|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001917288
007 cr mnu|||uuuuu
008 071119s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002024
Evanescent field absorption sensing using sapphire fibers
h [electronic resource] /
by Michael Grossman.
[Tampa, Fla.] :
b University of South Florida,
Thesis (M.S.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 56 pages.
Advisor: Nicholas Djeu, Ph.D.
ABSTRACT: This thesis explores the application of coiled sapphire multimode optical fibers for evanescent wave chemical sensing in both the visible spectrum and the near infrared. As has been suggested in the literature pertaining to silica fibers, bending converts low-order modes to high order ones, which leads to more evanescent absorption and thus a more sensitive chemical detector. By coiling the fiber many times, it was expected that even greater sensitivity would be attained. Experiments were performed to investigate the sensor response to different solutions and to characterize this response. In the first of three experiments, the large absorption peak of water at 3m was examined in order to compare the sensitivity of a straight fiber versus a coiled one. In the second experiment, the effect of increasing the number of coils was investigated, as was the response of the sensor to varying concentrations of water in heavy water. In the third experiment, methylene blue dye was used to investigate the extent of adsorption of dye molecules on the sapphire fiber and its persistence.
t USF Electronic Theses and Dissertations.
Ev anescen t Field Absorption Sensing Using Sapphire Fib ers b y Mic hael Grossman A thesis submitted in partial fulllmen t of the requiremen ts for the degree of Master of Science in Ph ysics Departmen t of Ph ysics College of Arts and Sciences Univ ersit y of South Florida Ma jor Professor: Nic holas Djeu, Ph.D. Dennis Killinger, Ph.D. Myung K. Kim, Ph.D. Date of Appro v al: April 10, 2007 Keyw ords: b er optics, c hemical sensor, infrared, ev anescen t eld, sp ectroscop y c r Cop yrigh t 2007, Mic hael Grossman
DEDICA TION T o Dr. Jermaine L. Kennedy for setting the standard.
A CKNO WLEDGMENTS I w ould lik e to thank Dr. Djeu and the committee mem b ers for their v aluable time, Dr. Da vid Rabson for a crash course in L A T E X and the inimitable Da v e Sissler and Jimm y Gamez.
T ABLE OF CONTENTS LIST OF T ABLES ii LIST OF FIGURES iii ABSTRA CT v CHAPTER 1 INTR ODUCTION 1 CHAPTER 2 THEOR Y AND NUMERICAL CALCULA TIONS 7 2.1 The Beer-Lam b ert La w and Estimate of Ev anescen t Absorption 7 2.2 Deviation from Beer's La w 12 CHAPTER 3 EXPERIMENT AL METHODS 14 3.1 Near IR Exp erimen t: absorption of w ater near 3 m 14 3.2 Absorption of H 2 O in D 2 O 17 3.3 Meth ylene Blue Dy e Adsorption 18 CHAPTER 4 DISCUSSION OF RESUL TS 23 4.1 3 m IR Exp erimen t 23 4.2 1.9 m IR Exp erimen t 27 4.3 Meth ylene Blue Adsorption Exp erimen t 28 CHAPTER 5 CONCLUSION 33 REFERENCES 35 APPENDICES 37 App endix A Data for gure 4.6 38 App endix B Data for gure 4.8 41 App endix C Programs for SR510 Lo c kin Amplier 44 C.1 DOS and QBASIC 44 C.2 Lin ux and BASH 46 C.2.1 setSerialSignal.c 46 C.2.2 queryserial2.c 48 C.2.3 wrap2.sh 53 App endix D Oriel 77250 Mono c hromator Calibration and Use 55 i
LIST OF T ABLES T able 4.1 A tten uation Data (compared with theory in gure 4.11) 32 T able D.1 Mono c hromator Calibration Data 55 ii
LIST OF FIGURES Figure 1.1 Self-consistency: the t wice-rerected w a v e repro duces the original. 3 Figure 1.2 A high-order mo de mak es more rerections p er unit length. 4 Figure 2.1 Ev anescen t Absorption vs. Refractiv e Index for a b en t b er in w ater. Normalized to a long, straigh t 75 m b er with n cl = 1 : 3i and n c = 1 : 7. 8 Figure 2.2 U-shap ed region considered in equation 2.4 9 Figure 2.3 Angles Used In Determining The Eectiv e Cladding Index Of Sapphire Fib er In Air 10 Figure 2.4 Eect of Bend Radius on Sensitivit y ( = 75 m ) 11 Figure 2.5 Fib er absorbance as predicted b y P a yne & Hale and according to Beer's La w 13 Figure 3.1 Sc hematic blo c k-diagram of 3 m absorption exp erimen t 14 Figure 3.2 Nylon sp o ol, trimmed along dotted lines 16 Figure 3.3 Nylon blo c k and cuv ette 18 Figure 3.4 Sc hematic Blo c k Diagram of Meth ylene Blue Dy e Exp erimen t 19 Figure 3.5 Measuremen t of Meth ylene Blue Ev anescen t Absorption with coiled b er 21 Figure 3.6 Measuring the n umerical ap erture of a lens 22 Figure 4.1 Absorption Sp ectrum of Straigh t 39cm 75 m Fib er In Air and W ater 23 Figure 4.2 Absorption Sp ectrum of Tw o T urns of R = 1 cm 75 m Fib er 24 Figure 4.3 ln I air =I w ater vs. w a v elength: straigh t 75 m compared to 2 turns coiled R = 1 cm 24 iii
Figure 4.4 Absorption of W ater In The Near IR (after Hale & Querry) 25 Figure 4.5 1.9 m absorption of w ater with 8 lo ops R = 0 : 5 cm 3 trials 26 Figure 4.6 Dilution of D 2 O with w ater: 8 lo ops, R=0.5cm 60 m b er 26 Figure 4.7 f(g) (Eq. 13 of P a yne & Hale) compared to D 2 O data & Beer's La w 27 Figure 4.8 ln I air =I w ater vs. n um b er of lo ops at 1.85 m 28 Figure 4.9 Resp onse of straigh t b er in Meth ylene Blue. One sample equals one second. 29 Figure 4.10 Ev anescen t Absorption vs. Concen tration for coiled and b en t 60 m b er 30 Figure 4.11 f(g) (Eq. 13 of P a yne & Hale) compared to exp erimen t 31 Figure 4.12 f(g) (Eq. 13 of P a yne & Hale) compared to Degrandpre & Burgess' Data 32 Figure A.1 99.6 % D 2 O & 99.8% D 2 O (reference) 38 Figure A.2 99.1 % D 2 O & 99.8 % D 2 O (reference) 39 Figure A.3 98.2% D 2 O & 99.8 % D 2 O (reference) 39 Figure A.4 96.6% D 2 O & 99.8 % D 2 O (reference) 40 Figure B.1 A TR sp ectrum: 2 lo ops, R=0.5cm 41 Figure B.2 A TR sp ectrum: 4 lo ops, R=0.5cm 42 Figure B.3 A TR sp ectrum: 6 lo ops, R=0.5cm 42 Figure B.4 A TR sp ectrum: 8 lo ops, R=0.5cm 43 Figure D.1 W a v elength(nm) vs. Mono c hromator Dial (0-999) 55 iv
Ev anescen t Field Absorption Sensing Using Sapphire Fib ers Mic hael Grossman ABSTRA CT This thesis explores the application of coiled sapphire m ultimo de optical b ers for ev anescen t w a v e c hemical sensing in b oth the visible sp ectrum and the near infrared. As has b een suggested in the literature p ertaining to silica b ers, b ending con v erts lo w-order mo des to high order ones, whic h leads to more ev anescen t absorption and th us a more sensitiv e c hemical detector. By coiling the b er man y times, it w as exp ected that ev en greater sensitivit y w ould b e attained. Exp erimen ts w ere p erformed to in v estigate the sensor resp onse to dieren t solutions and to c haracterize this resp onse. In the rst of three exp erimen ts, the large absorption p eak of w ater at 3 m w as examined in order to compare the sensitivit y of a straigh t b er v ersus a coiled one. In the second exp erimen t, the eect of increasing the n um b er of coils w as in v estigated, as w as the resp onse of the sensor to v arying concen trations of w ater in hea vy w ater. In the third exp erimen t, meth ylene blue dy e w as used to in v estigate the exten t of adsorption of dy e molecules on the sapphire b er and its p ersistence. v
