“119190;; .. 3333’ 11 I, 1 ,1, 3'33"?” 31111,, 1‘. 11 ..'.'5'I'1.'5'55555'-";5§l1~ 15"51' .. :1 11‘ 11 31111111.. 11 31‘» a "5"-" 1 1 l'n‘v'; 6;,1355' "'3'. a ;_‘- _:.L__ 3—— (-3 :{):~ _ 7.2:“;— -.. _____.. ”'5' 9135' a . 15“" 551'? ' 555,133,}, W153 :33.- 3.}, {fay 1.3133333 35‘" 333;. r u 3'1. 5‘" ' 3. 3111.331, 11137- 1:E"""'"'1115~;i, _ _ 5 555351" '1'1'155'5I555 535' "hi-15555555 55551551 1 1111 ~11 . 1111",“ , {.11- 1551:1115""51, 11" '1'. 1"" 13311311115"“ 1 1.1.5.5 . - "1.1...11111111‘11' 3'3, 3' 3 31.5111" 5 . 5,1. 5 '3, 55,1. ', ,.'11 ."'»"""' "' ' 5 '5'53515‘~1555'5'5'5"55'11"'33'13,, 3.155155 ' 0 ~—.-‘ 3. H my 31‘- .- ’4 ....- .. o— - _ _.g.—.J' £33 _._--» p...‘ “.3...— 5:195} ~11- J: U '5 “"313,- It """'3""; . 551,5 .1551- ‘ I 3'. l 5' , 15”, :‘1‘ U) 1 I 1 I 345,5155 ,5, | (,A'ag 5'555'5'5f'751 I: 5' ‘ H ' v . . ~ 1 3, 55,,5 50,1,1, :53. 55-1,. , ,. :2 , .. . . '.' .-.; 5' I ”5”. - 3.5!. (3.1,;;;5 2"" (‘5', '5'35 051:1. x, .u‘ 4". ‘ _ I; 'y‘ I . : . ‘-:o . u .' . . v . ' 1 n , ,I 111131;“.13I31‘3. '. 1 51”,; 1_.._j ..'~,-I'L"1:z:1; 3,5, .. , .. , . . , .. ‘ ' " "-'o""11'1'f'5":11'1'1%1'1331'"‘31")"""""‘1""°'?‘~1.-3.3-; ' '1'535‘1'“ : ‘ I.‘-':'.:,:' .. .- . ' 3 ‘ 3 1 ‘ , 1 ':,',,1.: 5:55:1555555“. ,~ :.:"‘IJ'XI:‘: J - 1. ' . $24,," _ '. .3... " .' . "'.',-, a... ' . o . l‘: ‘ 3. ' ' '55'5555'1151'5" '514355'5H1' """' ‘ '55F':"55"5'5'55"' 7'3!" ' .- " - ' ’1': . - .3 .. ;‘ ,.- *- .;3<:_,., I ' ~‘ -. . - ',. . -. .. ... , - . ‘ 5.; , j. ." , -( . ‘ 1". '3... -, .3... ‘_ . 1;; . . , 2 2‘1". '. . _' . .' ‘ : . v 1.5,}, -::,5~'=5"'t55", "55'3'5"'5511'5"'.' ’ '5'1' 155:5 :53“ 5 ,,.._,J; L . 3h: I . .. .' ' ~ 3. .: ”59’” . H. .<- .:. 2' . . , _ ‘ . . H 15‘ .' .' " ' ' '1" . '5 "117"“: 3'. '.'..'. " 5L " '31" '1'3'; v ‘ _ 5 :" :1. ‘ ,: ‘ .1 "1.. .-‘ , , , 5.19"" ,. ; 2,21. 5'31'555. 5‘2"" ~I <' .5"' 1 ~,;.', , '.‘ ; ~:. : 35%: , ...~ :' , , 1:' .z-z: -, 1.1. . ' .- 1‘ ‘ -.T=- 1.35131; 2. ;.- 5_ . -,|, .' _ 1. ’10- '155'". ,: ‘5 :.. , f‘ ;.;l : 1.,” .3 .. '5, ,. _; , .152" . . . I. . . . q 5,; 5.35“, 7.3" , ”3,; ;. ., .-, , 1. 3 .. '2' . :- :_ ;. . -.- 9,3. - , t. . 1.1 .2, -. I1, ""5' , 3,, , “1,; 1; .. , 3:25:15“. -, , -. - ' _ 3 3. v '2' .: :3... 13;. an, :v. ‘ . . .’ u':"": -' ,5": . "::_. I,.,::.:' 3:“ .- .3. 3:35, ‘ - ,:::,;::.V,.,S‘I '. 1* .3 ‘55-"',, :53... E.‘ . . :a‘ , -.":_‘r' _o . . ' '3' 5 1'43" ' ' " '3" (33:11." ' ' ;:.' . “ ..I.1 :.. : 2T1 , ‘ '. ‘1 ..1 .. . '531-‘5 “I .- I.; .1“ ‘3 .. . :3 u. _ .:.f.1; . . , 4, 5;“: ; _ ' .. r. ._ _‘. 3 1' ' -; :-~.:? 5- :2 :v' .=~-11~ r- _- 1 .. '1 ... ‘ ~, -, :—~-.: f1 2.11: .1 . . ‘53-‘33": L's-1'37.- ~ . . “=31 -..-.‘1-..'.. '3: “1-.- ~: ‘ “4 ' ' ‘ 1 Tr""':17~1?.'-1~' :;. . ‘ W. -- .. 5 5 ,5: ,1 2 ;, 5.,3, ' ‘: '. 1 , ._" ,'-.."' "'.‘..j' r, ‘51:“. >,.. _ . ‘ ,_l..-., .,.',5. .u 5 a . . -‘ . "' . ;'1 .1", ' ." 1 """.1:'1. .1. 1'.' "1:'.' '5.'" 1" - , ,L ,5: .3' - . .' ,3'. ' . I . " . ., - ' I1, ,‘_ “.1 , ' . ,‘ 1,; .'1. , ,,.v,;,,.,:ly,5:v'] . 275.251 '3 , t: 3, :.:,:._:.I ' .' ~1 ,4 ‘1 . 'Jt ‘ “3:5, ..... .1 - H .' ' ' "' ' ' "' ' " " ' " "" "1.‘n".‘1' '1' . "."1~ ' 3.2-. ‘ '1"' ' '.' '5 .04: '2 ‘ . .' '.' . 1 ' " "11". ". '1 ' '1' " "" ' '1' "3 . "-. '5. '.1'.-.'m"f3‘ 7. ~-:' ' ' ":11 ‘1". -.‘ " '7 3 1 . r: " ' , : ‘ ' ‘ -. 3 1:22.: :~ 2: .:- t. 531:. .. ' ' : l " ' ' ‘ 1 ‘ ‘ 1 .' 33‘s... 2-H . 3.. 1. - 3 '::r=f.;:‘.,. .13; ' 3.3-: '.x;...;. ~11 5 ' ' ' ' 1' ' . ' i . 31' 1 .' 3- ~91: . :~' 1. .,.-- . ‘j1 . :1 . ’[1'3' 1'. " 3% ' if. " "L9 ' 11.. '1; i~’~li-':."~"5'=‘-rf;. ‘2‘7-31',r'5:-f :-~ 1: 3"? '1 "N313" 5 11 3 ' 1. '1. ‘1 ‘. :3: 1-"-'I:‘-;. L1;.1. 1} -, .‘- ,13—1 1 ',.-: .2; . :' ' ' "'1 ,‘11'7 ' . 1,1,, '_ '.‘.,::.' ',_ ' ""5? 1 3' '. " ,, 1' ""5, 1,'.,'-'." "1 5 ." ; ,,5'_".'A‘f::_, ;._' ' 512" 2 ‘ ,zI';‘j ,- ' , -. _-- :.‘ I. .1. :', '.;, 1.1.11... .' 1f. ‘ ‘ 1' .,1,,,, 5,3 :3 3, 1'5 . .-.‘ 1 '. . . ' " 3. .‘ 1" 1- ,1 1, ‘ ' ‘ , ‘ , . 1555'; ' . .1' .' '2.-‘|I‘. . . '1 1 1: 1 - 3 11', . 5 1 '.' f V -2 '. 5 ' 1 0 . _',. . 3 1. ' , , . ”5"":5 ,..5,' 5'I1',51'.511,t ""5'1' 15555},;|.5’|',"15.,3- . -'.‘ 1' -' .',, 1:, . : , . . 1.1 ,.1 551' 35"15I"'5,'1'£t5' 51'5" 5:533:555‘11:5.'|' 5'” '.‘- ' . ' ' 5 '5', .1 ,1 ' . ;,| " 5;" t " ' 2' 1'1 ' '.5‘:,5), ‘15,. "5 5 5 55., , “:11 u ,,.I .' 1. 1- .. ‘ :"-,'é:;."'1"' ,5511 1113'“, «‘05.. '11.,» $151,:- 55,3115, I. . . ., . .1-.-.o _ . ,I -.f1:.', :1 . 5h, 1 ,. 501055, -,,,'1- '5‘ . ,3" ,.. .. 55' , , : . . . : ' -‘ ‘ 11. '5 5' . ‘: 1 -‘. .. 1 '. 1‘ :1. "1" ' "" ' ‘ ""' 3., " '5‘” 1'1“”. 311.1,11'5'5'55 5'5’. ' " ._ . 1 3"' ., 1 3‘ 1'1 1 ' ,1 [‘1‘1 5'1'.111 , ' . I5 5' ' 3..“ . " ' ' " ' 5 ' 5'5'5.5"1"' 1|" '5', 5, 1 5'5 W5 '5' "" "" ' 1"" ""5:'.'")"'5""551""' '5" '5" 5" 5" ' "|"5"" 5' " ""5"'5"' 5 5, 5'55 "5 555555", :"1 5 ‘ 1“1""1 ' 1 .1 1 .'.',111 51,1 5"51'51 '5 '551. 2315'1. ' 5' ,‘ 1"" 1' ‘ .'""|11' '115'51'5 l5, "' 555.1', |."' ,1555', """""55'5555"5"‘"1'1"5',555'55'5"555555555 ,55'55 1. , , ‘ "‘ . , l | ,1“le "1 ‘1~' ’ 1 1 "‘5' " .1'1' '1""5"1"11' '5.'1,1,|5,111, ' 5'55" "'15 5 15 ,1 .1 11.Ii:1,,,,;,1.. 111-11,... 1555 ..., W5 ,. ,. : . . , 1115'. 111.. .1' . 1 1 '31. . ,1, .- , 55,15“, 555,5;- ‘ 5'..,'1‘ ., ".T "'71"'1'"" ,1515' 5555'! ,5“ “"555” 55555551555 ' ' . ‘ . .u' 115'. . 5' ' ' 5 55555 ..|I5I'. '15.:"" "15.555 '55'5'5'L 55555.115" 'L on: A. “*w m ,..-,....._~! I. 114F333; W MiChigan 3) ca u? a, University This is to certify that the thesis entitled Two-Photon Spectroscopy of Linear Polyenes presented by Howard L. B. Fang has been accepted towards fulfillment of the requirements for _2h14l._—degree in 4mm awn; Major professor Date 8/26/77 0-7 639 TWO-PHOTON SPECTROSCOPY OF LINEAR POLYENES BY Howard Lih-Bao Fang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1977 ABSTRACT TWO-PHOTON SPECTROSCOPY OF LINEAR POLYENES BY Howard Lih-Bao Fang Linear, all-trans polyenes have received considerable attention because of their importance in photochemical and biological systems. Recent calculations and experi- ments indicate that, like the electronic ground state, the lowest excited singlet state in this centrosymmetric chromophore has A 9 the normal electric-dipole absorption spectrum because symmetry, and thus is forbidden in the transition violates the u«+ 9 selection rule. However, 1A 4+ 1Ag transitions are allowed in two-photon absorption, 9 and it appeared that two-photon spectrosc0py would enable direct observation of these hidden excited states. Two types of linear polyenes were investigated in this research: diphenylsubstituted polyenes and retinyl polyenes. The two-identical-photon excitation (TIPE) spectrum of all-trans diphenyloctatetraene (UFO) in EPA at 77°K has been obtained over the spectral range 10500- 12700 cm'l. Near infrared radiation from a tunable dye laser pumped by a Q-switched ruby laser served as the ex- citation source. A sharp origin and distinct vibrational Howard Lih-Bao Fang structure were observed in the TIPE spectrum of DPO. The 1 1 below the excited Ag state origin is about 2100 cm- 1Bu origin and is essentially coincident with the one- photon fluorescence origin. A low-lying 1Ag state in DPO has been confirmed. Several retinyl polyenes of the vitamin A family (all- trans retinol and anhydrovitamin A), which are relevant to the mechanism of vision, have also been investigated in EPA at 77°K by TIPE spectroscopy. Each shows structure suggesting the presence of a 1 1 Ag excited electronic state below the lowest Bu allowed state. The excited 1Ag state origin of anhydrovitamin A is about 1600 cm.-1 below the 1Bu origin. The spectrum of retinol is too broad to establish the origin. The existence of these forbidden states appears to explain the anomalous fluorescence properties of these polyene molecules such as the huge Stokes shift, the very long fluorescence lifetime and the abnormal solvent shift behavior. To My Parents ii ACKNOWLEDGMENTS I wish to express my most sincere gratitude to Professor George Leroi for his direction, encouragement, and patience. Without his help, this work could hardly have been completed. I am also deeply grateful to Professor Schwendeman, who served as my second reader, for his valuable comments and guidance. Professor El-Bayoumi was most generous in lending me a fluorimeter instrument for some experiments. I would also like to thank Dr. Robert Thrash for his help in constructing the experimental apparatus, and for his suggestions and understanding. Other group members and our secretary, Mrs. Naomi Hack, deserve thanks for their friendship and stimulation. Financial support from MSU and NSF is greatly appre- ciated. Finally, I wish to thank my wife, Julia, for her understanding and support while busily engaged in her own music career. iii TABLE OF CONTENTS LIST OF TABLES . C O O O C O C O O O O O O O O O 0 v LIST OF FIGURES . . . . . . . . . . . . . . . . . vi Chapter I O INTRODUCT ION O O O O O C O O O O O I O O O 1 II. A GENERAL SURVEY OF TWO-PHOTON ABSORPTION IN RELATION TO EXPERIMENTAL CONSIDERATION 15 A 0 Motivation 0 O O O O O O O O O O O O C 15 B. Experimental Methods . . . . . . . . . 17 (1) Direct Measurement . . . . . . . 18 (2) Indirect Measurement . . . . . . 26 (3) Parametric Mixing Measurement . . 34 III. EXPERIMENTAL APPARATUS . . . . . . . . . . 39 A. Absorption Spectra . . . . . . . . . . 39 . Emission Spectra . . . . . . . . . . . 39 . Ruby Laser System . . . . . . . . . . 40 . Dye Laser System . . _. . . . . . . . 52 (1) Ruby Laser Pumped Dye Laser . . . 58 U0!!! (2) Nitrogen Laser Pumped Dye Laser . 65 (3) Flashlamp Pumped Dye Laser. . . . 72 E. Materials . . . . . . . . . . . 76 F. Two-Photon Excitation (TPE) Spectrometer . . . . . . . . . . . . . 81 IV. THE TWO-PHOTON EXCITATION SPECTRUM OF DIPHENYLOCTATETRAENE . . . . . . . . . . . 88 A. The Reason for Choosing DPO to Study . 88 B. Results and Discussion . . . . . . . . 93 C. Conclusion and Implications . . . . . 102 V. THE RETINYL POLYENES . . . . . . . . . . . 105 A. Introduction . . . . . . . . . . . . . 105 B. Results and Discussion . . . . . . . . 119 C. Summary and Conclusions . . . . . . . 130 MPERENCES O O O I O C O O O O O O O O O O O O O O 133 Table LIST OF TABLES Page Comparison of Vibrational Intervals Observed in TPE Spectrum with Those Observed in Fluorescence Spectrum for Diphenyloctatetraene . . . . . . . . . . . 102 Figure 10 11 12 LIST OF FIGURES Excitation Energy Diagrams for Butadine by Several Approximate Molecular Orbital MethOds O O O O O I I O O O O I I O O O 0 Upper: A Diagram of Molecular Orbital Wavefunctions for Butadiene. Lower: Electronic Configurations of Butadiene . Absorption and Emission Spectra of DPH in EPA at 770K 0 O O O O O O O O O O O I O 0 Absorption and Emission Spectra of DPO in EPA at 770K 0 O I O O O O O O C O O O O 0 Ground State and Four Lowest Excited States of Octatetraene: (a) Hartree- Fock Calculation; (b) PPP Semi-empirical Calculation with Singly Excited CI Included; (c) PPP with Doubly Excited CI Included . . . . . . . . . . . . . . . . (a) Two-Photon Absorption (TPA) Effect (b) Some Scattering Effects might Compete with Direct TPA Measurement . . . . . . . Two-Photon Direct Absorption Measurement Two-Photon Absorption Signal in 1-chloronaphthalene . . . . . . . . . . . Ruby Burst Spot after Thermal Blooming EffeCt. O O O O O O O D O O O O O O O 0 Dual Beam Oscillogram of TPE Measurement Fluorescence Peak Height as a Function of (b/ T) O O I O O O O O O O O O O O O O O O Resonant and Nonresonant Processes in Three-Wave-Mixing. (a) Nonresonant Processes; (b) Raman Resonance; (c) TPA Resonance . . . . . . . . . . . . . . . . Vi Page 12 20 22 23 25 29 31 35 Figure 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 A Side-View Cross Section of the Ruby Laser Double Elliptical Cavity. . . . . . End-View of Double Elliptical Cavity . . Flashlamp Excitation Pulse Forming Net- work. 0 O O O O C O O O O O O O C O O O . Block Diagrams of the Ruby Power Supply and Triggering Systems . . . . . . . . . Automatic Firing Circuit . . . . . . . . Transverse Pumping by a Cylindrical Lens set 0 O O O O O O O O O O O O O O O O O I The Tuning Range and Relative Output Power of Several Dyes . . . . . . . . . . Dye Laser Tuning System . . . . . . . . . Circuitry of the Blumlein Type N2 Laser . (a) Side-View and (b) End-View of N2 Laser; (c) Output Window of N2 Laser . . N2 Laser Pumped Dye Laser . . . . . . . . Two-Dye-Laser (N2 Laser Pumped)-Photon Induced Fluorescence Measurement . . . . Flashlamp Pumped Dye Laser . . . . . . . Sample Materials . . . . . . . . . . . . Absorption Spectra of All-trans Anhydro- vitamin A (taken at 77°K in EPA) Before (——) and After (---) Lengthy Exposure to Room Light at Room Temperature . . . . . Two-Photon Spectrometer of a Ruby Laser Pumped Dye Laser . . . . . . . . . . . . Fluorescence Housing for TPE Measurements Energy Gap Between the Absorption Origin and the Fluorescence Origin of Diphenyl Polyenes as a Function of the Number of Carbon-Carbon Double Bonds . . . . . . . vii Page 42 43 45 47 48 57 60 62 67 68 71 73 75 77 80 83 84 90 Figure 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Diffuse Band Appears on the Low Energy Side of the Absorption Origin of DPO at High Concentration (C = 2 x 10'4M) . . . Two-Photon Excitation Spectrum of DPO in EPA at 77°K. o o o o o o o o o o 0 0 o 0 Square Dependence of Two-Photon Induced Fluorescence of DPO . . . . . . . . . . Absorption and Emission Spectra of A11- trans Retinol in EPA at 77°K . . . . . . Absorption and Emission Spectra of A11- trans Retinal in EPA at 77°K . . . . . . (a) The Absorption and Emission Spectra of Axerophtene in EPA at 77°K (ref. 33). (b) The Absorption Spectra of Retinol (-—) and 5,6-dihydroretinol (---) in Ethanol of Room Temperature (ref. 132) . Absorption and Emission Spectra of A11- trans Anhydrovitamin A in EPA at 77°K . Torsional Potential for the B-ionylidene Ring in Retinal (Ref. 124b) . . . . . . One Photon Excitation Spectrum of A11- trans Retinal in EPA at 77°K . . . . . . The Absorption Spectrum of All-trans Anhydrovitamin A in EPA at 77°K as Concentration Increases (~5 x 10'5M) . . Two-Photon 2Excitation Spectrum (in unit of ln(If/ID))and the Lowest Energy One- Photon Absorption Band (in unit of rela- tive absorbance) in EPA at 77°K . . . . TPE Spectrum of All-trans Retinol in EPA at 77°K. O O O O O O O O C O O O O O O 0 Square Dependence of Two-Photon Induced Fluorescence of All-trans Retinal . . . Two-Photon Excitation Spectrum of All- trans Retinal in EPA at 77°K . . . . . . viii Page 95 97 99 108 109 111 113 114 117 121 122 126 128 129 CHAPTER 1 INTRODUCTION A linear polyene is a chain of conjugated carbon- carbon double bonds with an unbranched n electron system. The spectrosc0py of this conjugated system in the near UV and visible regions can be described by considering only the n electron structure of the molecule. Consider- able attention has been paid to the electronic spectros- copy of polyenes because of their importance as chromo- phores in photochemical and biological systems}-3 Polyenes are involved in many different kinds of photo- chemical and biological functions including cis-trans isomerization, plant and animal coloration, photosyn- thesis, phototrOpism, and vision. All of these processes involve interaction of the polyene molecules with light or with excited states of other molecules. For example, carotenoid polyenes are known to be active in photo- 4,5 synthesis, and ll-cis retinal in rhodopsin is the visual pigment in vertebrate retina.6'7 However, the mechanism by which light energy absorbed by carotenoids is transferred to chlorOphyll in order for photosynthesis to proceed is still not perfectly understood. Also, the molecular basis of how an excited state of rhodopsin creates a visual response is an unsolved problem in sensory physiology.8 The photochemical interconversion of cis-trans isomers of polyenes is of fundamental importance because it is the most basic step in most chemical and biological transformations. A basic understanding of all these aspects depends on an understanding of the electronic structure of the molecules involved. The first systematic study of polyene electronic spectroscopy, including the measurement of absorption and fluorescence spectra, investigation of the solvent dependence of spectral properties, vibrational analysis of electronic bands, and the establishment of trends with number of double bonds and end substitution, was made by Hausser, R. Kuhn, A. Smakula and co-workers in the l930's.9'14 The primary spectral feature is a strongly allowed transition in the near UV or visible region whose intensity increases (hyperchromicity) and whose energy decreases (batho— chromicity) as the chain length increases. As the chain length becomes very long, the transition energy approaches an asymptotic value of almost 18000 cm-l. Several theoretical calculations based on either neglecting electron-electron interaction or taking such electronic interaction terms into account in an approxi- 2'3'15-17 Consider the mate way have been carried out. example of trans-butadiene, which is the simplest proto- type of linear polyenes. we may compare the calculated electronic state energies for butadiene in different molecular orbital approximations. These approximate MO methods include Huckel MO theory (Mulliken),16'17 18 Free Electron MO theory (Kuhn and Baylissls), the 17'19 method with and without singly 20) Pariser-Parr-Pople excited configuration interaction (Sidman , and an ab 21) initio calculation (Buenker and Whitten . The first five energy levels in different MO theories are compared to the limited experimental values22 in Figure l, where B = -4.854 eV has been used for the Hfickel calculation, and L = 5.4 A as the one dimensional box length for the Free Electron model. These parameters were chosen to fit the energy of the first 1Bu state to the experimental value. Butadiene has four w electrons moving in a fixed core with nuclei and o electrons. In this idealized planar geometry, the four n electrons form a centrosym- metric chromophore and belong to C2h symmetry, as is assumed for all-trans linear polyenes. Linear combina- tions of these four n atomic orbitals form four molecular orbitals which are alternately au and b9 in character,23 as shown in Figure 2.