......v... .. .. .U..H.H_...U...”. . .x. . ...... . ,.......:...... .. . . .I.. . ......2... ........y.... . ...... .. ti... ......zf. ._ ..r.............x25... .. ... 3...... ...“...i: . ....fi... ......t... . ......7. 1... ............... ...... ...... .1 7.52.7“... .... Z ...... .... ...... . , ... .. ......Z... . ...; . .... . .. w . . at... . , .. . . 1 .. .. . .....z... .... Z... : .... ................ 2.. .. _. .. ..m. .. ”£2.54 7—). _ .79} ....E. I. .... v . ....r...$....__.m....§.r ...:z. .. . . . ...... ...... . .. p ......22... . . ...—17......» 7.3 .. 1 . ...; ...: ... 12.2.... .3.... .....Z: .. . 4.24-1— 3...... .... . Z .... .... 1...... . .. . . ... ......i ...... .... ... L .. ...... ...... .. ... ... ..r.... . .. .. .......L.,1 . .... a T . .A ..u.. 3.3,. . ... i. . 3...... .2 ...; 1...... ...... . . ......51 ...... .... ... . .....I... . . .1 .....J..7. , . . . ... ...: .... .... . .3334”... .4; , ._ I . . . : ........;r..u....... . .... .. . I . . ...:n} I... 3.-.... .....rr ...... . .... .x a ... ...I . ...? .. .... .. 1.7;... .v... .....a). .73... .37.. 5...: to. . . .. . 1...... . . i. . . .4... . . -.. . .. ... .Z. .....fu..;..; 4.1.u1133-l..4.. . ...... .....ru... LS... .... . 2:37.. .. ...} :3. (.1... I. 2...... ...}...i . 3. _. .3 ......x .. ...: . . .... . .. . .2. .. _. . 1...... 1.... 5.1.} h . ...v... ...... . .... ....”qu ”Quit. I; . ...... _. ,. 5:... .. 1.3.... .3.... 1.5.2; 1.... .....1 ...,» r... . . .13... .... ..x . $.11... AI , u . v a r 1...... . ... ... . ..Z... 1.3.1.... ...... 2...... .. .. .3 . .... . .....VJL ...).14 . r J 1... 9.1.5.: . a ...a .. it... ......)..1... .........,..... . ...? ....Z. ...:L 5...... . .... ... r. 2.7:... . . . ...} .. . ...: Z...E..I...:, . .3.. L... ...f. ...... . c .....5. . .... .... I c... 1...... = .s. .w: 1.3.3.: . ..3.... ...... . .. . .. ...... . .. . , a .... ......a .a..............p..&... figs»... ...... . .1133... . ...: . .. $.53“. k?» .51 J6... .. .1.....u..§...fi.......§@g.a.e.... , ... . . .3 2. -..: . ...... mm» Hrflhmissrii 7r. .1». . . a;.......fl..3...fx:t. ..-! . . ..r ’ . .... 4: .. . 37.2%...»5. ...»... . . .... . . . . , k ...}...ihrfi‘. .. ... x. . . 1...... 7.03%”... . 2 u? . .2. .1 , m... l :L. . . f. A. ...... ...“.r. .. a.n.:r......;».... .3 ;‘ ...... . ..., .... ...... J... ......qnxxhr. 1 ... (...-.. _, , . .15 .... ... 11:». y 5... : J sci... . . 1" ; .. . .. “...“... ...Vfl... . ......h. .r .. ...) ... ... .. . .7:.?......> .. 3.3. £2... . . z... ...! . .. .... . s .... .... figurmrumigwr ... ., 1. 5.7. «A... 2.5.. ... V b. ......u. v.1. ...! fit... a .11.... .... . V... ,. ... ...... .2... #3:?“ ..4fliam..fifi...y.%.wr. ... .5...» 4.2;}... ammo”: ...... . 2 .... . ".....nv......5..,.rrn a... w... I .. .x.......«.r 1.17;... ...... ...... Trii... .L 1. WJ. iv. L .. fizzy: ... ...... ...?f .... a»)? 1.3:... . . .....I... . . ......Iv? flux: :53 ... . L .... Lit . $1. .. ...—35.....- m... :1... amiss. 5.77:). .... .. a. . .5. .r . .~ L o. .. . . r ’43.. z 3.1... .3.... r. u :1 P; Z. r... . .3 ... .1 ...)t...s£ : .Iz. x351. v ..rr, ..l; 11.11.»..24. ..I... . .I/rl. . ..r..r. .. ...)?fvzi.!:ri 1:... I I. 3 y ,1: .91.... . ... .5}... ilrwvolf. ......frfmfffra 1v . :7... but}. .I .| f. "13 This is to certify that the ‘QE, thesis entitled ': , RELAXATION PROCESSES IN ELECTRONICALLY EXCITED MOLECULES (A) INTRAMOLECULAR EXCIMERS (B) EXCITED STATE PROTON TRANSFER presented by PHAEDON AVOURI S has been accepted towards fulfillment of the requirements for \ Mdegree in «(Ah—Mg .-. 2/7. (cwqf W’ (32un Major érofessor . . DateMi’W 0-7 539 ’ -\‘I l- I . BINUEX’I’NB’ ‘ BnAéYBf'Nerns' I lmsylrmrqytmuzmu ,, ABSTRACT RELAXATION PROCESSES IN ELECTRONICALLY EXCITED MOLECULES (A) INTRAMOLECUIAR EXCIMERS (B) EXCITED STATE PROTON TRANSFER BY Phaedon Avouris In the FIRST PART of the thesis various aspects Of intramolecular excimer and exciplex behavior are investigated. The technique of time resolved fluorescence spectroscopy is used to study the dynamics of the association and dissociation processes of the intramolecular excimer formed in the excited state of l,3-Bis(a1pha-naphthyl)propane (1,3DNP). The specific rate constants of these processes are determined in.media of different viscosity. It is found that the intramolecular excimer equilibrium is characterized by a high equilibrium constant cvlo4) and that both the association and dissociation rates are viscosity dependent. Through coupling of transient kinetic data with steady illumination data and model calculations, various thermodynamic quantities describing the intramolecular excimer interaction are determined. A method for obtaining the radiative lifetime 1; of the excimer is developed, and the 7; value is discussed in terms of the excimer geometry. Other factors like the effects of oxygen and solvent polarity are investigated in the system 1,3DNP and also in 1,3-dipheny1propane (1,3DPP). Evidence is presented for intramolecular triplet excimers in.viscous solutions and microcrystalline aSgregates of 1,3DPP. — ‘v-fll- c—u—I—li'i' ‘ ‘ 'mzzafi; Phaedon,Avouris Other intramolecular systems studied are the systems ©(CHZ)n'@N n = 1 (PyCHéP), n = 3 (1,3PyPP). These systems, under various conditions, display a'variety of interactions. In neutral alcohol solutions, the absorption spectra of 1,3PyPP do not show signs of ground state interactions vflflle the room temperature fluorescence spectrum shows intramolecular nflxed-excimer formation completely analogous to the 1,3DPP case. As the hydrogen bonding ability or acidity of the solvent increases, strong hydrogen bonding or protonation of the pyridinic nitrogen changes the interaction drastically. Under these conditions the absorption spectra of 1,3PyPP and PyCHéP show ground state interactions while the broad emission bands show extremely large Stokes shifts (14,000-15,000 curl). These emissions are interpreted as charge-transfer emissions. This is supported by the out of plane polarization of the fluorescence and by thermodynamic arguments. Evidence is also presented that the emitting charge-transfer and exciplex states are not identical in these intramcle- cular systems. In low temperature rigid glasses the Stokes shifts are much smaller, demonstrating the importance of molecular and solvent Franck- Condon state relaxation. Phosphorescence arises from a localized excited state of the phenyl group. Mcre subtle manifestations of the exciplex interaction (no exciplex emission) are Observed in the case of rigid and non-rigid phenylcarboxylic acids. From the study Of fluorescence and phosphorescence quantum yields and lifetimes, it appears that in the flexible molecules interaction of the carboxyl group and the aromatic ring results in enhanced intersystem crossing. The SECOND PART of the thesis is devoted to the study Of excited state proton transfer reactions, especially the various acid-base reactions in which 7-azaindole (7A1) is involved. Initially, the excited state , , .. we‘ve “A, ‘rmuagz Phaedon Avouris acid-base properties of 7AI are investigated through absorption and emission studies and the excited state pKa* and pr* are determined. Time resolved spectroscopy is used to study the kinetics of double proton transfer in 7AI hydrogen bonded dimers. 7AI deuterated in the N1 position is studied in a frozen 3-methy1pentane matrix at 77°K. The analysis of the non- exponential character of the decay curves gives the forward and backward reaction rate constants. The phenomenon is clearly demonstrated through time resolved spectra in the nanosecond time range. The effect of isotopic substitution on the emission spectra is used to obtain the rate constant for proton transfer in non-deuterated 7AI. From the kinetic data at 77°K and the effect of temperature on the fluorescence spectra down to 4.2°K, it is concluded that quantum mechanical tunneling must be the mechanism by which the proton transfer occurs at very low temperatures. When 7AI is dissolved in alcohols, a new emission band appears in addition to the molecular fluorescence. The intensity ratio of the molecular band and the new emission is concentration independent. Through the effect of solvent deuteration on the fluorescence quantum yield of 7AI and other model compounds strong evidence is found that the new emission results from a tautomeric species resulting from double proton transfer between the excited 7AI and an alcohol molecule. Finally, the various mechanisms deactivating the lowest excited singlet state of 7AI in H20 are studied throngh solvent isotope effects, temperature and pH studies. RELAXATION PROCESSES IN EIECTRONICALLY EXCITED MOLECULES (A) INTRAMOLECULAR EXCIMERS (B) EXCITED STATE PROTON TRANSFER By Phaedon Avouris A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1974 lif_l-me; TO MY MOTHER AND IN MEMORY OF MY FATHER ii The author I II. Ashraf Ill-Bay the course of ti Iwould a1: whose lectures understanding. on my comittee Finally 31 Liedtke for nu Miss Li Yang f ACKNOWLEDGEMENTS The author would like to express sincere appreciation to Professor ‘M. Ashraf El-Bayoumi for his guidance, encouragement and friendship during the course of this investigation. I would also like to express my gratitude to Professor J. F. Harrison whose lectures and informal discussions greately contributed to my scientific understanding. I also thank Professors G. Leroi and A. U. Khan for serving on my committee. Finally special thanks go to my good friends and colleagues Dr. R. Liedtke for numerous discussions and help with computer programing and Miss Li Yang for encouragement and help with the proton transfer experiments. IIST (F IABIES LIST OF FIGURES CIIIPIER I. EXCI IIIIRODUC'IIC GROUND STA'J EXCITED STI TRIPLET MO CHARGE-TM THE ENERGI THE DETAIL TRIP THEORETIC “RAPPER 2. SI 11 INTRODUC 1: 3"DINA Kit Th. Ef TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES CHAPTER 1. EXCIMERS-EXCIPLEXES-CHARGE TRANSFER COMPLEXES INTRODUCTION GROUND STATE INTERACTIONS-DISPERSION FORCES EXCITED STATE INTERACTION-THE MOLECULAR EXCITON MODEL TRIPLET MOLECULAR EXCITONS CHARGE-TRANSFER INTERACTIONS THE ENERGIES OF THE NORMAL AND CHARGE-TRANSFER STATES THE DETAILED FORM OF THE WAVEFUNCTIONS-THE CHARGE-TRANSFER TRIPLET STATE THEORETICAL DESCRIPTION OF THE EXCIMER STATE CHAPTER 2. STUDY OF THE INTRAMOLECULAR EXCIMER INTERACTIONS IN 1,3-DINAPHTHYLPROPANE AND 1,3-DIPHENYLPROPANE INTRODUCTION 1 , 3 -DI NAP HTHYLPROPANE viii 12 14 15 l7 I9 25 25 Kinetic Treatment of the Excimer Formation and Dissociation Process in 1,3DNP The Radiative Lifetime of the Intramolecular Naphthalene Exc imer Effect of Solvent Viscosity on the Intramolecular Excimer Formation Process The Effect of Oxygen on the Excimer-Monomer Relative Yields 55 (a) Enhanced Intersystem Crossing (b) Energy Transfer to Oxygen Importance of the W0 Mechanisms for the Quenching of the Monomer Emission iv 28 48 51 57 58 mu; 0? CONTENTS NAPIER 2---conti 1,3-DIPHENYI Solven Polar Origit IntraJ Mecha Sing] CREE 3. IN TH CA IIITRODUC'I 1. (PHEN‘ Ele Abs ABLE OF CONTENTS---continued IHAPTER 2---continued 1,3-DIPHENYLPROPANE Solvent Effects on Excimer Luminescence Polar Nature of the Intramolecular 1,3DPP Excimer Origin of the Activation Energy Intramolecular Triplet Excimer Formation Mechanisms of Triplet Excimer Formation Singlet and Triplet Intramolecular Excimers of 1,3DPP at 77°K CHAPTER 3. INTRAMOLECULAR EXCIPLEX-CHARGE TRANSFER INTERACTIONS IN THE SYSTEMS (PHENYL)-(CH2)n-(PYRIDINE) AND AROMATIC CARBOXYLIC ACIDS INTRODUCTION I. (PHENYL)-(CH2)n-(PYRIDINE) SYSTEM Electronic States and Spectral Properties of Pyridine and Pyridinium Ion Absorption Spectra (A) Absorption Spectra of V-Picoline, Toluene, and 1,3-DPP (B) Absorption Spectra of (Phenyl)—(CH2)n-(Pyridine), n = 1,3 Emission Spectra-Fluorescence Phosphorescence II. MANIFESTATIONS 0F CEARCE-RESONANCE INTERACTIONS IN THE CASE OF [O-(CH2)3- .14”! IN THE GAS PHASE MOde of Fragmentation of Alkyl-Benzenes in the Mass Spectrometer III. INTERSYSTEM CROSSING ENHANCEMENT THROUGH CHARGE-TRANSFER INTERACTIONS Introduction Intramolecular Exciplex-CT Interactions in Aromatic Carboxylic Acids 59 64 66 68 69 74 78 78 80 83 91 99 115 118 118 123 123 125 _, ___-c; mm OF CONTENTS mm 4, EXCIT THE 7 INTRODUCTIOT THE HTDROGE] (1) Th E (11) ...—JD‘Af‘é (III) PROTON TRA GROUND ANT PROTON TU] ISOTOPE E Prim Soh Dew PROTON '1‘ SYSTEMS In! DYNAMIt HYDROG? vi ABLE 0F CONTENTS---continued mAPTER 4. EXCITED STATE PROTON TRANSFER REACTIONS INVOLVING THE 7-AZAINDOLE MOLECULE INTRODUCTION THE HYDROGEN BOND (I) Theoretical Models of the Hydrogen Bond Electrostatic Model Valence-Bond Theory Charge-Transfer Theory SCF-MD Theory (II) Potential Energy Curves The Double Minimum Potential (III) The Hydrogen Bond in the Excited State PROTON TRANSFER REACTIONS IN HYDROGEN-BONDED SYSTEM: GROUND AND EXCITED STATES PROTON TUNNELING ISOTOPE EFFECTS Primary Kinetic Isotope Effect Solvent Kinetic Isotope Effect Deuterium Isotope Effects on the Rates of Non-Radiative Transitions PROTON TRANSFER REACTIONS IN 7-AZAINDOLE HYDROGEN BONDED SYSTEMS Introduction-Absorption Spectra Indole 7-Azaindole Excited-State Acid-Base Properties Indole 7-Azaindole DYNAMICS OF BIPROTONIC PHOTOTAUTOMERISM IN 7-AZAINDOLE HYDROGEN-BONDED DIMERS: TIME RESOLVED SPECTROSCOPY STUDIES Biprotonic Phototautomerism in 7-Azaindole Hydrogen-Bonded Dimer Excitation Delocalization in 7AI Dimer Kinetic Analysis 132 132 133 133 133 135 137 138 140 140 145 147 150 154 154 157 159 163 164 172 173 173 176 177 177 182 184 mm OF CONTENT mm 4""00nt Rate ( Activ: Kinet Tunne EXCITED ST ‘ EVIDENCE ( THE INTER Quen . cum 5. E)! I (I) Expe] . (II) 801* (m) Sp mummy BLE 0F CONTENTS---continued AFTER 4---continued Rate Constants for Proton Transfer Activation Energies for Proton Transfer Kinetic Isotope Effect Tunneling vs Proton Transfer EXCITED STATE COMPLEX 0F 7-AZAINDOLE WITH ALCOHOLS: EVIDENCE OF EXCITED STATE DOUBLE PROTON TRANSFER THE INTERACTION OF 7-AZAINDOLE WITH WATER Quenching Mechanisms-Solvent Isotope Effects AFTER 5. EXPERIMENTAL (I) Experimentally Studied Molecules (II) Solvents (III) Spectral Measurements BLIOGRAPHY 186 188 190 192 196 201 201 215 215 217 227 Table 1. Kinet Glyce Table 2. Rate D188: at 2.‘ Table 3. Ther Tableh. Effe Table 5. Effe of 1 Table 11. - Table 12. TOle 13. Table 14. Table 15. Lb 1e mle mle mle mle able able able LIST OF TABLES Kinetic Data for the 1,3DPP Excimer in Ethanol- Glycerol Mixtures at 25°C. 37 Rate Constants for 1,3DNP Excimer Formation and Dissociation Processes in Ethanol-Glycerol Mixtures at 25°C. 42 Thermodynamic Data for the 1,3DNP Excimer Interaction. 43 Effect of Solvent on the Emission Properties of 1,3DPP. 59 Effect of the Medium on the Frequency and Intensity of the 7-Picoline Absorption Spectrum. 86 Effect of the Medium on the Frequency and Intensity of the Toluene Absorption Spectrum. 88 Effect of the Medium on the Frequency and Intensity of the 1,3DPP Absorption Spectrum. 88 Effect of the Medium on the Frequency and Intensity of 1,3PyPP and 1,3Py+PPC104' Absorption Spectra. 92 Effect of the Medium on the Frequency and Intensity of the PyCHZP and Py+CH2PC1O4' Absorption Spectra. 92 Stokes Shifts and Effect of Excitation Wavelength on the Emission Maxims of 1,3PyPP and PyCHQP in Different Solvents (in cm'l). 114 Relative Intensities of Various Fragments of Butylbenzene and 1,3DPP in the Mass Spectrometer. 120 Effect of Viscosity on the Fluorescence and Phosphorescence Quantum Yields and ISC Rates of Toluene and Ethyl-Phenyl Acetate 128 Effect of the Substituent X on the Absorption and Emission Properties of x-@CH2-COOH. 128 Luminescence Properties of Rigid and Non-Rigid Phenyl- akylcarboxylid Acids in Ethanol. 130 Absorption Spectral Shifts for Indole in Different Media (cm‘l). 171 viii LIST OF TABLES“ Table 16. AbsorP Relati Table 17. AbsorT Media. Table 18. Rate 1 indol: Table 19. Fluor llethy ix IST OF TABIES---continued able 16. Absorption Spectral Shifts for Indole in Different Media Relative to 3MP. able 17. Absorption Spectral Shifts for 7-Azaindole in Different Media. able 18. Rate Constants for the Double Proton Transfer in 7-Aza- indole Dimers at 77°K. able 19. Fluorescence Quantum Yields of Indole, 7-Azaindole and Methyl Derivatives in Water and Deuterium Oxide. 171 172 187 205 Figure 1a. Sche in T Figure lb. For: Arm Figure 2. Room (1.2 Figure 3. The Figure 4. The Flu! Figure 5. Qua Gro Figure 6. The N01 Et' Figure 7. Th Mo Figure 8. De E} Figure 9. T ( A FiElite 10. Figure 11. FiElite 12. Ifigure 13. Figure 14. ure ure ;ure :ure gure gure gure gure gure gure gume Sure gure 1a. lb. 10. ll. 12. 13. 14. LIST OF FIGURES Schematic Energy Diagram Showing Exciton Splitting in Molecular Dimers. Forster's34 Exciton Model of Excimer Formation in Aromatic Hydrocarbons. Room Temperature Fluorescence Spectra of 1,3DNP (1.2 x 10-414) in EtOH. The Gauche-Gauche Conformation of 1,3DNP. The Rise and Decay of the Intramolecular Excimer Fluorescence. Qualitative Potential Energy Diagram for Dimer in Ground and Excited States. The Effect of Viscosity on the Relative Intensity of Monomer and Excimer Fluorescence Bands of 1,3DNP in Ethanol-Glycerol Mixtures. The Effect of Viscosity on the Relative Intensity of Monomer and Excimer Fluorescence Bands of 1,3DNP. Determination of the Activation Energy for Intramolecular Excimer Formation in 1,3DPP. the the Total Emission Spectra of 1,3DPP; (1) 3MP Glass at 77°K (2) iP Glass at 77°K, and (3) After the iPoGlass was Allowed to Melt (Estimated Temperature 115 K). Emission of Concentrated Solutions of 1,3DPP in iP. Symmetries and Energies of the lowest Singlet States of Benzene, Pyridine and Pyridinium Ion. Absorption Spectrum of 7 Absorption Spectrum of T-Picoline in Water (__), Absorp- tion Spectrum of v-Picolinium Perchlorate in Water (--). Absorption Spectra of Toluene in Ethanol. -Picoline in Methylcyclohexane. 20 20 34 35 36 40 53 54 61 73 75 82 84 89 J LIST 01P FIGLRES‘" Figure 15. AbSOT Figure 16. AbSOT in M Figure 17. Abso and Nate Figure 18. Abs PVT Figure 19. INT M0 Figure 20. R0: in Figure 21. I. I] Figure 22. R. W. 1 Figure 23. R 1 T Figure 24. . Figure 25. FiSure 26. Figure 27. Figure 28. PiSure 29. xi T OF FIGURES---contined ;ure ;ure gure gure gure gure gure gure gure gure ,gure .gure Lgure Lgure 15. 16. 17. 18. 19. 20. 26. Absorption Spectra of 1,3DPP in EtOH. 90 Absorption Spectrum of 4-(3-Pheny1-Propyl)-Pyridine in Methylcyclohexane. 93 Absorption Spectra of 4-(3-Pheny1-Propy1)-Pyridine (__) and the Corresponding Pyridinum Perchlorate (--) in Water. 94 Absorption Spectra of Benzylpyridine (__) and Benzyl- pyridinium Perchlorate (--) in Water. 95 INDO Calculations. Energies of the Lowest Unoccupied MO and Charge Densities. 98 Room Temperature Fluorescence Spectra of (1) 1,3DPP in EtOH, (2) 4-(3-Phenyl)-Pyridine in EtOH. 100 I. Room Temperature Fluorescence Spectra of 4-(3-Phenyl- Propy1)-Pyridine in EtOH + HC104. II. Total luminescence Spectrum of the Above Solution at 77°K (Excitation at 265 nm). 104 Room Temperature Fluorescence Spectra of 1,3PyPP in Water (__), Room Temperature Fluorescence Spectra of 1,3Py+PPClO4‘ in Water (--). 106 Room Temperature Fluorescence Spectra of Benzylpyridine in Water(__), Room Temperature Fluorescence Spectra of Benzylpyridinium Perchlorate in Water (--). 107 Total Luminescence Polarization Spectrum of 1,3Py+PPC104- EtOH at 77°K (__). Phosphorescence Polarization Spectrum of 1,3Py+PPC104' in EtOH at 77°K (--). 111 (__) Total Luminescence Spectrum of Benzylpyridinum Perchlorate in EtOH at 77°K. (--) The Phosphorescence Spectrum under the Same Conditions. 112 (__) Phosphorescence Spectrum of 1,3PyPP in EEOH at 77°K. (--) Phosphorescence Spectrum of 1,3Py+PPC104 under the Same Conditions. 116 Portions of the Mass Spectra of 1,3-Diphenylpropane and Butylbenzene. Energy of the Electron Beam 70 eV. 122 Potential Energy Diagrams for the Motion of the Proton 142 in a Hydrogen Bond A-H---B. Hypothetical Potential Energy Curves for the Formation Of a H Bond A-H~--B in Ground and Excited States, (1) 146 wg) we, (2) wg we. FIST OF FIGURES- Figure 30. Mode “gum 310 hi! 01:14 on' Figure 32. Ind Figure 33. 7-1 Figure 34. Ca' Figure 35. Ca Figure 36. ET of Figure 37. E1 0: Figure 38. R ( 1 Figure 39. i 1 l"lgure 40. - Figure 41. Pisure (+2. Pisure 43. Figure 41.. FiEure 1+5. FiElite 46. 1Nam 47‘ 3T OF FIGURES---continued gure gure gure gure gure gure gure gure gure gure gure .gure .gure Lgure igure Lgure igure igure 30. 31. 47. Model Barrier for Proton Tunneling Calculations. Primary Kinetic Isotope Effect. The Vibrational Origin of the Effect of Deuterium Subsititution on Reaction Rates. Indole Vapor Absorption Spectrum. 7-Azaindole Vapor Absorption Spectrum. Calculated Charge Densities for Indole. Calculated Charge Densities of 7-Azaindole. Effect of pH on the Fluorescence of an Aqueous Solution of Indole at 25°C. Effect of pH on the Fluorescence of an Aqueous Solution of 7-Azaindole at 25°C. Room Temperature Absorption (left) and Fluorescence (Right) of 7-Azaindole in 3MP at 1.0 x 10‘5M (__) and 1.0 x 10-2M (--). Biprotonic Phototautomerism in 7-Azaindole Hydrogen-‘ Bonded Dimers. Corrected Fluorescence Excitation Spectra of 7-Aza- indole in 3MP at 25°C. Time Resolved Emission Spectrum of 7AI-d1 in 3MP at 77°K. Arrhenius Plot for Estimation of the Barrier to Double Proton Transfer in 7AI Dimers. Effect of Deuterium Substitution on the Fluorescence of 7AI (IO-2M) in 3MP at 77°K. Corrected Room Temperature Fluorescence Spectra of ZAI in EtOH (__) and 7AI in EtOD (---). Room Temperature Fluorescence Spectra of 7AI in EtOH (__) and N-Methy1-7AI in EtOH (---). Application of the Benesi-Hildebrand Equatign in the Determination of the 7AI-EtOH Complex at 25 C. Room Temperature Fluorescence Spectra of 7AI in H20 (--) and D20 (__). Both Solutions Have the Same Optical Density. 153 158 165 166 167 169 174 175 179 181 185 189 191 197 199 202 206 LIST OF FIGURES- Figure 48. V311 Fleas Flu: as; Figure 49. Que 1 / .1. Figure 50. Del "Ar Figure 51. BI Figure 52. Th Ou xiii 3 OF FIGURES---continued me 48. Ire 49. Lre 50. re 51. re 52. Variation of the Solvent Deuterium Isotope Effect Measured by the Ratio of the Room Temperature (25°C) Fluorescence Quantum Yields of 7AI in H20 and D20 as a Function of pH(pD). 208 Quenching of 7AI's Fluorescence in Water by Hydroxyl Ions. Ig7l vs -1og(OHT) Where 10 is the Intensity at Neutral P . 210 Determination of the Activation Energies for the 'Eetivated Quenching" of 7AI's Fluorescence in Water. 211 Block Diagram for Time Resolved Spectrophotometer. 221 The Basic Idea of Time-To-Amplitude Conversion. The Output Pulse Amplitude is Proportional to t t 223 stop' start' A molecul ent molecular state. Such r aESTegute) is dissociates a inverse case This case is bound-excite (excited Ste t11911111111 Phe Probabi in 19241 whr discharge 01 helium mole of mercury. high in 19 Beside large orEat berg. In of pyrene CHAPTER 1 EXCIMERS-EXCIPLEXES-CHARGE TRANSFER COMPLEXES INTRODUCTION A molecule in an electronically excited state may behave as a differ- molecular species. This can lead to associations unique to the excited :e. Such a case is the "excimer" association where the dimer (or higher :egate) is stable only during the lifetime of the excited state and aociates after the emission of the photon. The excimer represents the arse case of absorptiOn followed by dissociation of the excited state. a case is manifested by continuous absorption spectra (ground state id-excited unbound) while excimers show continuOus emission spectra :ited state bound-ground unbound). Continuous emission spectra provide main phenomenological manifestation of excimer formation. Probably the first observation of the excimer emission was that of Lyman .9241 who observed a diffuse-appearing band at 600.3 A in a transformer :harge of helium. Sommer in 19272 attributed this emission to the .um molecule, He2*. Another of the early discovered excimers was that ercury. The excimeric emission of mercury was discovered by lord Ray- h in 19293 and the assignment to Hg2* was done by S. Mrozowski in 19374. Besides the excimers found between excited and ground state atoms, e organic molecules (e.g. aromatic hydrocarbons) form molecular exci- 5 . In 1954 Forster and Kasper were studying the concentration dependence Yrene fluorescence when they observed that by increasing the concentratiOn of pyreue, the while a new bro intensity red £ to an excited I years this beh the phenomenon Excited 5 called "excip‘. where it was carbons by ar 5000 cm'1 to to an excite: Simple atomi The obs manifestatir We mention 1 (1) The app Changes in “We is c that the m (2) The em state equi (3) From 5 the dialer ‘10“; thi °bserved (10) The ‘ pyrene, the fluorescence emission band (0,0 w27,000 cm‘l) was quenched, la a new broad and structureless emission started appearing with a peak :ensity red shifted by m 6000 cm'l. The authors ascribed this new emission an excited state association between two pyrene molecules. For several lrs this behavior was considered somehow unique to pyrene. Birks6 showed 0 phenomenon is common to most aromatic hydrocarbons and their derivatives. Excited state associations between different kind of molecules are lled 'hxciplexes". They were discovered by Leonhardt and Weller7 in 1963, are it was found that upon quenching the fluorescence of aromatic hydro- :bons by aromatic amines, like anilines, a new broad band appeared about )0 cm"1 to the red of the hydrocarbon emission. This emission was assigned an excited complex of the hydrocarbon and the amine. Exciplexes between nple atomic systems like (KrAr)* were also reported recentlys. The observation of a broad strutureless emission band is not the only iifestation of excimer formation. There are several criteria from which mention the following: 1 The appearance of the broad emission band is not associated with any inges in the absorption spectra. Even after prolonged irradiation no inge is detected in the absorption spectrum. This supports the idea it the new emission is not due to photoproducts. i The emission spectra show an "isostilbic" point indicative of the excited rte equilibrium. 1 From Stern-Volmer relations the ratio of fluorescence intensities of : dimer and monomer is expected to be proportional to monomer concentra- tn; this is actually observed. No concentration-dependent changes are erved in the absorption spectrum. The excimer formation is viscosity dependent and shows the characteristics of diffusion control1e (5) The decay of the “ It the excimer dissoci described by a sum 0f lished in the excited summon lifetime. (6) Thermodynamic dat of association proces Several other results and we will talk abor Excimer formatir molecular crystals. Fluorescence and or molecules like napht‘ Of an m, in whi dicular to one anoth Structures of a in pairs (pyrene, benzc (d-perylene , COW“ crystals of the B 12‘ Interactions (811811 further relaxation formation to occur . at very low tempera 10‘50 kbar show par and Polyvinylnaphtl Excimers and e Efusion controlled reactions. he decay of the monomer is usually described by a single exponential. e excimer dissociation rate is significant then the monomer decay is ibed by a Sum of two exponentials. Finally if equilibrium is estab- d in the excited state then both monomer and excimer will decay with mon lifetime. hermodynamic data always yield negative entropy changes characteristic fig sociation processes. 'l, al other results point to an excited state reversible association e will talk about them in the text. 1 Excimer formation is not unique to fluid phases but is common in ular crystals. Stevens9 proposed a general relationship between escence and crystal structure of aromatic compounds. 'Elongated" ules like naphthalene, anthracene, phenanthrene, favor structures A type, in which adjacent molecules are oriented almost perpen- ar to one another. 