CHAPTER 1 INTR ODUCTION Guiding ligh t b y means of in ternal rerection w as rst demonstrated in the 1840s b y the Ph ysicists Daniel Collo don and Jacques Babinet b y shining ligh t up from the b ottom of a foun tain and observing the ligh t to b e guided within the curving jet of w ater 1 . The English Ph ysicist John T yndall, in a more w ell-kno wn demonstration rst p erformed in 1854, sho w ed that it w as p ossible to guide ligh t b y shining sunligh t in to a stream of w ater p ouring from a tank. When view ed from the side, the ligh t w as observ ed to follo w the curv ed path of the falling w ater.[2 ] It did not tak e long for other scien tists to forsee applications for guided ligh t. In 1880, William Wheelding paten ted a metho d of guiding ligh t to illuminate homes using mirrored pip es, whic h he rather aptly called "ligh t pip es." Unfortunately for Wheelding, his in v en tion w as obscured b y Edison's in v en tion of the incandescen t ligh t bulb.[2 ] It w as not un til the latter half of the t w en tieth cen tury and the in v en tion of lasers that the application of b er optics 2 as comm unication and sensing devices w as realized.[3 ] Optical b ers are dielectric w a v eguides made usually of fused silica, plastic, or, in the case of these exp erimen ts, sapphire( Al 2 O 3 ). Although primarily dev elop ed for comm unications, optical b ers ha v e also b een used in a wide v ariet y of noncomm unications applications. These sensing applications can b e divided in to t w o broad categories: in trinsic and extrinsic b er optic sensors.[4 ] An in trinsic sensor is 1 This phenomenon can still b e easily observ ed in p erson b y an y one driving past man y p osh suburban neigh b orho o ds after dark. 2 The term "b er optics" w as coined b y Dr. Narinder Singh Kapan y the "F ather of Fib er Optics," in 1956.[2 ] 1
one in whic h the in teraction b et w een the guided ligh t and the quan tit y to b e measured o ccurs inside the b er itself. An extrinsic sensor is one in whic h the in teraction tak es place outside the b er. Some examples of in trinsic b er optic sensors include b er optic acoustic sensors whic h are used as h ydrophones in passiv e sonar systems and b er optic gyroscop es for inertial na vigation systems. The former senses sound w a v es b ecause sound w a v es inciden t up on the b er cause it to deform, causing a phase shift in the ligh t. The latter sensor exploits the Sagnac Eect to measure rotations. (When t w o ligh t pulses tra v erse a circular, rotating lo op in opp osite directions, they will in terfere b ecause the pulse going along with the lo op has to go farther than the pulse mo ving against the lo op's rotation.) Extrinsic optical b er sensors ha v e b een used to measure div erse ph ysical phemomena suc h as h umidit y temp erature, and pressure.[5 ] As will b e discussed b elo w, a b er optic c hemical sensor is an extrinsic c hemical sensor. Sapphire is a desirable material for b er optic c hemical sensing due to its signican tly higher melting p oin t than silica, allo wing the sapphire c hemical/temp erature/pressure sensor to op erate at higher temp eratures than the silica b er. Inside an optical b er, ligh t ra ys that strik e the core-cladding in terface at angles greater than the critical angle are totally in ternally rerected. Asso ciated with eac h totally in ternally rerected ra y at an allo w ed angle is a mo de The ra y represen ts the normal to the w a v efron t. As the electromagnetic w a v e propagates along the b er, there is a phase shift of 180 degrees at eac h rerection. If the w a v e repro duces itself after t w o rerections, it is called an eigenmo de, or just a mo de of the w a v eguide.[7 ] This condition is called self-consistency .[3 ] Fig. 1.1 depicts a ra y of w a v elength propagating along a b er of diameter d The unrerected, original w a v e m ust b e in phase with the rerected w a v e; since B C is a w a v efron t, the t w o w a v es m ust ha v e a phase dierence that is an in teger m ultiple of 2 (Eac h rerection causes a phase 2
q B A C d q l Original Wave Twice-reflected wave _ Figure 1.1. Self-consistency: the t wice-rerected w a v e repro duces the original. shift of .) 2 AC 2 2 AB = 2 q (1.1) where q = 0 ; 1 ; 2 ::: Since AC AB = 2 d sin (1.2) and 2 (2 d sin ) = 2 ( q + 1) (1.3) w e can write ( q + 1) = m and sin m = m 2 d ; m = 1 ; 2 ; 3 ::: (1.4) where m is the rerection angle and m indicates the m th mo de. The rerection angle increases with mo de n um b er. 3
Length L Low-order mode High-order mode Figure 1.2. A high-order mo de mak es more rerections p er unit length. A low-or der mo de (small m ) strik es the core-cladding in terface at a large angle as measured from the normal to the in terface, while a high-or der mo de (large m ) strik es the in terface closer to the critical angle (Fig. 1.2). A t eac h rerection, an exp onen tially diminishing, ev anescen t w a v e p enetrates in to the cladding. It is this ev anescen t eld that is the basis of b er-optic c hemical sensing. A tten uated total rerection (A TR) sp ectroscop y b y means of b er-optic ev anescen t eld absorption sensors has a wide range of c hemical sensing applications in industry and in science.[8 ] T ypically suc h a sensor exploits the in teraction b et w een the exp onen tially deca ying electromagnetic eld of the guided w a v e and a ruid surrounding the unclad b er. By measuring the transmitted p o w er through the b er, it is p ossible to determine the c hemical comp osition of the surrounding ruid. The p enetration depth, d p is dened as the distance from the core-cladding(ruid) in terface at whic h the ev anescen t eld drops b y a factor of e from its v alue at the 4
in terface.[10 ] d p = 2 p n 21 sin 2 sin 2 c (1.5) where is the free-space w a v elength, n 1 is the core index and is the angle of incidence at the core-ruid in terface measured from the normal to the in terface, and c is the critical angle. Insp ection of equation 1.5 sho ws that the greatest p enetration depth, and th us the most ev anescen t absorption, o ccurs for incidence angles near the critical angle. In other w ords, it is the high-order mo des whic h p enetrate farthest in to the surrounding ruid and are most strongly atten uated that mak e it p ossible for an optical b er to b e a c hemical sensor. Lo w-order mo des, striking the core-ruid in terface at shallo w er angles, do not p enetrate as deeply and b ounce few er times as they are in ternally rerected along a giv en length of the b er. Unfortunately in the case a sapphire b er sensor, whic h has a core index of refraction of appro ximately 1.7, immersed in a ruid of index 1.3-1.4, ra ys with angles of incidence close to the critical angle are incompatible with lo w-loss propagation. While this lo w sensitivit y compared to bulk-absorption metho ds can b e exploited, e.g., insensitivit y to scattering b y susp ended particles in the analyte ruid, it is desirable to increase the sensitivit y of the b er. Sev eral approac hes to increase sensitivit y ha v e b een rep orted in the literature: tap ering the sensing region, selectiv e ra y launc hing, and b ending.[10 ][12 ] The exp erimen ts carried out in this researc h aimed to further increase sensitivit y b y not merely b ending the b er, but coiling it in to man y compact ( R 0 : 5 cm ) lo ops. It has b een demonstrated exp erimen tally using silica b ers that coiling results in con v ersion from lo w to high order mo des. Since sensitivit y is directly prop ortional to length, there will b e an impro v emen t in sensitivit y b oth due to b ending and from ha ving a longer sensing region. Coiling man y lo ops also results in a m uc h more 5
spatially compact prob e than can b e obtained with a straigh t b er. In addition, due to mo de scram bling from irregularities in the sapphire b er, it is p ossible that high-order mo des will b e created as the ligh t propagates to w ard the sensing region. 6