~ From these Mo’s, four singly- excited singlet electronic configurations can be ob- tained. The wavefunctions for the ground state and the four excited states are also sketched in Figure 2. In all calculations, the lowest singlet excited state of trans butadiene is lBu‘ The transition from the i S1o- .. a, _ 8 "0 °. .1 ,0 . _ § I—~ :. .. . _. . LL] —— .-—-—r'. . . H —' — 2: Lu 5.. OJ _ — ab ab 1 initio‘ initio “ PPP2 PPP2 PPP3 ‘ States INOCI) (lstCI) Huckel lfiflf) (NoCI) (lstCI) (lstCI) Experimental 1 Ag(wo) 0 0 0 0 0 0 0 0 1Bubbl) 9.04 9.03 6 6 6.3 ‘ 6.3 6.21 6.0 1Ag(w2) 11.34 10.99 10.85 14.4 7.8 7.7 7.87 7.2 1Ag(w3) 12.29 12.64 10.85 9.6 7.8 7.9 8.51 18110114) 14.58 14.59 15.71 18 9.4 9.4 9.50 1) R. J. Buenker and J. L. Whitten, J.C.P. 49, 5381 (1968). 2) J. W. Sidman, J. C. P. 2:, 429 (1957). 3) R. Pariser and R. G. Parr, J. C. P. 21. 767 (1953). 4) R. S. Mulliken, Rev. Mod. Phys. 14, 765 (1942). Figure 1. Excitation Energy Diagrams for Butadiene by Several Approximate Molecular Orbital Methods. till ‘Agulp ground state singly excited states iltl doubly excited state Figure 2. Upper: A Diagram.of Molecular Orbital Wavefunctions for Butadiene. Lower: Electronic Configurations of Butadiene. ground state (1A9) to 1 Bu is strongly symmetry allowed and corresponds to the intense lowest energy absorption band normally observed in polyenes. The next two states (02 and 03) are 1Ag, and correspond to excitation from the highest filled to second unfilled orbital and from the second highest filled to lowest unfilled orbital. Their energies are very close (they are degenerate in Hfickel and PPP MO theories) and transitions to these states from the ground electronic state are symmetry forbidden. The linear combination of these two states, 02 + 03 (1A3) and $2 - 03 (1A;) are also excited elec- tronic states. One of these two 1A states, the 1A+, 9 9 becomes allowed if the polyene chain contains cis double bonds which destroy the inversion symmetry of the mole- 24 cule. It usually gives rise to the so-called "cis peak", having a transition energy between those of the 1 two lowest B states (01 and $4). The other linear u combination, 1 , had not been observed in polyene spectra until recently,gwhen Hudson and Kohler reported its determination in diphenyloctatetraene.25'26 All the theoretical calculations just presented in— dicate that the lowest excited singlet state in linear polyenes is of Bu symmetry, which corresponds to the strongly allowed bands observed in the visible or near UV regions. The polarization character predicted for such transitions is also in agreement with the experi- ments.27 However, an examination of the fluorescence properties of linear polyenes shows several discrepancies between the expected and observed results. These dis- crepancies are: (l) a large "Stokes shift" between the lowest energy peak of the absorption spectrum and the highest energy peak of the emission spectrum (examples of diphenyloctatetraene and diphenylhexatriene in EPA at 77°K are shown in Figures 3 and 4); (2) an extremely long intrinsic fluorescence lifetime, which is much longer than would be expected from the accepted formula28 relating the lifetime to the measured absorption intensity; (3) an abnormal solvent shift behavior, where the absorp- tion peak is sensitive to changes in the solvent properties but the fluorescence peak is almost unaffected. Several possibilities have been postulated to explain the anom- alous behavior, which is common to linear polyenes. The very long intrinsic fluorescence lifetimes suggest that the transition is symmetry forbidden, since the in- trinsic fluorescence lifetime is inversely proportional to the absorption oscillator strength. However, the magnitude of the absorption oscillator strength of most polyenes is very large (f m 1.0) which makes this assump- tion very unlikely. Moreover, it does not account for the other two discrepancies mentioned above. The lack of overlap between absorption and emission spectra could be a consequence of an electronic transition involving a large change in the molecular geometry which gives WAVENUMBERS (x103cm'1) 20 25 30 (HTVDS 88V) ALISNHINI NOISSIHH l °K 1,6 Diphenyl-l,3,S-hexatriene in EPA at 77 o o .0 . 1 o o .0 o C .0 z 1 “i O o c 8 1 I 0’. 0'... .0 0..., l .00... o". J ' d o 0...... 5.... o a... 0. ‘ :- a .0. O... o... u.....ooooooooooooo. . on... .00... ".1210...- O I O. .. . . C... 1 “0000000000000... 0.00'Coooooogooloooooocoooo HDNVHHOSHV 500 460 WAVELENGTH (NM) 300 Absorption and Emission Spectra of DPH in EPA at 77°K. Figure 3. con .xonn us ibt P id on SDNWUOSEV 10 rise to the Franck-Condon forbiddeness. Franck-Condon forbidden electronic transitions usually exhibit long progressions in one or more vibrational modes with smoothly varying intensity, because as the vibrational quantum numbers become higher, the vibrational energy levels become more closely-packed. However, for most polyenes, the absorption and fluorescence spectra show only three or four quanta, and this vibrational structure is sharp and discretely varying in one or two normal modes. Furthermore, the onsets of both absorption and emission are sharp and strong. Therefore, this possibility is not likely, either. All these discrepancies can be explained by a third possibility: the existence of a low-lying symmetry- forbidden electronic state. This forbidden state is close in energy to the strongly allowed 1 Bu state, but lies below it. The observed fluorescence comes from this symmetry—forbidden state through vibronic coupling with allowed neighboring states. The intrinsic fluorescence lifetime is long because the fluorescence comes from a symmetry-forbidden state. The Stokes shift is large because the absorption and emission involve two different excited electronic states, and the fluorescent origin lies below that of the 1 Bu absorption state. The solvent shift behavior is different for absorption and emission spectra because the upper states involve two different 11 dipole moments, which give rise to different interactions with the environment. If the existence of this low- lying symmetry-forbidden state is correct, the lowest 1 excited singlet state is no longer the Bu state, but is of Ag symmetry. Transitions to such states are normally forbidden by one-photon selection rules because the 1 electronic ground state is Ag, and only u4a~g transitions are permitted for centrosymmetric chromophores.29 This explanation suggests that a fundamental modification of state energy ordering in polyene molecules is needed. Several recalculations of the polyene energy ordering 30-32 have been carried out. The calculations have in- cluded not only singly excited configurations, but also configuration interaction with both singly and doubly excited states.21 One interesting result is that the inclusion of at least double excitation lowers the energy 1 of the Ag- state to an extent that it becomes the lowest excited singlet state. A typical example, calculated 30,32 for octatetraene by Schulten and Karplus using a double CI PPP method, is shown in Figure 5. Ab initio molecular orbital calculations for butadiene with doubly 21 excited configuration interaction included also indicate 2A3, although the calcu- the lowering characteristic of lated excitation energy for the 1Bu state is about 4 eV higher than the experimentally observed value. Although the transition from the ground state to Eflr' B 1 ._____JL_L.. Bu €3- .”‘?h:? x. . 1A + 1 -. Bu g 13 'x A41" H ""H-u. % L6 1 1 ().. .._f:L___. ...... . Ag HF 5 ° (9) (b) 5+0 (C) Figure 5. Ground state and four lowest excited states of octatetraene: (a) Hartree-Fock calculation: (b) P-P-P semi-empirical calculation with singly excited CI included; (c) P-P-P with doubly excited CI included. (reference 30) Butadiene Hexatriene Octatetraene .8. 8+0 s 0.000(1A ) 0.0003(1A ) 0.000(1A ) o 8 8 8 l 1 1 5.465( Bu) 5.408( Ag) 4.633( Bu) 4 1 1 1 7.371 A 5.860 8 6.591 A S (g) (u) (g) 7.624(1A ) 7.836(1A ) 6.672(1A ) s 8 8 8 1 l 1 9.582( Bu) 9.408( Ag) 7.48S( Bu) 7 a) The ground state energy is lowered b) The ground state energy is lowered c) The ground state energy is lowered 3+0 §_ 5+0 .000b(1A ) 0.000(1A ) 0.000c(1A ) g ' 8 8 1 1 1 .731 A 4.112 8 4.422 A (g) (u) (g) 1 1 l .182( Bu) 5.926( Ag) 4.830( Bu) 1 l 1 .630( Bu) 6.058( Ag) 5.268( Bu) 1 1 l .204 A 7.010 B 5.890 A (g) (u) (g) by 0.554 eV relative to S. by 0.817 eV relative to S. by 1.068 eV relative to S. 13 1 the A; state can be observed as cis peaks when some cis isomers are present, the 1 A; state remains forbidden even in the cis isomers of polyenes. Recently, low tempera- ture, high resolution polyene absorption and fluores- cence spectra have been obtained by Hudson, Christensen 25'33-37 using a mixed crystal technique. and Kohler, The compounds isolated in Shpolskii hosts included di- phenylhexatriene,37 diphenyloctatetraene,25 2,10-dimethyl- undecapentaene,34 2,4,6,8-decatetraene,36 and 2,12-dimethyl- tridecahexaene.3S These studies reveal a series of weak absorption lines on the red shoulder of the strongly allowed transition, which roughly show a mirror image of the emission spectra. In order to more easily observe such low-lying symmetry-forbidden transitions, a method which does not abide by the one-photon optical selection rules is necessary. Specialized techniques such as electron impact spectrosc0py (EIS) and magnetic circular dichroism (MCD), which do not follow Optical selection rules, might be used to locate optically-forbidden states. 38 and MCD39 have insufficient resolution to 1 However, EIS identify the low-lying- A; states of linear polyenes. Two direct and promising methods to observe such states have been developed in this laboratory: two-photon 40'41 and preresonance Raman excitation pro- 42,43 spectroscopy file spectrosc0py. The unique even-parity selection rule of two-photon spectroscopy and the electronic 14 transition enhancement of resonance Raman excitation pro- files reveal important information regarding such "hidden" states. In this thesis, the observation of a hidden electronic state by two-photon excitation spectroscopy will be described. (This work contains two parts: (1) the two—photon excitation (TPE) spectroscopy of all-trans diphenyl sub- stituted polyenes such as l,8-diphenyloctatetraene, and (2) the TPE spectrosc0py of all-trans retinyl polyenes such as retinol (Vitamin A) and anhydrovitamin A. The TPE spectra of these compounds in EPA (ether-isopentane- ethanol, 5:5:2 by volume) have been obtained at 77°K over the spectral range 10000-12500cm-1. Near infrared radiation from a tunable dye laser pumped by a Q-switched ruby laser served as the excitation source. The medium- resolution TPE spectra show the existence of low-lying 1Ag states in each compound studied. The presence of an excited singlet state in linear polyenes at an energy below that of the lBu state not only explains the apparently anomalous fluorescence properties of polyenes but also provides a fundamental modification of electronic state ordering in polyenes. Calculations which include doubly excited configuration interaction are apparently able to account for the pres- ence of this low lying state, but the present theoretical picture is by no means complete. CHAPTER II A GENERAL SURVEY OF TWO-PHOTON ABSORPTION IN RELATION TO EXPERIMENTAL CONSIDERATIONS A. Motivation The high light intensities available from powerful lasers have opened several new nonlinear spectroscopic techniques which involve a nonlinear response of a medium to applied irradiating fields. Here nonlinearity means that the electric polarization or the induced electric dipole density of the medium is a quadratic, cubic or higher order function of the electric field amplitude. The response functions to the fields can be calculated quantum mechanically by higher—order, time-dependent 44,45 perturbation theory. An early example is the cal- culation of two-photon absorption (TPA) intensity by Goeppert-Mayer in 1931.46 Compared to many other second order and higher order nonlinear optical processes, two- photon spectroscopy is relatively amenable to both ex- perimental observation and theoretical interpretation. Indeed, TPA serves as a prototype for many other non- linear processes. The limitations of linear optics have made the study of two-photon spectroscopy inevitable. Although one- photon spectroscopy is very successful in helping 15 16 understand matter, it explores only a fraction of the possible eigenstates. There are still many forbidden transitions which cannot be detected. However, allowed transitions in TPA cover a much wider range, usually that range inaccessible to allowed one-photon absorption. For example, in centrosymmetric molecules TPA permits a direct examination of those states having the same parity as the ground state, whereas one-photon transitions to such states violate the Laporte symmetry rule. Addition- ally, polarization analysis provides more information for two-photon than for one-photon processes. Polarization studies in random gases, liquids and glasses are unin- formative in one-photon absorption because of an averag- ing out of polarization effects. The one-photon transi- tion probability w(l) from ground state Ig> to final state . The result is =‘% (A-K*)(Efg°figg). Since X'A* = l for all polarization (linear, circular or elliptical), the ab- sorption then is completely independent of polarization. In TPA, the two-photon transition probability w(2) is proportional to IX'gfg-filz, where X and K are the polari- zation vectors of the two absorbed photons and gfg is the 17 two-photon transition tensor. The average value of is a linear function of two variables IA-filz and |1°;*|2, where ;* is the complex conjugate of 3.47750 A complete polarization study of two-photon transitions by suitable selections of polarization vectors 3 and A provides enough information to permit an unequivocal identification of the symmetry species of the excited states in a random system. Therefore, TPA in randomly oriented phases can uniquely identify the symmetry of eigenstates of matter in situations where the direction of the polarization