'Disk-Shaped" molecules on the other hand, favor tures of a B type, in which the molecules are either arranged in (pyrene, benzo g,h,fl perylene, a-perylene ) or stacked in columns :ylene , coronene, ovalene). Excimer fluorescence occurs only in als of the B type. The absorption spectra do not show any strong actions (slight Davydov splitting may be observed). In the crystal, er relaxation is needed from fire ground state arrangement for excimer :ion to occur. Thus a—perylene no longer exhibits excimer fluorescence 7y low temperatureslo, while crystals of type A under pressures of kbar Show partially excimeric emissionll. Polymers like polystyrene ‘1YVinylnaphthalene exhibit excimeric emissionlz. xcimers and exciplexes can be formed in a variety of other situations, for examPle‘ a. Radical ion annihi b. Triplet-Tr iplet at c. Photodecompositim The literature Birksm and Stevens1 While experimet well established, t] complete. The most amixture of "exc it fbllwing the molec discussed together transfer interacti‘ and ground state i' discussed. xample: dical ion annihilations (Chemiluminescence) A-"+A-+ ——————> (AA)* iplet-Triplet annihilation 3A*+3A* ...—....» 1(AA)* wtodecomposition of dimers, like dianthracene. The literature on excimers has been reviewed extensively by Forster1§ Ff :14 and Stevensls. " While experimentally the phenomenon of excimer association appears 3; established, the theoretical description of the phenomenon is not a .ete. The most commonly used description of the excimer state is as :ture of 'hxciton resonance and charge-transfer states". In the l Ming the molecular exciton and the charge-transfer models will be lssed together with a model which considers both exciton and charge- :fer interactions. Since excimer luminescence reflects both excited :round state interactions, the forces present in both cases will be Lssed. GROUND STATE INTERACTIONS-DISPERSION FORCES Consider two ground state molecules of the same kind (labeled as A ) that are weakly interacting. The ground state wave function of the m is described by: WG = ¢a¢b molecular electronic overlap is neglected so ¢a¢b need not be anti- tric with respect to intermolecular electron exchange. The Hamiltonian e dimer is given by: H=Ha+Hb+Vab Vab is the intermolecular perturbation potential (electrostatic interactions). The e landb is then EG = (la where E8 = Eb is the element on the left 1 V is expressed as ab V=‘V 'FV mono-mono Inc This point-multipole molecules a few A a] nultipole expansion shown by Dalgarno a truncated at some c and better when R :‘ For neutral in the only important cimImstances the The interam where Xe is the 8 land in the Same is EC coordinates in t Where , t he X component :ractions). The energy of the ground state of the system of molecules Id B is then EG = <¢a¢b IHI ¢a¢b> = Ea + Eb + <¢a¢blvabl ¢a¢b> re E8 = Eb is the energy of the isolated molecule while the matrix mnt on the left gives the van der Waals interaction energy. Usually is expressed as a point-multipole expansion: = Vmono-mono+vmono-di+vdi-di+unad-quad+vdi-quad+voctu-octu+°""" : point-multipole expansion does not strictly apply to the case of :cules a few A apart. As a result of this the use of the complete :ipole expansion will lead to a divergent series in (R‘l) as has been m by Dalgarno and lewisl6. This series is asymptotic, so that, when mated at some order, it gives an approximate value which becomes better better when R increases. For neutral molecules the monopole interactions are zero. Usually only flmportant term is the dipole-dipole interaction term, in these :umstances the point dipole approximation is used. The interaction energy between two molecules A and B in the point >le approximation is v = ez/R3(Xaxb + YaYb - zzazb) (1) :e Xa is the sum of the X coordinates of all the electrons in molecule 1d in the same way for Xb, Ya etc. The first order perturbation energy E(1) = eZ/R3(+ - 2) (2) :e , etc. are the expectation values of the sums 0f the :dinates in the respective molecules <¢alXal¢a> etC- BY d3f1n1t1°n X component of the dipole moment of molecule A is pxa = e (3) So the first order er For centrosymnetric lene (D211) the matri odd parity of the di T0 get an inter (2)_ 4 6 ‘ E - e /R iii-QC where a: is the i the ground state we orientations of the vanish so (5) is W (2) 4 6 E =e/R Z i,j#)( For isotropic 111016 o the first order energy can be written E0”) = l/R3(uxau - 2n ) (4) xb + uyauyb zauzb or centrosymmetric aromatic hydrocarbons like benzene (Déh) and naphtha- ene (D2h) the matrix elements of the form of (3) vanish because of the dd parity of the dipole operator. To get an interaction energy we have to go to second order, then (2): 64/R6 il j¢0< xa mxgo + Yingo io on E 22a Zb ) /(Ea 0+ Eb Ea l-Ebj) (5) io i 0 = < > etC. Xa walxal ¢a . 0 here ¢: is the i excited state wavefunction of moleCule A and ¢a 18 be ground state wavefunction. If the interaction is averaged for all rientations of the molecules, cross products involving XY: Y2 and X2 anish so (5) is written: 2)_46 ionoZ 102j02+ zi°2j°2/nao+n -E)(6) -e/Ri%#0(xa xb +3:a Yb 43 Z], )( bo' Eai bj or isotropic molecules (not aromatic hydrocarbons) . . r 2 iX:°|2 = |Y10|2 = lzlolz = 1/3|F10l '0 2 330 2 :0 2 jo 2 (7) J = = = 1 3 Fb be I le I le | / l l 0 (6) can be written: . . 2 E(2 )= 2e 41/31623- ¢0(|F:°l2lFJ°| )/(Ea o+ Ebo - Eai- Ebj) (8) London17 simplified further the above expression through the expression or the static polarizability ko 2 o= 2.22/3 E301? I )/(Ek- E0) (9) rom studies of the frequency dependence of the Polarizability it was ound that the most important terms (contributions) arise from states or which Ek- E0: I the ionization potential of that molecule. So approxi- ately (9) is Wkolz (10) a = (2e 2M and (8) may be writte'. ll(2 Expression (11) gives and B, while “A and c forces are always at! to the case of aniso E(2). For an anisotro a second order tensc is the cooridinates an optically inacti Gtomatrically 3 is be transformed by : quasi-scalar tense: ‘ tonal elements (or x and the directions °E the molecular 1 comes from the ex h. + E = 1 WhiCh leads to de l he matrix repres “9° cases or (1) Random 01’ 1811 (8) may be written as E(2)= -3IAIB/2(IA+IB)'(aAaB/R6) (11) ession (11) gives the dispersion energy for two isotropic molecules A B, while 9A and dB are the isotropic polarizabilities. The dispersion es are always attractive. This expression can be easily generalized he case of anisotropic molecules18 (2)_ ‘ a: I. 6 E — -%(IAIBtr[TaATuB])/(IA+IB)R (12) For an anisotropic molecule the polarizability is not a scalar but cond order tensor a a u xx xy xz a = a a a yx yy yz a azx azy zz he cooridinates are rectangular ' becomes a symmetrical tensor for a Ptically inactive molecule, i.e., six independent elements are required. etrically 3 is a polarizability ellipsoid. The symmetric tensor can ransformed by appropriate rotation of its coordinate system into a i-scalar tensor in the matrix of which the three pairs of non-dia- 1 elements (ax ,a ’azx etc.) are all zero. The diagonal components ayy’ and azz are then the principal polarizabilities of the molecule the directions along which that are measured are the principal axes he molecular polarizability ellipsoid. T which appears in equation (12) . + . s from the expression for the field E created by a d1PO1e u at a P0111t E = 1/R3si/Rrri/Rm- E}= 1/R3(3fi/R a n/R-r) h leads to define T = 3(fi/R a i/R) - l. In an orthogonal basis xyz + + . matrix representing R/R Q R/R is easy to obtain. TWO cases are important for us Random orientation of the two interacting molecules, (2) Stacked aromatic 1 in the first case we is the isotropic pola we can get the inter: l=3(loh)/1 SO TraceuuATaB) = ( A X/ If the stacking is Tr: 3° (12) becomes Ed For aromatic hydrc Polarized, the “x: out of plane P0181 tions that lie hi the fougwing Va] ( 1 2) Stacked aromatic molecules 1 the first case we can use equation (11) where a = (“xx + ayy + azz)/3 ; the isotropic polarizability. In the case of stacked configuration a can get the interaction as follows: xx xy xz Yx YY yz zx zy 22 = 3er Emz - '1' = 3/n2 l O 0 0 l O 0 0 l J Trace(TaATaB) = (3x2 /R2+(3y2-1)2 aAyyuBW/R2+(322-l)“AzzuBzz/R '1)2 u xx“ Bxx A z ‘i ‘ B e R ,. y A X/ Sthe stacking is vertical then x=y=0 and -- ‘I 2 2 o = - 1 = 4a a (e = 90 ) Trace (T‘ATaB) (3cos 6 ) aAzzaBzz Azz Bzz > (12) becomes 6 = + R (13) Edisp IA 1B“ Azz“ “Bzz/(IA IB) >r aromatic hydrocarbons with Aflfl*low excited states, which are in plane )larized, the “xx and “yy principal polarizabilities are larger than the 1t 0f Plane polarizability “22 which is determined mainly by no* transi- 9 . ions that lie higher in energy. For example leFevre and Sundaram give 1e following values 23 23 . 3 Compound “xx = ayy(x10 ) azz(x10 ) in cm Benzene 1.12 0.73 Triphenylene 3.89 1.54 Coronene 5.68 2.07 In conclusion we see th (1) For nonpolar ground second order and depend (2) For planar aromatic favorable orientation 1 figuration repulsive f( one would expect that 1 with configuration to l EXCITED STATE The exciton conce in 193120 in connectic heat in rare-gas solic‘ exciton model to the 1 tale. Since that tin book "Theory of Moles area of molecular and tally using exciton t and its applications to Consider in the and an unexcited molt “9 have the mo mole! lance/h emission <1)‘ and becOmes excited, 0“ the other hand if he nclusion we see that: or nonpolar ground state molecules the interaction appears in the i order and depends on the polarizability 3r planar aromatic hydrocarbons the stacked orientation is the least able orientation from the dispersion energy standpoint. In this con— ation repulsive forces (closed-shell interactions) are maximal. Thus )uld expect that ground state interactions particularly in the sand- :onfiguration to be repulsive as it is usually assumed. EXCITED STATE INTERACTION-—THE MOLECULAR EXCITON MODEL fhe exciton concept was introduced into solid state physics by Frenkel £120 in connection with the transformation of excitation energy into .n rare-gas solids. In 1948 A. S. Davydov21 applied the molecular in model to the problem of the electronic states of naphthalene crys- Since that time and especially after the publication of Davydov's "22, a great variety of problems in the Theory of Molecular Excitons f molecular and solid state physics have been approached theoreti- using exciton theory. For a recent review of the exciton theory 8 applications see the article by Philpott23. Here we are going sider in the most simplified form the interaction of an excited unexcited molecule on the basis of the exciton model. Say again e the two molecules A and B (of the same kind) in proximity (Dis- A <1). Now we consider that molecule A absorbs a photon emission comes excited, then we can write the wavefunction of the system as = a V1 ¢a ¢b other hand if the excitation is localized on b the wavefunction will ‘12 =¢a¢b1¢ i and *2 are degenerat 1 the system. Proper lin stationary states of ti sentations of the dimer perturbation theory. terminant: where H a: “at (* denotes electronic the determinant are and the eigenvectors In both stationary 6 molecules, A and B, The node correspond: an excitation node. transition mOments Evaluating E' and E E' = E as a g" a p; a - 1“ equations 16:17 am Is forces and a 10 d W2 are degenerate states and do not describe stationary states of ystem. Proper linear combinations of these two localized states are onary states of the dimer and belong to different irreducible repre- tions of the dimer point group. This is a simple case of degenerate rbation theory. The first order energies can be found from the de- nant: H - E H as ab = O Hab be - E Has = be = <¢a*¢b IH| ¢a*¢b> Hab = Hba = <¢a*¢b IHI ‘ba¢b*> notes electronic excitation, not complex conjugates), the roots of eterminant are l = + E Haa Hab II = _ H E Haa ab he eigenvectors ' = '% k + k (14) r 2 waob ogb) '% (15) n = _ 7' w 2 (oa*ob sash c) th stationary exciton states W' and w" the excitation is on both ales, A and B, i.e., the excitation is collective or delocalized. ode corresponding to the minus sign in the exciton wavefunction is citation node. At the excitation node, the phase relation between ition moments on the respective molecular centers changes sign. sting E' and E" we get: E' = E39: + Eb +<¢as¢tJvab1 ¢a4¢b> +<¢as¢blvabl¢£ba<> E" = Ea* + Eb +<¢ a% d Vafl¢ aw b> _<¢ 3% d vad¢ 4’b*> ‘ * * resents again the van der Jations 16&17, the term<¢a oblvflbhg ¢b> rep (16) (17) forces and almost always results in a lowering of the energy 0f the system. If energy leve photon resonance and th tern leads to repulsion always attractive. The splitting" A and describes the effe the dipole approximati X l * .. ‘hl an. >- d; , * _ ahh*i the exciton splitting he = - 2 \Ma ‘ ‘ where Ma represents t' z €05 is 60898 represen moment of as M makes a L max111111111 exciton intep in a perfect way, "9. will be given then b Experimentally M is t r is the molar exr 6.3. . u" I"lgure In t . hough in the molec th allowed sPectral t. by Ev a1 . “atmg the ll . If energy levels exist betwen the two states involved in the n resonance and the ground state then a Situation may arise that this . 2 leads to repulsron 4. In the ground state van der Waals forces are s attractive. The last term in equations l6&l7 gives the "exciton ting" A5 == M: : x component of the transition moment of A <¢D prb |¢b*>= M: : x component of the transition moment of B, etc., exciton splitting term is given by the general expression A5 = - [Ma Iz/R:b (2cos ezcoseg-cose :coseg-cosezcoseg) 2 M3 represents the transition moment in the free molecule, and cosefi, ;, cog): represent the cosines of the angles which the transition It of a, Ma makes with the x, y, z axes of the coordinate frame. The sum exciton interaction will appear when the two molecules are stacked perfect way, "perfect sandwich" configuration. The exciton splitting be given then by At = - Ma2 /R:b imentally M is obtained frOm the absorption of a dilute solution of ompound under consideration using die equatiOn M2 = 3.97 x 10'20e2I2d3/3 the molar extinction coefficient and e is the electronic charge in .. Figure la describes the case of a "perfect sandwich" dimer. Al- h in the molecular dimer two exciton states theoretically result from xciton splitting, both of these may not necessarily be observed as ed spectral transitions. Let us examine the spectral selection rules aluating the matrix elements of the electric dipole operator between L '1 _\ t-I‘“ . ,3 the ground state and the the transition moment v« M M or 14"" =l<¢a¢b|Ma+1fl>H Orthogonality and norm 1 or Similarly For the "perfect sandw and Thus, the transition 1 the transition moment strength (f) for elec transition to V" 15 f In conclusion we see interaction energy re lated molecule energ correction is in fir gurathn is the 1&& In Our dismiss for simplicity that t here is no electr. spin free, PhOton 12 e ground state and the stationary exciton states of the dimer. Thus, e transition moment vector of the dimer is given by M' = Gab II = (“I + w") M GIMa Mbl L Ml,u=< M+ ~k + 'k) 22 [ eaebl a ublwa ¢b_¢a¢b ) 1/ thogonality and normalization properties of the wavefunctions lead to: . a e M' =<¢ IM |¢ *>/2 + <¢ [M |¢ *>/2 a a a a a a L I: + 2 M (Ma Mb)/2 ;, . n: _ 2 milarly M (M.a Mb)/2 r the "perfect sandwich" dimer M' = 2-%(M - ) = o a Mb d M" = 2-%(M + ) = 2M /2% a Mb a us, the transition moments for the dimer are given as superpositions of a transition moments for the individual molecules. The oscillator rength (f) for electric dipole transitions to V' is zero, while the ansition to W" is equal to f" “IM”2| = |2M 2I or f" = 2f a monomer conclusion we see that for non-polar molecules in the ground state, the teraction energy will be given as a second order correction to the iso- ted molecule energy. In the excfl ed-ground state case, however, the traction is in first order. For aromatic hydrocarbons the stacked confi- ration is the least favored in the ground state. TRIPLET MOLECULAR EXCITONS In our discussion of the simple molecular exciton theory we assumed ? simplicity that there was no electronic overlap between A and B. If are is no electronic overlap and since the Hamiltonian is practically in free, photon resonance of the form \ t | l J is not allowed because allow electronic Wet] cribing the interacti< fish :1. ¢a*(1)¢b (2: .f ,ae(1)ob(2 where H(l,2) = 1/r1,2 spin wavefunctions. is the same integral i.e., it is the 001114 and ¢b(2)¢b*(2). Si' ordinates of electro has non-zero value c this Coulomb term v: exchange integral w] the "overlap charge has a non-zero valu C1ition is satisfied (”“1619 decreases 1 tegral will be Sigp The concept 0 e I Xplain cer tain mu Crhistalszs. Since in ‘ mic and molecula e o l3 3A*1B <-————->1A3B* I II Ts not allowed because it involves a spin forbidden transition. If we hllow electronic overlap the picture is changed. The matrix element des- sribing the interaction between states I and II is given by =f¢a*(1>¢b(2>u(1,2>¢a<1)¢be(2>xa<1>xb(2)xa<1)xb*<2>dv<1,2> 1].sgf(1)¢b(2)H(1,2)¢b*(l)fa(2)Xg*(1)Xb(2)X$*(1)X3(2)dV(1.2) (18) Hab here H(l,2) = l/rl 2, ¢'s are the spatial wavefunctions and x's are the 5 1pin wavefunctions. The first term in the right-hand side of equation 18 s the same integral as the one which arises in the absence of exchange, -e., it is the Coulomb integral between the charge distributions ¢é*(l)¢ (l) a nd ¢b(2)¢£*(2)° Since H(l,2), being the function of only the space co- rdinates of electrons, does not affect the spin function, the first term as non-zero value only when Xa*(l) = Xa(l) and Xb(1) =XB*(2). Therefore his Coulomb term vanishes in the present case. The second term is the xchange integral which represents the electrostatic interaction between he "overlap charge distributions" ¢;*(1)¢£*(1) and ¢a(l)¢%(l), and which as a non-zero value when 32f(l) = Xb*(2) and Xa(1) = Xb(2). The last con- ition is satisfied in the present case. However, since the electronic verlap decreases rapidly with intermolecular distance, the exchange in- egral will be significant only if the molecules are very close. The concept of the spin multiplet state exciton has been introduced to xplain certain multiplet structures occuring in the optical spectra of ionic rystalszs. Since that time numerous investigations of triplet excitons in onic and molecular crystals have appeared. For a description of the triplet rciton properties see for example the work by Avakian and Merrifield26. The theory 015 Ch? weloped by Mulliken 11 of this theory with e! of the subject see the hulliken and Pearson} The quantum-mach molecular complex in for normal state), th of the Schroedinger w Mulliken repress state wavefunctions d: accenter (A) With co: and interacting. F01 In the case the writes a lhre the no-bond wav (1 due to possible clas wuetu notions ¢D an attractive (dispersi in in and o A which 1 upon cOmplexing. The function it Pair obtained by re Pr eviously unoccupi b etween the odd ele Th e‘I’N function is 14 CHARGE-TRANSFER INTERACTIONS The theory of charge transfer complexes in its present form was de- aped by Mulliken in 195227. Here we are going to discuss some aspects :his theory with emphasis on electronic spectra. For excellent reviews the subject see the articles of McGlynnZB, Brieglebzg, M'urrel3O and liken and Pearson31. The quantum-mechanical description for the electronic structure of a ecular complex in its ground electronic state is given by ¢N(N stands normal state), the wavefunction which is the lowest-energy solution the Schroedinger wave equation for all the electrons in the complex. Mulliken represented in for a 1:1 complex in terms of the normal- :e wavefunctionsepD and ¢A of the electron donor (D) and the electron eptor (A) with corrections because the two partners are close together interacting. For this purpose, ¢D and ¢A need not actually be known. In the case that donor and acceptor are even-electron systems Mulliken i:ea ‘1’N = worms.) + bd>1(D+-A-) ‘ no-bond dative the no-bond wavefunction ¢0(D,A) is an antisymmetrized product of the functions e and e D A to possible classical electrostatic forces and to quantum mechanical each first corrected for any polarization effects ctive (dispersion) and repulsive (exchange) forces together with changes and ¢A which result from changes in the internal geometry of D and A complexing. The function ¢1 is of Heitler-London type for the hypothetical D+-A— obtained by removing one electron from an M0 in D, putting it into a ously unoccupied MO of A, and possibly forming a weak chemical binding en the odd electrons now situated on the two components of the complex. iN function is normalized to unity which means that a2+b2+2ab301 = 1 where 301 =(¢ 0W1) . A they should belong to t If the ground stat to quantum theory print can be called a charge- ‘pv The coefficients 13* an cited-state wavefuncti pv is normalized so 8* quiremeut (Nd/v) = 0 and h*: b. Since for (a2>> b2) and so (bec. is mostly dative (a"‘2 essentially amounts t THE ENERGIES To obtain the e: comPleat is of course approximate expressi Since both “(N and av Pmblem is reduced 5 Where NOting that H10 = . 1‘0 ots Which are to 15 ere $01 =(¢o[¢]> . An important property of db and (#1 is of course that ey should belong to the same group theoretical symmetry SPeCies. If the ground state of the complex is described by ¢N, then according quantum theory principles there must be also an excited state ¢V Which n be called a charge-transfer (CT) state given by W V m coefficients b* and a* are determined by the requirement that the ex- = -b*¢0(D,A) + a*¢1 0 follows that a _— a, b __ b. If $01 0 then a ._ a 1d b*:: b. Since for loose complexes the ground state is mostly no -bond 12>) b2) and so (because of the orthogonality requirement) the excited state 5 mostly dative (a*2>> b*2). Excitation of an electron from wN to WV ssentially amounts to the transfer of an electron from D to A. THE ENERGIES OF THE NORMAL AND CHARGE-TRANSFER STATES To obtain the exact solution of the Schroedinger equation for the mmlex is of course not feasible. For this reason we have to use the >proximate expression for ¢N and WV and apply the variation principle. ane both ¢N and ¢V are two component linear variation functions the 'oblem is reduced simply to the solution of a two dimentional determi- .nta1 equation: H00 ' E H01 ' SOlE H10 ' S01E H11 ' E ere H00 = E0 = <¢OIHI¢0> H11 = E1 = (annual) H01 = E01 “‘(‘i’oml‘l’fl 10 = E10 =<"’1'H""o> ting that H10 = H01 the solution of this determinant for for E has two H :ts which are to be identified as EN and EV Here 8' audh 2E1 - to (A >0 u The lower energy is EV’ even ifA <0. A the appearance of a ne band. The frequency < hch ‘ Knowing the energies 1 cients Ma and b*/&* h 5 b/ a = and We b*/a* In this way the wavef 2 1f (t/Z>>>sosl, and : for weak complexes, ' them The Position of the 2 and 30 + s1 . Clan is the energy 0f ton Olmstion of the For a series 0 equation 19 Can be IIIlIIIIIIIIIIIIIII--==----=="""——“" 16 i E = {%(E0+E1)-301E0Li[(%A)2+BOBIJ2}/(l-8012) fire 60 5 E01 ‘ E0501 (“0‘ O) 3 a E (Bl< 0) 1 01 ' E1301 and A 5E1 - E0 (A >0 usually but sometimesA <0). The lower energy root is always called EN (normal state), the upper is Ev, even ifA <0. As we said the complex formation is associated with the appearance of a new absorption band the so called charge-transfer (CT) )and. The frequency of the CT band is 2 a 2 ho = EV - EN = 2[(A/2) + 8081] /(1 — 301 ) CT (mowing the energies EN and Ev we can now obtain the ratio of the coeffi- :ients b/a and b*/a* A s b/a = -(EO-EN)/(E01-SOIEN) = -(EOl-SOIEN)/(El-EN) ‘nd **5 b*/a* = (E01'301Ev)/(E0'Ev) = (El-EV)/(E01-SOlEV) in this way the wavefunctions can be written ~\"' _ 2 wN - (so + Ao1)/(1 + 2Aso1 + A ) — 'k 7': *2% w —(¢1—x¢O)/(1-2x301+h ) V f (A/2)§>BOBI’ and if 3012< is antisymmetric with respect to the exchange of any two of its lelectrons, whereas Vmod(A) is antisymmetric in its NFM electrons. Besides this "local" antisymmetrization the operatoréA makes ¢O(D,A) intisymmetric in all the N electrons of the complex. The above wavefunction :an be written in terms of the MD's of the donor and acceptor (modified by :he presence of each other, but only slightly in weak complexes) ¢O(D,A) =9?m(1)a(1)wd(2)e(2)wd'(3)a(3)...wa(M+1)a(M+1) wa(Ml-2)B(M+2)wa'(M+3)a(M+3)... ] Ere wa refers to one of the MO's on the donor, w ' to another, and so on, d dee wa and w ' refer to MD's on the acceptor. In ¢0(D,A) the donor MO a H which supplies the electron in donor-acceptor action contains two elec- :rons, while the acceptor MD into which the electron will go is unoccupied. + - . The dative function4>1(D -A ) can be written as: + ' I u % 4» (D -A ) = (‘P'+ V)/ [2(1+S )1 ‘1” =\1 % and MS = i 1 functions can be described as o . and wT,-1 =69”13+ “1E.- THEORETICAL DESCRIPTION OF THE EXCIMER STATE In this section we will outline the usual approach in describing the excimer state. We will consider the case of naphthalene excimer which is of immediate interest to us. The first to consider the case of naphthalene excimer theoretically was Th. Forster in 196234. He attributed the excimer binding to be due to exciton coupling. His general model for the excimer formation process in the case of aromatic hydrocarbons was as follows: (1) The lowest excited singlet state of the monomer should be a 1Lb state, which has a small transition moment to the ground state 1A. The relatively long lifetime of this state allows excimer formation to occur at reasonable monomer concentration. (2) The 1Lb state should be adjacent to a higher energy 1La state, which has a large transition moment to the ground state. The large splitting of the 1La state in contrast to the usually small splitting of the 1Lb state results in the lowering of one of the 1La exciton components below the lowest 1L component, figure 1b. It is the transition from this lowest energy L component to the ground state which Forster attributed the excimer a fluorescence. Shortly after it was realized that the exciton coupling can not account for the excimer binding energy. For the naphthalene excimer 2 3 in order to account for the binding energy usingAe =I M I/R , the physically ia*¢b ¢ Figure 18. Sche} M0161 arr-sh 7' 20 .l« 2 2(4)451*4)b + ¢a¢b*) ¢ a*¢b -L 2 2(¢a*¢b - ¢a¢b*> Figure 1a. Schematic Energy Diagram Showing Exciton Splitting in Molecular Dimers. 1 + [—— L I a 1 /’ La , 1L: 11:1) \ ”’ \ 1 - 1L” 8 1A 1A Monomer Dimer Sure 1b. Forster's34 Exciton Model of Excimer Formation in Aromatic Hydrocarbons . unacceptable R of 1.82 Slifkin35 conside in the excimer emissic hydrocarbons he found the excimer emission ' charge transfer state A=e1ectron affinity, and R is a term resul charge resonance stat In the case of x approximation38, C i: than the energy of at high to be accounted and AT and the grc The failure of exillein excimer beh; figuration interact adoPted by most sub I n a fixed" state pro and exciton-resonar Murrelu and i axeimer. 