CHAPTER 2 THEOR Y AND NUMERICAL CALCULA TIONS 2.1 The Beer-Lam b ert La w and Estimate of Ev anescen t Absorption The Beer-Lam b ert La w describ es the atten uation of a b eam of ligh t as it passes through an absorbing medium of length l I 1 = I 0 exp ( l ) (2.1) where I 0 and I 1 are the inciden t and atten uated in tensities, and the bulk absorption co ecien t is giv en b y = c l og 10 e (2.2) in whic h is the molar absorptivit y and c is the molar concen tration. When applying Beer's La w to an unclad length l of b er exp osed to an absorbing ruid, the sym b ol r will b e used in place of and b e referred to as the ev anescen t absorption co ecien t of the ruid. It has b een sho wn that r dep ends on the angle of incidence of the ra y as measured from the normal to the core-cladding in terface.[14 ] r ( ) = n 2 cos cot 2 n 21 cos 2 c p cos 2 c cos 2 sin 2 (2.3) where is the w a v elength, is the b er radius, n 1 is the core index, c is the critical angle, is the bulk absorption co ecien t, n 2 is the index of the analyte, and is the sk ewness angle. If is held xed, r ( ) is maxim um for sk ewness angle equal to 7
2 whic h corresp onds to meridional ra ys. Th us, only a t w o-dimensional analysis of meridional ra ys will b e attempted. It has b een sho wn that the ev anescen t absorption Figure 2.1. Ev anescen t Absorption vs. Refractiv e Index for a b en t b er in w ater. Normalized to a long, straigh t 75 m b er with n cl = 1 : 3i and n c = 1 : 7. co ecien t of a b en t b er of radius (see gure 2.2) is, as a function of the analyte ruid index n 2 :[10 ] r ( n 2 ) = K Z 2 0 Z 2 1 cos 3 d dh (1 n 2c cos 2 ) 2 p ( n c =n 2 ) 2 sin 2 1 Z 2 0 Z 2 1 sin cos d dh (1 n 2c cos 2 ) 2 + K Z 2 0 Z 2 1 cos 3 d dh (1 n 2c cos 2 ) 2 p ( n c =n 2 ) 2 sin 2 1 Z 2 0 Z 2 1 sin cos d dh (1 n 2c cos 2 ) 2 (2.4) 8
R ncln22r h q j d ncstraight fiber in air bent fiber in fluid Figure 2.2. U-shap ed region considered in equation 2.4 where K = n c 4 ( n 2c n 2cl ) (2.5) and 1 = arcsin ( R + h ) n cl ( R + 2 ) n c and 2 = arcsin R + h R + 2 (2.6) 1 = arcsin ( R + h ) n cl R n c and 2 = 2 (2.7) In these equations, is the b er radius, R is the radius of the b end, n c is the refractiv e index of the b er core, and n cl is the cladding index. In gure 2.1 ab o v e, equation 2.4 w as n umerically in tegrated using Mathematic a n c w as 1.72, and the results for 0 : 5 cm and straigh t b ers of radii 50 and 75 microns are plotted in gure 2.1. r 0 ef f is the ev anescen t absorption co ecien t for a straigh t sapphire ( n = 1 : 7) b er of radius 75 microns with w ater ( n = 1 : 3) as the surrounding ruid. N. B. the in tersection of 9
air sapphire a q q Figure 2.3. Angles Used In Determining The Eectiv e Cladding Index Of Sapphire Fib er In Air the lo w est curv e with the v ertical axis at (1 ; 1 : 3) in Fig. 2.1; all of the curv es are divided b y this n um b er. In the case of a sapphire b er in air, the minim um incidence angle for a guided ra y is not giv en b y arcsin 1 n c Instead, it is determined b y the launc h condition. Applying Snell's La w for a ra y inciden t at the b er with angle in gure 2.3, giv es sin = n c sin (2.8) Using Snell's La w again for a ra y striking the core-cladding in terface giv es the eectiv e cladding index, n ef f cl = n c sin (arccos (sin n c )) (2.9) 10
In the n umerical in tegration, the angle of incidence of the ra ys w as generously estimated to b e no greater than 30 degrees from the b er axis. The resulting eectiv e cladding index w as used for n cl in the in tegration of equation 2.4 ab o v e. The sensitivit y increases with decreasing b end radius and b er radius. The inv erse dep endence on the b er radius mak es in tuitiv e sense b ecause decreasing the p erp endicular distance b et w een the b er w alls results in more b ounces p er unit length of b er, i.e. more opp ortunities for ev anescen t w a v es to in teract with the analyte ruid.[9 ] F rom gure 2.4, where all other parameters are held constan t and only the b end radius is v aried, it is eviden t that reducing the b end radius can signican tly enhance sensitivit y Figure 2.4. Eect of Bend Radius on Sensitivit y ( = 75 m ) 11
2.2 Deviation from Beer's La w It has b een p oin ted out that the Beer-Lam b ert La w o v erestimates b er absorbance (log P out P in ) and a more accurate description whic h tak es in to accoun t the v arying amoun ts of ev anescen t absorption suered b y dieren t mo des has b een dev elop ed.[15 ] Th us, the total p o w er is not divided ev enly among the mo des. An estimate of the atten uation along the b er is made b y taking in to accoun t the dieren t fractions of the total p o w er in eac h mo de and summing o v er all mo des. The fraction of p o w er in the cladding for the th mo de, is giv en b y : = N p 2 N 2 (2.10) where N is the total n um b er of mo des. The atten uation of the th mo de after a distance l is: exp ( l ) (2.11) where is giv en b y Eq. 2.2. The p o w er out of the b er is giv en b y: P out = P in N N X =1 exp ( l ) (2.12) F or a m ultimo de b er, N is large so the sum can b e written as an in tegral: P out = P in N Z N 1 exp ( l ) d (2.13) The in tegration ab o v e can b e simplied b y the substitution suggested in Eq. 12 of reference . x = 2 p N ( N ) (2.14) 12
1e-04 0.001 0.01 0.1 1 10 100 0 0.2 0.4 0.6 0.8 1Absorbanceg Beer's La w f(g) Figure 2.5. Fib er absorbance as predicted b y P a yne & Hale and according to Beer's La w Setting the lo w er limit of in tegration to zero and the upp er limit to innit y the atten uation along the b er can b e written: P out P in f ( g ) = 2 Z 1 0 (2 p 1 + x 2 1 p 1 + x 2 2 x ) exp ( g x ) dx (2.15) where g = 2 cl V log 10 e (2.16) and l is the length of the sensing region, V is b er optic v-parameter ( 2 ( N A ) ) and c is the molar concen tration. Equation 2.15 w as in tegrated n umerically using Maple and the result is sho wn in Fig. 2.5 along with the exp onen tial atten uation according to Beer's La w. 13
CHAPTER 3 EXPERIMENT AL METHODS This c hapter describ es in some detail ho w the exp erimen ts w ere set up and interfaced with computers. See the app endix for the programming details of comm unicating with the SR510 Lo c k-In Amplier. 3.1 Near IR Exp erimen t: absorption of w ater near 3 m Chopper Lens Collimating Focusing Lens IR Source fiber coiled SR510 Lock-In monochromator Lens 1:1 Focusing PC InSb IR Detectorserial Figure 3.1. Sc hematic blo c k-diagram of 3 m absorption exp erimen t The purp ose of this exp erimen t w as to determine ho w m uc h more sensitiv e a coiled b er w as than a straigh t b er b y measuring the ev anescen t absorption co ecien t of w ater at w a v elengths from 2.5 to 3.3 m. The b er w as appro ximately 50cm long, with a radius of 75 m and an input tap er of 150 m. An infrared source w as c hopp ed at 150Hz, collimated with a plano con v ex lens, and fo cused in to a mono c hromator 1 1 Oreil mo del #77250 SN 142. F or calibration and other details, see App endix D 14