vectors can be independently varied. TPA then becomes a powerful tool to supplement one-photon spectroscopy. B. Experimental Methods Since the first observation of TPA by Kaiser and 51 who showed that the simultaneous absorption of Garrett, two ruby photons by Eu++ in crystalline Can leads to blue fluorescence, a considerable amount of work has been done, both on the study of the absorption process and on its application to spectroscopic purposes. The basic quantity of TPA is the two-photon cross section, 6, which relates the two-photon transition probability per unit (2) time, w , to the product of the photon fluxes F1 and F2 by w(2) = 6F1F2. At the time of writing, there are three general ways to observe two-photon cross sections: 18 (1) direct measuremen (2) indirect measurement, and (3) parametric mixing measurement.57’58 (1) Direct Measurement: In a direct measurement,59 two light beams are direct- ed onto the sample, one a continuous probe beam and the second a powerful, pulsed pump beam, and the intensity change of the probe beam is monitored during the pulse of the pump laser. The fraction of the probe beam intensity absorbed is related to the two-photon absorption cross section 6 by 6 = (AI/I)/C1F, where C is the sample concentration in molecules/cm3, l is the length of the sample cell in cm, and F is the photon flux of the pump laser in photons/cm2 sec. The units of 6 in this case 4 1molecule—1. An absorptivity of 10-50 are cm sec Photon- in these units is known informally as l maria. Ideally, two-photon direct absorption measurements should provide absolute values of the 6's. However, the technique suffers several substantial experimental dif- ficulties which make direct measurement experimentally impractical. The limiting factors are: (l) relatively low sensitivity, which prevents the study of molecules in dilute solutions, (2) inevitable side effects in the sample which increase the uncertainty in the determination of 6, and (3) uncertainties in the spatial and temporal properties of each individual laser beam which strongly l9 influence the 6 value.60 The TPA cross sections for most organic molecules are on the order of a few marias. In order to overcome such small values of 6, a powerful, Q-switched pump laser is necessary. Even for a good probe beam detector, a realistic value of the intensity change would be at least g; = 10.2 (a 1% dip in the probe beam). If C = I 1022 molecules/cm3 (a neat fluid sample) and l = 1 cm, the corresponding photon flux of the pump laser must be F = 1026 photons/cm2 sec, which is 20 Mw/cm2 in the vis- ible region. In other words, even for a megawatt pump laser, a highly concentrated sample is required for direct measurement of two-photon absorption. This pre- cludes the possibilities of observing direct absorption in solutions. As the power of the pump laser increases, several side effects such as stimulated scattering and thermal heating of sample may occur. These effects include the production of stimulated Raman scattering which depletes 61 the laser intensity in the sample, self-focusing which 62,63 59 enhances the laser intensity, sample filamentation, 59'64 All increase the and the formation of shock waves. uncertainty in determination of TPA 6's. In addition, many competing processes are of the same order of magni- tude as TPA.65'66 Some competitive effects are shown schematically in Figure 6(b). In many cases, it is w2 (1)2 (.01 “’2 1\ ml m1 ml _J h——. \‘~—_______———""F“-—_________,.I’ (a) (b) Figure 6. (a) Two-Photon Absorption (TPA) Effect (b) Some Scattering Effects might Compete with Direct TPA Measurement 21 impossible to distinguish, or to estimate the relative size of these competing processes. Therefore, the experi- mentally determined value of the two-photon cross section contains contributions from unwanted processes, and is not an absolute TPA cross section. The third uncertainty in determination of 6 arises from some characteristics of the lasers used. Any er- ratic temporal or spatial behavior of the excitation source can strongly affect the experimental value of ’the observed cross section. This intrinsic uncertainty of the laser beam is serious and cannot be completely resolved. It affects all kinds of TPA measurements, not solely direct measurement. Further discussion will be found in the next section, on the indirect measure- ment technique. An experimental example of the determination of 6 at a single wavelength is schematically shown in Figures 7 and 8. The sample was pure fluid a-chloronaphthalene in a 1 cm long cuvette. The 4765 A line from cw argon ion laser served as the probe beam. The pump beam was a Q- switched ruby laser. In order to obtain an entire two- photon direct absorption spectrum, the single wavelength probe beam should be replaced by a tunable light source (either a stable cw dye laser or a conventional light source). Careful protection of the detector (the photo- diode P2 in Figure 7) against the huge noise spike coming thEmmDmflmmmm no go .r as. as. .2 «as 2 =8 ._ a 2 a2 ”.2 :53 Load: .03. .33 38?... O D 2 it} fr 5 23 .xpw>mo sommfi +h< cry opwm:w :oflmuo cm mcflomam xn toumcfiswfio on :mo poop mach .oomhu Hmcmfim any so comm on E3 2: cm? as mowocosoohm “mob poms; wooed . a Eu . u x P. 1 _ oflsuoHoE :ouo; \uomnV 33?: H e a n-Oel \flfi mzmq<292mmw\uom:ooH mqHw\>EH Cw 4fit\>m cw >mDm 24 from the pump laser, and of the sample from overheating by the probe beam was required. Several ruby light rejection filters followed by a narrow-band 4765 A spike filter and a lOOu pinhole were placed in front of the photodiode to eliminate pulsed ruby light at the de- tector. An electromechanical shutter which was syn- chronized with the triggering system of the ruby laser was placed in front of the Ar+ laser to avoid overheating the sample by the probe beam. Just before the ruby was ready to fire, the shutter was opened. Then, the probe beam passed through the sample and filters and finally was focused by a lens through the pinhole onto the sur- face of the photodiode. Without the shutter, the heat generated in the sample, which has very low but finite absorptivity, caused a refractive index gradient which made the sample behave like a divergent lens. This phenomenom is well known as the thermal blooming ef- fect.63 When the ruby laser pulse passed through this divergent lens, the intensity distribution of the original Gaussian mode was slightly deflected, and took on a ”donut" shape with less intensity in the center. A burst spot obtained by inserting a piece of film at position(o) in Figure 7 looked like: 25 Figure 9. Ruby Burst Spot After Thermal Blooming Effect. The thermal blooming effect definitely affects the mea- surement of two-photon cross sections because the photon flux in the middle of a donut-shaped distribution is much lower than that in a Gaussian distribution. In other words, the sample molecules sitting along the center of the probe beam cannot see the pump laser efficiently. Therefore, the measured value of 6 will be slightly smaller than the true value. The two-photon spectrum of neat a-chloronaphthalene 47a was studied by Monson and McClain using the direct measurement technique. Three two-photon absorption bands were observed, at 42600, 37700 and 34400 cm-l. A complete polarization study was done to assign the symmetry of the transitions.47-49 The two bands, 42600 and 37700 cm-l, correspond to 1A + 1A and 1B1 + 1A transitions which 9 9 9 9 are totally forbidden in one-photon spectroscopy. The 1 weak absorption at 34400 cm' is probably due to a vib- . . l 1 . . . . ronically induced BZu + Ag trans1tion. The combination of a 4765 A Ar+ laser and a 6943 A ruby laser described 26 in Figure 6 gives an energy of 35400 cm.1 which falls in the vibronically induced lB2u + 1Ag transition. The measured two-photon cross section at this wavelength, 1.1 maria, is in agreement with McClain's value.4*7a (2) Indirect Measurement: Indirect measurement monitors consequences of laser- induced nonlinear processes initiated by primary two- photon absorption. Among two-photon induced secondary 54,55 67,68 processes are optical emission, ionization, 69,70 71 photodetachment and photochemical reactions. Each process has its own setup and detection system. A typical example is the two-photon excitation (TPE) in- direct measurement. It is based on the experimental study of the energy dependence of two-photon fluores- cence spectra induced by tunable lasers, and is the technique used in determining the low-lying forbidden electronic states of polyenes described in this thesis. The usual experimental arrangement72 is the following: the sample under study is irradiated by light from one or two lasers, at least one of which is tunable; part (but not all) of the two-photon induced fluorescence is collected by an optical system, and after apprOpriate filtering (either selected Optical filters or a mono- chromator), is detected by a photomultiplier. Compared 27 with the TPA direct measurement, TPE is much more easily detected. This is readily understood upon substitution of typical experimental parameters into the TPE rate equation: w<2> = QéNFlFZ, where Q is the fluorescence quantum yield and N is the number of sample molecules in the active region. For an ordinary pulsed dye laser with power :20 kw in the visible region, the photon flux is 024 2 correspondingly about 1 photons/cm sec if the dye laser is partially focused into a volume of approximately 0.1 cm3. The number of sample molecules in a 10.3 M solution within this region of focus is almost 1017. Even if the cross section of the sample is only 0.1 maria, the rate of two-photon fluorescence induced from this single laser pulse would still be w(2) = Q x 10'51 x 1017 x (1024)2 = 1014Q. Therefore, TPE should be relatively easy to study under the influence of a tunable, moderate power (m 50 kw) dye laser. However, TPE still has a serious problem, the stability and reproducibility of the laser pulse, mentioned earlier in relation to TPA direct measurements.60 The temporal characteristics of a laser pulse, which strongly influence the experi- mental data, will be discussed from a theoretical stand- point in the following paragraphs. In the time domain, the ordinary shape of a laser pulse can be described by a Gaussian beam: I(t) = a exp (-t2/2b2) where a is a constant and b is the root-mean- 28 square deviation of t from the mean of the Gaussian distribution. Since two-photon induced fluorescence can be related to 12(t) by a time correlation function, one can apply the well-known deconvolution technique73'74 to the TPE fluorescence signal giving by: F(t) i/d: Iz(t')K(t-t')dt' where K is the fluorescence correlation function. Usually the fluorescence function is simply an exponentially decaying function.73 Then the correla- _ I tion function can be represented by: K(t-t') = % exp(- £7$—), where r is the measured fluorescence lifetime of the sample molecule. Substituting the input Gaussian beam I(t) and the fluorescence correlation function K(t-t') into the integral of F(t) and evaluating leads to the relationship: 2 t 2 = -a— -t/T - E. - _b_ 2 b I F(t) T e {_mEGXP (b 21’ ‘1' fiJdt 2 2 2 Z(t) 2 ___ .47: b e-t/r eb /4T [_21_ + 1 f 8-2 dz] «Zr 0 32. Zr error function, which can be evaluated from mathematical 75 where Z(t) = (%|- ). The integral here is the standard tables. The fluorescence response F(t), then, is de- pendent on three variables (t,b,r). For each particular choice of b and r, (bi' Ti), the fluorescence response function F(t,bi,ri), shows a peak in the time domain. This peak is designated F (t,bi,ti), and is the fluores- max cence peak height seen on the oscillOSCOpe. If a TPE 29 experiment is monitored by observing both the laser and fluorescence signal on a fast oscilloscope, where the upper trace shows the laser pulse and the bottom trace shows the fluorescence pulse, the oscillogram looks like: 2b. . 1 —9’ K— 1 u Laser Trace I T Fluorescence Trace F max Figure 10. Dual Beam Oscillogram of TPE Measurement. The measured fluorescence peak height Fmaxit’bi'Ti) is strongly dependent on the values of bi and Ti. If the laser pulse shape, bi' changes, the fluorescence peak height will change too. In other words, the fluores- cence intensity will depend strongly on the laser pulse shape. If the laser pulse is not reproducible, the measured fluorescence intensity will fluctuate from 30 pulse to pulse. A quantitative plot of Fmax(t,b,r) as a function of the ratio of the laser pulse width to the fluorescence lifetime, g., has been calculated by Swofford 23 31.60 and is shown in Figure 11. When the g» value is greater than 2 (that is the laser pulse width is at least twice the fluorescence lifetime of the sample), only a small variation appears in fluorescence peak intensity F As the 2* value max’ becomes smaller and smaller (the laser pulse width ap- proaches the fluorescence lifetime), large fluctuations are expected in Fmax' In order to obtain reliable TPE measurements, two conditions are required: (a) a stable pulse laser, and (b) the pulse width of excitation laser must be longer than the fluorescence lifetime of the molecule under study. This indicates that subnanosecond pulses such as mode-locked laser pulses cannot be used to induce the TPE peaks. As the laser pulse becomes really sharp, the pulse peak will be a short spike and cannot be detected by a photodiode and oscilloscope because it is faster than the risetime of these detectors. Therefore, any erratic temporal behavior of the tunable laser source can introduce serious errors into the observed TPE spec- trum . In the TPE experiments described in this thesis, the fluorescence lifetimes of the samples studied (di- phenyl polyenes and retinyl polyenes) are shorter (<8 31 L J A l A j A l O 0.5 1.0 1.5 2.0 b/T -—<) Figure 11. Fluorescence Peak Height as a Function of (b/T). 32 nsec) than the pulse width of the dye laser, which is about 30 nsec. The ratio ?§ is thus greater than 4, and the fluctuation of the fluorescence peak height caused by the variation in the pulse width of the dye laser should be negligible. The measurement of the peak height of both the laser beam and the two-photon induced fluorescence can be easily made from a photograph of a fast oscilloscope trace. This method is relatively simple but suffers difficulties if the pulse width of the dye laser fluctuates markedly. This is particularly true for a pulsed dye laser which has relatively poor beam quality. Dye laser fluctuations arise mainly from the quality of the excitation source, if the other parameters of the dye laser are fixed. Al- though the pulse shape of a ruby laser, such as that used in this work, changes from pulse to pulse, under the con- ditions of a good quality ruby rod and a uniform flash- lamp pumping system, the variation is not great. The primary problem is the Q-switching dye of the ruby laser. If the Q-switching dye is not well adjusted, or the Q- switching dye cell is not in an appropriate position, the ruby output might contain either multiple pulses or sharp spikes, superimposed on a single pulse. These spike pulses are usually very short compared with r, and increase the fluorescence signal by an unpredictable amount. Thus, short duration laser fluctuations produce 33 false peak heights in fluorescence measurements. A method.”