Elforlight Ltd, UK. The Nd:YAG laser worig at 5 wavelegth has a pulse duratio of s ad a repetitio rate P of KHz. The laser light was spatially filtered by a iris ad the expaded to a parallel bea which was rastered over the tissue object by D Galvaoeters. The itesity ad the stability of the laser bea was oitored ad calibrated by a photodiode DETA, Thorlabs, NJ. A achroatic les with a focal legth of 5 was used as the objective les. O the Mylar protective layer, oe couplig pad layer was used to optiize the couplig of the b photoacoustic sigal fro the saple ito the icro-rig resoator detector. A tuable laser TLB-6, NewFocus, CA provided the light source for the icro-rig resoator at a wavelegth tued to the axial slope of the resoace pea of the icro-rig s trasissio spectru. A low oise photodiode 8-FC, New Focus, CA was used to record the chage of the itesity of the light through the icro-rig resoator which reflects the wavefor of the photoacoustic sigal. The photo-detector has a DC output gai of V/A ad AC output gai of 4 V/A with oial -6dB electrical badwidth of 5Hz-5MHz. Usig the DC output, the icrorig s trasissio spectru ca be easured. Throughout the experiet, a coercial calibrated Oda trasducer HNC-5, Oda, CA with -db badwidth of Hz- MHz was utilized to realize covetioal PAM as a cotrol to evaluate the perforace of the PAM. The PAM with Oda trasducer shared the sae optical focusig ad scaig copoets with icrorig based PAM. The Oda trasducer operated o a reflectio ode at a sae 5

121 distace fro the target as the icrorig resoator worig o a trasissio ode. Figure 6.4 left pael scheatic of a icrorig based PAM syste based o a icrorig resoator. iddle pael axiu aplitude projectio MAP iage of the USAF resolutio teplate group 7. right paels A-lie sigals alog the Z axis of the iages of the USAF resolutio teplate with a icrorig based PAM right upper pael ad covetioal PAM with Oda trasducer right lower pael. The iset at upper left shows a scaig electro icrograph of a polystyree icrorig with radius used i this experiet. The icrorigs used i this study have a size of 6 i diaeter ad the polyer waveguides have a cross sectio of.4. A sigle-ode ad a ulti-ode optical fiber were aliged with the iput ad output of the straight waveguide, respectively, ad the fixed usig UV curable epoxy. Polyer icrorig device is favorable i photoacoustic icroscopy because of its ultra-low oise ad broad badwidth. A uch lower oise equivalet pressure NEP copared with other types of optical resoat structures [97] has bee deostrated ad further iproveet of NEP is possible by desigig uch higher-q device ad/or by couplig ore power ito waveguides. Fro optical poit of view, the detector s respose tie will be liited by the cavity s photo lifetie which is the tie required for eergy to decay to e its origial value ad is 6

122 give by Q [98]. Iverse of the respose tie will give the cut-off frequecy. So our curret device s frequecy respose ca be up to 6.7 GHz. Fro acoustic poit of view, a axiu odulatio frequecy of 57 MHz ca be estiated cosiderig the.4--thic PS waveguide ad the acoustic ipedace of the claddig, polyer, ad the substrate [99]. If assuig strog acoustic reflectios fro the rigid substrate ad egligible reflectios fro the claddig-polyer boudary, a approxiated forula gives a quic estiatio o the badwidth liit: si l P l, 6- l where P is the ea distributio of stress across the thicess l of the sesig fil due to a orally icidet plae acoustic wave with wave uber. The photoacoustic icroscopy with icrorigs has the potetial to achieve higher axial resolutio if ore broadbad sigals ca be detected, which requires a photodetector with higher frequecy respose tha used i the curret experiet. The lateral resolutio of the icrorig based potoacoustic icroscopy was easured by iagig a USAF resolutio teplate T--P-TM, Applied Iage Ic, NY. The iddle pael i Figure 6.4 shows the axiu aplitude projectio MAP iage of the resolutio teplate, where the 6 bar eleets of the group 7 ca be resolved with the gaps up to.9 with odulatio trasfer fuctio MTF value of 4%. Fittig the MTF to 5% yields a lateral resolutio of.5. The PAM with Oda trasducer shows the sae lateral resolutio because they shares the sae optical focusig ad scaig architecture which deteries the lateral resolutio of the syste. I tissue iagig, the lateral resolutio will deteriorate ildly due to the optical scatterig. However, whe the 7