1“ the follow The M0 Pictm‘e of w . e cons lder Only t 21 unacceptable R of 1.8A must be used. Slifkin35 considered the possible implication of charge resonance states in the excimer emission. By plotting I-A-lLb vs hoD for a number of aromatic hydrocarbons he found a linear relationship which made him to suggest that the excimer emission is a charge-transfer emission. The energy of such a charge transfer state can be expressed as I-A—QiR (I=ionization potential, A=electron affinity, C=Coulombic interaction between the resulting charges and R is a term resulting from the interaction of the two possible degenerate :harge resonance states 1 In the case of naphthalene I-A=8.5036 or 8.7437. Using the point charge 3 ipproximation38, C is approximately 3 eV at 3 A. So I-A is over 2 eV higher :han the energy of excimer fluorescence 3.13 eV39 an energy which is too iigh to be accounted by the resonance energy between the structures AfA- ind A-A+ and the ground state repulsion. The failure of either the exciton or the charge-resonance model to :xplain excimer behavior in general led to the introduction of the con- iguration interaction model by Konijnenberg40. This model WhiCh has been dopted by most subsequent authors, considers the singlet excimer state as y'hixed" state produced by configuration interaction of charge-resonance nd exciton-resonance state. Murrel41 and McGlynn42 used this approach to study the naphthalene xcimer. In the following we will just outline how such a treatment goes. he MO picture of naphthalene molecule (D2h SymmECry) is as follows if e consider only the most important MO's Molecule A Let us consider the one charge resonance 1La monomer state an produced by t he excl to the lowest vacant 0f the two zeroth-m Molecular exciton 9 two wavefunctions where the bar denot can be expressed as Finally for the ch Linear co“Hilario: Irreducible repre: 22 Symmetry species to 6 3 3————— b2g w 6 2 2 lowest vacant MO blu 1 I I 91 I I highest occupied MO b3g w 9 a 0+.— 0-.-O—- U Molecule A Molecule B Let us consider the three lowest energy excited states of the excimer: one charge resonance state and two molecular exciton state, one due to La monomer state and one due to 1Lb monomer state. The 1La state is produced by'the excitation of an electron from the highest occupied orbital to the lowest vacant orbital m1 + ”2 while the 1Lb state is a resultant of the two zeroth-order degenerate excitation: wl + w3 and mo + w 2. Molecular exciton states of 1L parentage may be expressed by the following a two wavefunctions __ _, ___ WA =i wlmlelezl - lwlwlelezl = — — - ‘ 5 VB I “1wze1e1l I“’lmzel II where the bar denotes 3 spin. Molecular exciton states of Lb parentage can be expressed as __ __ ._l_ + _ _ ‘_ _ 9 I *0 = lwiw1eoezi ' l“’1‘*’1"o"2i iwlwlelesi 'wlwlel 3 WD = |91§lwoaiI ' |elern0w2| + le191%“?l ' |61°1w1w3i Finally for the charge resonance states WE = leaéeléll ' lwiezeleii VF = lewlwzel‘ ' lewlwzell Linear combinations of the above wavefunctions transform according to irreducible representations of the D2h group 1 : - Y La B38 VA B grin Charge resonance state From these results one transform in the same exciton states transf resonance states. To determine the separability approxim Hamiltonian of the s where 1/r is the e W Ecora= bacon“) Whl V electmn w . For the D211 ex vanishing configure and between B ‘ l 3g(e. the solution of Se “here f°r example “cum and Murre involved and esPe both congluded t] the Othel- exCime 23 1Lb 32g: IyC - YD B3u‘ Vc ' in Charge resonance states B3g: WE - WF BZu: WE + VF From these results one sees that for a D2h excimer the 1La exciton states transform in the same way as the charge-resonance state while the 1Lb exciton states transform differently from both 1La exciton and charge- resonance states. To determine the energy of the excimer state the validity of the a-n separability approximation is usually assumed. In this approximation the Hamiltonian of the system can be written as H = Hcore +vZu1/r where llrvu is the electrostatic repulsion between n-electrons v and u and Hcore= EHcore(V) where Hcore(v) represents the effect of the core on electron v . For the D2h excimer, from our previous considerations, there is non- vanishing configuration interaction between {B2u(exc)> and iB2u(CR)> and between |B3g(exc)> and |B3g(CR)> states. So the problem reduces to the solution of secular equations of the form H11 ' E H12 ' S12E = 0 - - E H12 SIZE H22 where for example H12 = <(exc) iHi (CR)> = CR > $12 <(exc)| ( ) MCGlynn and Murrel used different methods to evaluate the matrix elements involved and especially matrix elements of the type <(ech Hi(exc)> but both concluded that the lowest excimer state is of Bag symmetry and that - + the other excimer states in order of increasing energy are B2u’ B2u and Both calculation + big. semiempirical techniqu concepts do not give I A "supermolecule entity, molecule, wit] Moreover the ab initi appears technically u initio calculations r 5p/3s, 1p) and H(Ss/ or two-electron inte molecules" of von N structed from the w molecule maybe use Up to now only performed. In this naphthalene excimer tion interaction on 0 on again 3 A. 24 ng. Both calculations give an interplanar distance of about 3 A. These semiempirical techniques although useful in providing us with "physical" concepts do not give results of quantitative significance. A "supermolecule" approach where the whole excimer is treated as one entity, molecule, with the technique of molecular orbitals would be desirable. Moreover the ab initio approach is needed. At this moment this approach appears technically unrealistic. Buenker and Peyerimhoff43 performed ab @w initio calculations on naphthalene. Even with an inflexible basis C(lOs, 5p/3s, 1p) and H(Ss/ls) they had to compute 0.89 billion electron repulsion or two-electron integrals. Probably an approach like the "molecule in molecules" of von Niessen44; where the wavefunction of a molecule is con- structed from the wavefunctions of fragments structurally related to the molecule may be useful for constructing excimer states. Up to now only semiempirical supermolecule calculations have been performed. In this way Azumi45 used a simple Huckel treatment for the naphthalene excimer while Lim46 used Huckel orbitals and limited configura- tion interaction among several states. The calculated interplanar distance was again 3 A. The first intr . styrene in liquid s polystyrene shows tensities of the tw These findings led styrene occurs intr phenyl groups. Let the form (pheny1)-l molecular excimer 1 pane chain (n = 3) nnnu CHAPTER 2 STUDY OF THE INTRAMOLECULAR EXCIMER INTERACTIONS IN 1,3-DINAPHTHYLPROPANE AND 1,3-DIPHENYLPROPANE INTRODUCTION The first intramolecular excimer reported was the one formed by poly- styrene in liquid solutions47’48. Under these conditions the spectrum of polystyrene shows monomer and excimer bands except that the relative in- tensities of the two bands are independent of polystyrene concentration. These findings led the authors to conclude that excimer formation in poly- styrene occurs intramolecularly by the association of excited and unexcited 49 phenyl groups. Later Hirayama studied a series of diphenylalkanes of the form (phenyl)-(CH2)n-(phenyl) and showed that only when n = 3 intra- "n = 3 rule"). The pro- molecular excimer formation can take place (The pane chain (n = 3) allows for parallel sandwich excimer configuration and does not have the configurational instability of n = 4,5, and 6 alkane Chains. Since that time several cases of intramolecular excimer formation have been reported in aromatic polymers like polyvinylnaphthalene12 and Polyvinylcarbazole50 and simpler cempounds like (4,4)-paracyclophane12. Intramolecular excimer fluorescence has also been observed in various biological and model systems. The dinucleotides of cytosine (CpC) and thymine (TpT) and also polycytosine (poly C) all show excimer emission in solid solution at 85°K, when the pH of the solution is such that the bases are stacked, as determined by optical rotatary dispersionSI. Synthetic 25 models of the form (a used through their ex stacking interactions (calf thymus) is atru spectrum of the dinu adenine-thymine exci only the thymine d (stirrer) as the ini this case the import intersystem crossin that will be studio Intramolecular preferred conf igura i is a transient spec. not susceptible to his study of the pp tion with an intern i hit. Conclusion hitched to hit 26 models of the form (adenine)-(CH2)3-(N—isopentanyl-adenine) have been used through their excimer properties to obtain information about the stacking interactions in tRNASZ. Finally the fluorescence spectrum of DNA (calf thymus) is structureless and closely represents the fluorescence spectrum of the dinucleotide ApT and is accordingly identified51 with the adenine-thymine exciplex 1AT‘k. One of the most important photomutations namely the thymine dimerization is considered to involve selfassociation (excimer) as the initial step or involvement of the 1AT* exciplex53. In this case the important photophysical result is the enhancement of Sl__’T1 intersystem crossing because of the excimer/exciplex interaction a subject that will be studied further in Chapter 3. Intramolecular excimer have been used to get information about the preferred configuration of the excimer in fluid solutions. Since the excimer is a transient species existing only during the excited state lifetime is not susceptible to the usual methods of structural analysis. Ferguson54 in his study of the pyrene crystal fluorescence Suggested a sandwich configura- tion with an interplanar spacing less that the normal graphite distance of 3.5 3. Conclusions drawn from study of excimers in crystals cannot be generalized to the solution cases because of the constrains imposed by the lattice. Others like Birks55 have suggested that the excimer geometry may be one where one molecule is displaced considerably from the other, along one of the molecular axes, so as to minimize the repulsion between corre- SPOnding carbon atoms. Chandross and Dempster56 studied a series of 1,3- dinaphthylpropanes and showed that only the symmetrical aa— and @g-dinaph- thylpropanes which allow for perfect sandwich configuration of the naphthyl groups, show significant excimer formation. This appears as a strong support of the "perfect sandwich" excimer configuration in fluid media. The use of intr under various conditi case:(l) The geometry theoretical predicti radiative properties unimolecular process be used. The high c like to study the e Recently excimers f 57 , micelles and biol be promising in thi possibility is the expected to be obs quenching), (3) Pi intonation about i rotational relaxer in the shady of po In this chop! the and and? 27 The use of intramolecular excimers to study the properties of excimers under various conditions has several advantages over the intermolecular case:(l) The geometry of the excimer is known and this is important for theoretical predictions about the excimer state and its radiative and non- radiative properties, (2) Since the intramolecular excimer formation is a unimolecular process (independent of concentration) a dilute solution can i be used. The high concentration poses solubility problems if one would like to study the excimer behavior in viscous media or at lower temperatures. Recently excimers found an important use as probes of the microfluidity of micelles 7 and biological membranesSS. Intramolecular excimers appear to be promising in this area because of the concentration problem. Another possibility is the search for triplet intramoleCular excimers which are expected to be observable in viscous/low temperature media (collisional quenching), (3) Finally the study of intramolecular excimers will provide information about the effect of the connecting chain on the translational~ rotational relaxation of the interacting chromophores which is important in the study of polymers and biopolymers. In this chapter we are going to study various aspects of intramolecular excimer and exciplex behavior. The technique of time resolved spectroscopy is used initially to study the dynamics of the association and dissociation processes of the intramolecular excimer formed in the excited state of 1,3- BiSQI-naphthyl)propane (1,3DNP). The specific rate constants of these pro- cesses are determinedin media of different viscosity and the translational relaxation nature of the phenomenon is demonstrated. Through coupling of transient kinetic data with photostationary data and model calculations, various thermodynamic quantities describing the intramolecular excimer interaction are determined. The behavior of the intramolecular excimer in a wider viscosity with the aim of us for obtaining the netry of the exci diative deactivati molecular oxygen i and the results quenching action. The excimer dependent of othe another intranole that the monomer/ energies of the with the polarity viscosity and the nation energies ‘ Finally the possibility of t exist in viscous 28 a wider viscosity range is studied through steady illumination techniques with the aim of using 1,3DNP as viscosity probe of biomembranes. A method for obtaining the radiative lifetime is developed and the fact that the geo- metry of the excimer is known allows speculation on the mechanism of ra- diative deactivation of the excimer state. The relative efficiency by which molecular oxygen is quenching the monomer and excimer fluorescence is studied and the results are discussed in terms of the proposed mechanisms of oxygen's quenching action. The excimer interaction has been considered in the literature as in- dependent of other solvent properties besides the solvent viscosity. In another intramolecular excimer system 1,3-diphenylpropane (1,3DPP) we found that the monomer/excimer fluorescence intensity ratios and the activation energies of the association process are solvent dependent and correlate with the polarity of the solvent. The relative contribution of the solvent viscosity and the methylene chain rotational barriers to the observed acti- vation energies is also analyzed. Finally the system 1,3-diphenylpropane is used to investigate the possibility of triplet excimers. Indeed triplet excimers are shown to exist in viscous solutions and in microcrystalline aggregates. 1 , 3-DINAPHTHYLPROPANE Kinetic Treatment of the Excimer Formation and Dissociation Process In our kinetic treatment we consider the following reaction scheme: kDM M*+M ZN" kMD / kFM kIM kFD kID_ .z \, l \. M+th M M+M+th M+M Where M signifies the "monomer", in this case a single naphthalene ring The meaning of the rate constants is as t . . and D the intramolecular exc1mer. follows: Obviously Where TM and TD ar. respectively. For 1“ Us assume that 5'function at tin and at a SUbseque [the] and [Dr] res [iii] and [Dr] are This system of t and Die} can be from where A and B a: substitute equa 29 follows: kFM: fluorescence of monomer k : internal quenching of monomer kDM: excimer formation kFD: fluorescence of excimer kMD: dissociation of excimer kID: internal quenching of excimer Obviously km + km = kM = 1/rM kFD + kID = 1(1) = l/TD where TM and TD are the experimental lifetimes of the monomer and excimer respectively. For convenience we group the rate constants as follows: kFM+kIM+kDM=x kFD+kID+kMD=Y Let us assume that the solution is excited by a light pulse which is a 6-function at time t = O. The pulse produces [M*]O excited "monomers" and at a subsequent time t the concentration of monomer and excimer is [M*] and [D*] respectively. [D*]at t = 0 is zero. The rates of change of [M*] and [D*] are as follows: d[M*] /dt = -x[M*] + kMD[D*] d[D*] /dt = -Y[D*] + kDM[M*] (21) This system of two differential equations in two dependent variables {PM} ‘ and [D*J can be easily solved. let us assume a particular solution of the f0rm [M*] = Ae->\t and [D*] = Be->\t (22) where A and B are constants and x,is a parameter to be determined. If we substitute equation 22 into 21, we get: -)\Ae'-)‘t = -XAe-)t + kMDBe->‘t ->.Be'>‘t = -YBe-)t + kmAe')‘t 0r A(X—)) - kMDB = 0 II C B (er-)0 — kDMA (23) The system of equat: determinant is zerc From the above detr )‘1, The particular sol Satisfy the initie is necessary to he combination of pa: The system of eqt the secular de ten and From equatiODS l [1 At t = 0) [Dir] 0f M’ [NO is i SOlVing equati. Then the Conce Now let Us f1, 30 The system of equationsZ3has a nontrivial solution for A and B if the determinant is zero X -A -k = 0 (24) -kDM Y -A From the above determinant, we get "1,2 = %[x + Y Him-x)? + 4kDMkMD ] (25) The particular solution is not an acceptable solution because it does not satisfy the initial conditions of concentration, at t = 0, [D*lfi£0. It is necessary to have a general solution which may be written as a linear combination of particular solutions -X t -> t _ l 2 [m] clAle + czAze (26a) * >‘1t+ B xzt [D ]= ClBle c2 2e (26b) The system of equations 23 splits to two systems one for each root of the secular determinant 24 A1(X-)1) - BlkMD = 0 (27a) B1(Y-).1) - AlkDM = 0 (27b) and A2(X-x2) - szMD = 0 (28a) 32(Y-A2) - AszM = 0 (28b) From equations 26a, 27a and 28a: * -Xfi + dzt/ [M] = c 1km) B 1e t/(x- -->.1 ) cszDBz e (x-xz) (29) At t = 0, [Dr] = 0 so c131 = ~c2B2 or c1'= -c2'. The initial concentration of me, [W10 is then = ' x- - 'k X- [M‘k]0 c1 kMD/( )1) c1 MD/( )2) (30) Solving equation 30 for cl' we get cl' = -[Mklo/ know-h) (31) Then the concentration of M* at any time t will be: t [Mk1= cl'kl‘fmee)1":/(X-)1 ) — ci'k MDe )2 /(x-)2) (32) NOW let us find an expression for [D*]. From equations 27b and 28b, we get that [on] from aquatic it] = by using the exp ind Equations 32 and time t. The qua molecule of M) ”Sins equations Defining the q and USing equa The solutiOn 4 are torn Ways 1. me Our 6 is not in 8ft Because give“ by the Where Ilt) j 31 get that A1 = (Y -‘)l)Bl/kDM and A2 = (Y—A2)Bz/kDM ._ ' _ _ ‘ - So [Mflo — c1 (Y AlmDM c1 (Y kzkaM or cl' = kDM[M*]0/(}2 - )1) (33) [D*] from equation 26b is given by _ 'Mt "Mt [D*] — cl'e - cl'e (34) by using the expression for cl' from equation 34, we get = “ht _ ‘kzt - [13*] kDM[w]0(e e )/(,\2 )1) (35) Equations 32 and 33 give the concentration of monomer and excimer at any time t. The quantum intensity of monomer fluorescence (per initial excited molecule of M) at time t is IM(t) = kFM[M*]/[M*]0 (36) Using equations 30 and 32 we can write equation 36 as: 1M“) = kFM(x2-X)[e':1t + \2. Equation 40 is a transendental equation and was solved with a computer iterative technique. Knowing and we can use equation 25 to generate the followin ’\1 /\2 g system of two sin kD in equations of monomer in t1 from decay expet experimental co similar to that equations 42 gi where and Results The measx rate constant: Using th manic quantit quantitiea of State leading tiOn energy : Pane in etha EDM = 4.0 kc to ethanol 5 which can b£ Which by us 33 system of two simultaneous equations + =1: +k +k +k A1 'A2 M D PD - = k k + k k A1A2 DM D M MD kD in equations 42 is essentially)1 and kM which refers to the decay DM + kMkD (42) of monomer in the absence of the excimer formation process, was obtained from decay experiments on a-methyl naphthalene (model) under the same experimental conditions. kM of a-methyl naphthalene is assumed to be similar to that of kM of dinaphthylpropane. Solution of the system as equations 42 gives kDM = A ' kMD (43) km = (AkD - BWku - km) (44) where A=>‘l+)‘2-kM—kD mm B=A1*2-H$D Results The measured values of A1,’A2, kM and kD along with the calculated rate constants kDM and kMD are collected in Table 1. Using the kinetic data thus obtained we can calculate some thermody- namic quantities for the excimer formation reaction and thermodynamic quantities of the assumed equilibrium between reactants and the transition state leading to the excimer. In our treatment we make use of the activa- tion energy for the excimer formation process of 1,3-Bis-(a-naphthyl)pro- Pane in ethanol which has been found by Chandross and Dempster to be EDM = 4.0 kcal/mole56. The thermodynamic quantities to be computed refer to ethanol solution at 25°C. The excimer formation is an activated process which can be represented generally by an Arrhenius form of equation kDM = ADMexp(-EDM/RT) (45) which by using the determined values of kDM and EDM gives us a value of [ .f.. L HUZWUWHm—oadl rh-quhz— Fl o 0 3 Figure 2 34 INTENSITY FLUORESCENCE ”:V/ g: 1% \, ~——-u—.g~7ar: f: .‘n ‘ r 300 II \ X \ X2 I \ I \ ( \ ( \ l \ I ‘ I‘ I \ I \ I \ . \\ I ‘ \ l \ l l \ ’ \ I \ l \ l \ I \ l \ I \ ’ \ I \ ’ \ Q \ I \ I \ I \ , \ I \ \ \ \ \ \ \ x \ \ \\ fi I ‘ I I 400 500 gen m) . -4 Figure 2. Room Temperature Fluorescence Spectra of 1,3DNP (1.2x10 M) in EtOH. (__) Air Equilibrated, (--) Nitrogenated Solution. Figure 3 35 7 ‘.A'J ‘ - 4:- fun-.3 ;__ei,..ttt. ." Ni ’\ .! F1Sure 3 . The Gauche-Gauche Conformation of 1,3-dinaphthy1propane ' I cowusHom Hoamaum AHV .oocoomouosHh Hoafioxm HaasowH 36 E22. .ousuxHE Houwozfiwufiosmaum HHH Amv ofimnucH msu mo >moon tam omwm osa mCO— .e ouswfim 0:! AllSNEIlNl Table l. K M _—_——- Solution 1. ‘ Solution I \ Solution 37 Table 1. Kinetic Data for the 1,3DPP Excimer in Ethanol-Glycerol Mixtures at 25°C. Solution I. Ethanol Solution 10-4M at 25°C, Viscosity = 1.2 cp. 7 _ x1 = 3.10 x 10 s 1 _ 7 -l _ 8 '1 A2 — 12.96 x 10 S kDM - 1.2 x 10 s 7 — 4 -l kM = 1.12 x 10 s 1 kMD = 1.6 x 10 S -l 7 kD = 3.10 x 10 s Solution II. Mixture 25% Glycerol, 75% Ethanol at 25°C, Viscosity = 3.6 cp. A1 = 3.17 x 107 s'1 - 7 _ 9\2=8.27x107s1 kDM=6.4xlOSl 7 ..1 3 -1 kM = 1.89 x 10 s kMD = 3.0 x 10 s 7 -1 k = 3.17 x 10 s D Solution III. 1:1 Mixture of Ethanol-Glycerol at 25°C, Viscosity = 20 cp. )‘1 = 2.68 x 107 5'1 _ 7 -1 9‘2 = 6.32 x 107 s 1 kDM = 4.9 x 10 s 7 -1 3 -1 kM = 1.41 x 10 s kMD = 2.9 x 10 s 7 -l x 10 s H N e 0" 00 k D 38 ADM = 9.8 x 1010 s-1 for the preexponential factor. A more informative description of the process is through the transition state theory in the form that Eyring, Evans and Polanyi put it = 0 ‘ u _ '4: kDM kT/h exp(AS:M/R) exp( AHDM/RT) (46) In equation 46 we have assumed that the transmission coefficient is one and we must also realize that this equation strickly applies to gas phase. A more detailed discussion of that point will be given later. AS1 and AH# are the entropy and enthalpy change for the process of going from the reactants to the transition state. An expression connecting the experimentally obtained activation energy to the enthalpy of activation can be easily obtained as follows. Equation 46 is valid for both constant- volume and constant temperature systems. In our case we are dealing with a constant pressure system. Differentiating the logarithm of equation 46 at constant pressure we get ;# 2 = + T (47) (aha/611)}, (AHDM RT)/R Also from the Arrhenium equation 2 = T (48) (alnk/BT)P EDM/R BY Comparing equations 47 and 48, we see that = ‘1: + (49) EDM AH13M RT Using equation 49 and EDM = 4.0 kcal/mole, we obtain Aug; = 3.5 kcal/mole with this result and kDM’ using equation 46 we obtain - -1 A83; = -10 cal mole 1 deg Let us now turn to the thermodynamics of the process M* + M;:2D*. The equilibrium constant of that reaction is = (50) Ke kDM/kMD and from thermodynamics the Gibbs free-energy change will be 39 ZXGDM = —RTane (51) equation 51 gives in our case AGDM = -5.3 kcal/mole. It would be very informative if we could determine separatly thenAH and AS of the reaction. Unfortunately we do not have a value for the binding energy of the excimer which is equal tolsH. In the following we will try to estimate this binding energy. Figure 5 is a qualitative diagram showing the excimer energy as a function of the distance between the rings. From this diagram we see that ~0, 0 ~max _ = _ + hc(vM VD ) AH ERep (52) ~0 _ . . . where V ’0 is the frequency in cm 1 of the (0,0) tran51tion in the monomer, ~maX . VD is the frequency of the maximum of the exc1mer fluorescence and ERep is the energy of the ground state configuration generated by the vertical (Franck-Condom) emission process of the excimer. . -163 ax . 33’0 for 1-methy1 naphthalene in ethanol 13 31,500 cm . F: 1S from our data 23,800 cm-l. ER is more difficult ot get. First we have 6P to know the configuration of the excimer. The intramolecular excimer con- figuration is a stackede~D configuration where the two naphthalene rings 2h 0 H a H are forced by the steric requirements of the methylene chain to a sandw1ch O . 56 . . . _ configuration with an interplanar distance of 3 A Wthh 13 also in agree ment with all the theoretical calculations on the naphthalene excimer. The best way to calculate E is by semiempirical methods using parameters Rep that fit experimental data64. In this study ERep is estimated by using an interatomic potential (referred to as "6-exp") 0f the form —6 (53) = _ ,, - ..r.- Vij Aijrij + BlJexP( C11 13) and finally obtaining the intermolecular potential by summing the various interatomic potentials (54) = V.. Vmolecular ij lJ 4O Energy—a ERep. M+M Figure 5. Qualitative Potential Energy Diagram for the Dimer in Ground and Excited States. 41 The values for the parameters A, B and C of the "6-exp" potential used here are those given by Rae and Mason65 which are appropriate for aromatic hydrocarbons °6 AH---H = -36 A kcal/mole AC-o-C = -535 A6 kcal/mole ACWH = -139 Z6 kcal/mole a. BHMH = 5010 kcal/mole it chc = 74460 kcal/mole : BC.”H = 9411 kcal/mole in CHMH = 3.74 2'1 Came = 3.60 11‘1 cc”.H = 3.67 3‘1 A computer program was written to calculate the intermolecular potential 0 as a function of the interplanar distance the result for r — 3 A is ERep = 0.15 eV. For intermolecular excimers of aromatic hydrocarbons, the 66 . . usually assumed value for ERep is between 0.2-0.3 eV . It 18 of interest 67 . . . to note that a similar method by Kitaygorodsky after the modification 0f Mantione for stacking configurations gives an even less repu131ve po- tential EReP(3A) = 0.05 eV. This method was used extensively for predicting configurations68 and the importance of van der Waals forces in the binding energy of "charge—transfer" complexes by Mantione69. This method appears to understimate the repulsive potential and so we will use the reSults of the ”6-exp" potential. So using the above mentioned values for 171960 and vrfiiax and ERep = 0.15 eV we obtain from equation 52 AH = Binding Energy = -18 kcal/mole NOW we can use this estimated value to try to analyse further the excimer formatiom and dissociation process. We understand of course that the 42 quantities thus to be obtained have the uncertainty of both the originally obtained data and of the above estimate of the excimer binding energy. The numbers are then only indicative. The entropy change AS will be AS = -oac - AH)/T = ~21 cal mole-1 deg-1 Let us now consider the dissociation process. If the binding energy is -18 kcal/mole then die activation energy for the dissociation process will be .- EMD = -AH + EDM (55) y because the difference of the activation energies of the dissociation and association process equals the binding energy. In this way EMD = 22 kcal/mole Proceeding in the same way as for the association process we estimate AMD = 3.8 x 1020 s‘1 AHMD = 21.4 kcal/mole ASMD = +32 cal mole~l deg—l A summary of the kinetic and thermodynamic results is given in Tables 2 and 3. Table 2. Rate Constants for 1,3DNP Excimer Formation 3nd Dissociation Processes in Ethanol-Glycerol Mixtures at 25 C. ...1 '1 o Médium Viscosity (cp) kDM(s ) kMD(S ) TemP-( C) 4 . Ethanol 1.2 1.2 x 108 1.6 x 10 25 7 3 EthanOl-Glycerol (3:1) 3.6 6.4 x 10 3.0 x 10 25 7 3 Ethanol-Glycerol (1:1) 20 4.9 x 10 2.9 x 10 25 43 Table 3. Thermodynamic Data for the 1,3DNP Excimer Interaction. 3.5 kcal/mole 4: “in :t: _ -1 -1 ASDM — ~10 cal mole deg 2&1th = +21 kcal/mole ASMD = +32 cal mole.1 deg”1 AHDM = -18 kcal/mole -1 -1 ASDM — -21 cal mole deg Discussion Before discusing the data let us first understand the meaning of the equations used. The result of the classical theory of reaction rates is written in the form k = Vexp(~E*/RT) (56) where V is a frequency factor which is in principle calculable from spec- troscopic frequencies of the normal molecule and which should lie in the 12 14 -1 . . . . . range 10 to 10 s . E* is the experimental activatlon energy. Adapting the idea that a moleCule must have an energy E* or higher to react and that energy is localized to a particular set of normal coordinates essential for . . 