using a bicon v ex lens of fo cal length 2.54cm. F rom the output slit of the mono c hromator, the ligh t w as 1:1 fo cused ( d i = d o = 2 f ) on to the tap ered end of the b er using a one-inc h fo cal length bicon v ex lens. (All lenses used w ere made of Calcium Fluoride, to a v oid infrared absorption in the lens material.) The output end of the b er w as moun ted on a 3D translation stage directly in fron t of a cry ogenically-co oled InSb Detector and carefully p ositioned in fron t of the 1mm diameter detector face. Care w as tak en while lling the detector with liquid nitrogen to a v oid splashing the sides of the detector; a mere few microns of condensed w ater v ap or on the detector windo w will result in signican t absorption. No lens w as used to fo cus the ligh t emanating from the end of the b er on to the detector; after m uc h trial and error it w as determined that one-to-one fo cusing failed to increase the amoun t of ligh t inciden t on the detector face. The InSb detector w as connected to a preamplier and then a Stanford Researc h mo del 510 lo c k-in amplier whic h w as connected to a PC via a serial p ort. The output v oltage of the InSb detector w as sampled at a rate of appro ximately 5Hz. (See App endix C for details ab out the in terfacing. Surprisingly the do cumen tation for SR510 w as helpful in writing the programs.) Before p ositioning the tap ered end of the b er, an indium an timonide infrared detector and iris w ere used to nd the fo cus. Once the fo cus w as found, the tap ered end of the b er w as moun ted to a 3D translation stage using a b en t pap erclip padded with duct tap e. The tap ered end of the b er w as then slo wly scanned across the narro w ed iris to maximize the ligh t coupled in to the b er. The b er w as coiled around a 1cm-radius n ylon sp o ol whic h had b een trimmed to minimize con tact b et w een n ylon and b er. Sev eral lo ops of string held the b er in place. It w as only p ossible to coil the 75 m b er in lo ops of radius 1cm. A ttempts to b end the b er more tigh tly brok e the b er. The trimmed sp o ol w as placed in 15
a shallo w p yrex dish whic h w as lled with deionized w ater. The length of straigh t b er exp osed to the ruid w as appro ximately equal to the length of coiled b er (2 1 cm 2 = 12 : 6 cm ). 2cm Figure 3.2. Nylon sp o ol, trimmed along dotted lines A t b oth ends of the b er w ere placed blac k 5" b y 7" notecards with small pinholes just large enough to admit the b er to minimize the amoun t of stra y ligh t from the source that reac hed the detector. In addition, the en tire exp erimen t w as co v ered with a hea vy blac k tarp while the data w ere b eing collected. The tarp w as used to prev en t stra y IR sources suc h as h umans and hot cups of tea from disturbing the exp erimen t and to k eep air curren ts from the air conditioning system from disturbing the b er ends. The same precautions w ere tak en in the hea vy w ater exp erimen t describ ed in the next section. Data w ere collected using a home-made program that sampled the output v oltage of the lo c k-in amplier at appro ximately 5Hz and then recorded it, along with the elapsed time, in a plain ASCI I le. Using the mono c hromator calibration curv e, the w a v elength w as calculated. Gnuplot w as used to mak e all of the graphs and to manipulate the data. Care w as tak en to mak e sure that the InSb Detector did not drift during the course of b oth infrared exp erimen ts due to the long time, appro ximately 1 hour, required for the mono c hromator to scan from 1.5 to 3.5 microns. Before eac h sp ectrum w as 16
recorded, the magnitude of the zeroth-order p eak and the emission p eak close to 2.6 microns w ere recorded b y man ually cranking the mono c hromator dial. After the mono c hromator had nished scanning, the zeroth-order and 2 : 6 m p eaks w ere remeasured. The exp erimen t w as rep eated in the ev en t the v alues did not matc h. 3.2 Absorption of H 2 O in D 2 O The purp ose of this exp erimen t w as to in v estigate the resp onse of a coiled b er sensor to c hanges in the solute concen tration and the eect of the n um b er of lo ops on sensitivit y The H 2 O in D 2 O mixture w as c hosen to a v oid the problem of screening due to adsorption that w as presen t in the Meth ylene Blue exp erimen t describ ed b elo w. Also con v enien t is the fact that D 2 O do es not absorb at 1.9 microns while w ater do es so, but m uc h more w eakly than at 3 microns. Th us, the problem of to o m uc h absorption leading to w eak transmission ma y b e a v oided. The exp erimen tal setup w as v ery nearly the same as in gure 3.1, except that a magnetic stirrer w as placed under a 50mL b eak er to ensure adequate mixing of D 2 O with the w ater. The sp o ol holding the b er w as moun ted as in gure 3.5. The b eak er w as co v ered with clear plastic wrap to minimize ev ap oration. Tw o, four, six, and eigh t lo ops of 60 m sapphire b er of radius 0.5cm w ere used to obtain the absorption sp ectra displa y ed in gure 4.8. A micropip et w as used to add small quan tities of deionized w ater to 50mL of hea vy w ater. An absorption sp ectrum near the 1.9 m o v ertone band of H 2 O w as recorded at eac h concen tration of deionized w ater. 17
3.3 Meth ylene Blue Dy e Adsorption The purp ose of this exp erimen t w as to assess the exten t of adsorption of Meth ylene Blue 2 a common absorption standard, on sapphire b ers. It has b een rep orted in the literature that there is deviation from linearit y of absorption when silica b ers w ere used due to adsorption of the dy e to the b er. In addition, the p ersistence of adsorption w as also in v estigated. A length of sapphire b er measuring 8cm w as 0.25" hole cuvette Pinhole to admit fiber Figure 3.3. Nylon blo c k and cuv ette placed in a glass cuv ette of v olume appro ximately 5 mL. Custom-made, press-tted n ylon blo c ks w ere used to allo w the ro w of ruid through the cuv ette and admit the b er. The b er w as held in place b y w ax. The cuv ette w as moun ted with its 2 3,7-bis(Dimeth ylamino)-phenazathionium c hloride T etrameth ylthionine c hloride, CAS#61-734, c hemical form ula C 16 H 18 n 3 C l S 18
Diverging Lens HeNe Laser Cuvette drain 25.5cm Converging Lens 3cm Meter Optical Power Newport 835 Analog Out BNC cable Radio Shack Multimeter model: 22-812 9-pin Serial Cable Meterview Software PC Running 8cm fiberin cuvette reservoir Figure 3.4. Sc hematic Blo c k Diagram of Meth ylene Blue Dy e Exp erimen t long axis v ertical, to prev en t bubbles from b ecoming trapp ed. The photo detector w as moun ted v ery close, less than 1cm, from the end of the b er. (The end of the b er protruded only ab out 5mm from the n ylon blo c k.) It is essen tial to minimize the length of b er b et w een the sensing region and the detector; due to the mo de scram bling caused b y imp erfections in the b er, the high order mo des whic h ha v e in teracted with the sample are scattered b elo w the critical angle and out of the b er after only 10-20cm. Excess straigh t b er after the sensing region actually diminishes the sensitivit y The input end of the b er w as held in place using clip fashioned from a pap er-clip and padded with duct tap e to a v oid breaking the b er at the p oin t of con tact. A biconca v e lens w as placed in fron t of the helium-neon laser in order to 19