123 saple is optically thi i.e. withi oe ea free path, the deterioratio of the lateral resolutio is isigificat []. To quatify the axial resolutio, typical A-lie sigals extracted fro the iages of the USAF resolutio teplate were used as approxiatios of axial poit-spread-fuctios PSFs. The right upper pael i Figure 6.4 depicts the axial PSF of icrorig based PAM; while the right lower pael i Figure 6.4 depicts the axial PSF of covetioal PAM with Oda trasducer. Accordig to ayleigh criterio, the PSFs show that this iitial experiet of icrorig based PAM based o the icrorig resoator provided a axial resolutio of 8, close to the lateral optical resolutio achieved; while PAM based o the Oda trasducer gave a uch larger axial resolutio of 5. Figure 6.5 MAPs o XY,XZ, YZ plaes of the ex vivo iages of the vasculature i a ouse bladder wall acquired with AOPAM upper row usig icrorig ad covetioal PAM usig Oda trasducer lower row. The bladder excised fro CD ouse CD, Charles iver, MA was iaged ex vivo. The MAPs Figure 6.5 of the iages of the vasculature of the ouse bladder 8

124 acquired with icrorig based PAM upper row ad covetioal PAM with Oda trasducer lower row show the sae views i the saple, sice the icrorig based PAM ad PAM shared the sae optical scaig ad the saple was ot oved betwee scas. MAPS o the XY plaes acquired with icrorig based PAM ad PAM reder a cosistet structure of the bladder vasculature with sae resolutios. The differece of the cotrast distributios betwee the is due to the differet positios of the icrorig resoator ad oda trasducer, oe worig i trasissio ode ad aother worig i reflectio ode. MAPS o the XZ ad YZ plaes acquired with PAM idicated a tail-trail fro icroscale vessels alog the axial directio, iitially thought to be due to the liited badwidth of the Oda trasducer. MAPS o the XZ ad YZ plae acquired with icrorig based PAM idicated o tail-trail effect. The icroscale capillary et with apparetly large vasculature hidig i it ca be clearly redered. All the iages i Figure 4. were acquired without averagig ad provig axiu cotrast to oise ratios of 6 db for AOPAM ad 9 db for PAM. The coercial Oda trasducer has a high sesitivity providig oise equivalet detectable pressure NEDP value of 9 Pa. The results of icrorig based PAM give a estiatio of the sesitivity of icrorig resoator of NEDP value of 9 Pa, coparable with Oda trasducer, uch higher tha other optical resoat structures available for acoustic easureet with NEDP o the order of agitude of hudreds of Pascal. Ad this NEDP value is quiet close to the characterize device NEP aroud Pa i sectio 6.. 9

125 6.8 Coclusio Ultra-high Q polyer icrorig device has bee realized by both switchig the worig wavelegth to ear visible wavelegth where polyer aterial has lower absorptio loss ad further reducig the device s sidewall roughess to iiize the optical scatterig loss. The device s Q as high as 4x 5 has bee easured with device s radius = ad agai the detector s NEP has bee characterized to be.4pa over -75MHz frequecy rage. We have iproved our device s NEP by over fourfolds. With our device s high sesitivity ad broadbad respose, it greatly iproved the photoacoutsic icroscope s axial resolutio to 8. Also high sesitivity, saller size devices =, ad = have bee developed to iprove the iage resolutio i bea forig iagig applicatios.

126 Chapter 7 Coclusio 7. Achieveet i high sesitivity bio-cheical sesors I this thesis, a pris coupled silica icro-tube bio-cheical sesor platfor has bee deostrated for the first tie to y owledge. This platfor ot oly overcoe the device s reliability proble by usig a thic wall silica icro-tube as a sesig eleet, but also realize high sesitivity detectio by picig up higher order resoace odes i the silica icro-tube resoator. Copared to other deostrated silica icro-tube sesors by usig fiber coupled ethod with a wall thicess aroud hudred aoeters to few icros, our platfor has realized high sesitivity sesig experiet with silica icro-tube wall thicess aroud. Eve thicer wall also ca be used i our sesig platfor because of the ature of the pris coupled ethod which allows us to choose differet orders of resoace odes i the icro-tube resoator. By tuig the icidet agle to pic up differet resoace ode which ca be used to realize differet sesitivity sesig fuctios. I the bul sesig experiet, we have successfully deostrated device s sesitivity aroud /IU at the icidet agle aroud 7 o ad this is typical belog to a sesitivity of a resoace ode with oly evaescet wave i the liquid regio low idex regio. The optical wave i this ode is well cofied at the ier ad outer