-79 which overcomes pulse shape fluctuation problems involves the simultaneous measurement of the fluorescence peak height from a standard "reference" of already known absolute cross section and fluorescence quantum yield, and that of the sample. If the ratio of fluorescence peaks from the unknown sample and the standard reference is taken shot by shot, one will get effective cancellation of the pulse width irregularities of the dye laser. This will be true if the fluorescence lifetimes of unknown and reference are similar. However, the use of this method requires accurately known two- photon standards, which are unavailable at the present time. Also, if the lifetimes of unknown and reference are somewhat different, the correction procedure begins to breakdown. Although the fluorescence peak height is sensitive to the laser pulse shape, the total number of fluores- cence photons emitted from the sample still should be accurately proportional to the square of the total number of laser photons. Therefore, if it is possible to integrate both the upper trace and lower trace on the oscilloscope through fast integrator circuits, one should 6“ fF(t)dt I f12(t)dt very accurately. With high repetition rate pulsed dye be able to determine the TPE cross section, lasers and fast integrating devices such as the boxcar 34 integrator, several compounds have been studied in this way.80'81 (3) Parametric Mixinngeasurement: 44'57'82 has been Recently, a great deal of attention paid to the creation of light at w3 = 2w1 i ”2 when two intense laser beams at ml and “2 interact with matter. These experiments have been called four-wave mixing (4WM) for the case w3 = Zwl + “2' and three-wave mixing (3WM) for the case w3 = Zwl - wz. In both cases the in- duced polarization in the material is a cubic function of the electric field amplitudes. In cartesian coor- dinates, the ith component of the polarization is written in terms of the third-order susceptibility:44 P.(2w iw)= 2 (3) _ i l 2 jki xijk£( w3,wl,wl,iw2)Ej(w1)Ek(wl)E£(iw2) where i, j, k, 2 = x, y, z. The third-order nonlinear susceptibility is a fourth-rank tensor and contains information regarding the interference between various resonant and nonresonant processes. In 3WM, for centro- symmetric materials, there are three primary contributions to K(B): The nonresonant electronic processes (Figure 12a), stimulated Raman scattering when wl-wz corresponds to a Raman-active mode (Figure 12b), and two-photon 35 wl w2 ml :Lmz “ti" “Ir-r " " " wr- -':f-- '5'-r- “x” m .0 w (0 cu m “i 2 L1 “3 m1 003 1 2 1 3 ml (”3 x. ___;L. “is 21:. L Nonresonant Raman TPA Processes Resonance Resonance (a) (b) (C) Resonant and Nonresonant Processes in Three-Wave-Mixing. (a) Nonresonant Processes; (b) Raman Resonance; (c) TPA Resonance Figure 12. 36 absorption when 2w corresponds to the energy difference 1 of a TPA-allowed electronic transition (Figure 12c). Then the total third order susceptibility can be repre- (3) .. (31 (3) (3) ' XNR + xRaman + xTPA' The second case (Figure 12b) has also been called sented as X CARS (coherent anti-Stokes Raman spectroscopy). When the frequency difference wl-mz is tuned through the region of a Raman active transition (ml-w2 is tuned from below to above wR), the observed intensity at frequency w3: (3)|2, should show _ 58 1’w2 ' wa' In other words, a plot of the shape of the 3WM suscept- I(2w1-w2), which is proportional to IX a minimum on the high frequency side of w ibility X(3) XE wl-wz in the vicinity of a Raman resonance should show a dispersion curve which has a maximum and minimum. The reason for this behavior is that the real part of the Raman susceptibility is positive for ml- 2 < ”R and negative for wl-wz > wR. According to formal perturbation theory, if neither frequency w nor ”2 is 1 close to any real electronic transition of the material the nonresonant contribution, Xég), will be a slowly varying function. If the Raman resonance contribution is the only resonance process, the dispersion minimum (3) R contribution of the Raman term. However, if when tuning is caused by partial cancellation of by the negative frequency ml the value of Zwl approaches a two-photon allowed level, a sudden change in the shape of the minimum 37 will occur. Therefore, the position and magnitude of this minimum can probe the TPA resonance contribution to K(B). The application of 3WM in the determination of two-photon absorption is based simply on this idea.58'58’83 In contrast to traditional TPA methods, the mixing technique does not involve any change in the quantum state of the material; the ground state is both the initial and final state. (In both direct and indirect measurements, the final state is either a real two-photon allowed state or a vibronic induced state.) Therefore, the competing processes represented in Figure 6(b) will not influence the measurement of K(B). For this reason the 3WM technique can provide a more accurate two-photon cross section than the other methods. However, the 3WM technique has some disadvantages too. In addition to several experimental difficulties, 3WM can be used only to determine TPA information in the vicinity of a Raman active mode. It is useful only for a selected wave- length.84 Several good reviews have been written on one or another aspect of the two-photon process, including those of Peticolas54 (1967), Jortner85 (1969), Gold53 (1969), Frohlich86 (1970), Worlock87 (1972) and Mahr56 (1975). The theoretical aspects of TPA are well covered in these reviews, and will not be repeated here. Two-photon ef- fects in the microwave region are not reviewed here because 38 the electronic region of absorption is emphasized in this work.88 CHAPTER III EXPERIMENTAL APPARATUS A. Absorption Spectra All absorption spectra were recorded on a Cary 15 spectrometer. For low temperature measurements, the sample was dissolved in EPA (~ 10-6M.) and placed in a square quartz cell with 1 cm width and 0.75 mm wall thick- ness. The sample cell was immersed in liquid N2 in a cylindrical, partly-silvered Dewar which was placed in the sample compartment of the Cary instrument. This quartz Dewar was 24 cm long and 5.7 cm in outside diameter. It had 4 windows, 2 on each side, for light to enter and exit. Helium gas was blown over the surface of the liquid NZ to prevent bubbling. Nitrogen gas was also blown into the sample compartment to reduce moisture condensation on the windows. In order to obtain a homogeneous glass phase without any internal cracking and imperfection, the sample cell was slowly and carefully immersed into the liquid N 2. B. Emission Spectra Low resolution emission spectra were obtained with an Aminco spectrofluorimeter. High resolution fluorescence 39 40 spectra were obtained with a system comprised of the following components located in the laboratory of Prof. A. El-Bayoumi of the MSU BiOphysics Department: The excitation source was a 900 W Xenon lamp. Excitation wavelengths were selected by a 0.5 m B & L monochromator with a 600 l/mm grating blazed at 3000 A. The excitation illumination was focused on the sample whose fluorescence emission was collected at 90° from excitation and focused onto the entrance slit of the emission monochromator. The latter was a 1 m Spex 1700 II with a 1200 i/mm grating blazed at 5000 A. The emission detector was an air— cooled EMI 9558 QA photomultiplier tube which was normally operated at 1.1 KV. Signals from the phototube were fed to a PAR HR-8 look-in amplifier followed by a strip chart recorder, while the reference signals were provided by a light chopper placed between the excitation monochromator and the sample. For low temperature fluorescence measure- ments, samples were immersed in the same liquid N2 Dewar described in the previous paragraph. C. Ruby Laser System A melt grown Czochralski ruby rod (0.05% Cr203) with dimensions 5/8" diameter X 6 1/2" long was purchased from the Linde Crystal Products Division of the Union Carbide Corp. Both ends were cut at Brewster's angle (60° for the ruby-to-air interface) to avoid internal 41 self-lasing. The ruby was excited by pump radiation from two linear high energy flash lamps (EG & G FX-47C-6.5 Xenon flash lamps). The ruby rod and flash lamps were placed at the 3 foci of a double elliptical configuration to obtain more efficient coupling of the pump light to the laser rod. Polished Alcoa Alzak Aluminum lined the inner surface of the dual elliptical cavity to increase 89 A totally reflecting concave rear the reflectivity. mirror with 8 m radius of curvature and a 35% R partially reflecting mirror for the emergent beam formed the ruby resonant cavity. The dual elliptical cavity was made of brass and con- tained 2 end plates, a brass back template and a ceramic front template (Figure 13). The ceramic front template, made from Corning machinable glass ceramic, served as an insulator for the high voltage of the flash lamp trigger wires. Cylindrical quartz tubing surrounded both flash lamps and the ruby rod, and served to circumscribe the water cooling system shown in Figures 13 and 14. Viton O-rings between the flash lamps and quartz tubing pre- vented water leakage. The ruby was held with Teflon O-rings, each set in a groove in a brass cylinder which closely fit the ends of the ruby quartz tubing. A tiny amount of Torr Seal cement was used to seal the ends of the tubing to the two brass holders to prevent water leakage. A side-view cross—section of the ruby laser 42 .wufi>mu HmoHumHHHm cannon momma whom ecu mo cofluomm mmouo 30H>Impfim < .ma muomflm mumHmEma xomm mmmum uOHuDO o m uchH omm _i 0 e cccccccc 50.0.0.0... UwHuDO 0N3 11.3.... u........“..”" a“. . ”.”.u...... :1: . . com sham ocflulo scamme Hocaom mmmum muflz Heavens DCflHIO ............. coua> m_ umauso omm V/l/l/I’IV/l/I/I/I/l/l/II’ll/IIIllI?III/II n no N z mumHmEmB DAEmumulb 43 MINOR AXIS 3.4 CM LAMP MAJOR AXIS 3.8 CM _W R UBY Flowing Distilled H20 Alzak Aluminum Quartz Tubing I?igure l4. End-view of Double Ellipitcal Cavity. 44 cavity is shown in Figure 13. Two small pieces of metal foil (not shown in the figure) were also wrapped around the outside of the ends of the quartz tubing to cover the sealing zone and protect the cement from damage caused by heat and the high radiant intensity from the flash lamps. Flowing distilled water circulated through the quartz cylinders by a peristaltic pump to remove the heat generated by the lamps and the ruby. The two flash lamps were wired in series and were charged by a 5kv, 1280 uF capacitor bank drawn schemat- ically in Figure 15. In order to maintain a uniform in- put energy for the excitation of the ruby, several coil inductors were introduced into the circuit to impede sharp changes in discharging current. In addition, a constant voltage firing circuit was placed between the ruby power supply and the triggering system (Figure 15). The constant voltage and slow discharge rate gave a nearly uniform flash lamp energy input for the ruby pump radiation. The typical pulse width of the Xenon flash lamps was ~300 msec. Two kinds of triggering systems were employed. For an internal trigger, the lamp discharge was triggered by a high voltage surge through the lamp circuit, generated by a high voltage relay inside the ruby power supply. For an external trigger, the break- down was initiated by a high ac voltage provided by a Tesla coil discharge through the trigger wires surrounding .xuo3umz UCHEHOh moans coflumuwoxm.QEMH:mmHm Ima onsmflm :00 0.3... 807$... 9:309: [— [4‘ 33:0 05.: 0208050 ~ >733. .332. \ , 45 46 both flash lamps. When high time resolution was desired the external triggering was preferable because of the lower noise level. The high voltage surge in the internal trigger usually produced a rather high noise background. It was necessary to carefully shield and ground all electronic components. The discharging voltage of the flash lamps was controlled either by a manual variac on the power supply control panel or by the automatic firing circuit. Figures 16 and 17 show block diagrams of the ruby power supply and the automatic firing circuit. In order to compensate for the low efficiency of two photon processes, high power ruby laser radiation was required. Giant pulse techniques were used, which are based on blocking and unblocking (Q-switching) the optical path between the cavity mirrors within a desired period of time so that a higher population inversion can be built up.90 This enables most of the stored energy to be released in an intense, short duration pulse. Bleachable dyes served as passive Q-switches. The Q-switching dye, cryptocyanine dissolved in controlled concentrations in methanol,91 was placed within the optical cavity between the ruby and the totally reflecting mirror. The laser pulse characteristics are extraordinarily sensitive to the operating parameters of the component which make up the laser cavity. Each cavity configuration, distance between mirrors, mirror reflectivity, variation of the 47 External Triggering System r - - ----------------- I ' - i ' Automaticl To Flash ' Circuit '...‘....1. --.‘ .............. Pulse Forming Network Variable 0 - 5 kV A'A'A'A MOLE. Power Supply To Flash Lamps Cathode L I; 15v p------‘ Internal Triggering System Figure 16. Block Diagrams of the Ruby Power Supply and Triggering Systems. 48 Control Meter To Ruby : : Power I I. Sup 1y I I I I I I I L-- -------- - -J ": . V0 tage I : Divider Optically Isolated : : Comparator AC Switch I n ' u l ___; : : JLBIV[ 115 v L-- --J -r 1: _Tesla Coil—_l 1: - 5V Reference Level Potentiometer Figure 17. Automatic Firing Circuit. 49 Q-switch dye, etc. had its own impact on the laser power and pulse shape. The most critical factor for obtaining short pulse duration was the choice of the saturable absorber, its concentration and solvent.92'93 The Q-switch dye cell was 0.3 cm long with two parallel quartz windows. Mode locking was avoided by placing the cell in the laser cavity so that the windows were per- pendicular to the laser beam. The quality of the output pulse was determined by the time shape of the pulse and the burst spot on a piece of film. The Output pulse shape was measured by diverting a small portion of the radia- tion with a beam splitter through neutral density filters to a photodiode followed by a Tektronix 519 or 555 oscilloscope. The pulse width was very sensitive to the concentration of the saturable dye. At relatively low concentration, multiple pulses and a long overall pulse width was obtained. As the concentration of the dye was increased a single giant pulse was observed. When the concentration was further raised the pulse became sharper, but the flash lamp discharge threshold also increased significantly. The optimum concentration at 2.5kv firing voltage was 5 X lO-GM in methanol; under these conditions the pulse width (FWHM) was around 50 nsec. Below 1 x 10-6M, at least 2 or 3 pulses with different widths 5 were observed. Above 8 x 10- M, the power level obtained was near the ruby damage threshold. In fact, a 50 6M in methanol was cryptocyanine concentration of 5 x 10- not the best choice for obtaining a short pulse. Narrower pulses (~30 nsec) could be obtained by slightly increasing the concentration. However, at the higher concentration the dye solution degraded after the first few shots.89 After these few shots, the pulse duration time became longer and the power dropped quickly. Variation in the Q-switch characteristics definitely influenced the stability of output power. As described earlier in the discussion of methods for measuring TPA, temporal fluctua- tions of the excitation laser causes uncertainty in de- termining the two-photon cross sections. A stable and reproducible ruby laser was necessary for the TPE experi- ments described in this thesis. With 5 x lO-GM crypto- cyanine concentration the dye endurance was long, almost 30 shots without refilling, and the pulse width was reasonably reproducible. A second reason for using this slightly low concentration as the optimum was to control the pulse duration. As noted earlier, short pulses should be avoided in two-photon peak height measurements. For the ruby pumped dye laSer, the pulse width of dye laser was shorter than that of the ruby pulse by about a factor of 2. In addition to the peak height fluctuations which occur when the pulse width approaches the fluorescence lifetime of the sample, the TPE peak shows a very short spike when the dye laser pulse is really sharp. This 51 spike could not be observed when it was shorter than the risetime of the photodiode and oscilloscope detectors, thus producing inaccurate values of the emission intensity. Burst spots in Polaroid film were used to determine the mode structure of the ruby laser emission. An iris with diameter of 3 mm was placed within the cavity to select the TEMoo mode. The shape of the burst spot was very sensitive to mirror alignment. The output of the ruby laser was determined by the ratio of the total energy obtained by an energy receiver to the