spread the b eam to b etter ll the bicon v ex lens whic h fo cused the b eam in to the tap ered end of the b er. Rather than mo v e the laser or the input end of the b er, the bicon v ex lens w as moun ted on a mo v able stage to mak e small adjustmen ts. The Newp ort 835 Optical p o w er meter do es not ha v e an y straigh tforw ard w a y to in terface with a PC; it w as designed to w ork with a c hart recorder. Ho w ev er, the v oltage output can b e measured with a digital m ultimeter (Radio Shac k mo del 22-812) whic h is easily connected to a PC. Using a serial cable and the soft w are that came with the meter, it w as p ossible to sample the output of the p o w er meter at a maxim um frequency of 1Hz. T o obtain v arying concen trations of meth ylene blue, anh ydrous p o wdered meth ylene blue w as mixed with deionized w ater to pro duce 500mL of 500 M solution. This solution w as then diluted to concen trations 5 M to 350 M in roughly 50mL p ortions. The data w ere collected b y lling the cuv ette with deionized w ater, using the reserv oir to in tro duce the dy e in to the cuv ette, and then rushing out the cuv ette with ltered tap w ater. During eac h run, MeterView recorded the in tensit y eac h second, pro ducing a plot of transmitted signal vs. time. Collecting this data w as extremely time-consuming b ecause the b er had to b e remo v ed from the cuv ette ev ery t w o to three runs for cleaning. T o eliminate an y p ossibilit y of adsorption, the b er w as cleaned b y b eing placed in an acetone bath in an ultrasonic cleaner for 30 min utes. In order to in v estigate ho w coiling the b er aects the relationship b et w een the ev anescen t absorption co ecien t and concen tration, it w as necessary to c hange the design of the exp erimen t. Instead of sealing the b er in a cuv ette, it w as coiled on to a 1cm diameter n ylon sp o ol, and immersed in a shallo w 200mL p yrex dish con taining the dy e solution. The dish, rather than the b er, w as mo v ed up and 20
Bath Lab Jack Lens Conv. Lens Div. Laser HeNe RS 22-812Multimeter PC (MeterView) fiber Meter Optical Power Newport 835 Figure 3.5. Measuremen t of Meth ylene Blue Ev anescen t Absorption with coiled b er do wn to immerse the b er b y b eing placed on a lab jac k. (See Fig. 3.5.) The ev anescen t absorption co ecien t w as determined b y the ratio of the p o w er measured with the b er immersed in deionized w ater to the p o w er measured with the b er immersed in dy e solution. Fiv e trials w ere p erformed at eac h dy e concen tration. In addition, the theoretical expression dev elop ed in Eq. 13 of P a yne and Hale w as compared to exp erimen tal measuremen ts of the atten uation. T o do this, the eectiv e n umerical ap erture of the lens, n sin had to b e determined. (Since the n umerical ap erture of the lens is m uc h smaller than that of the b er, it limits the n um b er of mo des that can b e excited.) The sp ot size of the laser b eam w as measured using a card with a pinhole blo c king the face of the detector. The sp ot w as measured from the cen ter to where the p o w er drops b y a factor of e 2 F rom gure 3.6: = arctan (0 : 38 = 2 : 54) = 8 : 5 (3.1) N :A: = (1) sin 8 : 5 = 0 : 15 (3.2) 21
2.54cm 0.76cm q Figure 3.6. Measuring the n umerical ap erture of a lens V = 2 N :A: = 2 100 m 632 : 8 nm 0 : 15 = 147 (3.3) F rom the n umerical ap erture, the V-n um b er 2 N :A: w as obtained. In addition, the molar absorption co ecien t of meth ylene blue dy e w as determined using a HeNe laser and a cuv ette. The 100 m used ab o v e is the radius at the tap ered, input end of the b er; the b er radius tap ers o to 75 m after a short distance. P out =P in w as measured for a straigh t 75 m b er with dy e concen tration ranging from 5 M to 200 M. P out means the p o w er measured at the end of the b er when the b er is exp osed to the solute. P in is the p o w er measured at the end of the b er with no solute; the b er is just exp osed to deionized w ater. 22
CHAPTER 4 DISCUSSION OF RESUL TS 4.1 3 m IR Exp erimen t 0 5 10 15 20 25 30 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3mVw a v elength ( m) in air in w ater Figure 4.1. Absorption Sp ectrum of Straigh t 39cm 75 m Fib er In Air and W ater Comparison of the b ottom curv es in gures 4.1 and 4.2 clearly sho ws that there is more absorption with the coiled b er. Figure 4.3 sho w a plot of the logarithms of the ratios of the in-air to in-w ater sp ectra from gures 4.1 and 4.2. Insp ection of gure 4.3 indicates that coiling the b er do es in fact increase sensitivit y The data b ecome noisy near three microns b ecause there is so m uc h absorption that the transmitted ligh t is buried in am bien t infrared noise; th us the baseline of the in-w ater sp ectra is not zero. Considering only the p ortions of the sp ectrum to the left of 2.8 m and to 23
0 5 10 15 20 25 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3mVw a v elength ( m) in air in w ater Figure 4.2. Absorption Sp ectrum of Tw o T urns of R = 1 cm 75 m Fib er 0 1 2 3 4 5 6 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3lnIair Iw aterw a v elength ( m) straigh t coiled Figure 4.3. ln I air =I w ater vs. w a v elength: straigh t 75 m compared to 2 turns coiled R = 1 cm 24
0.1 1 10 100 1000 10000 100000 1000 1500 2000 2500 3000 3500 4000Absorption Co ecien tW a v elength (nm) Figure 4.4. Absorption of W ater In The Near IR (after Hale & Querry) the righ t of 3.1 m in gure 4.3, it is safe to sa y that coiling t w o turns ( R = 1 cm ) has more than doubled the ev anescen t absorption from a straigh t b er. It w as due to this dicult y that the 1.9 micron p eak w as c hosen for measuremen ts in the next exp erimen t. Although, gure 2.1 is only an estimate of the ev anescen t absorption of a b en t b er, it is nev ertheless encouraging that the ev anescen t absorption co ecien ts plotted for some w a v elengths in gure 4.3 exceeds those of gure 2.1. The absorption sp ectrum sho wn in Fig. 4.3 app ears to b e sk ew ed sligh tly to the righ t, with the p eak b eing a little bit b ey ond 3 microns, rather than at 2.95 microns as exp ected based on the data of Hale and Querry plotted b elo w in gure 4.4.[17 ] This ma y b e due to anomalous disp ersion asso ciated with the 3 m band; the refractiv e index is higher to the red of the resonance, leading to a smaller dierence from that of sapphire. 25
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 1.7 1.8 1.9 2 2.1 2.2 2.3V oltsmicrons T rial 1 T rial 2 T rial 3 Figure 4.5. 1.9 m absorption of w ater with 8 lo ops R = 0 : 5 cm 3 trials -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 2.5 3 3.5 4 L% w ater Figure 4.6. Dilution of D 2 O with w ater: 8 lo ops, R=0.5cm 60 m b er 26
1e-35 1e-30 1e-25 1e-20 1e-15 1e-10 1e-05 1 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6Absorbanceg Exp. Beer's La w f(g) Figure 4.7. f(g) (Eq. 13 of P a yne & Hale) compared to D 2 O data & Beer's La w 4.2 1.9 m IR Exp erimen t Figure 4.5 sho ws that there is go o d repro ducibilit y of the absorption sp ectrum; it w ould b e imp ossible to mak e an y meaningful calculation of r if there w ere drift in the detector system while measuremen ts are b eing tak en. The sp ectra from whic h the p oin ts sho wn in gure 4.6 w ere calculated are included in app endix A. The ratio of the curv es used to calculate the p oin ts in Fig. 4.6 w as computed at 2.8 m b ecause the absorption w as v ery strong at this w a v elength. F rom gure 4.6, it app ears that the ev anescen t absorption as a function of concen tration is linear up to at least r L = 3 : 5 for eigh t lo ops of 60 m sapphire b er coiled to a radius of 0.5cm. As sho wn in gure 4.8, the relationship b et w een the n um b er of turns and the ev anescen t absorption at 1.85 m is also linear up to eigh t lo ops. Fig. 4.7 sho ws m uc h b etter agreemen t b et w een m y exp erimen tal data and Eq. 13 of P a yne & Hale (Eq. 2.15 in this pap er) than with Beer's La w. Beer's La w o v erestimates the absorption of the H 2 O in D 2 O b y o v er ten orders of magnitude. 27