127 boudary of the silica icro-tube by the total iteral reflectio which results i the evaescet wave at the outside of the silica regio. By choosig a saller icidet agle, uch higher order ode ca be excited which could result i a uch larger sesitivity. At icidet agle at 5 o, we have observed a resoace ode with a sesig sesitivity aroud 6/IU which is represetig a resoace ode with ore tha 6 percet of the optical field located iside of the liquid regio. The optical wave i ode is still well cofied by total iteral reflectio at outer boudary, but due to the low idex cotrast betwee the silica ad water Such a high sesig sesitivity has ever bee observed i oral icro-tube based sesor systes except oe syste has bee deostrated usig extreely thi wall silica icro-tube fro zero to few hudred of aoeters. But the wet etchig process used to thi dow the wall thicess causes too uch roughess o the surface ad aes the device with very low Q factor. The silica icro-tube which was used i our experiet does t eed ay extra treatet ad it aitais the sae surface property as its coig out of the factory. With curret Q factor aroud x 4 ad high sesitivity aroud 6/IU, we ca clai that our device s refractive idex detectio liit is aroud.5x -6. Further iprove the Q factor by switchig the device s worig wavelegth fro ear I to ear visible or visible rag to reduce water absorptio ca greatly lower dow the device s refractive idex detectio liit to -8. I surface sesig experiet, first we characterized our device s bul refractive idex sesig property ad the cobied with the siulatio results to pic up oe resoace ode with relative high respose to the surface bidig. This resoace ode has bee used to sese the bidig of lipid oolayer, lipid bilayer ad self asseble layer-by-layer to the ier surface of the silica icro-tube. With 4-5 POPC lipid

128 ebrae with refractive idex aroud.46 absorbed o the ier wall, we ca observe the resoace pea shift aroud 44p. Ad i lipid oolayer experiet, we observed p resoace wavelegth shift which is the half of the shift i the lipid bilayer case. I the sesig electrostatic self asseble layer by layer experiet, the alterative coatig of PDDA ad Poly dye s-9 fil has bee observed with resoace wavelegth shift at 9p ad 4p, respectively. The reaso of this differece is that the refractive idex of poly dye s-9 is uch higher tha that of PDDA. Mie scatterig siulatio of the resoace pea shift due to the bio-cheical fil absorbed oto the ier wall agrees very well with the experiet results. We also observed resoace pea shift with the ebrae protei Aexi V ad Alaethici bodig to the lipid ebrae. With the preset Q factor 6 4 at.55μ wavelegth, we estiate that our devices ca detect the presece of. thic absorbed fil o the ier wall of the tube. The device's sesitivity ca be greatly ehaced by switchig the worig wavelegth fro ifrared wavelegth to visible wavelegth where water absorptio is iiized. I order to better uderstad the sesig property of the silica icro-tube, a theoretical study o various sesig properties, such as bulig refractive idex sesig property, surface sesig property, absorptio sesig property, have bee provided. We foud that the bul refractive idex sesig sesitivity icreases with the radial order uber i the evaescet sesig regie ad oscillates i the o-evaescet sesig regie. The o-evaescet ode is particularly suitable for bul refractive idex sesig ad the sesitivity ca achieve aroud 6/IU i our structure. The evaescet ode, havig a high electric field agitude at the ier boudary, is preferred i the surface sesig experiet with sesitivity as high as p/. Also, a

129 high Q resoator sesor is desirable for absorptio-based sesig. Also a coupled cavity platfor, which ca be realized by coatig a low refractive idex aterial at the ier surface of the silica icrotube, was proposed to achieve device s bul refractive idex sesig sesitivity above /IU. 7. Achieveet i high sesitivity ultrasoic sesors Aother iportat achieveet i this thesis is to further iprove polyer icro-rig resoators perforace i high sesitivity ultrasoic detectio. By cobiig polyer s high optical elastic coefficiet ad high deforability which ca provide a sesitive respose uder acoustic pressure with icro-rigs high Q factor, high sesitivity ultrasoic detectio has bee realized usig polyer icro-rigs fabricated by ao-iprit techique i our previous publicatios. But the fabricatio process which ivolved shallow old ad deep old fabricatio result i too uch variatios o the couplig gaps ad roughess o the sidewall of the rig waveguide. I order to ae ore repeatable process, we developed a siplified ethod to ae a silico oxide deep old with reasoable Q factor aroud 6. The siplified ethod replaced the lift-off process by directly etchig the etal fil with Cl gas based reactive io etchig process which ca geerated ore cotrollable gap ad reasoable sidewall roughess. With the ewly developed process, we ca repeatedly achieve our polyer icro-rig devices with Q factor aroud 6. By usig calibrated trasducer, our polyer icrorig detector has bee characterized with relative low NEPs aroud.4,., ad. Pa over -5, -5 ad -75MHz, respectively. The device s NEPs are coparable to 4