70 71 _ the reaction, Rice, Ramsperger and Kassel developed a quantum mechanical theory for unimolecular reactions. The difference between the classical and quantum mechanical rates was shown to be in the introduction of the entropy difference between the normal and activated molecule. So Rate quantum mechanically/Rate classically = exp0331/R) (57) The final form of the equation for the specific rate constant of a unimolecular process is k = v*.exp(As*/R )-exp(-AHl:/RT) (58) In equation 58, y* represents a weighted average of the internal frequencies of a species that has the configuration of the transition state. One would expect it to lie in the same range as V for the normal molecule but such an equivalence is purely ad hoc. So althoughnAihtcan be obtained experimentally there is no obvious way of partitioning the experimental frequency factor between v* andASii The transition state theory introduces a universal value of)’= kT/h = 6 x 1012 s-1 for t = 300°K or in terms of wavenumbers‘~200 cm-l . 13 - . . . . 13.~10 s 1 then the reaction 1S con31dered as normal and in this case , a quite low frequency. If the preexponential factor AS*2!0. Another way to understand the deviations from normality (classical behavior) is to look to another expression for the ratio of quantum me- chanical and classical rates72 e Rq.m. /Rclas.: QVib/Q'Vlb (59) 3 Equation 59 assumes that (I:I§I:71aIbIC) = l, where Ii is a principal moment of inertia. This assumption is not necessarilly true in our case. Qvib (the vibrational partition function) can be generally factored as a product of contributions from each normal coordinate Qvib = [3‘11 = [i] (1 _ e-hVi/kT)-l (60) It is only when the frequencyyi é kT/h that qi differs appreciably from unity (for yi>> kT/h, qi = 1), so that the qi and hence Qvib will be unity except for the contributions of low-frequency vibrations. If some of the frequencies of M (reactant) are lowered in changing to M=§ then qj‘lqij> 1 for those particular frequencies and we should expect an enhancement of the rate. If on the other hand some originally low frequencies of M are increased, then we may expect to find qizlclqi ( 1 and consequent falling of 45 the rate. According to Benson the greatest contributions may be expected from the very low frequencies arising from almost free internal rotations. Going back to our data we see that the excimer formation process has -1 ¢_ 1 and a negative AS - —10 cal mole- 1 a preexponential factor of ~1011 3 deg' . These results indicate a significant ordering during the transition from reactant to the transition state. The loss of entropy and a low value of A can be attributed to loss of rotational degrees of freedom of both the naphthalene rings and of the methylene chain. Similar results have been found in cases where the transition state involves some kind of a cyclic intermediate. For example the isomerization of vinyl allyl ether to n- pentaldehyde-ene-4 is considered as going through a six-membered ring complex: Ii: H H H I / CH = ,0 ”(ix /CH—C - H2 ”.0 ___) C32 2 \\ CH-;CH2 CH=CH 2 The observed preexponential factor is'le s-1 andZ>S* = -8ca1 mole- deg _173 The excimer formation process although clearly a unimolecular process has also characteristics of a diffusional process and so is dependent on solvent properties particularly solvent viscosity. Also the observed ac~ tivation energy can have contributions from the rotational barriers of the methylene chain and the diffusion process which is an activated process (solvent viscosity temperature dependence). These points will be considered later. Turning to the estimated binding energy (0.78 eV) we see that it is considerably high compared with the already reported values for binding energies of intermolecular excimers. For example the binding energy of 74 l-methyl-naphthalene excimer in n-heptane was reported by Birks to be 0.30 eV and by Selinger75 in 95% EtOH solution as 0.20 eV. Pyrene in 46 76 ethanol has one of the highest binding energies 0.40 eV . Although our binding energy estimate depends on the estimate of ERe , we should note P . ~00 "ma: . ‘1 . a _ = + that in our case the value of}/M VD AHhM ERep 1S 7700 cm which is higher than the value of 6500 cm_ in the case of l-methyl-naphtnalene . . ~1 . . intermolecular exc1mer and 6100 cm in the case of pyrene exc1mer; an indication of a higher binding energy for our case. The effect of the . ;-Ef7 5.x in methylene chain strain is expected to be the same in both excimer and A ,3“: v’r-J'f '47:" 5 t ‘ _ Franck-Condom ground states. Even if we use a higher value for E ep, say, 0.3 eV, the resulting binding energy is still higher (0.48 eV) than usual. The excimer binding energies are uSually obtained by substructing the activation energy of the association process obtained at lower temperatures from the activation energy of the dissociation process obtained at elevated temperatures. This is done under the assumption that kFD is independent of temperature which especially at elevated temperatures we do not believe it is true (vibronically induced emission). Cundall and Robinson77 has shown for the case of benzene excimer that kFD is indeed a function of temperature and by correcting for this dependence the binding energy was changed from 0.22 to 0.37 eV. Other reported binding energies in the literature seem to be quite low like the 9-methy1anthracene in benzene 78 1 which implies a ERepgi 0.6 eV an unusually high 0.20 ev,Av: 6850 cm“ value. An unusually high binding energy value for 1,3—dinaphthylpropane may be associated with the steric requirements of the methylene chain which may bring the aromatic chromophores in a more favorable excimer conformation. It is interesting to note here that when pyrene was studied in viscous paraffin oil solution79 the binding energy was found to be higher than in fluid solution (0.48 eV). In our case the methylene chain is expected to Play the role of a highly viscous medium locking the molecule in the excimer Configuration. 47 The high negative entropy of reactiontis = -21 cal mole'1 deg—1 seems to be characteristic of excimer formation. Compare for example the value of -21 cal mole.1 deg-1 for 2-methy1 naphthalene in ether, -19 cal mole--l deg‘1 for acenaphthene in tolueneSO,Z§S = —22.7 cal mole"1 deg'1 for pyrene . 81 . . . . in toluene . This entropy decrease is again a result of free21ng of various degrees of freedom which is expected for the association of two complex molecules and appears in other cases of association reactions like the di- 82 merization of acetic acid (hydrogen bonding) in benzeneAAS = —22.3 eu Finally the results for the dissociation process D*——+M* + M of the intramolecular excimer seem to be different than in the case of inter- molecular excimers. The reported values of the dissociation rate constant _ 6 -1 in the case of intermolecular excimers are in the orderl~10 s for example 6 _183 pyrene in 957° EtOH km = 7.0 x 10 s '1 to be compared with 1.5 x 104 s in our case of intramolecular excimer. The estimated preexponential factor 1020 s"1 is very high. A collection of A's for the dissociation process of excimers by Lewis and Ware84 shows a good linear correlation between logA and END (a kind of enthalpy-entropy compensation). Using his graph we obtain for our activation energy of 22 kcal/mole a preexponential factor of 20 - . . . 10 s 1 which seems to Support the consrstency of our kinetic data. Again Viscosity seem to be important for example for pyrene in cyclohexane E 16 -1 . . = 0.5 A, ' ' E = 0. re 5 eV and AMD 10 5 while in paraffin 011 MD 76 eV and AMD 1018 s-; It is obviOus in our case that the methylene chain prohibits the excimer dissociation process much more effectively than a viscous solution. -1 -1 An anomalously high entropy of activation ASS; = +28 cal mole deg for 18 '1 s ) was the intermolecular benzene excimer in fluid solution (A 256 x 10 reported by Hirayama85. The authors believe that this high value is due to a partial immobilization of the aromatic rings as well as a longer range 48 ordering of solvent molecules accompanying the excimer formation process. i In our case the high value of A or equivalentlyASMD may be due to both the methylene chain and the effect of the solvent. The Radiative Lifetime of the Intramolecular Excimer The lifetime of the excimer fluorescence has been a subject of dispute for a long time.Hoijtink86 showed that if the excimer has a center of sym- metry, excimer luminescence is forbidden. On the other hand the mere fact that we observe excimer fluorescence proves that the emission is not com- pletely forbidden or it is induced by some intra- or inter-molecular me- chanism. The observed fluorescence decay time of pyrene excimer in fluid solutions is much shorter (V65 ns) than that of the monomer (v380 ns). This led to some speculation whether the steric configuration of the excimers may not be the symmetrical overlapping one87’88 89. Unlike the case of Pyrene Mataga et al.90 showed that the fluorescence decay time of naphthalene andCI-methylnaphthalene intermolecular excimers are almost the same as the fluorescence decay times of the monomer. Mataga argues that because the pro- bability of the radiationless transition from the fluorescent state of the excimer may not be smaller than that of the monomer, due to the smaller energy gap betwaen the fluorescent and phosphorescent states as well as between the fluorescent and the ground state of the excimer, the radiative transition probability of the excimer fluorescence may not be larger than that of the monomer. A possible experimental support of enhanced intersystem crossing from the singlet excimer state has been discussed by Cundall et al.91 Since the fluorescence of naphthalene is due to an approximately forbidden transition Metaga concluded that the fluorescence transition of the naph- thlene excimer is also approximately forbidden. Another interesting experiment by Mataga was on the anthracene excimer. Mataga photodecomposed 7 4V“T% .\,_’ 1‘ ‘1: “8:“ . 9 49 the anthracene dimer in a cyclohexane matrix at 77°K and subsequently excited by a nanosecond light pulse. The excimer fluorescence decay time was 215 us to be compared with a ~29 ns decay time of the monomer. On the other hand the fluorescence decay of the anthracene dimer produced by the controlled cooling and softening of the matrix92 was onlyrv6 ns. Further the perylene excimer produced in the cyclohexane matrix at 77°K shows excimer fluorescence with a decay time of 77 ns which is much longer than the decay time of the monomer~5 ns. The above experimentstend to show some degree of forbidness for the excimer emission but the results are complicated by the unknown factor of non-radiative transitions. An obviously more useful quantity will be the radiative lifetime. A few radiative lifetimes have already been determined. They are of the order of 102-103 ns except for pyrene, for example’V900 ns for l-methyl naphthalene in 95% ethanol75,~«770 ns for 1,6-dimethylnaph- thalene in 95% Et0H75,cJ550 ns toluene in hexane93 and~l77 ms for pyrene in 95% Et0H83. 0n the basis of these results Birks supported an overlap picture of the pyrene excimer where the long axes are parallel but the short axes are displaced88 and also proposed that kFD is temperature independent93 There are several difficulties with the above results and interpretations. For example Mataga and coworkers94 obtained a lifetime (observed) of 200 ns for pyrene excimer in cyclohexane matrix at 77°K while Birks reported a smaller radiative lifetime of~v86 us in the same solvent at toom tempera- 95 . 77 . _ ture . Recently Cundall and Robinson have found that the exc1mer radia- tive decay process of benzene is temperature dependent with an activation 3 - energy of 0.145 eV and a preexponential factor of 2.4 x 10 s 1. Hirayama 85 and Lipsky also suggest that the rate constant for the radiative process 50 for benzene excimer may be temperature dependent. In the following we will try to determine the radiative lifetime of the intramolecular naphthlene excimer (1,3-Bis-(a-naphtyl)-propane). The intramolecular excimer has a more or less definite steric configuration because of the restrictions put by the methylene chain. As it has been shown by Chandross this configuration must be a perfect "sandwich" con— figuration with an interplanar distance of 3 A56. Because we feel quite confident about the excimer configuration the result may be helpful in deciding on the mechanism by which the excimer emission becomes allowed. In order to calculate the radiative lifetime we make use of the results of our transient kinetics experiments which we couple with data from pho- tostationary reaction kinetics. For steady excitation with light of intensity I0 einsteinsefi1 s-1 the rate equations are d[1M*]/dt = 10 - (kM + kDM) [1M*] + kMD [IN] (61) d[1D*]/dt = kDM{_1M*1 - (kD + km) [1%] (62) l , 1 Under photostationary conditions d[ Mxl/dt = d[ D*]/dt = O. From equation 62, we get 1 1 kDM [ Mk1 = (RD + kMD) [ Bic] (63) or [113*] / [11%] = kDM/(kMD + kD) 1 1 By definition chM = kFM [ 14*] /IO, CDFD = kFD[ 0*] /IO (64) so from equations 63 and 64, 1 1 'c = (DFD/GJFM = kE‘D[D~k]/(k1‘"1"1f Md) kFDkDM/(kFM(kMD + 1.13)) (65) °r k‘FD = ‘pFDkFM(kMD + kD)/¢E‘MkDM (66) In equation 66, kFM was obtained from quantum yield and lifetime measurements ofcx-methyl-naphthalene fluorescence through the relation 7F = TF°¢ The F' ratio of the quantum yeilds was obtained as the ratio of the integrated emission intensities of excimer and monomer after corrections for instrumental 51 response. Finally the lifetimes were expressed in a medium of unit static 2 2 . dielectric constant through the expression kFD(n =1)=kFD/n where n is the refractive index of the solvent. The radiative lifetimes thus obtained are ° = 00 ° = 0 . TFM 3 ns and TFD 72 ns As we see the radiative lifetime of the excimer is longer than diet of the monomer. Mataga96 argues that although a-methylnaphthalene does not have a center of symmetry the perturbation of the methyl group is not significant and so there is a center of symmetry for the electronic density of the fl'system and consequently the excimer emission should be forbidden. If we accept the fixed sandwich configuration for the intramolecular excimer then a possible mechanism by which the excimer emission becomes alIOWed is the mechanism proposed by Chandra and Lim46. According to their treat- ment the emission is induced by thermal excitation of tortional oscillations of the excimer components. During such tortional vibrations the excimer geOmetry is distorted and emission can occur. Due to their low frequencies several tortional levels may be populated at ambient temperatures. Effect of Solvent Viscosity on the Intramolecular Excimer Formation Process From our kinetic data we see that, as expected, both the rates of formation and dissociation of the intramolecular excimer depend on solvent viscosity. As has been pointed out by Melhuish and Metcalf97 although both rates depend on viscosity their ratio, the equilibrium constant of the re- action may be independent of the viscosity. In our case the equilibrium constant is shown to be 1.7 (i 0.3) x 104, approximately independent of viscosity. Besides the transient kinetic measurements we studied the in- tramolecular excimer system in ethanol-glycerol mixtures of variable vis- cosity. The results expressed on IM/ID (IM = maximum intensity of monomer, 52 ID = maximum intensity of excimer) vs macroscopic solvent viscosity are displaced in Figure 7. It appears that for higher viscosities (above 20 cp) there is a linear dependence of IM/ID on n while a more complicated dependence is found at low viscosities. Using the results of the Einstein-Smoluchowski diffusion theory98’99 the rate constant of a diffusion controlled reaction is given by k = 4nN x 10'3DpR where D is the sum of the diffusion coeffients of the two simultaneously moving particles, R is the sum of the interaction radii and p(é 1) is a factor introduced to describe the reaction probability per collision. The diffusion coefficient is usually approximated through the Einstein- Stokes equation D = kT/6nrn where n is the microscopic viscosity which is assumed to be the same as the macroscopic viscosity. The intramolecular case is more complicated but if one makes the reasonable assumption that the excimer formation rate depends linearly on the diffusion coefficient then using the equation 66 derived under photostationary conditions we get that IM/ID = C(T) kFMi‘D/kFD where C(T) is constant for constant temperature.’ In this way a linear dependence of IM/ID is expected. Not excluding some trivial factor the anomalous behavior at low viscosities may be due 00 . . . . 1 to the implicit assumption that microv1sc031ty =I kn amoucmonmm och .mwunuxwz HOH00haonaocmnum aw mznm H mo mwnmm monoommuonfim uwfiwoxm was “cacao: can mo hufimaoucH m>HumHom onu co muflmoomw> mo uoomwm one .o shaman AS 5 IHUZm4m><>> AllSNBlNl SONS-138380015 .mznmJ no steam munoomwuosam “938nm was umfiocoz can we zuwmnmfinH o>fiumem 93 do muwmoomw> mo uummwm 9.5. .n 95me 33:603.; on. o: 2: on 8 on 00 on on on on S o _ u - u 4 q . _ — A _ _ N 54 V: 6) PJ /wa 55 The Effect of Oxygen on the Excimer-Monomer Relative Yields During our study of the intramolecular naphthalene excimer we noticed large variations in the excimer-monomer relative yields depending on the extent of degassing of the solution. The excimer intensity being more sensitive to oxygen quenching. In the following we will try to investigate the behavior of the excimer-monomer system towards oxygen quenching more quantitatively using data we have already acquired from the transient kinetic measurements. The rate of the monomer formation in the presence of quencher (Q22 02) 1 1 '1 ‘ d [ M*]/dt = 10 — (kM + kDM + kQM[Q]) [w] + km[ 13*] (67) Under stationary conditions d[1M*]/dt = O = (k + k + k [0])[1w4l - k [10*] (68) M DM QM MD From our data we see that the repopulation of the monomer through excimer dissociation is negligible so we can neglet the second term on the right hand side of equation 68 then, ]_ . [Mic] = 10/(kM + kDM + kQM[Qj) (69) If there is no quencher in the solution then, 1 _ 70 [Mir] — IO/(kM + kDM) ( ) By definition (PM = kFM[1M*]/IO (71) so (bib/(112M:(kMfl‘DMH‘QMM)/(kn+knn) = 1+(k0 flip/(kml’knn) (72) since (1M ) l= kM + kDM (73) lifetime of monomer emission which takes into account the additional path (bin/(BM = 1 + 'MkalQ] (74) Equation‘74 is the familiar Stern-Volmer relation. The rate of excimer in of excimer formation the presence of quencher is — l 56 Under stationary conditions d[1D*]/dt = 0, so 1 l - [0*] = kDM[ M*]/(1/(d>;M/¢FM) = 1 + kQD[Q]/(kD + km) = 1+ TDkQDm (84) From equation 84 solving for k D we get, kQD =((¢§1§DFM)/(¢F§DF°M) — 1)/TD[Q) (85) In the same way from equation 72, kQM =(d)FM¢FM - 1)/(TD[Q]) (86) So the ratio of the quenching rate constant will be kQD/th; [(4) FQFM/ qDivbbrn) — 1] TM/«bm’mm ‘ 1) TD (87) Using our experimental data we get, kQD/kQM = 0.48 So our kinetic analysis indicates that the monomer fluorescence is more susceptible to oxygen quenching and the experimentally observed pre- ferential quenching of excimer fluorescence is due to the fact that the excimer lifetime is longer than that of the monomer*. In this way the *The excimer lifetime is 32 ns while the monomer lifetime including the excimer forming step is about 8.5 ns. 57 excimer fluorescence is quenched to larger extent due to collisional quen- ching. If we want to understand the difference in the quenching rates of the monomer and excimer we should consider the possible mechanisms of such a quenching. (a) Enhanced Intersystem Crossing During the collision of the excited naphthalene monomer with the ground state oxygen there is coupling of the spin momenta to give a tri- plet collision complex 3[}M* + 32;(02)] so the spin prohibition of the intersystem crossing process is removed62. The process can be described as follows: - 3 3 _ 1M*+32(0)—>M*+ 2(0) Initially both naphthalene and 02 are in the zero vibrational state. Using the Robinson-Frosch103 theory of non-radiative transitions the matrix element for the interaction between the initial and final state can be 3 = Hif N<>ifl> 02 where H,f is the matrix element for the exchange interaction between the 1 written as initial and final states of naphthalene and the other two factors are the appropriate vibrational overlaps. The electronic state of oxygen remains the same before and after the reaction so 0 =ll for oxygen because the two vibrations belong to the same orthonormal set of functions and con- sequently for O#=u 02 = 0. The vibrational overlap is known to be very sensitive to the energy gap between the states. When the vibrational quantum numbers are very different the overlap becomes extremely small. (b2 Energy Transfer to Oxygen 104 This process can be represented as follows .? in", < [.5 A" ’7." in! 58 114* + 3{(02) -—-——)3M* + 14 (02) g s where during the quenching singlet oxygen is produced. In ortance of the Two Mechanisms for the uenchin of the Monomer Emission When the total electronic energies of the initial and final states are nearly equal, very little electronic energy is converted into nuclear in. vibrational energy during a radiationless transition between the two , .mp1: . _. states. If the total electronic energy of the final state is significantly less than that of the initial state, the difference has to be made up by «n,— ’ vibrational excitation in the final state. Because the Franck-Condom factor for oxygen becomes exceedingly Small if the initial and final vibrational states of the oxygen molecule differ by more than a few quanta105 we can, for practical purposes, assume that most of the vibrational excitation energy goes into the large naphthalene molecule. 1 For mechanism (a) the energy gap between the naphthalene B3u state and the 3B2U state is E = 10700 cm-l. For mechanism (b), since the ex- citation energy for the process 3E;(02)—~—9}Ag(02) is'J7880 cm-1 we are left with only 2800 cm—1 to be distributed as vibrational energy. Since the Fraan-Condon factors are known to dominate the nonradiative transition rate, process (b) should be favored. _1106, 107 1 In the case of excimer, the D* state is atrv25400 cm . The excimer singlet state can intersystem cross to the naphthalene triplet ‘ -l excimer state which has been placed atrV18600 cm with an energy gap 1108,109 AE 216800 cm' . On the other hand, energy transfer is energetically impossible. Although process (a) is more favorable in the excimer case, the elimination of process (b) which is the most efficient one seems to be the cause for the slower quenching rate of excimer fluorescence. 59 1 , 3-DIPI—IENYLPROPANE Solvent Effects on Excimer Luminescence In our kinetic treatment of the excimer formation process we neglected other properties of the solvent except its viscosity which obviously affects the motion of the aromatic rings. In the course of our study of intramolecular excimer formation we found that the ratio of monomer to excimer fluorescence intensity in 1,3- diphenylpropane is solvent dependent. Table 4 shows this ratio expressed as the ratio of the intensities at the corresponding maxima. Table 4. Effect of Solvent on the Emission Properties of 1,3DPP Solvent IM/ID Agfiu{(nm) #fileDebyes) n(2(sodium D) Methanol 0.56 329 1.70 1.316 Ethanol 0.56 329 1.69 1.359 n-Propanol 0.56 329 1.68 1.383 n-Butanol 0.56 329 1.66 1.397 Methylene Chloride 1.46 327 1.60 1.444 p-Dioxane 0.71 326 h’O 1.420 Mathyl Cyclohexane 0.77 325 N0 1.421 Hexane 0.88 325 n40 1.372 (1) Gas phase dipole moments in Debyesllo. 111 (2) Refractive index for sodium D line From Table 4, we see a correlation between the observed IM/ID ratios and the polarity of the medium as expressed by the dipole moment of the solvent 60 molecules. Associated is a shift of the broad maximum of the excimer emission. The above data refer to non-degassed solutions. of oxygen was realized certain results were rechecked after exhaustive freeze-pump-thaw cycles. The results are as follows: Solvent IM/ID Ethanol 0.33 Cyclohexane 0.43 Methylene Chloride 0.86 So although the actural numbers change, the relative ratios for different solvents do not change significantly and the effect is qualitatively the same. The excimer formation process in 1,3-diphenylpropane is a kinetically controlled process and the observed IM/ID ratios signify changes in the rate of the excimer formation process as a function of solvent or some special quenching phenomena. We have previously derived the equation = + kDM ¢FDkFM(kD kMD)/¢FMkFD wh I I ere ¢FD/¢FM a: D/ M To test the assumption that the observed different IM/ID signify different association rates in different solvents, we studied that excimer formation Process as a function of temperature in ethanol and methyl cyclohexane trying to determine the corresponding activation energies. If the tempera- ture dependence of IM/ID is detennined by the temperature dependence of kDM then the activation energy d the association reaction is obtained by the following equation112 61n(ID/IM)/o(1/T) = -EDM/R Plots of ID/IM against l/T are shown in Figure 8. The plots are quite linear and the activatiOn energies are as following: After the effect / 03 T d Figure 8. Determination of the Activation Energy for Intramolecular Excimer Formation in 1,3-Diphenylpropane. Circles: Methylcyclohexane, Triangles: Ethanol. Ethanol EDM = 3.3 kcal/mole Methyl Cyclohexane EDM = 4.3 kcal/mole The activation energy in ethanol appears lower as one may have expected from the static ID/IM measurements. This behavior may be explained if the activated complex is polar. Polar characteristics for the benzene excimer have been suggested by Hirayama and Lipsky85 Chandross and Dempster56 found an EDM for l,3—Bis-61-naphthyl)-pro- pane (1,3DNP) in ethanol of 4.0 kcal/mole and Itoh et al.112 an activation energy EDM = 2.0 for/3, a'(9,lO-dicyanoanthracene)-(CH2)3-(naphthalene) (,3, a'-DCAN) and an E = 2.2 kcal/mole for B, fi'-DCAN in ethanol. Our DM value of 3.3 kcal/mole lies between these values. It can be argued that the transition state for DCAN is more polar as expected for an exciplex than that of an excimer and thus the lower activation energy. So 1,3DPP excimer appears more polar than the 1,3DNP excimer. We studied 1,3DNP in ethanol and hydrocarbon and we did not notice any enhancement in the polar solvent. We would like to know if the assignment of polar character to the 1,3DPP excimer could account for the observed activation energy difference. For this reason we are going to do a very rough calculation using a theory first developed by Kirkwoodllg. This theory gives the energy of an idealized dipole (M) in a continuous dielectric of constant 6 compared to that in a similar medium of unit dielectric constant. no = $733 . - 1)/(2e + 1)] (88) solV . . . . . , 114 Applying this equation to the tranSition state theory one obtains _ 2 3 2 3 ink/k0 — -(e - 1)/(26 + 1) (AR/rR - Mi/rngT (89) where k is the rate constant in a medium of dielectric constant 6, k0 is the rate constant in a medium of E = l, and “R andiqkare the dipole moments 63 of the reactants and activated complex respectively. The change of entropy appearing in the transition state theory expression for the rate constant can be decomposed as = + S AS ASintra A solv (90) where Asintra refers to the intramolecular entropy changes during the association process and ASSO1V refers to the effect of solvation. From classical thermodynamics we know that AG = AH - TAS and GAG/8T)? = -AS So ASsolv = '36/(26 + 1)2-(51ne/5T)P(#R2/rR3 - Hi/rtB) (91) The connection between the Arrhenius equation preexponential factor A and the preexponential factor of the transition state theory is A = e3D* (105) 3 1 Reaction 105 requires the mutual diffusion of M* and M. On the other 70 hand it is a well known fact that triplet states are very effectively quenched in fluid solutions and phosphorescence can only be observed in crystals or frozen glassy solutions at low temperature. In this way by freezing the solution to prevent collisional quenching we prevent the triplet excimer formation. we should note here that if the 3D* formation requires a minimal configurational change this may be achieved even at a highly viscous solution because of the long lifetime of the triplet state. The temperature was considered above as acting only through changes . 