0 0.1 0.2 0.3 0.4 0.5 0.6 0 1 2 3 4 5 6 7 8 9lnIair Iw aterT urns Figure 4.8. ln I air =I w ater vs. n um b er of lo ops at 1.85 m As concen tration increases, Beer's La w completely div erges from the exp erimen tal data. Ho w ev er, Eq. 2.15 is within sev eral orders of magnitude of the exp erimen tal data. It should b e noted that the signal-to-noise ratio in the measuremen ts used to pro duce gure 4.6 w as far from ideal, the p eak-to-p eak noise amplitude approac hing half the signal in some of the measuremen ts. In calculating the ratio of the signals, the noise w as considered random so that the signal amplitude w as measured from the cen ter of eac h trace. 4.3 Meth ylene Blue Adsorption Exp erimen t Figure 4.9 is a plot of the p o w er of ligh t transmitted through a 75 m straigh t b er with 8cm exp osed to ruid in the cuv ette. Up to 2100 seconds, deionized w ater is just sitting in the cuv ette. 50mL of 60 M dy e is in tro duced, whic h rushes out the w ater. A t 3000 seconds, ltered tap w ater is allo w ed to ro w through the cuv ette at a rate of 0.31 mL p er second. 28
Figure 4.9. Resp onse of straigh t b er in Meth ylene Blue. One sample equals one second. Due to the slo w mixing rate, the curv e in Fig. 4.9 b ends o v er shortly after the dy e is in tro duced. In other w ords, the initial, precipitous drop when the dy e is in tro duced (2100 seconds) slo wly lev els o as the dy e and w ater mix in the cuv ette. Similarly when the cuv ette is rushed, there is a rapid increase in p o w er and then a slo w approac h to the pre-exp osure p o w er lev el. There w as no visible blue staining of the b er, suggesting that there w as little, if an y p ersisten t adsorption of dy e molecules to the b er surface. Fig. 4.10 sho ws the v ariation of ev anescen t absorption with dy e concen tration. The data w ere obtained using a sapphire b er of radius 60 microns. The upp er curv e sho ws measuremen ts made with 1.5 turns of radius 0.5cm. The lo w er curv e sho ws measuremen ts made with half a turn of radius 2.5cm. The data in b oth graphs w ere obtained using the apparatus sho wn in Fig. 3.5. 29
0 0.01 0.02 0.03 0.04 0.05 0 100 200 300 400 500Absorption ( cm 1)Meth ylene Blue Concen tration ( M) 1.5 turns R=0.5cm 0.5 turn R=2.5cm Figure 4.10. Ev anescen t Absorption vs. Concen tration for coiled and b en t 60 m b er It is clear that the coiled b er is more sensitiv e than the b en t b er, but there is disagreemen t in the literature as to whether the deviation from linearit y in Fig. 4.10 is due to adsorption of the dy e molecules to the b er or the result of rejection of highorder mo des with increasing dy e concen tration.[17 ] [18 ]  [11 ] Since the index of refraction of the ruid increases with dy e concen tration, this reduces the n umerical ap erture q n 2cor e n 2f l uid of the sensing region. A narro w er acceptance cone admits few er high order mo des in to the sensing region, whic h reduces sensitivit y With silica b er ( n 1 : 4) this eect ma y b e signican t, but, on accoun t of the large dierence in index of refraction b et w een w ater ( n 1 : 3) and sapphire ( n 1 : 7), this reduction in n umerical ap erture is negligible for sapphire b ers. A more plausible explanation is the non-uniform distribution of p o w er in the mo des; as sho wn in Fig. 4.11, there is go o d agreemen t b et w een exp erimen t and Eq. 13 of P a yne & Hale. Another factor that con tributes to the non-linearit y is that the high-order mo des are not reac hing the detector after lea ving the sensing region, but 30
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1f(g)g theory exp erimen t Figure 4.11. f(g) (Eq. 13 of P a yne & Hale) compared to exp erimen t are b eing scattered out of the b er altogether due to the hexagonal cross-section and the man y irregularities in the sapphire b er. Figure 4.11 compares exp erimen tal measuremen t of the absorbance(ln P out =P in ) of a straigh t 75 m radius b er with a sensing region of 8cm with equation 13 of reference [15 ]. The n umerical in tegration w as done using Maple 10 The data in table 4.1 w ere obtained using a straigh t 75 m b er in the apparatus describ ed in Fig. 3.4 and plotted in Fig. 4.11. The absorbance is larger than predicted b y equation 2.15. This is p erhaps due to micro b ends in the b er con v erting lo w to higher order mo des. Once again, it should b e p oin ted out that all of the calculations assume that the cross-section of the b er is circular and that the b er is free from defects, but sapphire b ers are hexagonal in cross-section. This hexagonal crosssection will clearly cause the guided ra y to strik e the b er surface at dieren t angles, causing mo de con v ersion. This is to b e con trasted with the exp erimen tal results of Degrandpre & Burgess[19 ], sho wn b elo w in gure 4.12, whose silica-b er data sho w less absorbance than predicted b y equation 2.15. 31
0.1 1 10 0 0.5 1 1.5 2 2.5Absorbanceg f(g) Degranpre & Burgess Data Figure 4.12. f(g) (Eq. 13 of P a yne & Hale) compared to Degrandpre & Burgess' Data conc: m g = 2 cl V log 10 e P out =P in 5 0.0159 0.939 10 0.0319 0.906 20 0.0638 0.888 40 0.128 0.830 60 0.192 0.796 80 0.256 0.779 100 0.319 0.771 120 0.383 0.750 140 0.447 0.749 160 0.510 0.738 200 0.638 0.729 T able 4.1. A tten uation Data (compared with theory in gure 4.11) 32
CHAPTER 5 CONCLUSION Coiling sapphire b ers clearly enhances ev anescen t absorption, suggesting they are ideal for c hemical sensing applications in the near-IR (1-4 m). This range of w a v elengths corresp onds to o v ertone and com bination bands of a large n um b er of molecules. In the rst exp erimen t, it w as disco v ered that coiling more than doubled the sensitivit y of the b er as compared to a straigh t b er of the same length(Figure 4.3). Due to the researc her's inabilit y to b end sapphire b ers to radii smaller than 0.5cm without breaking the b er, it w as not p ossible to explore the increase in ev anescen t absorption with smaller radius coils. If a reliable, cost-eectiv e metho d can b e found to create smaller b er coils, the calculations in gure 2.4 sho w that it will b e p ossible to greatly enhance the sensitivit y p ossibly b y more than an order of magnitude. Gupta et al. ha v e b en t unclad silica b ers to v ery small radii, 0 : 1 cm b y heating the b er with a burner and then examining the b end under a microscop e.[10 ] P erhaps their metho d can b e adapted to man ufacture coiled sapphire b ers of a few turns. More exp erimen tal w ork is needed in this area. The ev anescen t absorption with the b er placed in H 2 O / D 2 O solution w as found to v ary linearly with the n um b er of lo ops (Figure 4.8). In addition, there w as a linear relationship b et w een the concen tration of deionized w ater in hea vy w ater and ev anescen t absorption (Figure 4.6). It w as also disco v ered that the theoretical expression dev elop ed b y P a yne & Hale (Eq. 2.15) agreed with exp erimen tal data m uc h more closely than Beer's La w. While these results are encouraging, it is necessary to 33
explore higher concen trations to determine whether there is deviation from linearit y and to further test Eq. 2.15. In examining the eect of the adsorption of meth ylene blue dy e to the sapphire b er, it w as observ ed that there w as 0.2% cm 1 absorption indep enden t of concentration up to ab out 100 M. This is lik ely due to adsorption of dy e molecules on the surface of the b er. A t higher concen trations, the absorption due to dy e molecules in solution b egins to dominate. In rushing the cuv ette Fig. 4.9 the b ends are most probably due to the time required for the dy e and w ater to mix, rather than from an y p ersisten t dy e adsorption to the sapphire b er. Since meth ylene blue dy e has b een rep orted to adsorb to silica b er, sapphire b er is less susceptible to dy e adsorption than silica b er. Equation 2.15, as w ell as equation 2.4, w ere deriv ed under the assumption that the b er is uniform, lac king micro-b ends or other irregularities, and circular in cross section. As has b een discussed ab o v e, neither of these assumptions holds for a sapphire b er and this is v ery lik ely the reason the sapphire b er sho ws more atten uation than predicted b y equation 2.15. F urther researc h is required to deriv e an expression analogous to equation 2.15 that tak es in to accoun t the hexagonal cross-section of sapphire b ers and micro-b ends. 34