pulse width observed with an oscilloscope. A calibrated Quantronix 500 energy receiver (5.4 mv/Joule) with Quantronix 503 control unit, followed by a chart recorder, was used to measure the total energy. The voltage threshold for the Q-switched ruby to begin lasing at the optimum concentration of the dye was 2.371“! on the 1280 uF capacitors, giving an in- put energy of ~3600 Joules. The energy output of the ruby determined by the energy receiver was usually ~l.2 Joules. If only 10% of the input energy was transferred into the ruby, the conversion efficiency of the ruby laser was less than 0.3%. The alignment of the ruby laser was very critical. A small He-Ne laser with a pinhole over the output was used to direct a 6328 A beam through the center of the ruby. The polarization vector of the ruby rod was parallel to the optical table. A polarizer was put in 52 front of the He-Ne laser to select the E vector of the 6328 A beam in the same direction. The ruby rod was rotated until the transmitted He-Ne beam was precisely parallel to the table. Cavity alignment was accomplished in the following order: The reflected light of the He-Ne laser beam from the front mirror of the ruby laser cavity was adjusted to be coincident with the pinhole by rotating the mirror micrometer mount. Then the same reflection from the Q-switch dye cell was adjusted to coincide with the incident light. Finally, the same procedure was used to align the back mirror. After placing the iris inside the cavity, the alignment was completed. D. Dye Laser Systems Many dye molecules are high gain laser active media. For a large dye molecule in a condensed phase, collisional and electrostatic perturbations caused by the surrounding solvent molecules broaden the individual vibrational modes of the dye in each electronic state. Each vibronic sub- level has superimposed on it many rotational states. There- fore there is a quasi-continuum of states superimposed on every electronic state. If the fluorescence band of the dye solution is utilized in a laser, the allowed transition from the lowest vibronic level of the first excited singlet state to some higher vibronic level of the ground electronic state will provide a broad tuning energy range, usually in 53 the tens of nanometers.94 Organic dye lasers can be optically pumped to laser threshold by either other lasers or by flash lamps. Although dye lasers are high gain lasers, in order to design a stable laser cavity, several oscillation conditions must be satisfied. (a) A fast rising, high power and short duration optical pump pulse is necessary to raise the dye to laser thres— hold. The risetime of the excitation pulse must be suf— ficiently short to satisfy the condition t < %2—, where ST ksT is the rate constant for intersystem crossing of dye molecules from the singlet state to lower-lying triplet states. For most organic dyes, the value of %2— is ST around 1 usec; thus the excitation pulse must be shorter than 1 usec.94'95 For a laser pumped dye laser, this condition is easily reached. Both Q-switched ruby pulses and N2 laser pulses are shorter than 50 nsec. However, a flashlamp-pumped dye laser satisfies this condition only by the utilization of a specially designed pumping 96,97 system, or by the addition of some triplet state quencher.”-100 The rise time requirement insures that the population of excited triplet states of the dye molecules remains low. The excitation pulse must pump sufficient dye mole- cules to the first excited singlet state to reach the laser threshold. Some excited singlet state population will go to triplet states through intersystem crossing. If the 54 increase in triplet state population is fast, two con- sequences arise. First, it dramatically reduces the popula- tion of the excited singlet state, and hence decreases the dye laser gain factor. Second, it enhances the triplet- triplet absorption rate, and hence increases the dye laser loss factor because some of the stimulated emission has been utilized for triplet-triplet absorption. When the triplet-triplet absorption rate becomes equal to the laser stimulated emission rate, the laser will definitely stOp. Thus, if the risetime of the excitation pulse is longer than l2—, stimulated emission cannot start because all of kST the laser light has been used for providing triplet- triplet absorption. One or more of several means of quenching the triplet state concentration rapidly enough so that t < i0 , such as chemical additives or rapid flow ST of the dye through the excitation region, must be provided in order to maintain the dye laser efficiency. (b) Although dye lasers are high gain lasers, certain con- ditions must be achieved to match the parameters of the dye resonator. The simplest form of a dye laser consists of a cuvette of length 1 containing dye of concentration n and of two parallel reflectors possessing a total reflectivity R (R = Rle, the product of reflectivities of both mirrors). The condition for laser action can be written in the form:94 exp(-oa(0)noz)n eprcremIGMIIL) ; 1, (3.1) 55 where 0a($) and Uem(€) are the cross sections for absorp- tion and stimulated emission at 0 respectively, and no and n1 are the populations of the ground and first ex- cited states. The first exponential factor represents the loss factor which is mainly due to reabsorption of the fluorescence emission by the long-wavelength tail of the absorption band. The second factor represents the cavity 1. 2 and contains only parameters of the loss factor. Usually we define a constant S = £n(%) which gives R = e-S£ cavity. The third factor represents the laser gain factor. This condition then becomes: Oemnlfi - canofl - s 1 Z 0 (3-2) If we assume that the cross-sections for absorption and emission are approximately equal (0a 2 Gem)’ and rearrange equation (3-2) to the form: + S/n 0a 4. 0a Gem < < 1, 5'3 I..I then S/n < Oem’ This allows a quick estimate of the feasibility of laser action for a given resonator (defined reflectivities and cavity length) and dye concentration. For any dye laser, the condition S/n < Gem must be achieved. (c) The behavior of a dye laser depends not only on the choice of coupling coefficient (i.e. mirror reflectivity), but also on the geometry of the cavity (the distance 56 between mirrors and shape of the cavity, i.e. plane, spherical, confocal, etc.). A parameter, the Fresnel number, which relates the stability of a laser to the 2 configuration of the cavity is defined as F = 3—, where Ad A is the output wavelength, a is the radius of the circular aperture of the laser beam, and d the distance between the two reflectors. The Fresnel number of a dye laser must not be too small or the laser will be un- stable.101 For example, if we expand and focus the ruby laser to form a uniform plane to pump a dye laser by passing the ruby radiation through 2 cylindrical lenses as shown in Figure 18, a ~ 0.15 cm and A ~ 8 x 10-5 cm. In order to obtain a reasonably large F value, suCh as 20, the distance between the cavity mirrors can be only d < 14 cm. In other words, the cavity length of the pulsed dye laser must be compact. For a N2 laser pumped dye laser, this requirement is even more critical be- cause the N2 laser beam has to be focused into dye solu- tion to a rather smaller region (a is smaller). A ruby pumped dye laser, a N2 laser pumped dye laser, and a flashlamp pumped dye laser have been built in the course of this work. In addition to the powerful ruby pumped dye laser, the N2 laser pumped dye laser has been used to detect two-photon induced fluorescence. However, difficulties in tuning the Nz-pumped dye laser limited its utility for TPA measurements to several restricted wavelengths. 57 1 ° cell k... Ruby laserl 3 o ’m’ K i Cylindrical lenses lé‘ 1n _9' f1 = f2 = 1" Figure 18. Transverse Pumping by a Cylindrical Lens Set. 58 (l) Ruby_Laser Pumped Dye Laser As shown in Figure 18, cylindrical lenses focused the ruby pulse into the dye cell. The quartz dye cell was 2" long and l" in diameter, and had windows tilted 10° from perpendicular in order to prevent self-lasing. In order to obtain a reasonable value for the dye laser Fresnel number, the ruby pulse was not simply focused on the front surface of dye cell. Rather, it was focused inside the dye solution at a depth of almost 1 cm, en- abling a larger aperture for the dye active region to be obtained. The burst spot of a typical dye pulse, obtained by focusing the pulse on a piece of film, looked like: .> The 15 cm long dye laser cavity was formed by a rotat- 102 o ! able grating (1200 l/mm blazed at 5000 A) and a fixed output mirror (a 100% R ruby mirror at 6943 A) which had 50% R at 900 nm and 40% R at 800 nm. Dyeslo3 used to cover the wide spectral range required for this work were DDI (l,l'-Diethyl-2,2'-dicarbocyanine iodide), DTTC (3,3'- Diethylthiatricarbocyanine iodide), Candela #4 (3.3'- Diethyl-2,2'-(4,5,4',5'-dibenzo) thiatricarbocyanine iodide), #10 (l,l'-Diethyl-2,2'-quinotricarbocyanine iodide), #11 (l,l-Diethyl-4,4'-quinotricarbocyanine 59 iodide), #17 (l,3,3,l',3',3'-Hexamethyl-2,2'~indotri- carbocyanine iodide) and Eastman IR-l40. DDI and DTTC were stable at room temperature, while the polymethine dyes (the Candela dyes) were stored in a freezer. The output power and tuning range for these dyes in their optimum solvents are shown in Figure 19. DDI, like cryptocyanine, lases only in a viscous solvent such as glycerol or ethylene glycol; the recovery time of DDI in glycerol after each shot was rather long. Candela dyes 10 and 11 are both weak. The output powers of these two dyes were an order of magnitude less than the powers of other dyes. The tuning range of DTTC and Candela #17 shifted to the blue by ~50 A when the solvent was changed from DMSO to acetone. The alignment method for the dye laser is critical and therefore requires a detailed description. First- the dye cell was filled with the appropriate solvent (DMSO or acetone). He-Ne laser light collimated through a pinhole which was sitting l m away from the output mirror of the dye laser was then passed through the dye cell at a distance only few millimeters (2~3 mm) from the left edge of the cell. The perpendicular adjustment knob of the dye cell was adjusted until the 6328 A beam inside the cell was completely parallel to the cell axis. Then the horizontal adjustment knob of the dye cell was turned until the reflected light from the front window of the 60 .mw>0 H0H0>wm mo Hmaom usmuso 0>H00H0m 0cm mmcmm mcwcss one .mH madman 15:. meozmgm>gz oooa com com com .- Godumuaoxm whom . z: om u HwBOm absn usmcH . . m memo U H.o I J ; 0H owe L . H HH man I; . x u ’l“‘\ H ovHImH L u cmzo 2H case ea In mcoumod 2H oeaumH Ha .OH .na .e mac (MN) HHMOd Lfldlflo 61 dye cell was lying with the incident beam in the same plane perpendicular to the floor. The He-Ne laser beam passed through the dye cell filled with pure solvent and hit the grating. The grating rotation was selected so that the first-order diffracted light was coincident with the incident light. This coincidence could be checked by observing the reflected first-order light from the grating which passed back through the cell and provided an image spot on the surface of the pinhole plate. The grating was further rotated and checked by this coin- cidence procedure for zeroth order light and second order light to make sure that the grating surface was precisely perpendicular to the dye laser pathway. After alignment of the grating with the dye cell, the dye laser front mirror was reinserted and adjusted so that the reflected light from the inner surface of the mirror coincided with the pinhole. (Note: there were 2 reflection spots from thetnu>surfaces of the mirror; the brighter one was chosen.) Finally, the grating was rotated to the desired angle and several drops of the desired dye solution were added to the dye cell. This completed the alignment procedure. The selection of rotation angles for the grating and the tuning method are discussed in the following section. Two He-Ne alignment lasers were used as reference beams to do the tuning. The first laser beam (laser 1 in Figure 20) passed through the dye cell and served as an 62 Scale Rotatable grating AS=£°(2A8) Ruby '1 ®>< Cylindrical lens S1 [:1 Front mirror set __5_ Pinhole He-Ne Laser 1 He-Ne Laser 2 Figure 20. Dye Laser Tuning System. 63 alignment laser, as described above. Light from the second He-Ne laser (laser 2), directed onto the grating from the side of the dye cell, served as a tuning re- ference beam. This beam from laser 2 hit the grating at the same point at which the dye laser would hit. The zeroth order diffraction light of laser 2 from the grating was reflected onto scale S, where it served as a marker. This scale was a circumference meter with its circular center at point a; i.e. the radius of circle 1 (the distance from point a to any point on the scale) was exactly 1 m. A rotation angle A0 of the grating would provide a rotation of 2A0 for zeroth order reflected light of laser 2 from the grating. Therefore, as the grating rotated by an angle A0, the reference beam scanned a distance AS = 1-(2A8) on the scale. From the value of AS, the angle A8 could be calculated. The tuning theory was simply based on this relationship. The tuning pro- cedure was as follows: First, the grating rotation was set so that the lst order diffracted light from the alignment laser (laser 1) was coincident with the pinhole. The grating at this position had an angle 01 determined from the well-known grating formula: 1 = 6328 x 10.8 = 1 2d sin 01, where the grove distance d = %§%fi-cm because this grating had 1200 l/mm. So, 61 = 22°l9' or sin 01 = 0.3797. When the grating was at this position, the zeroth order diffracted light of laser 2 fell at a point 81 on 64 the scale. When the grating was rotated to another posi- tion, the zeroth order diffracted light of laser 2 shifted to another point 52. Therefore AS = $2 - S1 = 1 , 2(02 - 01) or 02 = 01 + g; = 22°l9' + %%0' If we knew 02, we could calculate the corresponding wavelength 12 by using the formula 12 = 2d sin 02. However, the wavelength correspond- ing to rotating the grating an angle A0 from 01 can also be related directly to the only variable which could be observed on the scale, AS, by the formula 12 = sin(22°l9' + 9%0) - 2d. After substituting numerical values, the final formula was 12(A) = sin(22.3167° + AS x 0.2874) x 16666.6. The accuracy of this tuning technique was checked by divert- ing a small portion of the dye laser ouptut with a beam splitter to a Heath monochromator. The difference between calculated and observed values of the dye laser wavelength was zero in the 800 nm region and was only 1 A in the 900 nm ,region. Hence the uncertainty in wavelength was within 0.1 nm. The bandwidth of the ruby laser pumped dye laser was also ascertained in this way. The observed bandwidth throughout the range throughout the range 760-1050 nm was less than 2 A. The temporal behavior of the ruby and dye lasers were recorded using silicon photodiodes (Hewlett Packard PIN photodiodes) and a Tektronix 555 oscilloscope. The pulse durations of the dye lasers were around 30 nsec. shorter than the ruby pulse by about a factor of 1.5 to 65 2. Dye laser output power was in excess of 200 kw, High power ruby excitation light was avoided because as the dye power became too high it would cause damage to the grating surface if a telescope expander was not used. The tuning range was very sensitive to the alignment and dye concentration. As the dye concentration was slightly 94 If increased, the tuning range would shift to red. the alignment was not perfect, the tuning range declined in both wavelength extrema. For low-gain regions, such as those encountered near the limits of a dye's tuning range, the pulse width decreased and the pulses became sharper than those in the high-gain regions. As mentioned in chapter II, when the excitation pulses became sharper, the fluctuation of measured TPA cross sections increased. Therefore, the side wings of a given dye tuning range should be avoided in two-photon applications; this can be accomplished by increasing the overlap between dyes. If such overlap is not available, a slightly increased dye concentration will extend the tuning range to the red and will thus enable a broader spectral range to be covered at optimum pulse widths. (2) Nitrogen Laser Pumped Dye Laser A transversely excited (TE) Blumlein type104 nitrogen 0 laser, which emits radiation at 3371 A, was constructed. The circuitry for the-operation of the N2 laser is shown 66 in Figure 21. Two capacitors, C1 and C2, were inter- connected electrically by a coil of copper wire. A high voltage power supply (40 kv, 2 mA, EOS Model ILM-70) served to charge both capactiors to the same potential and the same polarity. No potential difference existed across the electrodes. As the voltage was increased the air between the electrodes of the spark gap (~2 cm gap) suddenly began to break down. Then the potential drop across capacitor C2 suddenly was released. Shortly after the onset of conduction in the spark gap the voltage dif- ference abruptly shifted to the center of the discharge gap between the two electrodes because the charge on capacitor C1 could not be drained away through a small coil of wire in a short period of time. For changes that occurred within nanoseconds the wire acted as an Open circuit. Finally the voltage across the capacitor C1 discharged abruptly through the 2 electrodes and excited the N2 molecules within the laser cavity. The experi- mental setup is schematically depicted in Figure 22. The capacitors were made from 4 pieces (2' x 3') of 3/32" thick epoxy circuit board which were clad with copper on both sides. The copper was etched from a 1" margin around the perimeter of each board, and two boards were combined to form one capacitor. The laser discharge channel was a gastight rectangular box of dimensions 2" x 4" x 72". It consisted of two strips of brass electrodes connected 67 spark gap F—DH coil of wire ___. L; (J + ~— Electrodes C1 C2 25 K0 I l '--—AVAVAV” II I F— Power supply Figure 21. Circuitry.of the Blumlein Type N Laser. 