130 state-of-art piezo-electric trasducer with a siilar size. By usig a board bad acoustic source geerated by photo-acoustic ethod, the device s frequecy resposes of over 9 MHz at db were easured. Such a broadbad detector could greatly beefit the high resolutio ultrasoic iage. Further iproved the device s NEP could play a iportat role i obtaiig good iagig quality, especially for i vivo PA iagig where a low laser eergy is required for safety reasos. We have developed a ew fabricatio process to ehace the device s perforace by iprovig the Q factor of the polyer icro-rig. I this ew developed process, two iportat steps have bee added to reduce the sidewall roughess of the fabricated silico old. The first iportat step is the resist reflow process which could help to harde the edge of the resist ad reduce the daage to the sidewall durig the reactive io etchig process. The secod step is the theral oxidatio process followed by buffered HF etchig process, which will help to sooth out the roughess geerated o the reactive io etchig process. By applyig those two steps i the old fabricatio process, the iprited polyer icrorig resoators with Q as high as 5 has bee easured i the air. I the water the Q factor is aroud -x 4 ad the reductio of the Q is aily due to the water absorptio loss ad reduced refractive idex cotrast iduced bedig loss. The device s NEPs have bee easured aroud 5, 74, ad 88 Pa for 5, 5, 75 MHz badwidths, respectively. Therefore we have iproved the NEPs early three ties as copared with our previous best result of Pa, which represets the highest sesitivity ultrasoud trasducer of siilar physical size. We fid that the device s Q is liited by the aterials absorptio loss, after a detail aalysis of various optical losses i our curret devices. Further iproved the device s Q factor still 5

131 will be possible by switchig the device s worig wavelegth fro ear I to visible or ear visible wavelegth rage where both water ad polyer aterial have relative low absorptio loss. We ca reduce the water ad polyer aterial s absorptio loss by switchig the worig wavelegth to 78 wavelegth rage, however the optical scatterig loss will icreased because it is proportioal to / 4. I order to further iprove the device s Q factor, we ot oly have to switch the worig wavelegth, but also have to further reduce the surface scatterig loss. So developig a process to fabricate icrorigs with a very sooth sidewall is the ey to achieve a polyer icro-rig device with very low NEP. The developed fabricatio process icludes a resist theral reflow process which has bee used i our pervious devices fabricatio process ad optiized etchig process. I the etchig process, a low plate bias was used to reduce the daage to the reflowed PMMA as ad the SF 6 gas ad C 4 F 8 gas flow were optiized to iiize the sidewall roughess. The cobiatio of resist reflow ad odified Bosch process for Si etchig are the eys to produce silico aster with sooth sidewalls. The iprited polyer icrorig device with radius = has bee easured with Q aroud 4x 5 ad the device has bee tested with NEPs of.5 Pa, 5. Pa, ad.4 Pa for -5, -5, -75MHz badwidth, respectively. We have iproved our device s NEPs by over fourfolds. With our device s high sesitivity ad broadbad respose, it greatly iproved the photoacoutsic icroscope s axial resolutio to 8. Also the devices relative sall size =, =, ad = ca be used to iprove the iage resolutio to 75, 5 ad 5 i bea forig applicatios, respectively. 6

132 7. Future wor 7.. Ultra-sall device The sall size device will greatly beefit the toographic iagig, eablig a uiforly high resolutio ad high cotrast over a large regio of iterest. Usig the sigle-eleet low-oise sall size detector with liear traslatio stages, we have show high-resolutio photoacoustic iagig applicatios by sythetic oe-diesioal ad two-diesioal icrorig arrays. But it is really hard to ae a saller size polyer icrorig device < with relative high Q factor, because polyer aterials relative low refractive idex typically <.6 is ot good at cofiig optical wave whe it starts to bed too uch. But if we ca hybridize the polyer aterial with high idex aterials, such as silico =.5 or silico itride =., it will be very proisig to realize devices size as sall as =5 or =.5 with relative high Q factor. Figure 7. shows a silico slot waveguide hybridize with polyer aterial with bedig radius =5 ad we ca clearly see that optical field is well cofied eve with a 5 bedig radius. With such a copact device size, we could further iprove the bea forig iagig resolutio dow to few icros. 7

133 Figure 7. shows a silico slot waveguide hybrid with polyer aterial with radius =5 7.. Ultra-high Q factor device Oly PMMA has bee deostrated with very low absorptio loss ad polyer icro-resoator ade by PMMA ca achieve Q factor up to 8 at visible wavelegth rage. But due to PMMA aterial s relative low refractive idex =.49, it will be very difficult to fabricatio a copact size device with ultrahigh Q factor. But ultra-high Q factor ca be realized with devices size above 5. Aother proisig way is to ae a ultrahigh silico itride slot waveguide rig resoator covered by PMMA. By pushig the device s Q factor up to 6, we could reduce our devices NEPs dow to Pa which could greatly lower dow the required pup power i real photoacoustic iagig applicatios. 8