REFERENCES  J. Ha y es, Fib er Optics T e chnician 's Manual Delmar Publishers, 1996.  D. R. Go, Fib er Optic R efer enc e Guide: a pr actic al guide to the te chnolo gy F o cal Press, 1999.  S. E. A. Bahaa and M. C. T eic h, F undamentals Of Photonics Wiley 1991.  K. Grattan and B. Meggitt, Optic al Fib er Sensor T e chnolo gy V olume 2: Devic es and T e chnolo gy Chapman and Hall, 1998.  S. Otsuki, K. Adac hi, and T. T ak ahisa, Sensors and Actuators 53 91 (1998).  J. L. Kennedy D. Henry and N. Djeu, SPIE Pro c. 4833 116 (2002).  J. Wilson and J. Ha wk es, Opto ele ctr onics: A n Intr o duction Pren tice Hall, second edition, 1989.  S. K. Khijw ania and B. D. Gupta, Optical and Quan tum Electronics 31 625 (1999).  D. W. Lam b et al., Marine and F resh w ater Researc h 55 533 (2004).  B. D. Gupta, H. Do deja, and A. K. T omar, Optical and Quan tum Electronics 28 1629 (1996).  V. Ruddy B. D. McCraith, and J. A. Murph y Journal Of Applied Ph ysics 67 6070 (1990).  B. D. Gupta, C. D. Singh, and A. Sharma, Optical Engineering 33 1864 (1994).  M. T ateda and M. Ik eda, Applied Optics 15 2308 (1976).  A. W. Sn yder and J. D. Lo v e, Optic al Wave guide The ory Chapman and Hall, 1991.  F. P P a yne and Z. M. Hale, In ternational Journal of Opto electronics 8 743 (1993).  D. Gloge, Applied Optics 10 2252 (1971). 35
 G. M. Hale and M. R. Querry Applied Optics 12 555 (1973).  Y. Bunganaen, Using Optic al Fib er Sensing for Me asuring Chlor ophyl l-r elate d Pigments in T urbid Water. PhD thesis, 2002.  M. D. DeGrandpre and L. W. Burgess, Analytical Chemistry 60 2582 (1988). 36
App endix A Data for gure 4.6 These absorption sp ectra w ere used to calculate the data p oin ts sho wn in gure 4.6. The ratios of the top curv e to the b ottom curv e w as computed at 2.4 m and 2.8 m. Then the ratio at 2.8 m w as divided b y the ratio at 2.4 m and the natural logarithm tak en. 0 0.001 0.002 0.003 0.004 0.005 2.2 2.4 2.6 2.8 3 3.2V oltsW a v elength (microns) 99.8 % D 2 O 99.6 % D 2 O Figure A.1. 99.6 % D 2 O & 99.8% D 2 O (reference) 38
App endix A (Con tin ued) 0 0.001 0.002 0.003 0.004 0.005 2.2 2.4 2.6 2.8 3 3.2V oltsW a v elength (microns) 99.8 % D 2 O 99.1 % D 2 O Figure A.2. 99.1 % D 2 O & 99.8 % D 2 O (reference) 0 0.001 0.002 0.003 0.004 0.005 2.2 2.4 2.6 2.8 3 3.2V oltsW a v elength (microns) 99.8 % D 2 O 98.2 % D 2 O Figure A.3. 98.2% D 2 O & 99.8 % D 2 O (reference) 39
App endix A (Con tin ued) 0 0.001 0.002 0.003 0.004 0.005 2.2 2.4 2.6 2.8 3 3.2V oltsW a v elength (microns) 99.8 % D 2 O 96.2 % D 2 O Figure A.4. 96.6% D 2 O & 99.8 % D 2 O (reference) 40
App endix B Data for gure 4.8 The follo wing absorption sp ectra w ere used to compute the p oin ts on the graph of ev anescen t absorption vs. n um b er of lo ops sho wn in gure 4.8. Eac h p oin t w as computed from the log of the ratio of the in-air sp ectrum to the in-w ater sp ectrum at 1.85 m. 0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 1.7 1.8 1.9 2 2.1 2.2V oltsmicrons in air in w ater Figure B.1. A TR sp ectrum: 2 lo ops, R=0.5cm 41
App endix B (Con tin ued) 0 0.002 0.004 0.006 0.008 0.01 0.012 1.7 1.8 1.9 2 2.1 2.2V oltsmicrons in air in w ater Figure B.2. A TR sp ectrum: 4 lo ops, R=0.5cm 0 0.002 0.004 0.006 0.008 0.01 0.012 1.7 1.8 1.9 2 2.1 2.2V oltsmicrons in air in w ater Figure B.3. A TR sp ectrum: 6 lo ops, R=0.5cm 42
App endix B (Con tin ued) 0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 1.7 1.8 1.9 2 2.1 2.2V oltsmicrons in air in w ater Figure B.4. A TR sp ectrum: 8 lo ops, R=0.5cm 43
App endix C Programs for SR510 Lo c kin Amplier The easiest w a y to exp erimen t with the SR510 is to use a terminal em ulator program suc h as Hyp erT erm in windo ws op erating systems or, preferably Minic om in Lin ux. Simply connect the PC's serial p ort to the lo c k-in, mak e sure the terminal em ulator is congured to matc h the dip switc h settings on the bac k of the lo c k-in, and start t yping commands. See the SR510 do cumen tation for the complete command set, but try t yping Q and P to Query the output v oltage and the Phase, resp ectiv ely The SR510 w aits for carriage returns, but y our PC/terminal ma y b e using hardw are ro w-con trol. Use queryserial2.c to toggle serial lines to get things w orking. The programs b elo w only send the query "Q" command to the lo c k-in and record the time elapsed, mono c hromator w a v elength, and v oltage in a le. The slop e and in tercept from the mono c hromator calibration line (see app endix D) are needed to calculate the w a v elength from the elapsed time. (The programs log the elapsed time to le for debugging.) C.1 DOS and QBASIC N. B. do not run this program in the basic in terpreter. The QBASIC in terpreter cannot tell time prop erly The program m ust b e compiled as a DOS executable le. R E M a s i m p l e p r o g r a m t h a t r e p l a c e s a c h a r t r e c o r d e r R E M i t l o g s t i m e a n d v o l t a g e r e a d i n g s f r o m S R 5 1 0 l o c k i n R E M i n a f i l e c r e a t e d a t r u n t i m e R E M b e s u r e t o r e c o r d m o n o c h r o m e t e r r e a d i n g s a t s t a r t a n d f i n i s h R E M i n p u t s t a r t w a v e l e n g t h i n n m P R I N T e n t e r t h e o u t p u t f i l e n a m e I N P U T d a t a f i l e $ 44
App endix C (Con tin ued) O P E N d a t a f i l e $ F O R O U T P U T A S # 2 R E M Make s u r e s e t t i n g s o n t h e n e x t l i n e m a t c h t h e l o c k i n O P E N c o m 1 : 9 6 0 0 n 8 1 F O R R A N D O M A S # 1 P R I N T s t a r t i n g f r e q u e n c y i n n m a s r e a d f r o m m o n o c h r o m a t o r : I N P U T w a v e l n s t a r t T i m e = T I M E R b $ = I N K E Y $ W H I L E N O T ( U C A S E $ ( b $ ) = Q ) P R I N T # 1 Q S L E E P 1 I N P U T # 1 x b $ = I N K E Y $ W H I L E N O T I N K E Y $ = " W E N D I F b $ = " T H E N y = 1 E L S E y = 0 E N D I F R E M M a g i c n u m b e r s b e l o w a r e f r o m m o n o c h r o m a t o r c a l l i b r a t i o n P R I N T T I M E R s t a r t T i m e ( ( ( w a v e l n 3 9 5 + 3 4 ) / 1 0 0 0 ) + 0 0 0 5 9 0 7 7 9 # ( T I M E R s t a r t T i m e ) ) x P R I N T # 2 T I M E R s t a r t T i m e ( ( ( w a v e l n 3 9 5 + 3 4 ) / 1 0 0 0 ) + 0 0 0 5 9 0 7 7 9 # ( T I M E R s t a r t T i m e ) ) x W E N D 45
App endix C (Con tin ued) C L O S E # 1 C L O S E # 2 C.2 Lin ux and BASH The BASH script wr ap2.sh calls setSerialSignal and queryserial2 to comm unicate with the serial p ort/lo c kin. These programs are messy but they get the job done. I wrote the shell script and the t w o op en source C-programs w ere do wnloaded from the w ebsites included in the source les b elo w and w ere sligh tly mo died to cop e with serial hardw are quirks of the SR510 and m y laptop (IBM Thinkpad T22.) C.2.1 setSerialSignal.c / The original lo c ation of this sour c e is http:// www.emb e dde d linuxinterfacing .c om /chapters /06/setSerialSignal .c This pr o gr am is fr e e softwar e ; you c an r e distribute it and/ or mo dify it under the terms of the GNU Libr ary Gener al Public Lic ense as publishe d by the F r e e Softwar e F oundation ; either version 2 of the Lic ense or (at your option ) any later version This pr o gr am is distribute d in the hop e that it wil l b e useful, but WITHOUT ANY W ARRANTY ; without even the implie d warr anty of MER CHANT ABILITY or FITNESS F OR A P AR TICULAR PURPOSE Se e the GNU Libr ary Gener al Public Lic ense for mor e details. 46