2 68 N2 —qu Brass shim stock Spark gap 7 coil of wire Power Supply To vacuum _ pump Coil of wire Spark gap Discharge Channel 'i? Circuit board (b) -€) To Vacuum Pump quartz window p ....... E Electrodes 3-... O a. {?-N2 (C) Figure 22. (a) Side-View and (b) End-View of N2 Laser; (c) Output Window of N2 Laser. 69 to the circuit board by brass shim stock. The head of each electrode was sharpened to a knife edge. These two electrode knife edges were carefully aligned to form a straight plane throughout the channel. The distance be- tween electrodes was 1” at the back end and was 1.1" at the output end of the N2 laser. The entire cell was assembled by cementing together strips of plexiglass with epoxy glue. End plates with tilted quartz windows (15° from perpendicular to the laser beam) were cemented on both ends of the channel. Each plate contained a hose connection which served as inlet and outlet for the flow- ing N2 gas. The outlet hose was connected to a vacuum pump and the inlet hose to a N2 gas tank. The optimum N2 gas pressure was between 50 and 70 torr. An Aluminum mirror was used as the back reflector. No front mirror was needed, since the pulsed N2 laser is superradiant. The switching device was a spark gap with an adjustable gap distance. The output power and the repetition rate of the pulsed N2 laser were determined by the gap distance. As gap distance became small, the firing voltage decreased and the repetition rate increased. For example, the repetition rate increased from 1 Hz to 20 Hz as the vol- tage decreased from 20 kv to 16 kv. High voltage (> 24107) should be avoided because it might cause the breakdown of the epoxy circuit board. 70 The total input energy was calculated by E = %-CV2, where the capacitance C of the circuit board was derived by measuring the area of the whole plate A and the thick- ness of the plate d. (C = 5&5, where the dielectric con- stant K of epoxy is about 4.5 and the permittivity e is 8.854 x 10-12 farad/m.) The capacitance in this con- figuration was about 0.035 pF and the total input energy at a firing voltage of lBlnr was almost 5.5 Joules. The pulse width of the N2 laser, measured by an oscilloscope, was between 10 and 15 nsec. Assuming 0.08% conversion efficiency of N2 laser,105 the energy output of each pulse should be approximately 4.4 mJ and the correspond- ing output power for a 15 nsec pulse will be about 300 kw. The 3371 A output of the N2 laser was focused into the dye cell to pump the dye laser. Various dye laser geometries were tested, the most reliable tuning method being the transverse pump configuration shown in Figure 23. The N2 laser light was focused through a f = 30 cm spherical lens and a f = 2.54 cm cylindrical lens to form an almost 1 cm wide plane entering the dye cell. The latter was an ordinary 1 cm2 commercial cuvette, tilted at an angle about 15° from vertical to prevent self-lasing. A quartz plate and a rotatable grating formed the dye laser cavity. In order to maintain a reasonable Fresnel number, the cavity length should not be longer than 20 cm. 106 For the most popular Hansch design of the Nz-laser Grating Figure 23. 71 Sperical lens vs; I = 2.54 cm Cylindrical . lens A Dye laser Quartz plate Dye cell N2 Laser Pumped Dye Laser. 72 pumped dye laser, a telescope expander should be inserted between the dye cell and the grating. However, for medium resolution measurements, the dye laser was still tunable without the expander. For eXample, with this setup, R6G can be easily tuned from 5750 A to 6050 A. Several measurements of two-photon induced fluorescence signals excited by two N2 laser pumped dye laser photons were made. In particular, some preliminary data demon- strating the fluorescence of a-chloronaphthalene induced by two dye laser photons at several different wavelengths have been obtained. Sample fluorescence was collected at 90°, passed through a Corning 7-59 filter and Spex singh monochromator, and detected by a RCA C31034 photo- multiplier (as shown in Figure 24). However, the spark gap of the N2 laser produced a tremendous amount of noise which completely precluded accurate measurements of these weak two-photon induced fluorescence signals. Either extensive shielding of the whole circuit board or the use of a thyratron tube in place of the spark gap must be incorporated in order for the N2 laser to be a viable dye pump for two-photon experiments. (3) Flashlamp Pumped Dye Laser A flashlamp pumped dye laser was also built. The excitation source was a EG & G FX38C-3 Xenon flash lamp connected to several low inductance, fast risetime 73 L JNZ laser Photodiode Upper beam f, Cylindrical lens C_:’ I slit Grating urn-gusavfiauuu::: :'.l':.°::-.:*.-.-.-::: :: Al -m i rro r Spex single monochromator Lower beam Figure 24. Two-Dye-Laser (N2 Laser Pumped)-Photon Induced Fluorescence Measurement. 74 capacitors (2pF + 2pF + luF in parallel) which were charged by a 5 kv power supply. In order to obtain a fast discharge time, the conductance of the wires con- necting the flashlamp to the electrodes of the capacitors must be high. Two copper plates were used to replace the ordinary low conductance wire, and the distance be- tween the flashlamp and the capacitors was made as short as practicable. An elliptical cavity with polished Aluminum inner surface was used to cover the flashlamp and the dye cell, which were placed along the two foci. An external triggering system consisting of a Tesla coil and a trigger wire wrapped around the flashlamp was used. The dye cell was a 3" long, 0.5 cm radius quartz tube, tightly fitted into Teflon holders into which quartz windows were also placed and through which the dye solu- tion flowed. A schematic view of the flashlamp pumped dye laser in given in Figure 25. A methanol solution of the dye (~1o‘3 M.) was circulated through the cell. A front mirror and a rotatable back grating formed the dye laser cavity. The dye output was detected by a photo- diode protected with neutral density filters, followed by a Tektronix 519 oscilloscope. A typical flashlamp pumped dye pulse was almost 2~3 nsec in half width. The threshold voltage was about 2.5 kv and the threshold input energy was % CV2 = 15.6 J. Assuming 0.02% conversion 94,95 efficiency of flashlamp pumped dye laser, the dye 75 momma who pmmfism QEMHcmMHm .mm musmflm use who cfl who mcflnsu a l—‘ 309:3 mcwamnumaaom Nuumso Hawo 0%0 Nuumso e was Iumuw H00H0£ coHHOB momma who H E: mmmo um HOHHHE m mmm 39 Q « ‘ HmmmHHH.I onwam so 1; _ _ hammsm — Hm3om >x mue 76 laser output energies per pulse should be 3 mJ and the corresponding output power for a 2 nsec pulse will be about 1.5 kw. This output power is low for two-photon measurements. B. Materials Structural formulas for the molecules studied in this work are shown in Figure 26. All-trans diphenylhexatriene and diphenyloctatetraene were purchased from Aldrich Chemical Co. and used without further purification. All- trans retinol and retinal were obtained from Sigma Chemicals and were used in most experiments without further re- crystallization. All-trans anhydrovitamin A was rather simply synthesized from all-trans retinol by the following reaction:33'107 l g of all-trans retinol was dissolved in 250 ml CHCl3. About 2 ml of a CHCl3 solution saturated with HCl gas (HCl gas bubbling through it for 1 hr.) was added drop- wise while stirring in darkness. Too much acid was avoided since vigorous dehydration might cause formation of some side products. The solution was stirred in the dark for 15 min. The reaction was then stopped by adding a small amount of water and drying with Na SO . The CHCl 2 4 3 77 \© @\\ 1,6-diphenylhexatriene ©\\\\ 1,8-diphenyloctatetraene w““ I All-trans Retinol \. ‘\ ‘\> ‘5 All-trans Retinal /’ I’ a” /’ All-trans Anhydrovitamin A Figure 26. Sample Materials. 78 solution was concentrated by using a rotating evaporator until the volume of the solution had decreased to a few mls. The concentrated solution was then transferred to a chromatography column (180 ml alumina with mesh number 80-200) where the anhydrovitamin A was separated from reactants by elution with a 5% (ether-hexane 5:95 by volume) solution. The anhydrovitamin A came off the column first and left behind any unreacted vitamin A. The anhydrovitamin A obtained from the first column con- tained several isomers and was therefore rechromatographed through a second column containing 100 ml Woelm alumina (activity 1) to separate the all-trans fraction. The carrier solution was an ether-hexane mixture with gradually increasing ether concentration (5-25%). The first pale yellow portion (~10%) consisted of unwanted cis isomers, followed by the all-trans portion (~70%) and a second cis isomer (~20%). The all—trans anhydrovitamin A fraction was dark yellow color and was concentrated by evaporation to a very small volume of about 1 ml. If a few mls petroleum ether were added to this con- centrated solution, beautiful orange needle-shaped crystals of all-trans anhydrovitamin A could be obtained by cooling the resultant solution for 48 hrs. at -78°C (dry ice in acetone). The crystals were dried by decant- ing the solution and blowing dry Helium gas over the solid, or by vacuum desiccation. Anhydrovitamin A is 79 stable for extended periods at room temperature only in oil solution. Other concentrates and the crystalline preparation will either absorb water vapor to reform all- trans retinol, or degrade to other cis-isomers. In a sealed cell filled with dry Helium gas and stored in a freezer, all-trans anhydrovitamin A will last for weeks. Its isomeric purity can be easily checked by its dis- tinctive, vibrationally resolved absorption spectrum. Anhydrovitamin A is the only compound in the retinyl family which has a well resolved electronic absorption spectrum. All other retinyl polyenes show remarkably diffuse spectra when compared to those of unsubstituted or diphenyl-substituted polyenes of approximately the same size. Also, the absorption spectra of the cis isomers of anhydrovitamin A are bluer than that of the all-trans compound. For example, after lengthy exposure of the isomer to ultraviolet light, the electronic ab- sorption spectrum shows a decrease in resolution and an increase in broadness in the blue region, as illustrated in Figure 27. Solutions for low temperature studies were prepared by using MCB spectrograde EPA (ether-isopentane-ethanol, 5 = 5 = 2 by volume). To avoid effects caused by sample decomposition or degradation, fresh solutions were pre- pared just prior to each experiment and handled under safelights. RELATIVE ABSORBANCE 80 3 days \ | I 'I r—-After I I 1 - 1 . I . A l . A . . J . l 300 350 400 (n00 Figure 27. Absorption Spectra of All-trans Anhydrovitamin A (taken at 77°K in EPA) Before (——) and After (---) Lengthy Exposure to Room Light at Room Temperature. 81 F. Two-Photon Excitation (TPE) Spectrometer Because of the even-parity selection rule, two-photon absorption is an ideal way to observe and locate low- lying "forbidden" 1Ag-H-J‘Ag transitions. At the present time, the TPE technique, in which two-photon induced fluorescence is monitored is still the best method for detecting TPA processes because of its simplicity and high sensitivity. (Direct absorption measurements are still in the early stages of development.) In order to use TPE to observe low-lying 1Ag states of all-trans linear polyenes, several conditions must be met: (1) A near infrared tunable and powerful dye laser is needed. The best candidates for the study of low-lying lAq states are polyenes of medium size (n = 4~6). (Examples will be discussed in chapter IV.) For these medium size polyenes, the absorption origins are in the range 3500 A to 4200 A. Assuming that the low-lying lAg states are 1000 cm.1 to 3000 cm.1 below the origin of the lowest one- photon absorption, the two-photon tuning range of interest 1 to 11000 cm-l, which is in the near is from 14500 cm- infrared region. Therefore a ruby laser pumped dye laser is an excellent choice for the excitation source. (2) The polyene molecule selected for study must be fluorescent, and have a reasonably high two-photon cross section. (3) The noise level of the excitation source 82 must be low. Although TPE is a sensitive technique, after appropriate filtering (filters or a monochromator) only part of the selected fluorescence will be detected by a photomultiplier. If the noise is similar in magnitude to the signal, no spectrum can be obtained. This is the reason why an external triggering system was used for the ruby laser, since it provided a much lower noise level than was possible with internal triggering. The experimental apparatus is schematically depicted in Figure 28. Near infrared radiation from a tunable dye laser pumped by a Q-switched ruby laser served as the excitation source. Cylindrical lenses focused the ruby pulse inside the dye cell, which was within a resonant cavity formed by a rotatable grating (1200 1/mm) and fixed output mirror. The ruby laser and dye laser have been described in sections C and D of this chapter. The fluorescence housing, shown in Figure 29, con- tained a Spherical collection lens (f = 5 cm) which focused the two-photon induced fluorescence radiation emitted at 90° from the dye laser beam onto the photo- multiplier. The lP28 photomultiplier was normally operated at 1020 v. Three Corning colored glass filters, CS 4—76, 4-94, and 4-96, were mounted between the collection lens and the PMT. Fluorescence in the 380- 560 nm region, only, reached the phototube. Another filter, CS 2-60, was placed outside the entrance of the .ummmq 0%9 comssm momma whom 0 mo Houwsouuommm couonmIo3B .mm musmah H mmOU unceeouo: H0306 NZ .wHH cfl Hamo mamfimm (<0— mnsuummm Houuflamm Emma Hmuaflm mama mcfimSUOM 83 mcflumum HOHHHE pmumoo owuuomawflo _ >2; QmIUZ>>mIO 84 Dye laser CSZ-60 filter Filters CS4-76 . CS4-94 7‘» _ CS4-96 qu. N2 Dewar ,.L PMT 'r-‘L - . Quartz ample cuvette Foam rubber N2 purge - Figure 29. Fluorescence Housing for TPE Measurements. 