130 in the viscosity of the medium. Hoytink et al. suggest that there is an intrinsic (apart from diffusion) activation energy for triplet excimer formation. This activation energys.400 cm-1 although small will be im- 0 131 partant at 77 K. Different interpretations have been discussed by Baldwin and Siebrandl32. II. Generation of triplet excimer through intersystem crossing from the singlet excimer state. 1 1 1 iSC 3 M* + M ’ D* "“‘”‘_’ D* (106) Indirect evidence for the existence of this process comes the study of the temperature dependence of the singlet excimer lifetime. This dependence in most cases has been reproduced by the inclusion of a single temperature- dependent rate parameter kID in the form kID = AID~exp(~EID/RT) The computed frequency factors AID are some orders of magnitude lower than 14 -1 . 12 . Characteristic values in the range 10 ~10 s for unimolecular reactions and the process has been described as a spin-intercombination process or 15 kID = kISC D . Mechanism II depends on the formation of a singlet excimer formation process is again diffusion controlled and in fact is expected to be much more sensitive to the fluidity of the medium because of the short 71 lifetime of the singlet state. Another complication may be that the intersystem crossing may be followed by dissociation 1 isc 3 3 Dir-,—-—-> D*———-> M* + M (107) This behavior has been reported for pyrene in ethanol at 20°C by Melinger and Wilkinson334 who measured the instantaneous population of the molecular triplet state by flash kinetic spectrophotometry. Although these authors treat the over-all process as one of dissociative intersystem crossing le'r —-———> 314* + M (108) their results are consistent with an excimer intersystem crossing yield 15 6 _ of 0.12 and a rate constant of 2.3 x 10 s 1 . If process 107 takes place in a viscous environment, recombination of the products can give 3D*. Finally it has been suggested that the geometrical requirements for singlet and triplet excimers are different and probably in a rigid matrix the necessary spatial rearrangement is not possible to take placel33. It is almost obvious, but after overlooked, that in order to observe triplet excimer formation the viscosity of the medium should be high enough such that triplet quenching is minimized and low enough such that diffusion of the excited triplet chromophore to form an excimer dimer is probable during the lifetime of the excited state. A way to enhance the possibility of triplet excimer formation is the use of a "double molecule" like 1,3DPP. Toluene has quite a long triplet lifetime (in rigid medianub sec.). The COmbination of long lifetime and the enhanced collision probability are ideal. Since the optimum viscosity is unknown one should scan a large range of viscosities (temperatures). 1 A degassed lO_3M solution of 1,3DPP in 3-methylpentane (3MP) at 77°K exhibits a toluene type emission with fluorescence maximum at 285 nm and 72 phosphorescence maximum at 390 nm with typical vibrational structure. It should be noted here that the viscosity of 3MP at 77°K (~1012 cp) is high enough to prevent diffusional relaxation before emission. The lifetime of this phosphorescence is 6.5 sec. similar to that of toluene. The relative intensities of fluorescence and phospha:escence are also similar. The emission observed from a 10'3M degassed solution of 1,3DPP in isopentane(iP) at 77°K (~106 cp) varied somewhat depending on the rate of formation of the glass. Usually a monomer fluorescence MF and a mono- mer phosphorescence MP are observed as in 3MP, however the phosphorescence spectrum is relatively diffuse compared t) thatobservedin 3MP at 77°K. Figure 9 shows both spectra together with the emission spectrum re- corded after the sample was allowed to warm up slightly; the estimated temperature is about 115°K. The phosphorescence spectrum in the latter case is shifted to longer wavelengths with a maximum at 420 nm. When the sample was aIIOWed to further warm up and the viscosity of the solution decreased the intensity of phosphorescence decreased rapidly until it was completely quenched and only monomer fluorescence is emitted. At Such viscosities, diffusional relaxation leading to singlet excimer formation during the lifetime of the excited monomer singlet state does not occur. Excimer fluorescence at 330 nm (DF) began to appear and dominated the spectra in the neighborhood of room temperature at appreciably lower viscosities. The energy gap between the maximum of DP and the 0,0 of M iSrVSOOO cm-l. The 420 nm emission was interpreted as excimer phosphorescence. The me- chanism of the excimer formation in this case (singlet excimer is not pre- sent) is obviously the direct association of a triplet excited phenyl ring with a singlet ground state one: 3 3 mfi + 1M ———-———9 D* 73 > M l: U) z UJ F— E UJ o z UJ o (D UJ g E D 1M:! .J 200 300 400 500 A(nm) Figure 9. Total Emission Spectra of 1,3-dipheny1propane; (l) 3MP Class at 77° K, (2) i? Glass at 77° K, and (3) After the 1? Glass Was Allowed to Melt (Estimated Temperature 115° K). 74 Since the emission is observed during the softening of the glass matrix and disappears upon warming, photoproducts could not be responsible for it. This maximum correlates with electron-beam induced emissions from alkyl benzene liquids. It is also consistent with Castro and Hochstrasser's reported emissions from halobenzene crystals. These were disputed by Lim and Chakrabarti as photoproducts however the maximum of the emission reported by Lim is far too low in energy ( 8000 cm"1 from the 0,0 of MP) to correspond to triplet excimers. Singlet and Triplet Intramolecular Excimers of 1,3-Diphenylpropane at 77°K Concentrated solutiOns of 1,3DPP in rigid glass or iP at 77°K under conditions where microcrystals are formed exhibit an excimer fluorescence at 340 nm and excimer phosphorescence at 425 nm, Figure 10. Upon slight warming the phosphorescence band maximum shifts to 420 nm similar to that observed in fluid media. Phosphorescence decay measurements at different wavelengths show that there are two components a long-lived one at short wavelengths and a short-lived component at longer wavelength. This appears to be a unique case where both singlet and triplet excimers are observed. In the following we will try to analyze the mechanism of excimer production in a disordered solid, like the microcrystalline solid we used. I. Singlet excimer: There are may kinds of defects that can be realized in a disordered Crystals, such as vacancies which are regions of lower density than the bulk material, or overcrowded regions of higher density. The initially Produced singlet exciton wave will propagate till it meets such a defect. The loss of translational symmetry and the loss of the strict phase relations of the excited regular lattice result to the immobilization of the exciton. Grain boundaries may also have the same effect. if" 1:57 <5.”— 7‘ «I v _, mam .-'_ 3 ,,.finx,\.. . . . . mlllflhtthhwr. Iv , llqllll .Suemfim .5 Eng 3 egofle mus waeaam 75 Houw< ANV .Monm um meumzuoouofiz aHu .mw :H mmmm.H mo mCOHuSHom woumuucooaoo mo cowmmfifim .oH osamwh 2E; 000 com oov com com _ _ _ a _ _ _ _ AllSNELLNl HONHOSEINIWH‘I 76 An important type of defect will be an excimer site that is two molecules forced in close proximity and held by the strong crystal packing forces. There are obviously two ways that the singlet excimer can be formed: (1) Free propagation of the exciton till it meets the appropriately oriented molecular pair then the energy decreases adiabatically and the exciton is trapped with the formation of a singlet excimer. (2) Trapping of the exciton at another defect site with subsequent longer range energy transfer to the excimer site. This mechanism which is con- sidered more effective for impurity trapping at low temperatures by Hochstrasser134 can only operate in our case if the "excimer pair" is strongly interacting in the ground state and the transition energy is appreciably reduced. We believe that trapping of singlet excitons of either the Frenkel or the Forster-Dexter type at the "excimer site" is the reason for observing only excimer emission in our microcrystalline system. We also believe that an analogous situation appears in various polymers like polystyrene where the emission appears almost completely excimeric. II. Triplet exciton: Although the probability of intersystem crossing in crystals of aromatic hydrocarbons is lower than in the free molecule (solution) case still a significant amount of the absorbed quanta will ultimately reside in the lowest triplet state135. Analog0usly to the singlet exciton the triplet excitation can propa- gate in the crystal in the form of a triplet exciton. This was shown by studies of Avakian and Merrifield136 and by the dynamic experiments of 13 . . Ern et al. 7. The motion of the triplet eXCiton can be best characterized 77 by a relatively slow random-walk, hopping mechanism138. In this way trapping of the triplet exciton may give rise to triplet excimers. Another possibility of course is direct intersystem crossing from the singlet excimer state. In this case even if the triplet excimer state is -dissociative at the ISC energy recombination may be expected to take place. CHAPTER 3 INTRAMOLECULAR EXCIPLEX-CHARGE TRANSFER INTERACTIONS IN THE SYSTEMS (PHENYL)-(CH2)n-(PYRIDINE) AND AROMATIC CARBOXYLIC ACIDS INTRODUCTION In this chapter we are studying intramolecular exciplex interactions. Initially we study the systems 4-(3-phenylpropyl)-pyridine {3F(CHQ)3«@£N (1,3PyPP) and 4-benzylpyridine ©4112-pr (PyCHZP). These systems depending . on the solvent properties display a variety of interactions. In neutral alcohol solution, the absorption spectra of 1,3PyPP do not show signs of ground state interactions while the room temperature fluorescence spectrum shows intramolecular excimer formation completely analogous to the 1,3DPP 'case. As the hydrogen bonding ability or acidity of the solvent increases (H20, EtOH + HC104, H20 + H0104) strong hydrogen bonding or protonation of the pyridinic nitrogen changes the interaction drastically. Analysis of the absorption spectra of 1,3PyPP and PyCH2P shows significant ground state interactions while the broad emissions show extremely large Stokes shifts of~d4,000-15,000 cm-l. Since upon protonation of the pyridinic nitrogen the electron affinity of the pyridine segment is expected to increase drastically (see also INDO calculations) the interaction is now of the charge-transfer type. This fact is also verified by the out of plane Polarization of the fluorescence. Thermodynamic arguments lead us to believe that in certain cases complete electron transfer is possible. It is interesting to note that while Py+CH2P shows strong CT-exciplex 78 79 interaction, no excimers are observed in diphenylmethane, illustrating the difference in geometric requirements of excimer and CTeexciplex interactions. Upon changing the excitation wavelength, shifts are observed in the emission. The emission bands arise from molecules associated in the ground state (charge transfer interaction) or from molecules associated in the excited state (exciplex interaction). This fact is also proved by study of the excitation spectra of these emissions. Although one would expect that the relaxed exciplex and charge-transfer states to be the same, the rotational barriers of the connecting chain appear to inhibit this rela- xation resulting to a composite emission. Finally the emission spectra at low temperature show the importance of solvent relaxation in these systems. The phosphorescence of 1,3PyPP and PyCHzP under all conditions is arising from the lowest locally excited toluene triplet state suggesting that the 3CT state is dissociative. It is interesting that the emission of the relaxed singlet state appears at lower energy than the triplet state emission. In the even-odd system .(CH2)3-@ unlike HN@>— (CH2)3-@ the two structures where the positive charge is on either ring, are degenerate and charge-resonance is expected. This phenomenon is manifested in the differ- ence of fragmentation patterns of butylbenzene and 1,3DPP in the gas phase inside the mass spectrometer. Finally exciplex interactions may not be manifested by the appearance of new emission bands but may show up in a more subtle way. Aromatic car~ boxylic acids of the general form <:>P(CH2)n-COOH show enhanced intersystem crossing and changes hafluorescence/phosphorescence intensity ratios and lifetimes compared to the compound lacking the carboxyl group. This ' 77" 777757 »- al-”. ‘ 1.. 8O phenomenon has been interpreted as an enhancement of spin-orbit coupling as a result of exciplex interactionszoa. In order to test further this point, namely, that the effect is due to the intramolecular exciplex interaction of carboxyl group and the aromatic ring, compounds that have the carboxyl group rigidly held and do not allow for intramolecular re- laxation are studied. The observed fluorescence and phosphorescence quantum yields and lifetimes support the above mechanism. I. (PHENYL)~(CH2)n-(PYRIDINE) SYSTEM Electronic States and Spectral Properties of Pyridine and Pygidinium Ion The spectroscopic properties of pyridine have been under investigation for many years. The sharp absorption, in the vapor phase, that starts at Jc o 139 2800 A was assigned by Kasha to an n,W' transition. This transition 7' 0 overlaps the lowest Wyflc absorption which starts around 2600 A . The presence of many progressions built on a strOng 0,0 band in die case of * . . . . the n,fi transition, indicate a dipole allowed tran51tion 31' The low intensity of this band was attributed by Orgel141 to the small overlap of the n and n-orbitals. Finally the rotational fine structure of the 0,0 142 140 band is consistent with a B1 state . Sponer and Rush place the l l —1 _ 143 l 1 B2( Lb) state at 38,350 cm and finally El-Sayed places the A1( La) at 49, 750 em'l. An interesting feature of the pyridine absorption spectra is the , dramatic changes induced by the solvent (hydrocarbon vs hydrogen bonding protonic solvents)l44. The n,n* band disappears (blue shift in hydrogen bonding solvents is a characteristic of n,r* states) and thevrr* absorption intensifies and apparently sharpenslas. This subject will be discussed in more detail later. In an acidic environment the lone pair on the nitrogen is protonated 81 to give the pyridinium ion. E. Clementi146 showed that the lone pair on the nitrogen is partly delocalized on the neighboring carbon atom but re- tains 1.41 electrons in the sp2 hybrid on the nitrogen nucleus. The energy of the 1A1 state of pyridinium was placed at 5.5 eV and that of 1B2 at 4.8eV by Brown and Heffernan147 In Figure 11, we show the relativepositions of the lowest states of benzene, pyridine and pyridinium. The energies of the benzene states are those given by D. S. McClure148 Considering the emission properties, we find that pyridine is a non- fluorescing compound. Cohen and Goodman149 suggest that the fluorescence quantum yield is less than 10'5. 0n the other hand the intersystem crossing 150 yield for pyridine is about 0.3 . The most probable deactivation process of the lowest singlet seems to be internal conversion to the ground state. Pyridine besides being non-fluorescent is also non-phosphorescent. Singlet- triplet absorption studies by Evans151 using high pressure oxygen as a perturbation, indicate that the observed triplet state of pyridine is n-«* in nature. This state should correlate with the lowest triplet state of benzene (a molecule that exhibits strong phosphorescence). Pyridium ions, like pyridine, are not emittinngz. This is true for cases where there is no charge-transfer interaction between the pyridinium cation and the associated anion (like in the case of perchlorates). has shown that the absorption spectrum of l-methyl-pyridinium iodide shows an additional shoulder at long wavelength which is interpreted as a charge- transfer absorption of a complex between the pyridinium cation (acceptor) and the iodide ion (donor). Briegleb et al.15 found an emission from the above compound which they ascribed tothe reverse charge-transfer process. 153 Kosower rffiEE? AW’v-j; y, fink j . 82 AE(eV) Benzene Pyfidine Pyfidhfium 7a 6-« __._'B ' o- _ 'A, __ 'A ._ 'A Figure 11. Symmetries and Energies of the Lowest Singlet States of. Benzene, Pyridine and Pyridinium Ion. 83 The above discussion on the spectral properties of pyridine apply to its methyl derivatives (picolines) and especially to the 4-methyl- pyridine (V-picoline) which represents more accurately our system. Absorption Spectra (A) Absorption Spectra of y-Picoline_(§;M§thyl-Pyridine) As we see from Figures 12 and 13 the lBl(nr*)(—-1Al transition which appears as a long wavelength tail in the hydrocarbon solution is blue shifted and disappears under the 1B2(w#*)4——lA transition in water. An assumed 1 hydrogen bond strength of'V6 kcal/mole might be expected to correspond to 14 a shift of around 2000 em'l . 155 Hochstrasser has measured the dipole moment of pyridine in the 1Bl(n7r*) state to be -1.0D to be compared with the ground state dipole 157 156. Simonetta considers the di- moment of +2.19D (opposite directiOn) pole-dipole interactions to be important in blue-shifting the 1B1(n1r*) ’state. Let us now consider the behavior of the rr* states as a function of solvent polarity and nitrogen protonation. Table 5 shows this behavior. The 132(rw*) is formally allowed in C2V symmetry but still is quite weak which shows that the replacement of a C-H group (toluene) by a N (y-picoline) is a small perturbation. Therefore we expect that the inten- sity of a such a transition will be sensitive to an environmental pertur- bation. The intensity increases on going from hydrocarbon to water to acid and the extent of perturbation increases in the same way. The intensity in- crease, is also accompanied by a small blue shift. At this point we do not know if this shift is real or it is artificially caused by the blue shift of the overlaping 1B1(mr*). An important thing to note is that no new bands or long wavelength tails appear as a result of nitrogen protonation. 6 .. 4 _. Flo x U 2 _ O I l l I I I 200 220 240 260 280 300 Mnm) Figure 12. Absorption Spectrum of ‘y-Picoline in Methylcyclohexane. 85 -3 6X10 I 220 l 240 300 A(nm) Figure 13. Absorption Spectrum of y-Picoline in Water (__), Absorption Spectrum of Y-Picolinium Perchlorate in Water (--). 86 Table 5. Effect of the Medium on the Frequency and Intensity of the ’Y-Picoline Absorption Spectrum. 1 1 B 4—— A1 Solvent 1‘; (cm-1) 6 max max MCH 39,039 1705 H20 39, 139 1844 H20 + H0104 39,573 3898 1A-+~— lA l l MCH 49,652 6815 H20 49,383 4759 H20 + H0104 45,872 4481 87 This spectroscopic behavior has also been predicted theoretically by Mataga and Mataga158 using a semiempirical SCF MO technique involving only the r electrons, and later by Janet Del Bene and H. H. Jaffe159 by an all electron CNDO method. Del Bene and Jaffe ascribe the above effects to the polarization of the r system induced by the hydrogen bond or protonation. In pyridine, the charge density ontherfitrogen atom is 5.24 electrons, of which 1.16 belong to the r system. In the pyridinium ion, the total charge density on nitrogen decreases to 4.94 electrons, even though the w electron density increases to 1.46. The tendency of the system to relieve the posi- tive charge on the nitrogen atom that results from the addition of a proton gives rise to a highly polarized r cloud. Another interesting feature of the spectra is the apparent sharpening on going from hydrocarbon to hydrogen bonding solvents. In the nr* state an electron is removed from the lone pair so the hydrogen bond is expected to dissociate or to greatly weaken160. If the hydrogen bonded nw* state is dissociative for the N...H bond then the absorption to the 1B1(n7r*) will be continuous with no discrete vibronic bands. Since 132(wt*) and 1B1(n1r‘k) are overlapping and in protic media 1B2 is increased in intensity then one may expect a "sharper" structure. The behavior of the 1A1(ww*).—_}A1 band is the opposite of the 1B2 (WT*)*‘~lA1 band. Increasing the polarity and acidity of the medium we observe a considerable red shift and a decrease in intensity. This be- havior has been also observed by Zanker161. Another molecule that we should study before examining the spectra of our systems for ground state interactions is of course toluene. Table 6 shows some numbers. The 0,0 transition of the 1Lb state of toluene was placed by Ginsburg 162 .,:rx- 88 Table 6. Effect of the Medium on the Frequency and Intensity of the Toluene Absorption Spectrum. Solvent 5 (Cm-1) 6 max max 11b MCH 38,100 260 EtOH 38,197 290 1 La MCH 48,100 7900 EtOH 47,985 8400 at 37,477 cm“1 while the 0,0 for the 1Lb state of v-picoline was placed by Stephenson145 at around 38,500 cm-l. From the above data we see that toluene's absorption extends to the red of the 7-picoline absorption and that in the case of the 2500 A band of the systems that we are interested in (phenyl)(CH2)n{pyridine), toluene's contribution to the extinction coefficient is small while in the case of the 2000 A band toluene's absorption is most important. Finally let us consider the absorption spectra of 1,3-diphenylpropane as a model system where there are no specific ground state interactions between the two aromatic rings (Table 7). Table 7. Effect of the Medium on the Frequency and Intensity of the 1,3DPP Absorption Spectrum. Solvent Fmax(cm-l) emax 1Lb EtOH 38,197 473 1L EtOH 48,309* 17,000 a *The vibronic component with the largest Franck-Condon factor in the spectrum of toluene appears at 47,619 cm” in the case of 1,3DPP. 1:5? . v 89 Toluene in Ethanol 9L- . I Las—A 260 - IL ‘_IA 6 _ b ‘0 280— ’30 x P >< U U 200 — 3.. 220 - 240- ‘ h—T—-—-—.T__ 200 220 I I I l I 220 240 260 280 300 A(nm) Figure 14. Absorption Spectra of Toluene in Ethanol. 90 l,3-diphenyl propane in Ethanol 400 _ 18}— 'La‘_ IA l4 ~ 300 — 'o x T 10 — ... s 200 _ 3 6 _ l00 — 2 _ l l 200 220 I I I I 220 240 260 280 300 A(nm) Figure 15. Absorption Spectra of 1,3-Diphenylpropane in Ethanol. 91 The first absorption band in 1,3DPP is not shifted compared to toluene but shows a hypochromic effect emax(l,3DPP) = 473 to be compared 25 (to- luene) = 580. More important is the reduction in intensity of the 0,0 band, I /Io 0 = 1.19 in 1,3DPP and I / = 1.08 in toluene. Since max ’ max I0,0 the intensity of the 0,0 band is a measure of the diviation of the electronic symmetry from D6h(benzene) this may mean that the introduction of the second phenyl group counterbalances to some extent the effect of the side carbon 0' chain. The second absorption band of 1,3DPP shows a better resolved vi- 33 1,: 4’ bronic structure, a slightly altered vibronic structure and a small red shift (~36O cm'l). One would expect some differences between 1,3DPP and toluene in the 1B2 state. If we consider the polarization of the 1Lb(lB1) state we find that it is X polarized while the lLa(1B2) is Y polarized (see cooridinate system) so any difference between the model (toluene) and the actual system (1,3DPP) will be accentuated in the 1B2 state. Also the dispersive interactions will be stronger in the more polarizable 1B2 (exci- ton interaction will show the same behavior). The sharpening may be due to the relative rigidity of the 1,3DPP rotamers (steric interactions, selec- tion of the TT configuration) compared to the free rotating group of toluene. Y , jfilgAbsorption Spectra of (Phenyl)-(CHq)n-(Pyridine), n = 1,3. Table 8 shows the effect of solvent and protonation on the spectra of 4'(3-phenylpropyl)-pyridine (1,3PyPP). The first band of neutral 1,3PyPP appears normal a superposition of the absorptions of toluene and y-picoline. There is a slight hypochromicity in MCH which seems to be an inductive effect and a slight hyperchromicity in H20 which may be due to mixing of local states with charge-transfer states. 92 Table 8. Effect of the Medium on the Frequency and Intensity of 1,3PyPP and 1,3Py+PPc10; Absorption Spectra. 1St absorption band 2nd absorption band Compound Solvent 3fiax(cm_1) Emax 95ax(cm-1) e 1,3PyPP MCH 39,139 1760 48,660 15,325 EtOH 39,060 1860 48,640 13,229 H20 39,139 2300 48,540 15,180 1,3Py+PP EtOH 39,535 5180 48,160 10,400 H20 39,620 6020 48,146 12,145 The second band appears again normal, the band is composite of tran- Sitions of 7-picoline and toluene in this region. The most dramatic changes appear in the spectrum of 1,3Py+PP. Looking at the spectrum of 1,3PyPP in water and water + HC104 we see then in the acid medium the abSorption has a weak but obvious red tail. Also the hyperchromicity is considerable (emax expected aroung 4200). The second absorption band appears anomalously blue shifted and hypochromic. The spectra of benzylpyridine show the same behavior: Table 9. Effect of the Medium on the Frequency and Intensity of the PyCHZP and Py+CH2PC1O4' Absorption Spectra. Compound Solvent $Eax(cm'1) ‘max Py-CHz-P EtOH 39,000 2290 H20 38,986 2580 . Py+-CH2-P EtOH 39,308 5835 H20 39,370 5940 93 1.1 12 — 7 o x u; 8 - 4 - 0 1 r l I 200 220 240 260 280 Anm Figure 16. Absorption SpeCtrum °f 4-(3-Phenyl-propy1)-Pyridine in MethyICyclohexane. 6x10'3 94 220 Figure 17. Absorption Spectra of 4-(3-Pheny1-Propy1)-Pyridine (__) and the Corresponding Pyridinium Perchlorate (--) in water. 95 16 — IEXIo'3 I 1 I 200 220 240 Figure 18. F fi 260 300 A (nni) Absorption Spectra of Benzylpyridine (__) and Benzylpyridinium Perchlorate (~-) in Water. 96 The sec0nd band of benzylpyridine was not studied because it extented further to the blue beyond our instrumental capacity. The red tail in the case of benzylpyridinium is even more pronounced. We assign this red shifted band to an intramolecular charge-transfer tran- sition frOm the phenyl ring to the pyridinium cation. This intramolecular charge-transfer transition appears stronger in the case of Py+-CH2-P than in 1,3Py+PP, at 280 nm the apparent extinction coefficients are as follow: 1,3Py+PP in H20 6: 144 EtOH e= 96 + _ Py CHZP in H20 e— 560 EtOH e= 468 This difference in intensity may be due to a stronger CT interaction in benzylpyridinium and/or a higher association (conformational restric- tions). Also the interaction appears to be stronger in the more polar water. The blue shift of the second band may be due to perturbation by the CT state. A literature search showed that similar behavior have been reported in related systems. Kosower163 noticed a long wavelength shoulder (a809OA) in the absorption spectrum of 1-(4-methoxabenzyl)-4-carbomethoxypyridinium CH3O©-CH2-N@-COOCH3 Later,S. Shifrin164 studied the absorption spectra of N-(B—p-x-phenyl- perchlorate ethyl)-3-carbamoylpyridinium chlorides with x various electron donating + groups (NHZCO4ESN-CH2-CH2-<:)-x) and observed red shifted shoulders which 16 a8cribed to CT transitions. Finally de Boer 5 studied a series of compounds of the general formula x»©-(CH2)n-I\.Y (n = 1,2,3) and found CT bands with maxima between 280-290 nm. 97 The appearance of the charge-transfer transition in acid media only, we believe, is a manifestation of the increased electron affinity of the pyridine ring upon protonation. To test this idea further we performed some all-electron semiempirical SCF calculations using the INDO (inter- mediate neglect of differential overlap)method of Pople166. The descrip- tion and analysis of the technique can be found in J. A. Pople and D. L. Beveridge, ‘hpproximate Molecular Orbital Theory", MeGraw-Hill, 1970. We used the same nuclear geometry for pyridine as that used by Clementi in his ab—initio SCF calculation167. The N---H distance in the case of hydro- gen bonded pyridine-water was assumed to be 1.75 A following Hbffmann168. Finally the N+-H distance was found by energy minimization to be 1.08 A. Figure shows the net charges and the energies of the lowest unoccupied orbital 6b2(wh) in various situations. As we see from Figure 19, the first antibonding orbital of pyridine 6b2("4) has a positive energy which in good agreement with the value cal- culated by Clementi. Hydrogen bonding lowers somehow the orbital energy by’V0.4 eV (9 kcal/mole). Finally protonation has a dramatic effect giving an orbital energy of -5.15 eV. Of course this result is probably quite high but it is indicative of the large change of orbital energies upon protonation. Another interesting result is that the positive charge in the pyridinium ion is spread to the ring on the ortho and para positions while the nitrogen still remains slightly negative. Hydrogen bonding appears to delocalize about 4% of a positive charge to the ring, this is 14 in good agreement with experimental results from N nuclear magnetic re- sonance experiments where the positive charge was estimated to be 7 i 3% Of an electronic chargel69. 