App endix C (Con tin ued) / # include < sys / io ctl. h > #include < fcn tl. h > #include < termios .h > / we ne e d a termios structur e to cle ar the HUPCL bit / struct termios tio; in t main ( in t argc, c har argv) f in t fd; in t status ; if (argc != 4) f prin tf("Usage: setSerialSignal p ort DTR R TS n n"); prin tf("Usage: setSerialSignal /dev /tt yS0 j /dev /tt yS1 0 j 1 0 j 1 n n"); exit( 1 ); gif ((fd = op en (argv,O RD WR )) < 0) f prin tf("Couldn 't op en % s n n",argv); exit(1); gtcgetattr(fd, &tio); / get the termio information / 47
App endix C (Con tin ued) tio.c crag &= ~HUPCL ; / cle ar the HUPCL bit / tcsetattr(fd, TCSANO W &tio); / set the termio information / io ctl(fd, TIOCMGET &status); / get the serial p ort status / if ( argv == '1' ) / set the DTR line / status &= ~TIOCM DTR; else status j = TIOCM DTR; if ( argv == '1' ) / set the R TS line / status &= ~TIOCM R TS ; else status j = TIOCM R TS ; io ctl(fd, TIOCMSET &status ); / set the serial p ort status / close (fd); / close the devic e le / gC.2.2 queryserial2.c / The original lo c ation of this sour c e is http:// www.emb e dde d linuxinterfacing .c om /chapters /06/querySerial .c Copyright (C) 2001 by Cr aig Hol lab augh This pr o gr am is fr e e softwar e ; you c an r e distribute it and/ or mo dify it under the terms of the GNU Libr ary Gener al Public Lic ense as publishe d by the F r e e Softwar e F oundation ; either version 2 of the Lic ense or (at your option ) any later version 48
App endix C (Con tin ued) This pr o gr am is distribute d in the hop e that it wil l b e useful, but WITHOUT ANY W ARRANTY ; without even the implie d warr anty of MER CHANT ABILITY or FITNESS F OR A P AR TICULAR PURPOSE Se e the GNU Libr ary Gener al Public Lic ense for mor e details. / / querySerial querySerial pr ovides b ash scripts with serial c ommunic ations This pr o gr am sends a query out a serial p ort and waits a sp e cic amount of time then r eturns al l the char acters r e c eive d The c ommand line p ar ameters al low the user to sele ct the serial p ort, sele ct the b aud r ate, sele ct the time out and the serial c ommand to send A simple hash function c onverts the b aud r ate c ommand line p ar ameter into an inte ger / #include < stdio.h > #include < sys /io ctl.h > #include < fcn tl. h > #include < termios .h > #include < stdlib .h > / These ar e the hash denitions / #dene USERBA UD1200 '1'+'2' 49
App endix C (Con tin ued) #dene USERBA UD2400 '2'+'4' # dene USERBA UD9600 9 '+'6' #dene USERBA UD1920 '1'+'9' #dene USERBA UD3840 '3'+'8' struct termios tio; in t main ( in t argc, c har argv) f in t fd, status whic hBaud result ; long baud; c har buer; in t i; for (i=0; i < 256; i++) f buer [i]=0; gif (argc != 5) f prin tf("Usage: querySerial p ort sp eed timeout (mS) command n n"); exit( 1 ); g / c ompute which b aud r ate the user wants using a simple adding hash function / whic hBaud = argv  + argv; 50
App endix C (Con tin ued) switc h (whic hBaud ) f case USERBA UD1200 : baud = B1200 ; break ; case USERBA UD2400 : baud = B2400 ; break ; case USERBA UD9600 : baud = B9600 ; break ; case USERBA UD1920 : baud = B19200 ; break ; case USERBA UD3840 : baud = B38400 ; break ; default : prin tf ("Baud rate % s is not supp orted ); prin tf ("use 1200, 2400, 9600, 19200 or 38400. n n", argv); exit (1); break ; g / op en the serial p ort devic e le O NDELA Y tel ls p ort to op er ate and ignor e the DCD line 51
App endix C (Con tin ued) O NOCTTY this pr o c ess is not to b e c ome the c ontr ol ling pr o c ess for the p ort. The driver wil l not send this pr o c ess signals due to keyb o ar d ab orts, etc. / if ((fd = op en (argv,O RD WR )) < 0) // j O NOCTTY)) < 0) f prin tf("Couldn 't op en % s n n",argv); exit(1); g / we ar e not c onc erne d ab out pr eserving the old serial p ort c ongur ation CS8 8 data bits CREAD r e c eiver enable d CLOCAL don' t change the p ort s owner / tio.c crag = baud j CS8 j CREAD j CLOCAL ; tio.c crag &= ~HUPCL ; / cle ar the HUPCL bit, close do esn 't change DTR / tio.c lrag = 0; / set input rag non c anonic al no pr o c essing / tio.c irag = IGNP AR; / ignor e p arity err ors / tio.c orag = 0; / set output rag non c anonic al no pr o c essing / tio.c cc [VTIME ] = 0; / no time delay / tio.c cc [VMIN ] = 0; / no char delay / tcrush (fd, TCIFLUSH ); / rush the buer / tcsetattr(fd, TCSANO W &tio); / set the attributes / / Set up for no delay ie non blo cking r e ads wil l o c cur. 52
App endix C (Con tin ued) When we r e ad we'l l get what's in the input buer or nothing / fcn tl(fd, F SETFL, FNDELA Y); / write the users c ommand out the serial p ort / c har temp; sprin tf (temp, "%s n r", argv); result = write(fd, temp strlen (temp )); if (result < 0) f fputs("write failed n n", stderr ); close (fd); exit(1); g / wait for awhile b ase d on the user's time out value in mS / usleep (atoi(argv) 1000); / r e ad the input buer and print it / result = read(fd,buer ,255); buer = 0; // zer o terminate so printf works prin tf("%s n n ,buer ); / close the devic e le / close (fd); gC.2.3 wrap2.sh # !/bin /sh 53
App endix C (Con tin ued) # simple script to log time and v oltage from SR510 lo c k in # serial in terface starttime =`date +% s` startfreq =375 # T oggle some lines to get p ort going ./setserial /dev / tt yS0 1 0 ./setserial /dev / tt yS0 1 1 ec ho "Start time: `date `" > data.log ec ho "Phase : `./queryserial2 /dev /tt yS0 9600 1000 P j tr n r' n n '` >> data.log while true ; do elapsed = $((`date +%s ` $starttime )) lam b da =` ec ho "($startfreq 3.9+4)/1000+0.000590779 $elapsed j b c l` ec ho $lam b da " n `./queryserial2 /dev /tt yS0 9600 1000 Q j tr n r' n n '` >> data.log done 54
App endix D Oriel 77250 Mono c hromator Calibration and Use The mono c hromator w as calibrated using a Helium-Neon laser and a Newp ort 835 Optical P o w er Meter. The laser w as aimed in to one slit of the mono c hromator and the p o w er meter w as placed at the other slit, ha ving b een set to detect 630nm ligh t. The mono c hromator w as man ually crank ed to nd the 1st, 2nd, etc. order p eaks and the dial reading of eac h p eak w as recorded. Since eac h p eak o ccurs at an in teger m ultiple of the fundamen tal laser w a v elength, the frequency is just the order times 632.8nm. The equation of the calibration line w as found using Mathematic a order w a v elength (nm) dial 1 632.8 152 2 1265.6 311.5 3 1898.4 471.5 4 2531.2 629.8 5 3164.0 790.5 6 3796.8 953.2 T able D.1. Mono c hromator Calibration Data 400 600 800 1000 1500 2000 2500 3000 3500 Figure D.1. W a v elength(nm) vs. Mono c hromator Dial (0-999) = 3 : 95( dial ) + 34 : 48 (D.1) 55
App endix D (Con tin ued) The mono c hromator has a small motor whic h turns the grating. A stop w atc h w as used to measure the rate at whic h the grating turns. That's it. There isn't ev en an on/o switc h. (If one is needed, try using a p o w erstrip.) It is the resp onsibilit y of the in v estigator to sync hronize the mono c hromator with a timer. 56