85 fluorescence housing to eliminate stray visible and UV light from the ruby and pass only the near infrared dye laser radiation. The dye laser radiation was focused in the middle of a 1 cm2 quartz sample cell which was immersed in the liquid N Dewar. The fluorescence housing 2 was purged with dry nitrogen gas to reduce moisture con- densation on Dewar windows. Helium gas was blown across the top of the Dewar to minimize bubbling. Dye laser intensity was measured by diverting a small portion of the radiation with several beam splitters through neutral density filters to a photodiode. The neutral density filters were adjusted so that the dye signal displayed on an oscilloscope did not exceed 1 v. The photodiode voltage was normally 40 v. Since the dye laser aperture was larger than the photodiode area, it was necessary to focus the dye laser onto the photodiode in order to measure the total dye intensity, rather than only some portion of the inhomogeneous dye laser beam. At high dye laser powers, a neutral density fitter was also placed before the sample cell to avoid saturation of the fluorescence signal. A typical experimental procedure may be described as follows: First, the ruby laser is operated under normal conditions, in order to optimize both temporal and spatial characteristics by adjusting the concentration of the Q-switched dye solution and the ruby alignment so that 86 a good pulse is observed on the oscilloscope and a uni- form burst spot is formed on the film. Second, the alignment and the concentration of the dye laser is ad- justed until a dye pulse is shown on the oscilloscope with suitable temporal characteristics and the desired tuning range. Third,-a sample cell containing a pure EPA solution is placed in the liquid N2 Dewar to check the noise background. Without the sample, a good EPA glass (immersed in liquid N at a very slow speed to pre- 2 vent cracking or other crystalline imperfection) ordinar- ily shows a noise signal of only a few mv on the lower beam of a dual beam oscilloscope. (With a sample such as DPO, a several hundred mV signal can be obtained.) This noise might come from either i§_§itg_second harmonic generation of the dye laser light in the sample cell, or two-photon induced fluorescence from some impurities in either EPA or the quartz cell.108 However, at high sample signal levels the noise can be neglected. After the blank background is checked, a fresh sample is placed in the sample cell and the measurement formally started. As mentioned earlier, the TPE spectrum consists of a plot of the two-photon induced fluorescence intensity as a function of excitation energy. Two-photon excita- tion requires that the fluorescence intensity be propor- tional to the square of the dye laser intensity. It is necessary to record the ratio of the measured fluorescence 87 intensity (fluorescence peak height, If) to the square of the measured dye laser intensity (square of the dye peak height, I3). Thus ln(If/Ig) is plotted point-by- point XE twice the incident dye photon energy to obtain the TPE spectrum. The square dependence of the fluores- cence intensity on the dye laser intensity was checked at several incident wavelengths by plotting 1n If !§_ln ID. Straight lines with slopes ~2 were obtained, indicating that second order processes were responsible for the fluorescence. . At each dye wavelength selected, several data points were taken. The repetition rate at which the ruby laser was operated was one or two shots per minute. Polaroid high speed 410 film was used to photograph each oscillo- gram. Typically, 3 or 4 hours were needed to measure a TPE spectrum over one dye tuning region. Usually three or four dyes are required to cover the spectral region of interest for the investigation of a low-lying 1Ag state. Therefore, at least 10 hrs., which does not include data treatment time, is needed to obtain one TPE spectrum for one compound. (Automation of the apparatuseo’al'108 would relieve much of the tedium involved in these ex- periments; this would be rather difficult to accomplish ,for the present instrumentation.) .The results described in the succeeding chapters justify the experimental effort expended. CHAPTER IV THE TWO-PHOTON EXCITATION SPECTRUM OF ALL-TRANS DIPHENYLOCTATETRAENE As noted earlier, several recent calculations3o 25,26,33-35,76 and experiments indicate that the lowest excited singlet state of linear, all-trans polyenes is 9 symmetry as the electronic ground state, is hidden in . 1 of Ag symmetry. Thus A state, which has the same the normal electricdipole absorption spectrum because the transition violates the u<+ g selection rule. How- ever, 1Ag'«++-1Agtransitions are allowed in two-photon absorption, and two-photon excitation (TPE) spectro- scopy thus may enable direct observation of this hidden excited state.59 A. The Reason for Choosing DPO to Study The TPE method, based on an investigation of the energy dependence of two-photon fluorescence spectra induced by tunable dye lasers, is a powerful and sensi— tive technique as has been described in chapter II. The primary requirements for organic molecules which are amenable to TPE measurements are: (l) reasonably large two-photon cross sections for transitions from the ground 88 89 state to the excited singlet states of interest; (2) a relatively high fluorescence quantum yield from the lowest excited singlet state; (3) no apparent wavelength depen- dence of the fluorescence quantum yield from excitation in the energy range of interest; (4) higher excited states are photochemically stable and exhibit no decomposition or photochemical reaction. The choice of a particular polyene or a particular set of polyenes for study should I '1'..- be considered first. There are several reasons why diphenyloctatetraene has been chosen for investigation. Consider first the Stokes shift, one of the anomalies in fluorescence spectra of linear polyenes. At least at relatively small values of n, the separation between the absorption origin and the fluorescence origin increases with polyene chain length. In the range 3 i “.5 6 the gap roughly approaches a linear dependence on chain length as has been plotted by Hudson and Kohler.26 Figure 30 shows the energy gap of the diphenyl polyenes as a function of the number of carbon-carbon double bonds. When the chain length is short (n < 3), the fluorescence origin is very close to the strongly allowed absorption (1Bu + 1Ag) origin. If the fluorescence comes from a low-lying hidden 1A excited state, the origin of the 9 transition (lAgs-IAg ) would be very close to the lBu-+ 1Agorigin. According to the two-photon selection rule, the vibronic symmetry of the excited states must 90 II” .4000 ' .’ ' / I I [I _1 .3000 ./ AE(cm ) / I I I I -2000 o’ I I I I I .1000 ,I' ’ .” ._— _— j 4 1 J I L 1 2 3 4 s 6 n Figure 30. Energy Gap Between the Absorption Origin and the Fluorescence Origin of Diphenyl Polyenes as a Function of the Number of Carbon-Carbon Double Bonds. 91 be g. Thus the origin of the lBu state should be symmetry— forbidden in the TPE spectrum. However, certain low fre- quency skeletal bending or torsional modes of the polyene chain in the 1Bu state are still two-photon allowed and may become observable in the TPE spectrum if their total vibronic symmetry is 9. For short polyenes, then, the 1Ag origin may be combined with or covered by low-fre- quency vibrational modes of the 1Bu + 1Ag transition, and the data cannot provide conclusive evidence of the existence of a low-lying lAg state. Another reason for not choosing short diphenyl substituted polyenes is that the energy states of the benzene ring substituents will interfere with the energy states of the polyene chain.109 In 109 110 stilbene and diphenylbutadiene, absorption char- acteristics of the polyene chain are severely perturbed by excitation involving the terminal phenyl groups. At least three or four double bonds are needed in the polyene chain before mixing between the benzene and polyene states becomes insignificant. As the chain length increases, the polyene absorption bond shifts toward the visible, which emancipates it from the bondage of benzene. (The 111 lowest lB2n energy state of benzene is at 4.9 eV, which is in the near UV.) When the chain length becomes 1 relatively long (n 3 6), the Bu + 1A energy gap in- 9 creases, which causes the low-lying lAg state to lie much closer to the ground state than in shorter fluorescent 92 polyenes. The increasing number of vibrational modes as the molecular size becomes larger, and the decreasing energy gap between the 1Ag excited state and the ground state might well provide a great rate of internal con- version of the excitation energy, thereby decreasing the fluorescence quantum yield. This may be the reason that longer polyenes, such as B-carotene and lycopene, are non- fluorescent molecules112 < 10-5). Therefore, only medium size diphenyl substituted (fluorescence quantun yield polyenes are good model candidates for two—photon excita- tion investigations. As a practical consideration, the solubility of the polyenes decreases as the chain length increases. This will severely influence the TPE measure- ments for longer polyenes. As unsubstituted or dimethyl polyenes are concerned, it is well known that short unsubstituted polyenes are nonfluorescent.113 None of the dimethyl polyenes are commercially available. They have to be synthesized through a Wittig reaction by combining a short polyene aldehyde with a short polyene phenylphosphonium chloride.114 Also, dimethyl polyenes are unstable even below 0°C. The unsubstituted polyenes are also fairly difficult to synthesize and are unstable as well.115 Octatetraene is explosive in pure form. Compared with these potential model polyenes, diphenyloctateraene (DPO) is stable at room temperature, 93 easy to purify, and available commercially. Another reason for choosing DPO is that the original motivation for this work was to understand the electronic states of the visual pigments. DPO is an excellent prototype be- cause it has about the same number of double bonds as the polyene chromophore in the visual pigments. The two- photon cross sectionIIEDPO is fairly large. According to the most recent 3WM (three-wave mixing) measurement, 6 is about 60 marias.116 28a The fluorescence quantum yield of DPO, 117 is rea- 0.15 in benzene and 0.09 in cyclohexane, sonably high. Also it shows no apparent wavelength de- pendence. The one photon excitation and absorption spectra of DPO are essentially identical over the range of inter- est.33a Therefore DPO is an ideal candidate for a TPE study. B. Results and Discussion The one-photon absorption and fluorescence spectra of DPO in EPA (either-isopentane-absolute ethanol at 5:5:2) glass at 77°K are shown in Figure 3. The absorp- tion spectrum was recOrded on a Cary 14 Spectrometer with sample concentration ~2 x 10-6M. The emission spectrum was recorded on a component system fluorimeter (see chater 3) with sample concentration ~5 x 10-6 M, the O exciting wavelength was 3500 A. The absorption transition corresponds to the 1Bu + 1Ag strongly allowed transition. 94 The absorption origin is at 24420 cm_1, and is followed by 1200 and 1600 cm—1‘ vibrational progressions. These correspond to C-C and C = C stretching modes in the ex- cited electronic state. At least three quanta of the C = C symmetric stretching vibration are present, as are several combination bands. The fluorescence origin is at 22280 cm-l, again followed by lines separated by about 1200 and 1600 cm-1. The intensities of the higher vibrational quantum numbers of the progression are greater in emission than in absorption. Thus, the mirror image symmetry is not precisely followed. The large Stokes shift (2140 cm-1) is a typical characteristic of polyene spectroscopy. As the concentration of DPO in EPA was gradually in- creased, a weak and diffuse band was found on the low energy side of the absorption origin (Figure 31). It was possibly due to some vibrational modes of a low-lying lAg state. This feature was not found on the high energy side of the fluorescence origin. The energy gap between the absorption and fluorescence origins is strongly solvent dependent. For example, the energy gap at room tempera- l -1 26 ture in methanol is 3200 cm‘ 1500 cm . In and in CS2 all cases it is the absorption origin which shows significant solvent dependence; the fluorescence origins hardly change. 95 .xxoee an «am he sense x m u 0v soeumuusmosoo seem as one no cwmwuo cofiuQMOmnm can no 00am amuocm 30A 0:» co mummmm< occm monumeo .Hm musmwm com omv oov Ashe meozmqm>mz _ a q .1 . _ . 1 . . — I} 1i (‘1' ceofluo mocmommnosam ©////© moaxv mamm222m>es _ . . . u _ II om Hm mm mm em mm 0 “HIE HDNVHHOSHV 96 According to Strickler and Berg's (Birks and Dyson's)28 formula, the intrinsic fluorescence lifetime of an ex- cited state can be calculated from the oscillator strength of the measured absorption spectrum, subject to certain approximations. This formula, which is related to the integrated absorption spectrum by an extension of the derivation of Einstein's A and B coefficients, can be approximately reduced to %_.= 1.3 f 32 where f is the oscillator strength and 5 is an average frequency for the absorption transition in cm-l. This relationship has been extensively tested, and the accuracy for most organic molecules is within 80%. However the intrinsic fluorescence lifetime of DPO in benzene at room tempera- ture is 49 nsec, which is almost 21 times longer than the 2.3 nsec value calculated from the formula.28 In cyclo- hexane, the measured intrinsic fluorescence lifetime of DPO is 69 nsec, also 17 times longer than calculated value of 4.1 nsec.117 As described in chapter 1, all these discrepancies support the hypothesis of the existence of a low-lying symmetry forbidden state in DPO. The two-photon excitation spectrum of DPO (~2 x lO-4M) in EPA at 77°K has been obtained over the spectral range 10500-12700 cm-l. A sharp origin and distinct vibrational structure are observed in the TPE spectrum, which is shown in Figure 32. These features were not discernible in the polarized TPE spectra of DPO in cyclohexane at 97 .xohb um 4mm cw one no Eduuommm cOwuouwuxm couonmIoze .Nm wusmflm . 22v Eczema? OD“. ON»... 0qu £3. 00? @5 m. a . u 0% in be do 2. m s. 0% to 235QO 292me 2051.1 02: co m . £54. L P b e L— b _ mm mm em 8 n 9323533 €20 no. a 98 room temperature reported by Holtom and McClain. How- ever, their polarization ratio measurements indicate that the low-lying state in this energy region has A.g symmetry. Two-photon excitation requires that the fluorescence intensity be proportional to the dye laser intensity. This necessary condition was checked at several incident wavelengths by plotting ln If yg. ln I straight lines D; with slopes ~2 were obtained. An example is shown in Figure 33. The lowest energy two-photon transition, 2hv = 22360 cm"1 is assigned as the origin of a 1Ag excited electronic state of DPO. It lies 2060 cm.1 below the origin of the first allowed 1Bu + 1Ag transition. This origin is about 80 cm.1 to the blue of the fluorescence origin, but this difference (about twice the experimental uncertainty) is considered minor. The essential coincidence of the 1Ag origin with the fluorescence origin confirms the suggestion that the fluorescence originates from a lower-energy hidden state. Two-photon selection rules require that the vibronic symmetry of the excited state be 9. Thus the origin of the lBu state is symmetry-forbidden in the TPE spectrum; a minimum is observed in that region (dotted arrow in Figure 32). Transitions to symmetric vibrational states 1 of the Ag electronic state account for the remaining 99 +- O r08 , r06 A J ‘01 o 114 _I 2 SAMPLE: DPO IN EPA A1771 0' ova = one AT 82801 | o I ! .1 l l A L I 1 j 0 O2 04 06 new.) Figure 33. Square Dependence of Two-Photon Induced Fluorescence of DPO. 100 peaks in the TPE spectrum (solid arrows in Figure 32). Assignments are based on comparisons with low-temperature, high resolution absorption spectra of isolated DPO26 34 and 2,lO-dimethylundecapentaene, and approximate correlations with the ground state vibrational frequencies obtained in emission or by Raman spectroscopy.26 The vibrational assignments are listed in Table l and dis- cussed below. (1) The low frequency region (up to 400 cm“1 from the origin) consists of a broad band at about 230-315 cm-l. This is possibly due to skeletal bending or to torsional modes of the polyene chain. The fluorescence spectrum 1 of stilbene shows a band at 211 cm- which has been assigned as the in-plane bending mode of the ethylene 109 bond. Also, 2,lO-dimethylundecapentaene in a nonane Shpolskii host shows a 220 cm"1 vibration mode in the absorption spectrum.34 The fluorescence spectrum of DPO in a pentadecane Shpolskii host shows several low fre- l -1 26 quency modes, at 154 cm- and 257 cm . Therefore this 230-315 cm"1 broad band in the TPE spectrum of DPO may be the counterparthfthose low frequency modes in the DPO ground state. (2) The next broad feature (550-780 cm—l) is attributed to vibrational modes of the benzene rings, since no corresponding vibrations are noted for the methyl sub- 34 stituted polyenes. Ground state vibrational intervals 101 1 have been observed for DPO in the fluorescence spectrum.26 Toluene has a normal mode at of 612 and 726 cm' 622 cm-1 which has been assigned to a ring mode by Dyck and McClure.109 (3) The relatively weak band at 980 cm“1 can be assigned either as a symmetric ring mode, or a C-H bending mode of 118,119 the polyene moiety, or the combination of these two possibilities. (The a symmetric stretching mode of ben- zene is at 992 cm-l.109) For all polyenes, including those without benzene substituents, the IR spectra show 1 119 a 995 cm- C-H bending mode. 1 (4) The remaining features, shifted about 1230 cm- and 1600-1760 cm"1 from the band origin, are due to c-c and C = C symmetric stretching vibrations, respectively. 