98 [6 (6b,)=+ 3.47.31] 01147 -0025 ol‘ 4 _________ - ' 7 __——__—_.__ [6129:3229] Etw.)=+s.oeev1 0055 0055 O -0.023 O ‘0.023 0J33 0J32 bl N~ l‘O.l76 {'0.l76 l I 'Il 0131 T O\H O 0.002 O\H 0.158 P4 I0028 [f(6b,)=-5.15efl Figure 19. INDO Calculations. Energies of the Lowest Unoccupied M0 Pyridine, Pyridine-Water Complex and Charge Densities. (Two Configurations) and Pyridinium Ion. 99 Emission Spectra-Fluorescence Figure 20 shows the emission spectrum of 4-(3-phenyl-propyl)-pyridine in neutral ethanol solution. The spectrum appears to be quite similar to the spectrum of 1,3DPP. The emission at around 284 nm is the emission from the toluene part of the molecule while the broad and structureless emission at 325.5 nm appears to be arising from an intramolecular 'heteroexcimer" ~ ~ D -1 state. The shift of the heteroexcimer emission is v v = 6520 cm , max M 0,0- blue shifted by around 460 cm'1 from the excimer maximum of 1,3DPP. This heteroexcimer emission is also viscosity dependent: addition of glycerol to the solution drastically reduces the dimer to monomer ratio and at about 70% glycerol (n = 80 cp) the dimer emission disappears. Under the same conditions excimer emission in 1,3DPP still occurs. This is easily under- stood because of the polarity of the pyridine and glycerol and the possi- bility of hydrogen bonding (solvation, reduction of mobility). Dispite the similarities With 1,3DPP there are certain important differences. First the emission from 1,3PyPP is blue shifted. According to Chandross170 one would expect stronger interaction between two molecules with different elec- tron affinities. The electron affinity of pyridine is about 1.7 times larger than that of benzene171 and so the charge-transfer interaction may be ex- pected to be stronger. Since the states of pyridine and benzene are almost degenerate one may expect also a molecular exciton interaction. If the lowest state of pyridine is the n,«* state and since it is well known that the n,r* is out of plane polarized there will be no exciton interaction in a face to face configuration. (The exciton interaction of course would have been negligible because of the small oscillator strength). We believe that ‘the lowest state in ethanol is a 1rr* because when 1,3PyPP is disolved in hydrocarbon where the n,w* state is known to be the lowest state no emission 1 >. ’2 09 2 Lu p. E m 2 0 2 Lu 0 U) Lu 3 E 3 .J I l I l I 7 280 300. 320 340 360 380 A (n m) Figure 20. Room Temperature Fluorescence Spectra of (l) 1,3-Dipheny1- propane in Ethanol (2) 4-(3-Pheny1)—Pyridine in Ethanol. 101 was observed. Apparently an exoenergetic energy transfer from the toluene to the nonemitting v-picoline part is taking place. We believe that the difference in ionization potentials and electron affinities may be important in cases where electron transfer is taking place while in our case the in- teraction is of the charge-exchange type and resonance betWeen the CT states may be important. Looking at the emission spectra of 1,3DPP and 1,3PyPP in EtOH, we see that the dimer to monomer ratios are different. ID/IM(1,3DPP) = 1.8 (Air equilibrated) ID/IM(l,3PyPP) = 1.3 (Air equilibrated) 172 A similar behavior has been observed by Hirayama for a series of intra- molecular excimers. We list some of his results: System ID/IM Phenyl-(CH2)3-Phenyl 2.45 0-Tolyl-(CH2)3-O-Tolyl 1.14 (Nitrogenated solutions) Phenyl-(CH2)3-O—Tolyl 0.48 Phenyl-(CH2)3-MrTolyl 0.30 Again we observe a lower ID/IM ratio in the case of mixed (hetero-) excimers. In the case of 1,3PyPP intramolecular excimer we may have expected a high ID/IM ratio. Let us consider the possible mechanisms of excimer for- mation in this system: I. P. + Py -—~——->E (PPy)*\ P+Py+heat\’ P+Py+hv II. P + Py* : (PPy)* 4% P + Py + th Since the pyridine part of the molecule is responsible for the largest part P+huM P+heat D Py + heat P + Py 102 of the absorption and since pyridine is not emitting then if mechanism II is operating the ID/IM ratio should be quite high. From our results it appears that mechanism II is not operating because the excimer formation can not compete with the efficient non-radiative processes in pyridine. Birks 7 discusses the previously mentioned results of Hirayama and concludes that the observed ID/IM ratios reflect the fact that the binding energies of the mixed excimers are less than those of the excimers. It is interesting to note that in the systems studied by Hirayama $3 is always the same. In our case we observe both spectral blue shifts and decreased ID/IM compared to 1,3DPP. These phenomena may just result from weaker interaction (binding) due to the out of resonance conditions. The same results seem to us that may be produced by another cause. It is expected that if one of the partners in the mixed excimer is not emitting then the mixed excimer non-radiative mechanisms may be enhanced. Cherkasov174 verified this idea experimentally. Efficient non-radiative decay of the excimer would obviously decrease the ID/IM but also may lead to blue shifts. Usually we assume that emission results from an equilibrium state wherethe solvent has relaxed com- pletely. This is usually true because of the very short dielectric relaxa- tion times of the common solvents. Ethanol has a relatively long relaxa- tion time~10a10 sec.,and one can visualize the case where the dielectric relaxation time and the lifetime of the molecule are comparable. In this case the solvent relaxation may not be complete and this will result in a blue shifts. Very recently this idea was tested experimentally by Weber175 who studied derivatives of naphthalene and indole under high oxygen pressures and in deoxygenated solutions. The oxygen quenched the molecular fluorescence and decreased significantly the lifetime. As a result of incomplete relaxa- tion the oxygenated solutions showed blue shifts of'v8 nm. It would be -Esrv 103 interesting to test this possibility in our system by attempting to measure the mixed excimer lifetime. Benzylpyridine (Py-CH2-P) under the same condition (neutral ethanol) A model for benzylpyridine will be diphenyl methane. As it was shown by Watson and El-Bayoumi176 gave no emission. diphenylmethane does not show any excimeric emission but only changes of the P/F ratio and the intersystem crossing rate constant compared to toluene. Earlier Hirayama49 had estab- lished the "three carbon chain" rule for excimer formation. So it is not surprising that benzylpyridine does not show any excimer emission. The very low quantum yield of monomer emission is undoubtly a result of energy trans- fer to pyridine. Figure 21 shows room temperature emission spectra of 1,3PyPP in ethanol solution after addition of perchloric acid. The different curves correspond to different excitation wavelengths. The change is dramatic. The spectra are very broad and structureless and show a huge Stokes-shift. There is also an excitation wavelength dependence. 1,3Py+PP in EtOH at room temperature emission (cm-1) Excitation (nm) amax 265 23,810 280 23,946 300 24,414 The shift expressed as 3M - iD ’* 13,400 cm-1 (at 265 nm). The shift 0,0 max 1 between the maxima of the spectra excited at 265 nm and 300 nm is’v600 cm- Benzylpyridine (Py+—CH2-P) in ethanol + HC104 shows the same behavior and the shifts are even larger: Py+-CH2-P in EtOH at room temperature Excitation (nm) Emax emission (cm-1) 22,727 260 . 1- witfif :7 ._ \‘ 327:“. ‘4 104 4-(3-phenyl-propyl)-pyridine l. Excitation of 265nm 2. Excitation at 280nm 3.Exchofion at 300nm EtOH + HCI 0‘ ., ~ :\'<:’f-r~r..-: :7 S” i *1. Ethanol + HCIO‘ at 77° K LUMINESCENCE INTENSITY 300 330 450 45'0 A (nm) I I l l I I l fi 360 380 400 420 440 460 480 500 A (nm) Figure 21. 1. Room Temperature Fluorescence Spectra of 4-(3-Phenyl-Propyl)- Pyridine in Ethanol + HClO . 1. Excitation at 265 nm, 2. Excitation at 280 nm, 3. Excitation at 300 um. II. Total Luminescence Spectrum of the Above Solution at 77°K (Excitation at 265 nm). 105 . . " . . -l Exc1tat10n (nm) ymax emiss1on (cm ) 270 22,830 280 22,988 290 23,364 pg,0 - umak~ll4,500 cm 1 (at 260 nm) and vmax(290) - Vmax(260)ev640 cm 1. From our study of the absorption spectra we know that upon addition of the acid a charge-transfer interaction is becoming important. It is natural therefore to ascribe these red shifted emissions to charge-trans- fer fluorescence. The shifts are unusually high. There are very few reports on charge- transfer fluorescence at room temperature. Rosenberg and Eimutis177 studied the CT fluorescence of several methylbenzene complexes with pyromellitic dianhydride at room temperature, the observed shifts were about 8000 cm-1. The largest shifts observed were in a system involving again a pyridi- nium moiety. Briegleb et al.154 observed very large Stokes shifts 10,000- 20,000 cm-1 for the CT fluorescence of the systems MePy+X- where X- various electron donating anions like 1-, SCN-, N3- etc. The temperature range was -120° to -l70°C(glass). Briegleb's interpretation for this enormous Stokes shift was "the result of an intermolecular configuration transformation in the excited state". Figure 22 shows the emission spectra of 1,3PyPP in water. The spectra resemble the spectra in acidic ethanol. 1,3PyPP in water . . ~ -1 Excitation (nm) Em1331on ymax(cm ) 260 22,727 270 22,883 280 23,810 106 Fluorescence Intensity I 300 400 Amm) Figure 22. Room Temperature Fluorescence Spectra of 1,3PyPP in water (__). l. Excitation at 260 nm, 2. Excitation at 270 nm. _ Room Temperature Fluorescence Spectra of 1,3Py+PPC10 in water (--). 1. Excitation at 260 nm, 2. 270 nm, 3. 280 nm, 4. 290 nm. 107 Fluorescence Intensfiy 560 A(nm) Figure 23. Room Temperature Fluorescence Spectra of Benzylpyridine in water L). Excitation 1. 260 nm, 2. 270 nm, 3. 280 nm. Room Temperature Fluorescence Spectra of Benzylpyridinium Per- chlorate in Water (--). Excitation l. 260 nm, 2. 270 run, 3. 280 nm. 108 4M .2D .- ~D «ID '1 .0,0 - ”ma£' 14,500 cm 1 and ”max(280) - ”max(26°)'”1100 cm . The same behavior is displayed by Py-CHz-P in water, Figure 22. Py-CH2-P in water Excitation Emission 3 (cm-1) max 260 21,930 270 22,222 280 22,624 1 1 and 7” (280) - 3 (250)~ 690 cm' . max max ' 7g,o -';§a§g 15,300 cm” Finally we added acid (HClO4) to the water solutions of 1,3PyPP and Py-CHZ-P and Figures 22 and 23 show the results. 1,3Py+PP in water -1 Excitation (nm) Emission 3 (cm ) max 260 22,371 270 22,472 280 22,624 290 22,988 . ~D — ND UD -1 VM - v (260)~’l4,900 cm 1, v (280) - v (260)’~250 cm . 0,0 max max max Py+-CH2-P in water Excitation Emission 353x(cm-l) 260 21,520 270 21,692 280 21,739 D _ ~D ~D - $M - 7 (260)'~15,7OO cm 1, u (280) - u (260)»u220 cm 1. 0,0 max max max From the absorption spectra study we found evidence of charge-transfer interaction in the case of 1,3PyPP and PyCHzP in EtOH + H0104 and in 320 + HC104. Weak charge-transfer interactions in neutral water solutions can not excluded since the low sensitivity of the absorption spectra compared to the emission spectra does not allow detection of small interactions. 109 Further the blue shift of the n,1* state, which appears as a red tail in non protic media, tends to obscure Such weak interactions. In the excited state the charge transfer interaction is expected to be stronger. Leonhardt and Weller178 pointed out that electronically excited molecules have a lower ionization potential and a higher electron affinity than molecules in the ground state. The following diagram helps to clarify this point. _—-1f—-——l)+u I 43""lt* IVZIEEom E00 E":- E “ E0» ' ID 0* R A A Eop E" _ __J Experimental studies by the same authors confirmed this idea. A flash-spectroscopic investigation of the system perylene-«SEH in methanol showed formation of perylene cations. Using for perylene ID = 7 eV and 1 = 23,000 cm- or 2.85 eV179, we get a lower limit of the pyridinium E0,0 electron affinity of —4.2 eV. In the same way for toluene ID = 8.8 eV and E0,0 = 4.6 eV which leads again to a lower limit of the electron affinity of -4.2 eV. The observation of perylene cations makes us believe that electron transfer is also possible in the systems we are considering. From the above discussion one may expect the existence of both charge- transfer complexes and exciplexes in the systems under study. Excitation Spectra (uncorrected) support this idea. The excitation spectra of 1,3PyPP and PyCHzP in water + H0104 acid solutions show two excitation bands, one corresponding to the molecular absorption and the other to the CT absorption. The ratio of the intensities (maxima) ICT/IM for the two compounds is differ- ent. For example, observing the emission at 450 nm, in the case of 1,3PyPP the ratio is 1.4 while in the case of PyCHZP is 3.1. This seems to suggest 110 that the 1,3PyPP system is less associated probably due to the steric res- trictions that lower the probability of association especially since the in- teraction potential (charge exchange) is a short-range one. Further the energetics of the binding are not expected to be favorable in the ground state of this system since there is no Coulombic interaction of the kind present in the usual D+A' complexes. In the case of PyCHZP the phenyl and pyridine rings are held in close proximity and association and dissociation are less meaningful. The importance of water in increasing the intramole- cular interaction can be easily understood if we consider its acid pro- perties, high polarity and finally high surface tension (hydrophobic bonding). The large observed Stokes-shifts in CT complexes and exciplexes are considered as a result of large configuration changes in the excited state plus solvent shifts due to the polar nature of the excited state. In our case we expect large configuration changes, especially changes in the solvent cage created by the electron transfer and the production of a phenyl cation from the hydrophobic and apolar phenyl group. The counter- ion (C104-) may play a role too. The INDO calculations discussed earlier indicated that the positive charge in the pyridinium moeity is delocalized on the ring carbons. If this is true then we may expect that on the ave- rage the anion will be situated above the ring in the "ground state" of the exciplex and probably in the case of the ground state weak CT complex. In the excited state the positive charge will be mainly on the phenyl creating a large effective dipole(cation-anion) leading to farther relaxation. The importance of relaxation is also shown by the low temperature, rigid glass spectra. Figure 24 and 25 show the total emission spectra of 1,3PyPP and PyCHZP in acidic ethanol glass at 77°K. In the rigid glass, relaxation is greatly prevented and thus the considerable blue shift com- pared to fluid solution. The fluorescence is overlapping with donor 111 >. t 09 Z “J I'- Z UJ 0 Z LU O U) DJ E 2 D _J I I I fi 200 300 400 500 600 A(nm) Figure 24. Total Luminescence Polarization Spectrum of 1,3Py+PPClO - in Ethanol at 77°K (__). Phosphorescence Polarization Spectrum of 1,3Py PPClO4' in Ethanol at 77°K(--). 112 LUNHNESCENCEINTENSWY I I 300 400 500 A(nm) F1Sure 25. (__) Total Luminescence Spectrum of Benzylpyridinium Perchlorate in Ethanol at 77°K. (--) The Phosphorescence Spectrum under the Sane Conditions. 113 phosphorescence as we will see later. The apparent maximum in the case of 1,3PyPP is at 350 nm a blue shift of‘r4800 cm'l. In the case of PyCHZP the apparent maximum is at 330 nm which means a blue shift of-76OO cm-l. This seems to suggest that re- laxation is more important in PyCHZP while the binding energy of the CT eomplex is less. This most probably is associated with the larger average distance of the phenyl and pyridine ring in PyCHZP compared to 1,3PyPP (assumed in a "sandwich" configuration). This larger distance will reduce the covalent character of the CT state. Note also that the absorption CT band is situated at‘V290 nm in acidic water solution of 1,3PyPP (maximum of excitation spectra) while the corresponding band for PyCHZP is atrvz75 nm. Polarization spectra at 77°K give additional evidence for the nature or the excited state. Figure 24 shows the polarization spectrum of 1,3PyPP -in ethanol + HClO4. The polarization ratio is (In - Ii)/(I" + 11) = -0.22 (fluorescence + overlapping red shifted, phosphorescence). By the use of a phosphorescope the phosphorescence (donor phosphorescence, see later) was isolated. The polarization ratio of this phosphorescence is -0.23. Since as it is well known, the phosphorescence of r systems is out of plane pola- rized, the fluorescence from 1,3PyPP under the above conditions appears to be out of plane polarized (The theoretical polarization for perpendicular absorbing and emitting states is -0.33) which is in agreement with the assigned CT character of the emission. The blue shift accompanying progreesive red excitation is somehow puzzling. From our absorption and emission data we observe that conditions that favor strong interaction or higher association (water + HC104 as solvent and in the case of PyCHZP) the shifts are less pronounced. 114 Table 10. Stokes Shifts and Effect of Excitation Wavelength on the Emission Maxima of 1,3PyPP and PyCH2P in Different Solvents (in cm‘ ). Solvent Compound ; (280)-; (260) Stokes Shift(260nm) max max H20 1, 3PyPP 1100 14, 500 PyCHzP 694 15,300 EtOH + acid 1,3PyPP 600 13,400 PyCHZP 260 14, 500 H20 + acid 1,3PyPP 250 14,900 PyCHzP 220 15, 700 This may suggest that the shift is associated with excitation of different species. Shifts of the emission spectra with excitation wavelength were observed in various cases, especially for quinine and related compounds}80’181 Note that in all cases red excitation resulted in red shifted emissions. The interpretations suggested usually involve emissions from slightly differ- ent species like rotomers or differently solvated species. Itoh and Azumi review these various possibilities and suggest that the shift arises from incomplete relaxation of the solvent Franck-Condon state182. The authors note that the "shift is more prominent in solvents which are capable of hydrogen bonding or protonation" although they don't accept protonation as the cause of the shift. In the case under study it is natural to believe that if the emission is resulting from closely related species, these species are "exciplex like" or ’bharge-transfer complex like" species. As we progressively excite with longer wavelength light the importance of CT complexes over exeiplexes is expected to increase. One may expect that in a usual inter- molecular complex the CT excited state and the exciplex state arising from 115 the same molecular pair are the same. This may not be true for an intra- molecular case. Recently Itoh et al.112 studied the intramolecular exci- plex and charge-transfer complex formations in (9,10-Dicyanoanthracene- (CH2)3-(Naphthalene) DCAN system. Their conclusion was that there are two fluorescent states in the potential energy surface of DCAN. The authors rationalized this as follows: the electronic interaction and therefore the geometrical arrangement between the DCA and naphthalene moieties are dif- ferent from each other in the Franck-Condom excited states of the exciplex and the CT complex, because the exciplex formation is the photochemical process but the formation of the latter is really the thermal process followed by the photoexcitation. The internal or geometrical conversion from their Franck-Condom states to the identical fluorescent state, if it is correct, is forbidden in our compounds by the steric factor concerned with the trimethylene chain. Such a reasoning could be helpful in our case if we assume that the CT state is blue shifted relative to the exciplex state. The reason for this being so is not easy to visualize however, especially because besides the geometrical reorganization of the solute itself the solvent reorganization is very important. This may explain why shifts (although smaller) are observed in the case of PyCHzP and the large shifts in plain water where a variety of solute-solvent interaction centers (dif- ferent degrees of solvation,hydrogen bonding) are expected. Mew from 1,3PyPP and m Figure 26 shows the phosphorescence spectrum of 1,3PyPP in (l) neu- tral ethanol glass at 77°K and (2) in acidic ethanol glass. As we see no shifts are produced by the addition of the acid and the position of the Phosphorescence maximum is the same as that of toluene's phosphorescence (~380 nm). 116 PHOSPHORESCENCEINTENSHW’ r 300 400 A(nm) Figure 26. (__) Phosphorescence Spectrum of 1,3PyPP in Ethanol at 77°K. (--) Phosphorescence Spectrum of 1,3Py PPClO4 under the Same Cenditions. 117 As we already have seen the charge-transfer fluorescence and the phosphorescence are strongly overlapping at 77°K. This appears to be quite a common case. In fact in the first systems to be studied, namely 1,3,5- trinotrobenzene complexes of polycyclic aromatic hydrocarbons in glasses at -l80°C, a phosphorescence emission band of the donor virtually coincided with the charge-transfer fluorescence band183’184. So the very resonable postulate was made that the total emission of complex was the phosphorescence transition of the donor. 2 As discussed by Nagakura and coworkerslas, the case where the 3LE state of the donor is lower energetically than the 3CT is the most common case. McGlynn and his associatesl86’187’188 accounted for the donor phosphorescence by proposing that the energy of the excited 1CT state could be lost by trans- fer to a lower 3CT state. Since this triplet level is dissociative at the energy of the intersystem crossing, McGlynn suggested that probably the complex would largely dissociate to yield the acceptor in the ground state and the donor in its lowest excited triplet state. Clearly in our systems the 3CT state is dissociative since there is no covalency (unparied spins) and no ion-pair Coulombic interactions but only the usual intermolecular forces. Measurements of the phorphorescence lifetime of 1,3Py+PPClO4- in ethanol glass gave a TP = 6.8 sec. in good agreement with the phosphorescence lifetime of toluene (6.5 sec.) under the same condition. The phosphorescence of PyCHQP is again donor phosphorescence. 118 II-MANIFESTATIONS 0F CHARGE—RESONANCE INTERACTIONS IN THE CASE OF (b-(CH2)3-d>+ IN THE GAS PHASE From our previous discussions onthe system 1,3Py+PPClO4-, we concluded that complete electron transfer may take place in the excited state. Badger 189,190 and coworkers studied the absorption spectra of benzene cation . .7 + (C6H6 ) and discovered long wavelength bands which they assigned as follows: 3*‘ + 1.80um-l C6H6+; l.08,um'1 (C6H6)2+ (charge-resonance band); 2.15um'1 (C6H6)2 714,7... I (corresponding to the l.80,um-l transition in the monomer). The monomer IR “7 band was assigned to a a—+T transition. The similarity between the bands ascribed to the benzene dimer cation and the bands of (2,2) paracyclophane cation led the authors to suggest a sandwich structure for (C6H6)2+° In a subsequent paper Badger and Brocklehurst191 studied the diphenyl propane and diphenylmethane cations. Both compounds gave absorption bands charac- teristic of the intramolecular charge resonance band. The interaction in diphenylmethane show again the difference between the geometrical require- ments for excimer and CR interaction. The charge resonance in species of the form (C6H6)2+ has also been demostrated by E.S. R.192. In this section we will study this intramolecular charge resonance interaction in 1,3-diphenylpropane cation intzhe gas phase as is manifested in the mass spectra of 1,3-diphenylpropane. Mode of Fragmentation d3 Alkyl D in the Mass §pectrometer The mass spectra of aromatic hydrocarbons show several characteristic features that reflect the influenece of the aromatic nucleus. Thus ioni- zation of alkylbenzene gives rise to parent ions of formula CnHZn-6+' De- + . composition produces chiefly ions of the type CnHZn-7 , corresponding to phenylalkyl ions; such data suggest that the benzene ring itself is stable 119 . 93 to electron bombardment and survives in many fragment ions . The most characteristic cleavage of alkylbenzenes occurs 3 to the aro- matic ring, i.e., by rupture of the benzylically activated bond. In to- luene, this results in the loss of one hydrogen atom to yield the ion + C7H7 (m/e = 91), which is also obtained in the higher homologs (ethyl-, propyl-, butyl-,... benzenes) by elimination of a methyl, ethyl, propyl, ..., etc. radical. So, B-cleavage .CHZ-é-R w C7H7+(m/e=9l) The most obvious formulation for C7H7+ would be that of a benzyl cationI but extensive studies using appearance potentials and isotope labeling <9... I II have shown that the actual structure is that of tropylium cation II. If the side chain is propyl or larger, another fragmentation process accom- panies this simple B-cleavage, namely'B-fission accompanied by rearrangement of one hydrogen atom, which results in the production of a neutral olefin molecule and the C7H8+(m/e = 92) .-C'H\2;('JHZ —.) .—CH3 m/e = 92 H‘CH-CH3 ' III or CH2 \ Qfl \, H2 H2 _ \ h‘)‘ m/e — 92 /CHCH3 H H H IV , + Since the energy requirements for the further decomposition of the C7H8 ion differ from those of toluene, the methylenecyclohexadiene formulation IV may be preferred. Another mode of fragmentation is the'Y-cleavage (@)-CH2-CH2-%-R ———-> @-CHz-CH2+ + R- m/e = 103 120 Table 11. Relative Intensities of Various Fragments of Butylbenzene and 1,3DPP in the Mass Spectrometer. Compound Ion (m/e) Relative Intensity Ii/Ifi+ n-Butylbenzene 134 (Molecular ion, MI) 1.00 91 3.73 E 92 1 . 84 103 0.12 ‘ 1,3-diphenylpropane 196 (14+) 1. 