1 Their frequencies are higher in the excited Ag state than in the ground state (see Table l). The same effect. 34,35 120 has been noted for undecapentaene, benzene, and 121 naphthalene, and explained on the basis of vibronic coupling between the ground and excited states.12]"122 The ground state vibrational intervals of DPO in the high resolution, low temperature fluorescence spectrum26'34 are also listed in Table l for comparison. 102 Table 1. Comparison of Vibrational Intervals Observed in TPE Spectrum with Those Observed in Fluorescence Spectrum for Diphenyloctatetraene. Ground State Vibrational Excited 1A. State Intervals from Fluores- Vibrationai Inter- Assignment cence Spectrum in valsfrom TPE Bibenzyl Host Spectrum in EPA Origin at 22112 cm'1 Origin at 22360 cm"1 19 164 255, 337 230-315 Skeletal Bending or Torsion 612, 726 550-780 Phenyl Ring 995 980 Ring Mode C-H Bending 1144 1230 C-C Stretching 1575 1600-1700 C=C Stretching C. Conclusion and Implications The origin of the transition to a low-lying "hidden" excited electronic state in DPO has been found through the two-photon excitation technique. It lies 2060 cm-1 below the 1Bu + 1Ag origin, and is essentially coincident with the one—photon fluorescence origin. Since it is a two-photon allowed transition, the symmetry of this state 103 is most probably 1Ag. Polarization measurements on the TPE spectrum of DPO at room temperature substantiate the 76 Ag character of this electronic state. The coincidence with the fluorescence origin supports the notion that the one-photon fluorescence originates from this state.26 It is concluded that a low-lying 1Ag state of DPO has been confirmed by this work. The location of this low energy state removes several discrepancies between the expected and observed fluorescence properties of DPO. The existence of this 1Ag state not only provides a fundamental modification of electronic state ordering over the traditional picture of electronic structure of linear polyenes but also contributes some implications in understanding the photochemistry of polyene molecules. 109 that It was pointed out by Dyck and McClure isomerization proceeded by two paths - one with an activa- tion energy in the singlet manifold and the other with a singlet-triplet crossing. There is evidence that the photochemistry of polyenes occurs in the singlet mani- fold rather than the triplet manifold. Since the internal conversion is extremely fast, the lowest ex- cited singlet state must be the one responsible for photochemistry. If the lowest excited singlet state is a symmetry forbidden 1A.g state, the lifetime of this state is long which is favorable for photochemical change. Additionally, a calculation of the bond orders for this 104 low energy 1Ag state indicates some possibilities for cis-trans isomerization. For example, the bonds in octatetraene in the 1Bu excited state have bond orders26 of 0.79, 0.48, 0.55, 0.59, 0.55, 0.48 and 0.79. The Cl-2 C2-3 C3-4 C4-5 CS-6 C6-7 C7-8 bond orders of 1Bu state are almost constant for all bonds except the terminal bonds which are slightly higher. The bond orders of octatetraene in-the low energy 1A 9 excited state are 0.64, 0.56, 0.41, 0.63, 0.41, 0.56, 0.64 Cl-2 C2-3 C3-4 C4-5 C5-6 C6-7 C7-8 which shows a strong central bond (C4-5) and relatively weak bonds (C3-4 and C5-6) on either side of the center. This indicates that cis-trans isomerization will occur preferentially at those two bonds (C3-4 and C5-6) in this low-lying 1Ag state. Obviously, an understanding of the connection between the observed photochemistry and the electronic characteristics of the excited states in linear polyenes depends on an understanding of the reordering of their electronic states. CHAPTER V THE RETINYL POLYENES A. Introduction It has long been known that vitamin A deficiency is accompanied by visual deterioration.124 It has also been found that ll-cis retinal, one of the isomers of vitamin A aldehyde, is the chromophore of visual pig- ments in vertebrate and many intervertebrate retina.125 All visual pigments in vertebrate retina are composed of a low molecular weight chromophoric molecule (retinal or its derivatives) and a protein (rod or cone opsin) . The combination of retinal and rod Opsin is called rhodopsin. The retinal-opsin linkage is a protonated Schiff base to the e-amino group of lysine in the pro- tein.126'127 Visual pigments of different species vary only in the nature of their protein, opsins, not in their chromophores. Spectra of visual pigments are usually broad and structureless, and absorption maxima of different_rhodopsins range from about 430 nm to 620 nm. This variation is entirely due to differences in the opsins of different species. Retinal has an ab- sorption maximum at about 370 nm in ethanol solution. Upon binding with opsin, the absorption maximum shifts 105 106 to the red. Many spectroscopic and theoretical models have been suggested to explain this large bathochromic shift. The initial step in visual excitation is light absorption by ll-cis retinal, the chromophore of the visual pigments. This light absorption by visual pig- ments initiates a series of events which by a still un- known mechanism leads to the triggering of the photo- receptor cells. Presumably, one of these primary events is a photochemical isomerization of the ll-cis isomer to the all-trans isomer of retinal, as postulated by Wald.128 The isomerization leads to the release of a transmitter substance that subsequently decreases the Na+ current across the photoreceptor cell membrane.125 Although the detailed mechanism of vision is still unknown, one experimental fact is that the optical excita- tion of the visual pigment results in a cis-trans iso- merization of the visual chromophore, a linear polyene, and provides a coupling between the electronic structure and molecular motions in this chromophore. An under- standing of this vibronic coupling depends on an under- standing of the electronic structures of retinyl polyenes. However, the spectroscopic information available on the excited electronic states of the retinyl family is quite limited due to their broad and vibrationally unresolved spectra. This diffuseness unfortunately precludes accurate assignment of the electronic origins. The discovery that 107 diphenyloctatetraene has a symmetry-forbidden excited singlet state below the lBu state responsible for the strongly allowed visible absorption suggests a reexamina- tion of the traditional picture of the electronic energy level ordering of polyene molecules. Retinyl polyenes have a very similar polyene structure to DPO, except for the presence of the B-ionylidene ring in place of one of the terminal phenyl groups, and a carbonyl or hydroxyl group at the other terminus. (See structures shown in Figure 26.) The existence of low-lying 1A.g state in 34-37 also indicates several dimethyl substituted polyenes that this level order might be general for all polyene molecules. Thus, a search for such hidden electronic states in the retinyl polyenes by the TPE measurement technique is worthwhile not only to reexamine the electronic energy ordering, but also to find a possible way to understand the vision mechanism. The absorption and emission spectra of retinol and retinal in EPA glass at 77°K are shown in Figures 34 and 35. All show a remarkably diffuse nature, and present substantial difficulty to spectroscopic analysis. Several possibilities have been suggested to explain the broad- ness. Some authors indicate that the diffuseness in * * retinal is due to the near degeneracy of the nu and NW 129 130 transitions. Blout and Fields found that the polyene aldehydes CH3 - (CH = CH)n - CH = 0, for n = 2,3, 108 .xohh um «mm cw Hocfiumm mGMHUIaam mo muuowmm coemmwsm 0cm coflumuomnd A s: a :Bzmamiz ems see emm com . em 0.53m LLISNHLNI NOI SSINEI HAILV'IHH IO - g.ee e< ouohncm msmuulaad mo mnuommm coflmmflem can cowumnomnfl A e: o zsuzmsm>OHQM£G¢ msmuulaad mo Esuuommm COwumHown< was .ov madman AEE 1525m><>> com Lomv oov . J . . . . q d . 4 . _ V m m. V N 3 fl: L \C \ \ \ \ i — r -l . p b ON AN NN MN VN DN A 120%58 mamm232m><>> .Mohh um flaw ca Amocmnuomnm w>wumHmu mo pass swv ocmm cofiuauomnd souonmlmso amumcm ”6030A 9.3 can QMH\MHV:H mo ”in: 9.0 5.50QO counumufloxm souonmloza .3. 055.3 A 82 v rpozum><3 8W 0% ox... omv o~.v 0.3 00v 0mm _ . . _ _ . _ Fl. n . NL l a o.o 4.3 m. A N $5 ...... V: .9 . n u m. X? 2 «mm 2. < 2.2539913 " no zambmam zoEonm 2051a oz: 0.0 . b — p _ 3mi°< . h — p m a mu mam mm msm vw mew mm mmw ow Anzufloc; 3323533 123 The most prominent feature in the TPE spectrum is a strong two-photon allowed band in the region 25120 - 26000 cm’1 1 , which falls in the minimum between the Bu + 1Ag origin (3970 A) and the second vibrational quantum in one-photon absorption spectrum. The TPE spectrum also shows a minimum at 3980 A, which is roughly the one-photon absorption origin. This arises because the origin of the 1Bu state is symmetry-forbidden in the TPE spectrum. On the red shoulder of 1Bu +~1Ag transition, several broad and diffuse two-photon allowed bands are observed, which are interpreted as vibrational structure of a low-lying excited 1Ag state. The sharp onset at 1 23500 cm- , the lowest-energy two-photon transition 1 - 1 (Zhv = 23500 cm ), is assigned as the origin of this A 9 1 electronic state. It lies about 1560 cm- below the origin of the first allowed 1Bu + 1Ag transition. No two-photon allowed peak is found at energies below this origin, and only background noise is observed in the region 21300-23300 cm-l. Up to about 1000 cm-1 above the origin, the TPE _sepctrum is broad and diffuse. This shows that anhydrovitamin A probably has a wide (or multiple) potential minimum in the lowest excited lAg state; this is entirely consistent with the diffuse fluorescence spectrum, mentioned earlier. Although the diffuse nature of the TPE spectrum precludes precise assignments for the 124 vibrational structure, several features still can be identified by comparisons with low temperature absorption spectra of stilbene, diphenylbutadiene, diphenyl- octatetraene, and 2,lO-dimethylundecapentaene, and through approximate correlations with ground-state vibrational frequencies obtained in fluorescence spectra. The broad band about 80-450 cm.1 from the origin probably involves low frequency torsional and skeletal bending modes. The assignment is based partially on analogy to the DPO fluorescence spectrum (in a pentadecane Shpolskii host) which has a strong vibrational mode at 77 cm-l, and several medium intensity modes between 257-410 cm-l.26 1 109 Stilbene also shows a strong 211 cm- normal mode, which has been assigned as the in-plane bending vibration of the ethylene bond, and several low frequency modes between 129-297 cm-l. Diphenylbutadiene shows ground state vibrational intervals at 140 and 360 cm-1.109 The fluorescence spectrum of 2,lO-dimethylundecapentaene 34 also exhibits several modes in this region (105, 256, 335, and 435 cm-1) and its absorption spectrum in a nonane Shpolskii matrix shows excited state vibrational modes in the range 220-400 cm-l.34 The next broad feature (710-1000 cm-l) may possibly be attributed to 118,119 CH out-of-plane bending vibrations; the IR spectra of all linear polyenes show such a fundamental vibration in the neighborhood of 995 cm‘1.119 125 Beyond this diffuse region, two distinct TPE features, a C-C symmetric stretching vibration (~1200 cm.1 from the origin), and a C = C symmetric stretching vibration (~1530 cm.1 from the origin), are observed. The fluorescence spectrum of all-trans anhydrovitamin A is broad and fairly structureless which precludes accurate location of the fluorescence origin. Its gen- eral shape differs markedly from the usual "mirror image" of the absorption spectrum (see Figure 37), which suggests that the emission originates from a different excited electronic state than the upper state of the absorption transition. The TPE spectrum definitely shows a two- photon allowed band on the red shoulder of lBu + 1A 9 origin, and this two-photon allowed band has a broad low frequency intensity distribution, similar in breadth to the fluorescence spectrum. Apparently electronic excita- tion into the strongly absorbing 1 1 Bu state can rapidly decay to two or more of the Ag excited state conforma- tional isomers, and the fluorescence emission from several such conformers is superimposed to produce the observed broadness. The two-photon excitation spectrum of all-trans 4 retinol (~5 x 10- M.) in EPA at 77°K is shown in Figure 42. The main feature is a strong two-photon allowed band in the region 23250-24390 cm-1, with the maximum at 23800 cm-l. This band is broad and structureless. The 126 .xonn no «mm :a Aocfluom mcouuuAAm mo souuomom mos .mo magmas :22 . rhozuz><>> 0mm Om? . omv. Om? Own. Om? 01V Omv quv OAS“. DOV :1 q q xexlh .7... 6mm 2_ .402me w......>F H T m; Antes Am mm mm . ATZU no. 1 23235335 vm , mm 127 one-photon absorption and fluorescence spectra of all- trans retinol in EPA at 77°K are both broad (Figure 34). The separation between the maximum of the one-photon ab- sorption band, 340 nm, and the TPE absorption maximum, 420 nm, is almost 5600 cm-1. If the 1Bu + 1Ag transition origin is assumed to lie at the red edge of the one- photon absorption band, 370 to 380 nm, then the separa- tion between the 1Bu origin and this two-photon allowed band is on the order of 2500-3200 cm-l. Due to its diffuse nature, little more can be said of this TPE Spectrum than that it shows the probable existence of a low—lying "1 Ag" state. In the region bluer than 4100 A, the TPE spectrum shows another diffuse two-photon allowed transition which might be attributed to the vibronically induced 1Bu + 1Ag transition. Figures 43 and 44 show the square dependence of two- photon induced fluorescence and the TPE spectrum of all- trans retinal in EPA at 77°K. As mentioned above, retinal is not a good candidate for TPE measurement because of its wavelength-dependent fluorescence quantum yield. Never- theless, it is interesting to note that the TPE spectrum shows a shift of the intensity maximum (423 nm) from.that of the one-photon absorption band (390 nm). This shift is similar to the one-photon excitation spectrum shown in Figure 39. 128 ZSF 2.0)- 15. 2 “(X b C? Log( loo: 2 Log( ID) V o o _1 )0 )- 05' SAMPLE: ALL-TRANS RETINAL IN EPA 77°K on: : CANDElA DYEA M 850011 I 1 l J.__ 0 05 10 L5 Log(b) Figure 43. Square Dependence of Two-Photon Induced Fluorescence of All-trans Retinal. 129 xoeA so