00 ’ 91 1.50 92 2.90 103 0.21 The comparison of the butylbenzene and 1,3-ciphenylpropane mass spectra is shown in Table 11. Discussion The most intense peak in the mass spectrum of n-butylbenzene is the one corresponding to the ion with m/e = 91 resulting from the previously mentioned fi-cleavage. The most abundant ion in the case of 1,3DPP is the ion with m/e = 92 resulting fronlfi-fission and accompanied hydrogen re- arrangement. This process is energetically more favorable in the case of 1,3DPP than in the case of Bu¢ because in the former the resulting double bond is stabilized through conjugation with the phenyl ring in the resulting styrene molecule. @CHZQCHz _____> C7H8+ + CH2=CH-@ m/e = 92 121 The ion with m/e = 103 is resulting from 7-cleavage. Again this process is energetically favored in the case of 1,3DPP because the resulting radical (benzyl radical) is stabilized through resonance with the phenyl ring. <:>-CH2-CH2-CH2<€§> <:>FCHé-CH2+ + CH24<:> m/e = 103 The most important ion for our purpose is the ion m/e = 91. As we see this ion is more abundant in the case of n-butylbenzene than in the case of 1,3DPP. Since its formation should not depend on the nature of the substituent on the V-carbon one may suggest that its reduced rate of for- mation from the parent ion in the case of 1,3DPP may be due to some sta- bilization of the parent ion by the presence of the second phenyl group. From our previous discussion an obvious explanation is the formation of an intramolecular complex of the type (A) which is stabilized by charge- ((0193 $4 I (A) resonance between the two phenyl rings. The ratio of the rate constants of fragmentation of 1,3DPP and n- butylbenzene to the various ions is given by the ratio of the relative abundances of the particular ion in the two compounds k(m[e=9l)l,3DPP = 0.40 k(m/e=9l)Bu kgm/e=92)1,3DPP = 1.57 k(m/e=92)Bu kgmge=103)1,3DPP = 1.81 k(m/e=103)Bu In a magnetic sector mass spectrometer a typical ion will spend about one microsecond in the ionization chamber between the instant of formation and its departure through the exit slit. It is accelerated to an energy .>o oh Boom nosuoon mom: can mo mnowuaom .um shaman ecu mo hmnwnm .ocwuswmeusm was onwaoumHmcoannum.H mo muuoomm o\E o\E .oao mop vm. .omo no, 00. — _ — __ *_— — . _. __ _ a: . @ie _ .x x .3 123 of three kilovolts in the next seven microseconds, spends about four mi- croseconds traversing the field-free region and the magnitic field, and in another two microseconds it arrives at the collector. Of course all the above times are estimates. For fragment ions to be formed in the ion source it is necessary that the decomposition reaction proceed with a rate 10+6 sec-1 or faster. If the rate of the reaction is in the neighborhood of 105 sec-1 the decomposition will occur in transit, and that part oc- curring in the field-free region between the electrostatic accelerating and the magnetic deflecting fields produces the metastable ions. we looked in the spectrum of 1,3DPP for metastable ions m/e = mfz/mi where mi = parent ion, m = m/e = 91 (fi-cleavage) and we did not find them E which means that the rate of formation of the m/e = 91 ion from the parent ion is 106 sec-1 or higher. In order that the proposed stabilization be effective the rate of intramolecular phenyl-phenyl collision must be comparable or higher to the rate of fl-cleavage which put a lower limit atrv106 sec'l. III-INTERSYSTEM CROSSING ENHANCEMENT THROUGH CHARGE-TRANSFER INTERACTIONS Introduction Exciplex (CT) interaction need not be associated with the appearance of a new emission band. The phenomenon of impurity quenching is considered as arising from the formation of non fluorescent exciplexes (dynamic quen- ching) or non fluoresced charge transfer complexes (static quenching). One of the most important quenching mechanisms upon complexation is 186 an enhancement of intersystem crossing. Christodouleas and MtGlynn have studied the intersystem crossing process in donor-acceptor complexes 124 of naphthalene with sym-trinitrobenzene (TNB), tetrachlorophthalic an- hydride (TCPA) and tetrabromo (TBPA)- and tetraiodo-phthalic anhydrides (TIPA), in rigid ether-isopentane solutions at 77°K. They compared the ratio ¢%/¢% of the naphthalene (donor) phosphorescence and fluorescence yields with that of a similar solution of naphthalene. ¢@/¢% increased by a factor of'VBO in the naphthalene—TNB complexes, an increase attributa- ble to CT interaction since there are no heavy atoms in the acceptor (TNB). ¢}/¢% increased by a factor of 103-5 x 104 in TCPA, TBPA and TIPA com- plexes, increasing in the order of the atomic number of the heavy atoms in the acceptor. Since that time numerous investigations have verified that charge- transfer interactions enhance the intersystem crossing. In several cases it has been shown, through laser photolysis studies, that the triplet state is generated "immediately" after the laser pulse and before the ex- ciplex reaches the excited thermally equilibrated state. This for example has been shown for anthracene and pyrene quenched by diethlaniline194 for pyrene and 1,2-benzanthracene quenched by dimethylanilinelgs. Independent support for the occurrence of this fast ISC mechanism comes from comparison of temperature effects on the triplet yields with those on the yields of exciplex fluorescence (nonpolar solvents) and ion formation (polar solvents);96 Regarding the mechanism by which the charge-transfer interaction en- hances ISC, McGlynn has proposed the following scheme197. The initially formed locally excited state (A1D*) can be described in a first approximation by the wavefunction NILE = a1’3 (4513+) + blq>(A1D*) b» a in which 1’3<1‘»(A'D+) and 1(A1D*) represent the contributions of the cor- responding zero~order charge-transfer and locally excited states. Fast 125 internal conversion to the singlet CT state may subsequently take place. 1 _ 1:3 - 'I' 1 1 WOT—c (AD)+d(AD*) c>>d Finally ISC will lead to 3mm = e1’3<1> (A'D+) + £34) (A3Dv'v) The efficiency of the IQICT - 3q’CT coupling arises from the mixing of the 1¢)(A'D+) and 3¢’(A-D+) states. Matrix elements of the form (14) (A-D )IHSO|3 (A'D+)> may be quite large mainly because the plane of symmetry which negates significant spin-orbit coupling of apt”: states of planar aromatics is destroyed by the very process of CT complexing. Some Pg character is mixed into the pi-MO's of D or A; one-center spin-orbit matrix elements are no longer zero and, in fact, should become dominant. In addition the perturbation gap between the two states is quite small, a factor that enhances the mixing. Finally for the same reaSOns 3¢’(A-D+) may mix significantly with other local singlets l‘bi' Second order mixing involving vibronic interactions is also contributing. Intramolecular Exciplex-CT Interactions in Aromatic Carboxylic Acids It is known that the absorption and emission properties of an aroma- tic chromophore incorporated in an amino acid molecule, are greatly in- fluenced by the presence of amino and carboxylic groups even though these substituents are separated from the chromophore by two methylene groups. In aqueous solutions this effect was manifested198’199’200’201 by the pH effects on the yieldsof fluorescent aromatic amino acids (phenylalanine, tyrosine, tryptophane) at room temperature. Addition and removal of a proton at the substituent occur due to changes in pH and are responsible for the observed changes in the fluorescence yields. This reflects an interaction between the carboxylic (or amino group) and the aromatic ring which is the emitting moiety. 126 El-Bayoumi and coworkers have undertaken a study of this quenching effects through various model compounds. Tournon and El-Bayoumi202 have studied the absorption and fluorescence spectra of a number of phenylcarboxylic acids of the general form(b-(CH2)D- COOH (n = 1 phenylacetic up to n = 4 phenylvaleric). It was found that the acids in their conjugate base forms have essentially the same fluo- rescence yields as toluene. When the pH is decreased, the fluorescence yield undergoes a sharp drop at a pH value of about 4.5, corresponding to the appearance of the un-ionized form of the carboxyl group. The extent of this fluorescence quenching decreases with increasing the separation between the phenyl and the carboxylic group, i.e., with increasing the number of methylene groups. The ratio of phosphorescence to fluorescence is higher for the protonated form than that for the corresponding anion form. The most dramatic effects are exemplified by phenylacetic acid. In going from phenylacetic anion to phenylacetic acid the oscillator strength of the l B E--1A benzene transition decreases byvaZZ while the fluorescence 2 1g yield decreases byr~70%. Since singlet-singlet nonradiative transitions from the lowest singlet to the ground state are expected to be negligible in these molecules, this decrease in fluorescence yield was considered by the authors as an indication that proronation of the carboxylate group increases the intersystem crossing rate. This is further supported by the larger phosphorescence/fluorescence intensity ratio for phenylacetic acid compared to toluene, the measured phosphorescence/f1uorescence ratios in ethanol glass being 2.05 and 0.96 respectively. The possibility that the quenching is due to some proton transfer reaction was eliminated by observing that the ethyl ester of phenylacetic acid has a similar fluorescence yield as phenylacetic acid. The most reasonable interaction appears to be of charge-transfer 127 nature. Considering the values of ionization potential and electron affinities of toluene and acetic acid molecules, the electron transfer would be from the aromatic group to the carboxylic acid group. Observation of charge-transfer absorption in phenylacetic acid would provide direct evidence for intramolecular interaction between the phenyl ring and the carboxyl group. Benzoic acid, where the carboxyl group is directly con- jugated to the phenyl ring, a charge-transfer band appears at 228 nm. By comparing the 1Bf;—-1Alg absorption bands of toluene, phenylacetic acid and phenylvaleric acids we see that a shoulder on the long-wavelength side of the absorption band appears in the case of phenylacetic acid (between 220 and 235 nm). This shoulder may be due to a CT transition. In a recent flash photolysis experiment203 of these acids, intramolecular interactions between the carboxyl and phenyl groups were also implied. The carboxylic acid is expected to have a higher electron affinity than the carboxylate anion. This would make the energy of the charge- transfer state lower for the phenylcarboxylic acid compared to the cor- responding anion. I In.a subsequent paper Tournon and El-Bayoumi204 studied the fluorescence and phosphorescence yields in media of different viscosities. Table 12 shows some interesting results. The strOnger fluorescence quenching in lower viscosity media at the same temperature indicates that intramolecular relaxation occurs during the lifetime of the excited singlet state. This is similar to intra- molecular exciplex interaction205 reported for naphthylalkyl amines, al- though in that case exciplex emission was observed. An additional evidence for the charge-transfer nature of the quen- ching mechanism came from the study of the luminescence properties of 128 Table 12. Effect of Viscosity on the Fluorescence and Phosphorescence Quantum Yields and ISC Rates of Toluene and Ethyl-Phenyl Acetate. 300°K 77°K 77°K 77°K 77°K -6 -1 Compound Solvent wk 0% ¢% (¢?fi°f) kISC x 10 s Toluene 3MP(2.6x1012) 0.14 0.27 0.29 1.08 13.2 1P(8.9x105) 0.14 0.23 0.17 0.75 16.2 Ethyl-phenyl— 3mr(2.6x1012) 0.04 0.17 0.33 2.00 18.8 acetate 7 3MP+iP(10 ) 0.03 0.12 0.11 2.09 28.0 (1:3) 5 iP(8.9x10 ) 0.02 0.02 0.06 3.00 188 The number in parenthesis after the solvent is the viscosity at 77°K in poise. 20 compounds of the form x4C:>-CH2-COOH 6. By changing the substituent x, the ionization potential of the donating moiety is varied. If a charge-transfer mechanism is operating an increase in quenching is expected as the ioniza- 207 tion potential of the donating moiety is lowered , i.e., in the series x = H, CH3, OH, OCH3. Table 13 shows the results in water at room temperature. Table 13. Effect of the Substituent X on the Absorption and Emission Properties of x-@~0H2-000H. x fc00_(x10'3) Af/f(‘7.) ...COO- °coon Aw/ocz.) LP. (eV) ‘H 5.3 18 0.06 0.02 68 9.24 -CH3 12 17 0.14 0.03 75 8.82 -0H 22 10 0.03 0.03 90 8.50 -OCH3 20 10 0.03 0.03 90 8.20 f = oscillator strength 129 The expected correlation is found and moreover it is shown that .A¢/0 and Af/f vary in opposite directions. This indicates that the inductive effect is mainly responsible for the variation of Af/f and the charge-transfer interaction for the variation of A¢%¢. In order to provide some more evidence for the importance of steric factors in determing the extent of intramolecular interactions between the phenyl and carboxyl groups and separate out the inductive effects we have studied some model systems. The study involves a comparative study of the luminescence properties at room and liquid nitrogen temperature of phenylacetic indan-l-carboxylic, phenylpropionic, indan-Z-carboxylic acids as well as the parent hydrocarbons toluene and indan. The carboxyl group is separated by one or two methylene groups in these acids but while rotation is free around C-C bonds involving the methylene groups in phenylacetic and phenylpropionic acids, no rotation around these bonds can occur in indan-carboxylic acids due to the five membered ring. This prevents the carboxyl group from assuming confor- mations that may bring it close to the ring in indan derivatives. The results are summarized in Table 14. From the study of the results in Table 14 the following observations are made: (1) The remarkable (70%) fluorescence quenching observed when phenylacetate anion is protonated does not occur in the case of indan-l-carboxylic acid although both acids have one methylene carbon between the phenyl and car- boxyl groups. This indicates that the carboxyl group in the indan deri- vative can not approach the phenyl group in a conformation favorable for charge-transfer interaction which may lead to fluorescence quenching. (2) The small but measurable fluorescence quenching effect of the carboxyl 1.41.. L C N, dz... 0... m .um—{Ame .. Cm... u ems .0530me woeoowouosemosm Hmunumc ecu mun e ... m .wEHuowHH oonoomohosmmosm wo>nomao one ma Moms 5 .maowom oSu mo mUHo«% onsumnoasou Boon one muwxomun cw mowuflunmsw osu MonmAe .%Ho>wuooewou .Menn an new wusumuwmfiou Econ um ooaoomouosam mo mwawwm Esusmsv ohm Monnm .M.oomwe 0 i 3 _ 1 mew as 3.0 2.0 30 A200 2.0 00008 oeesxoeemoéémefi 0.00 0.0 00.0 3.0 00.0 A800 0N0 8 cessxoeeeoiéeefi 0.00 0.0 2.0 2.0 0.0.0 3.0 0000 AuflOu e005 0.0a 0.5 was «0.0 80 A200 00.0 0000-0A900© 380083.320 0.2 0.0 0.0 00.0 :0 A200 00.0 0000-~00© 0383.995 0.: as 00.0 00.0 3.0 2.0 000© 00038. ems ms MonnAfie\Aev MonnAe Mennfie Mooomme mHsEpom wssoaaoo .Hoomsum cw mvwom swamxonumoazxwamaofim vwmwmunoz mam wwwfim mo wwHuHoQOHm mocoomoCHEsq .¢H oHan fi 131 group in the case of phenyl propionic acid (23%) is absent in indan-2- carboxylic acid. In both cases the phenyl and carboxyl groups are separated by two methylene groups. (3) The ratio of the fluorescence yield at liquid nitrogen temperature to that at room temperature is the same for indan and indan-2-carboxylic acid but is slightly different for indan-l-carboxylic acid indicating some but insignificant relaxation effects of the carboxyl group in the excited singlet state of indan-l-carboxylic acid. (4) For indan-l-carboxylic acid the room temperature fluorescence yield is lower than that of indan by about 26% and there is a slight phospho- rescence lifetime, indicating some interaction between the carboxyl and Phenyl group. (5) From the absorption spectra we find that the oscillator strength of the first absorption band decreased by 23% in indan-l-carboxylic acid and by 17% in indan-Z-carboxylic acid compared to indan. These results are similar to changes in oscillator strength in phenyl acetic and phenyl propionic acids compared to toluene and are interpreted in terms of the inductive effect of the carboxyl group. CHAPTER 4 EXCITED STATE PROTON TRANSFER REACTIONS INVOLVING THE 7-AZAINDOLE MOLECULE INTRODUCTION Excited state proton transfer reactions along with excimer/exciplex interactions comprise the most common adiabatic photoreactions; that is, reactions in which the products are generated in an excited state. This fact allows the study of these phenomena by fluorescence spectroscopy. In this chapter, we are concentrating on the various kinds of proton transfer reactions in which a model biological molecule 7-azaindole (7AI) is involved. Initially the acid-base properties of the molecule are dis- cussed in terms of its absorption spectra and from absorption and emission measurements the excited state pKa* and pr* are determined. As it was 208 , first shown by Taylor, El-Bayoumi and Kasha hydrogen bonded dimers Of 7AI upon excitation Show biprotonic phototautomerism. The technique of time resolved spectroscopy is applied here to study the dynamics of this double proton transfer. 7AI deuterated in the N1 position is studied in a frozen 3-methylpentane matrix at 77°K. Time resolved spectra in the nanosecond time range show clearly the time evolution of the phenomenon (development of the emission corresponding to tautomeric species). Analysis of the nonexponential character of the decay curves gives the rate constants of the forward and backward reactions. The rates for plain (non deuterated) 7AI are calculated through the observed isotope effect on fluorescence 132 133 spectra. Finally from the invariance of the relative intensities of dimer and tautomer fluorescence at 77°K and 4°K and the kinetic results at 77°K, is concluded that the reaction at low temperatures proceeds through quantum mechanical tunneling and not by a thermally-activated process. When 7AI is dissolved in alcohols a new emission band appears that has at 7 been ascribed to a tautomeric species arising from a double proton transfer involving the 7AI molecules and an alcohol molecule. Through the effect of “Iv-TH ,. . \‘r' solvent deuteration on the fluorescence quantum yields of 7AI, N-methyl-7- azaindole and a model compound for the tautomeric species strong evidence P is found in favor of the tautomerism mechanism. Finally the more subtle interactions between 7A1 and water (no new emission appears) are studied through solvent isotope effects, temperature and pH studies. THE HYDROGEN BOND (I) Theoretical Models of the Hydrogen Bond Electrostatic Model The hydrogen bond in this model is regarded as a special case of donor-acceptor interaction. In order to form a stable hydrogen bond A-H'~oB, the essential requirement of the acceptor is that the charge distribution of the A-H bond orbital is such as to leave the proton sufficiently unscreened. The essential condition to be met by the donor is that it have lone-pair electrons. Attempts have been made to represent the energy of electrostatic interaction of the hydrogen bond in terms of 209,210 dipole interactions or in terms of the interaction of point-charge models211’212. It has been known that the directional properties of lone-pair 134 orbitals, like those of bonding orbitals in the first row elements, are due primarily to the directional characteristics of the atomic p orbitals, and the directional properties of bonding orbitals and lone-pair orbitals centered on the same atom are closely interdependent. Accordingly, hy- bridization of atomic orbitals, which nearly always occurs to some degree in bonding orbitals, will also cause hybridization of the lone-pair atomic orbitals. An important consequence of hybridization which has been pointed out by Coulson213’214 is that the charge distribution of the orbital re- lative to the nucleus is therby made unsymmetrical, giving rise to an effective orbital dipole (or atomic dipole). The magnitude of the lone- pair orbital dipole can be obtained by determining the centroids of the charge of the hybridized lone-pair orbital. If the wave function for the hybridized orbital is given by w= (1 + 12)'15(s)+ >.4>(px)] where A is a hybridization parameter and ¢(s) and ¢(px) are the usual one-electron atomic functions, then the centroid of charge distribution is expressed by the following relation: E = 2)\f(s)x d>(px)dT/(1+)\2) = 2k/(l+)\2)'5/z(3)% (109) x in equation 109 is measured relative to the nucleus in the direction of the orbital. Slater's values for the effective nuclear charge 2 are used. If the hydrogen bondis regarded as resulting from an HrA group of a second molecule directed at a lone-pair orbital, then the dominant term in the interaction energy could be considered to be the interaction of the proton with the lone pair dipole. The values of the dipoles corresponding to the individual lone-pair hybrid orbitals are obtained by multiplying E by the total charge -2e. 135 _O Centroids of lone-pair orbitals x(A) N 0.339 Tetrahedral O 0.290 F 0.254 For example, considering the central hydride molecules NH3, H20 and HF, if the lone-pairs may be represented approximately by tetrahedral hybrid orbitals; and using fl1e above values for the centroids, one would con- clude that, relative to a common H-A group, the NH3 lone-pair would form stronger hydrogen bonds than the HF lone-pair, with the H20 lone- pair intermediate. For the quantitative calculation of the hydrogen bond energy, it is assumed that the total electronic charge of each localized orbital (either a bonding or lone-pair orbital) is concentrated at a point at the centroid of the orbital. To fix the centroid of the bonding orbitals, the observed dipole moment of the molecule is used and the charges are placed so as the give the correct dipole moment. The electrostatic description of the hydrogen bond based on point- charge model gives the correct order of magnitude for the energy of the hydrogen bond. It should be noticed though that calculations of the hydrogen bond energies based on the electrostatic model do not offer a severe test because there is parametric freedom in the placement of the point charges, and furthermore, the overlap energy as well as the dis- persion energy and polarization energy have been neglected. The de- ficiencies of electrostatic model in explaining several experimental aspects of the hydrogen bond have been pointed out by Coulsongls. Valence-Bond Theory 216 Coulson and Danielson gave a semi-quantum mechanical treatment of the hydrogen bonding O-Hy-oO as a four-electron problem. They considered 136 the following resonance structures for the hydrogen bonded system: \‘OA-H ogj: ¢l(pure covalent, no charge transfer) \\ - + // . . 0A H QB\ ¢é(pure ionic, no charge transfer) .. + \\9A H-OB<: ¢3(charge transfer, ions H-O bonding) The wavefunction for this hydrogen-bonded system takes the form \l/= <1> + + c11 c2‘1’2 c34’3 where the wavefunctions for the component structures are 4.1 = N1{ |.A(1):H(2)2¢B(3>IB<4>| - IZA<1)~H(2)~B(3)?B<4)|} .2 = N2 |xA<1>~A<2>~Bo>ch4>I 4,3 = N3 { I"A(1)"A(2)"H(3)"B“‘" - axA(1)xA(2)~H(3>xB<4)I} where rA and ”B are oxygen atomic orbitals and ”H is a hydrogen atomic orbital. For the choice of atomic orbitals, the ls orbital is used for the hydrogen atom. For the oxygen orbitals hybrids of the s and p orbitals are used. The appropriate values of the coefficients c1 and the energy E of the system are determined as usual by the variational method. The Hamiltonian is just an effective four-particle Hamiltonian. In the solu- tion of the problem experimental information is also used. The calculation is performed as follows: (a) The overlap matrix is calculated nonempirically using Slater's orbitals. (b) The diagonal elements of the Hamiltonian are calculated semiempirically taking into account the fact that Hii is just the energy of the pure state 01. Hii is composed of the following terms: (1) the covalent bond energy expressed by the familiar Morse formula; (2) the short-range repulsion energy expressed by the help of the empirical formula Kexp(-br); (3) the energy associated with the transfer of an electron from atom x to 137 an atom y; (4) the polarization energy. This is the Coulson-Danielson method. The methods used by Sokolov217 and Tsubomura218 differ in many points. Nevertheless, the conclusions reached by all these theories are similar and can be stated as follows: (I) The electrostatic forces(l), the short-range repulsion forces(2), and the charge-transfer forces(3), are all of the same order of magnitude but of different sign. Often effects (2) and (3) cancel each other which explains the success of electrostatic theories. (II) The amount of charge transferred from OB to 0A is found to be non- negligible (but small) for shortbonds and negligible for long bonds; thus the long bonds are essentially electrostatic. Charge-Transfer Theory This theory is developed in analogy to the charge-transfer theory of molecular complexes that we have already outlined. Since in both cases the charge migration is an important factor it may be hoped that the theories applicable to the study of charge-transfer interactions are, mutatis mutandis, applicable to the study of hydrogen bonding. The theory in the form put by Bratoz219 is as follows: again the system OA-H--°OB is considered. The four electrons in the hydrogen bonding are distributed over three orbitals the bonding and the antibonding OH orbitals, k and x and the 0B lone-pair orbital x3. The wavefunction representing 1 the hydrogen bond is W=cf1+cfb 4:: "' where 1 N1 [X1X1K3x3‘ 0 = N ' _ — — x z i x 2 2{ |"1"1"2"3\ ‘1 1 2 3} The first term ¢1, describes the situation where two electrons are placed on OAH and two on OB (no bond function). This accounts for the electrostatic 138 and shrot-range repulsion interactions. The second function o describes 2, the situation in which charge transfer has taken place from the 0B lone- pair orbital u3 to the antibonding OAH orbital x2. This function confers on the hydrogen bond 3 partially convalent character (dative bond function). The energy E and the coefficients c1 and c2 are obtained by the variational method. The diagonal elements are calculated again semiempirically. Al- though different authors have different computational approach, the results are similar: (a) The hydrogen bond energy can be decomposed as postulated by the VB theory. Al remarks concerning the relative importance of its components remain valid. (b) The contribution of the charge-transfer term increases as the ioni- zation potential of the lone-pair becomes smaller. (c) The special role of the hydrogen in the hydrogen bond formation is due to the smallness of the short-range repulsion term. This reflects the small "size" of the hydrogen. (d) In the charge-transfer process a fraction of an electron is placed on the antibonding OAH orbital v2. The OAH bond is thus weakened, its length increased, and the force constant kl, decreased. This is just what is observed experimentally. OH (e) The charge migration increases the polarity of the OA-H---OB systems and intensifies the le IR bands. SCF-MO Theory In this theory the complex AH-«vB is considered as a single large molecule and is treated by the SCF-M0 and CI techniques. The problem has been approached from a direct ab initio approach or by introducing drastic approximations. In the non-empirical valence bond calculations, 139 the nonrelativistic Schroedinger equation is solved using a nonorthogonal atomic basis in a multideterminant wavefunction. This approach has been used for the study of H-bonds220’221. However, this method is very difficult to apply to larger systems in which the number of determinants for a minimal calculation becomes enormous and the resultant wavefunction difficult to interpret. It appears that as a starting point in understanding hydrogen bonds, molecular orbital theory is the most satisfactory approach. The literature on the subject up to 1971 has been reviewed by Allenzzz. 223 For more recent references see for example the work of Del Bene . There are several generalizations that have been produced by these studies: (1) The stabilization of an Hrbonded dimer appears to arise primarily from the electrostatic interactions between the proton and the lone-pair of electrons. For example in the dimers ROH-'°O=CH2, the order of increasing hydrogen bond strengthparallelsthe order of increasing sigma electron withdrawing ability of R. (2) The hydrogen bonds (open chain dimers) are almost linear. (3) Although the electrostatic interaction appears to be the principal interaction, the dipole-dipole term is also important. Thus, in the dimers ROHoo-O=CH2 there is a tendency for the permanent dipole moments of the proton donor and acceptor molecules to approach a somewhat antiparallel alignment, in so far as possible within the orientational requirement for the directed lone-pair. (4) There is a small transfer of charge which takes place from the proton acceptor to the proton donor molecule, and it occurs through the sigma electron system. A very interesting investigation, through ab initio techniques,of the 140 2 double well picture of the hydrogen bond was performed by Clementi et a1. 24 The study was on the DNA base pair Guanine-Cytosine and the motivation was the theory put forward by Lowdin225 on the possible importance of quantum mechanical tunneling for the interconversion between different tautomeric forms. Although there was a noticeable shoulder where the second minimum might have been expected no second minimum was found. Clementi and co- workers gave various possible sources of error in their prediction of the lack of a doulbe minimum. The most important is that no simultaneous motion of two hydrogen bridges was considered. To test this possible error source, Clementi and coworkers carried out calculations on the formic acid dimer, which has two hydrogen bonds. First only the motion of a single hydrogen bridge was considered. Using the Guanine-Cytosine basis set only a single minimum was found. When a much larger double-zeta-plus polarization basis was used, a very pronounced shoulder was predicted but no minimum. It was only when the coupled motion of the two hydrogen bridges was con- sidered and a double-zeta basis set was used that the calculation yielded a double minimum. (II) Potential Energy Curves The Double Minimum Potential The bulk of physicochemical evidence presently available indicates that the H atom is localized near the donor atom in the majority of H bonds, even when A and B are similar atoms. This situation is represented by an unsymmetrical potential energy curve (Figure 28) where the lower of the two minima is located near the donor atom. The first experimental evidence for an unsymmetrical double minimum potential in a H bonded 141 system, was provided by (A) vibrational spectroscopy. In 1959 Bell and Barrow226 showed that the first and second overtone O-H stretching vibra- tions of ethanol, phenol, and p-nitrophenol exhibit a pair of bands in the presence of bases. The height of the second potential minimum was shown to depend on the strength of the proton donor and the acceptor strength of the base227. The same concept was used to interpret the split- .7 ting of the first overtone of the NHZ Symmetric stretching vibration band i . ., "(Henry'- in a number of intramolecularly H bonded ortho~aniline5228. Some of the basic criteria for the existence of a double minimum potential in a H bonded system, with the second minimum at about the)’= 2 level, are (1) splitting of the V02 but not the V01 band; (2) an increase in the magnitude of the splitting with increase in the strength of the H bond; (3) a reduc- tion of the splitting on deuteration; (4) a temperature independent inten- sity ratio for the pair of bands resulting from the splitting of V02, since both transitionsoriginate in the ground state. The band splitting is con— centration independent in an intramolecularly H-bonded system, since the effect originates in the monomer. Some of the strongest supporting evidence for a double minimum po- tential comes from (B) neutron diffraction studies. Hydrogen possesses only a single scattering electron and accordingly exerts rather an insig- nificant effect on the anay scattering. With neutrons, however, the scattering is by the nuclei of the atoms and it happens that the scattering by a proton is not much below the average of all the elements. As a result of it is possible by studying the diffraction of neutrons to produce projections of "neutron scattering density" in which the protons are quite accurately located in relation to the other stems in the struc- ture. For example neutron diffraction studies of KH2P04 show the proton 142 (a) symmetrical double minimum ......... potential energy < << n u n potential _b.o_r;ier. A R(A—B%+ B (b) (c) symmetrical single minimum unsymmetrical double minimum a. OLV V A R(A-B)-* 3 A R(A—B)—> 3 Figure 28. Potential Energy Diagrams for the Motion of the Proton in a Hydrogen Bond A-H- - ~B. ”hwy ‘1; 2:1,“? 1" w i a 4 ”T7 i“ 143 as an elongated shape between the two oxygen atoms which are separated by 2.49 $229. This distribution could be interpreted in either of two different ways: as a centrally located proton with very anisotropic motion or, more likely as a disordered distribution of the protons between two possible positions, one on each side of the mid-point between the oxygen atoms. In the case of ice the 0-0 distance is much longer, being 2.76 X and the two possible hydrogen possitions are much farther apart (about 0.74 2) so that the disordered protons appear as separate well-defined peaks. In the study of Peterson and Levy230 on deuterated ice, the magnitude of the deuteron peaks on the neutron-diffraction projections is consistent with a half-deuteron being associated, on the average, with each of two possible positions on the lines joining the pairs of oxygen atoms. Other indirect evidence comes from (C) dielectric saturation experimentSZBI. Piekara studied alcohol dimers and found a positive saturation effect. His explanation was as follows: in a system of dimers, transitions both from state 1 (left well) to state 2 (right well) and the inverse transitions leading from state 2 to 1 occur, as a result of which equilibrium 12:22 (O-H°--O;=iO'...H-O+) is established, so that the system contains a number of dimers with the proton shifted. When an external electric field is switched on, the dimers whose O-H---O bonds are directed approximately in accordance with the field produce l-—+2 transitions more readily, since for them the barrier height is smaller. This leads to a new state of equilibrium 15:32 more to the advantage of state 2, ”field-stimulated proton shift". As state 2 possesses an electric moment due to the hydrogen bond #2 that is greater than the moment #1 in state 1, the shift in equi- librium necessarily involves an increase in the dielectric permittivity due to the external electric field. In dimers, this increase greatly prevails 144 over the decrease from the usual saturation effect due to alignment of dipoles and, as a result, there is a positive saturation effect. Other indirect evidence on the double well picture comes from data on (D) compress ibility232 233 , 234, 235 , (E) thermal expansion the change in 236 change of intermolecular distance on deuteration and N.M.R. studie5237. bond energy ‘ 1"?“- 415 7 At this point it should be mentioned that the interpretation of the ‘f‘f—s e~mffth above experiments is not as decisive as might be desired. For example even one of the best evidence the split of the IR bands can be due to an V 238 overtone of the AH bending mode, enhanced by Fermi resonance with‘vstr One of the early views of H bonding was the concept of 'hesohydric tautomerism" according to which the H was thought to resonate very rapidly between two equally probable positions, one near the donor atom, and the other near the acceptor atom. This concept would require that the potential energy function for the H atom either posses two equal potential minima with a low barrier between the two (Figure 28-b) or be symmetrical, with one broad minimum (Figure 28-c). Hadzi and collaborators239 have classified hydrogen bonds of the OA-H---OB type into four groups. In group A are included the symmetric single minimum H bonds; in group B, the symmetric double minimum H bonds with a small potential barrier; in group C, the symmetric double minimum H bonds with a potential barrier higher than in group B; and in group D, the H bonds with an asymmetric double minimum potential curve. The H bonds of type A and B are uniformly very short, with R(OA"°OB) varying 0 from 2.4 to 2.6 A. Some typical single minimum H bonds are observed in KH maleate, NaH diacetate, KH dibenzoate, KH bisphenylacetate. The IR spectra are characterized by the absence of a\101(0-H) band in the region 145 above 1800 cm'lg the PMR signals are narrow and weak and remain unchanged at low temperature. The H bonded compOunds belonging to type B (KHZASO4, NH4H2P04, CaHPO4, NaHCO3) exhibit two O-H bands in the 1900-3000 cm-1 region, separated by 300—500 cm'lg the PMR signals are strong and narrow . 0 240,241 at room temperature and only slightly broader at -180 C . $1112 The Hydrogen Bond in the Excited State g A molecule in its lowest excited electronic state has a different electronic distribution than the molecule in its ground state. If the chromophoric portion of the molecule is involved in Hrbonding one would expect changes in its strength. In this way electronic absorption or emission spectroscopy may provide experimental evidence about the H- bonding. Kasha242 first discussed the effect of hydrogen bonding on absorption spectra, pointing out that absorption bands corresponding to nx* transitions should be 'blue shifted" in hydrogen-bonding media. Bayliss 18 and McRae 2 considered hydrogen-bonding interactions to be a special case 43 of dipole-dipole interactions. However, Pimentel2 pointed out that dipole- induced dipole and dipole-dipole interactions produce small solvent shifts compared with those due to hydrogen bonding. He discussed the importance of hydrogen-bonding effects compared with other solvent effects and pointed out the role of the Franck-Condom principle in hydrogen bonding. In the case when hydrogen bonding is stronger in the ground state than in the excited state (Figure 29) the hydrogen bond energy We < Wg and the exci- tation energy implied by the Franck-Condom Prindiple is labeled w. In the case where the hydrogen bonding is weaker in the ground state then we) wg. Solvent shifts due to hydrogen bonding can be formulated as follows: 146 ENERGY R(A..B) Figure 29. Hypothetical Potential Energy Curves for the Formation of a H-Bond A-HH'B in Ground and Excited States, (1) w) w , 8 e (2 ) wg < we. 147 Va - yo = A); = wg - we + we Vf - Vb = Ab% = Wé - We - wg When hydrogen bonding is stronger in the ground state We < Wg so A»%)>0, i.e., a blue shift which exceeds Wg - We by we will be observed in absorp- tion. Similarly in emission a shift less than Wg - Wé by wg will be ob- served (either a red or blue shift). When hydrogen bonding is weaker in the ground state Wg ('Wé, both emission and absorption spectra will show red shift. The well characterized H bonds have energies in the range 1- 7 kcal/mole (350-2500 cm'l). According to the above discussion, a blue shift in absorption may exceed the ground state hydrogen bonding energy, hence, the blue shift is expected in the range of 350~2500 cm'1 or larger than 2500 cm-1. But a red shift in absorption should never be as large as WE, so should not be larger than 2500 cm-1. So a red shift in the absorption spectrum indicates that the hydro- gen bonding is stronger in the excited state. This indicates that the base strength of the acceptor has increased in the excited state. The spectra of nitrogen heterocyclic bases in which the lowest transition is us* have been found to undergo red shifts upon hydrogen bonding. This is because of electron redistribution around the hydrogen bonded nitrogen, increase in charge density, which increases its basicity in the excited state. PROTON TRANSFER REACTIONS IN HYDROGEN-BONDED SYSTEMS: GROUND AND EXCITED STATES A proton transfer reaction in which a proton involved in a hydrogen bond A-H"°B is transferred from A to B, belongs to the broad classifi- cation of acid-base reactions. The transfer may occur along a potential energy curve similar to the ones depicted in Figure 28 by two mechanisms, direct transfer over the barrier and/or through quantum mechanical tunneling 148 through the potential barrier separating the two minima. The phenomenon of tunneling will be discussed more extensively later. The transition from a H bonded complex A-H-o-B to a proton transfer complex A-n-H-B+ in a nonpolar solvent is accompanied by significant changes. For example, in the IR spectra, the disappearance of V0H(A-H) bands and the appearance of V01(B+-H) bands have been observed in the IR spectra of mixtures of some carboxylic acids ranging from acetic (the weakest, pKa:= 4.76) to tri- fluoroacetic (the strongest, pKa = 0.23) with pyridine in CHCl3244. Evidence for ground-state proton transfers has also been found in ultraviolet absorption studies. Baba et a1.245 investigated the absorption spectra of the p-nitrophenol-triethylamine system as a function of amine concentration and found absorption bands corresponding to a hydrogen-bonded complex and a proton transfer complex. The first observation of excited state proton transfer seems to have been made by Weber246 who noticed that the fluorescence of 1-naphthylamine- 4-sulfonate changed color as the pH of the Solution was altered, although no corresponding change was observed in the absorption spectrum. These observations were later explained247 as an excited-state ionization of the compound due to the fact that it becomes more acidic in the excited state. At low pH values, the cation of the molecule fluoresced; at high pH, the anion was the emitting species. Since ionization occured only in the excited state, the absorption spectrum remained the same. Since Forster's observation numerous cases have been found of excited state 248 that the shift of the proton transfer. For example it has been found fluorescence maximum of beta-naphthol in C6H6 brought about by the addi- tion of triethylamine is quite large (4000 cm'l) and is almost equal to that due to the ionic dissociation in aqueous solution (4200 cm'l). 0n 149 the other hand, the shift of the absorption spectrum (11b) of beta-naphthol caused by hydrogen bonding with TEA is only 370 cm-1. These results suggest strongly that ion pair formation, due to complete proton transfer rather than a hydrogen bond formation, occurs in the fluorescent equilibrium state. Intramolecular proton transfers in the excited state have also been observed. Weller249 found two fluorescence bands in salicylic acid and its esters, one of which corresponded to a normal fluorescence. The other fluorescence was shifted to much longer wavelengths; be attributed this band to emission from a proton transfer tautomer in which the hydroxyl proton is transferred to the carbonyl group. Stokes shifts, Whidl have been attributed to excited state intramolecular proton transfer, have . . 250,251,252 also been reported for the hydroxynaphth01c ac1ds . 253 2 4 2 5 Weller ' 5 ’ 5 has performed a number of experiments concerned with the kinetics of Proton transfer in the excited state. He discussed methods of determining the excited-singlet-state dissociation constants, pK*, from spectroscopic data. Jackson and Porter256 have applied the same techniques to determine the pK of molecules in the lowest excited triplet state. Their results showed that the pK of the lowest triplet is usually close to that of the ground state rather than to that of the lowest excited singlet state. Recently nanosecond time-resolved spectroscopy has been used to investi- gate the kinetics of excited-state proton transfers. Brand et a1.257 studied the excited state ionization of beta-naphthol. Proton transfer rates of the order of 107 3.1 (room temperature) were found and the results agree very well with results obtained using Weller's photostationary experiments technique. Ware et al.258 applied the same technique to study the intra- molecular proton transfer in the excited singlet state of 3-hydroxy-2- naphtholic acid. 150 PROTON TUNNELING One of the most unusual and interesting results of quantum nechanics is the prediction of tunneling, i.e., the ability of a particle to exist in, or pass through, a region of space where its total energy is less than its potential energy. According to classical mechanics, such a phenomenon is impossible. The physical importance and consequences of tunneling have been recognized since the very earliest days of quantum mechanics. Hund259 discussed the probability of intramolecular rearrangements via tunneling in 1927. As early as 1932, Wigner260 discussed tunneling with a view aimed at chemical kinetics; in the same year, the tunneling mechanism responsible for the doubling of certain spectral bands of ammonia was clarified261. When proton transfer is very fast, such as in acid-base reactions in aqueous solutions, there exists the possibility that the proton in the incipient H bonded complex A-H°-°B tunnels through the po- tential barrier between the two minima. This possibility was pointed out by Bell in 1935. Since then, a great deal of indirect evidence has been accumulated which indicates the occurrence of proton tunneling in ultra- 265 262,263. Johnston264 and Caldin have provided re- fast proton transfer views of various aspects of proton tunneling in ordinary chemical reactions, and the significance of tunneling to the understanding of the hydrogen 227’266’267. Iowdin268 has discussed tunneling Ibond has been well-described from a biological viewpoint. In the following we will outline a simplified approach for obtaining the transmission coefficient for an arbitrary shape potential. To do so we use the so called JWKB method (after its proponents in quantum mechanicS, Jeffrey, Wentzel, Kramer and Brillouin). Let us start from the l-dimentional 151 Schroedinger equation*: (~h2/8r2m)82¢/6x2 + V(x) We introduce the notation pU’D'exp( ZPE/RT) (117) a * where VH and VD are the frequencies of the vibrations of the transition state which become translational motion (a)y AH and )hD are the stretching frequencies of the one-dimensional molecules AH and AD, and vg=and ngare the frequencies of the real vibration in the transition states of AHB and ADB (b) (a) <-—A H B—-) symmetric (b) +—-A H—94-B anti-symmetric ZPE = ZPE (reactants, hydrogen) - ZPE (reactants, deuterium) — ZPE (ac- tivated complex, hydrogen) - ZPE (activated complex, deuterium), where ZPE = zero point energy. westheimer has shown that if k1 = k2, i.e., if the force constants binding H (or D) to A and B are equal then fill/yr: (2)? VAD/VAH; (a? and vt/v:= 1 so the entire preexponential term of equation 117 is unity. If k1)> k2 then again the preexponential factor is approximately unity, since the ratio of the frequencies of the "symmetric"vibration now 'hancels" the * * ratio of the frequencies of AH and AD and the ratio VH /WD = 1. Re- gardless, then,of the ratio of the force constants, the preexponential factor estimated for this simplified model is near unity. The factor that determines the isotope effect is then just the difference between the zero point energies. Solvent Kinetic Isotope Effect "Solvent isotope effect" is a term frequently used in discussions of kinetic and equilibrium processes in light and heavy water for the part of the total isotope effect which is attributed to isotopic substitution of the solvent. It has been said to be due primarily to differences in struc- 2 2 6 ture of die two waters 74’275’ 7 . That differences exist is evident from differences in thermodynamic properties of the two liquids, in the heats 277,278 of hydration and solubilities of salts in the two waters Swain -7 ‘(FVL l, i“ , l5. _Ji E > 0(0) Earn kD >l 'l he vibrational origin of (he cllccl of dculcr ~ rium substitution on reaction rates. Figure 31. Primary Kinetic Isotope Effect. The Vibrational Origin of the Effect of Deuterium Substitution on Reaction Rates. 159 279 280 and Bader ’ proposed a model where the solvent isotope effect is ex- plained in terms of zero point energies associated with the hindered trans- 1ations and hindered rotations (librations) of H20 and D20 (as well as the internal vibrations). The hindered translations are of very low frequency (Vl70 cm'l) and are treated classically. The librational fre- quencies of H20 and D20 are different for example at 10°C EH20 = 710 cm-1 and VDZO = 530 cm'l. The conclusion is that solvent isotope effects can be calculated rather accurately by (l) determining the internal vibrations of any molecules which have protons which would exchange with water and become deuterons in D20 solution, (2) determining the librational frequencies of the water molecules solvating each molecule appearing in the reaction. As a crude generalization, it can be expected that ion-forming reactions will have a solvation isotope effect contribution to k corresponding to a net increase in HzO/kDZO > 1’ structure breaking (net decrease in librational frequencies) on going from reactants to transition state, while effects ( 1 are expected for ion-des- troying reactions, and effects close to unity are expected for reactions which do not produce or destroy ions. Solvation isotope effects will gen- erally be small, probably no more 20-301 except for small ions which can have a considerable effect on the water structure (librational frequencies). Deuterium Isotope Effects on the Rates of Non-Radiative Transitions There are two general approaches in the literature to the problem of non-radiative transitionszsl. In the time dependent approach, initially a set of zeroth-order states is selected which are not the true stationary states of the system, but which are eigen functions of an inexact Hamil- tonian. The difference between the true Hamiltonian and the inexact Hamiltonian provides a perturbation which causes ”radiationless transitions" . _. (2.7:? 160 between this initial set of states. According to the stationary state approach absorption occurs not to a non-stationary state but to a stationary state of the system. Any mixing of states due to an intramolecular per- turbation would occur before interaction with an external field. The most important theory using the time dependent approach is the theory of Robinson and Forsch282. The popularity of this theory is mainly due to its simplicity. teirjnnae7 . \ An account of the stationary state approach has been given in detail in a a» ‘-—.‘ 28 review article by Jortner, Rice and Hochstrasser . Vrc’ In order to understand the effect of deuterium substitution on the rates of non-radiative transitions, we will give a very brief account of the theory of Robinson and Frosch. We consider a molecule interacting with a solvent. The electronic states of the system (for fixed nuclei) are ¢¢ and db with energies Eg<'Eu, and with corresponding vibrational states X2(E) and xu(E) where E is the vibrational energy measured relative to the zero-point energies Eg and Eu respectively. Initially the system is in the state ¢hxu(0) which is an exact eigenstate in the Born-Oppenheimer approximation. If we go beyond this approximation there is a non-zero matrix element of the form Hiu = (m X€(E>lTN|¢uXu(°)> where TN is the nuclear kinetic energy operator. Hg is negligible, except u when E2: Eu - By. Its main effect is that it induces radiationless transi- tion from mu to cg. In polyatomic molecules, where one has a large number of states ¢Lx{(E) and these are broadened by solvent interactions, they merge into a continuum of state density pE. According to Robinson and Frosch the radiationless transition probability per unit time is then given by nr 2 2 2 (u = 4'” PElHtul “n (118) k 161 H;u can be separated into electronic and vibrational components by writing Hen = <¢€(E)lT€ulq’u(O» Tee = (XclTNIXu) So equation 118 can be written as knr 2F/h2 (Zu (u 2 where J? is the electronic factor and F is the Franck-Condom factor, u E describing the square of the overlap of the oantinuum of vibrational states <§(E) with ¢>(0). So while a radiative transition is a vertical transi- u tion between the potential surfaces corresponding to different electronic states, a radiationless transition is a horizontal transition, which in- volves crossing or tunneling from the potential surface of the initial electronic state u to the continuum of isoenergetic vibrational levels of the potential surface associated with the final electronic stateil. The deuterium effect on the rate of non-radiative transitions is hidden in the Franck-Condon factor. Since F depends on vibrational overlap it is sensitive to any isotopic substitution in the molecules which modifies the vibrational frequencies. Siebrand284 has derived the so-called "isotope rule" (blogFi(E)fiBE)E, = (#i/fl)%(510gF(E)/5E)Ei where i denotes the isotope (deuterium) and u is the reduced mass of the oscillator governing F(E) for E = E'. The lower Franck-Condon factor for 'moleculeS'where H.has been replaced by D (C-HE—ac-D) is due to lower fre- quency of the C-D vibrations relative to C-H vibrations so it takes more C-D vibrational quanta to make up a given energy gap with a poorer vibra- tional overlap as a result. This is of course true in cases where the C-H vibrations are important in accepting the energy. For aromatic hydrocarbons it has been shown that when the energy gap is larger than'v4000 cm-§ C-H vibrations are important but wheni(4000 cm"1 the C-C stretching modes dominate and the isotope effect disappears. 162 The most well studied non-radiative transition is the T1-—9SO inter- system crossing and this process is also the most sensitive to deuterium substitution. The result of deuteration being an increase of the phos- phorescence quantum yield and lifetime. For example Tim and Laposa285 found: (TT)D/( 7T)H Phenanthrene 4.13 Naphthalene 8.13 Triphenylene 1.49 The same authors and also Laposa, Lim and Kellogg286 did not find any change in the fluorescence lifetime and quantum yield. Due to the large energy gap between S1 and S0 one would expect a very large isotope effect if internal conversion is important. The absence of an isotope effect shows that at least in aromatic hydrocarbons the Sl——->S0 non- radiative transition is not important. This observation has been generalized to other classes of molecules although there are not experimental data for the proof of this assumption. In cases of partial deuteration of aromatic hydrocarbons it has been found that the triplet lifetime shows a significant 287 288 289 dependence on the position of the substitution ’ ’ . A deuterium isotope effect appears also in cases where not the molecule itself is deuterated but the medium in which the molecule is situated. The first case reported is by Hirota and Hutchison290 who found that the triplet lifetime of naphthalene (guest) in deuterated durene (host) is longer that in the case of protonated durene. Very large solvent isotope effects on the 291,292,293 luminescence of metal ions were found by Kropp and Winsdor For example for Eu(N03)3 the quantum yield increases 18 times going from H20 t0 D20. These solvent isotope effects have been explained within the frame of 163 Robinson and Frosch theory as a slowing of the radiationless transitions in the deuterated solvent due to smaller overlap integrals between the molecular vibrations and the solvent vibrations which eventually will accept the energy. Recently Ermolaev 3 has used the Forster-Dexter theory of energy transferzgs’296 to explain solvent isotope effects. The electronic energy degradation process was supposed to occur due to in- ductive-resonant interaction between and excited molecule (ion) and the vibrations of the solvent molecule297. Solvent isotope effects on the fluorescence quantum yields of several polar molecules were reported by 298 Stryer . The effect was attributed to excited state prototropic reactions. This interpretation has been challenged for at least some of the mole- cules studied by several investigatorszgg’BOO. This point will be discussed later in more detail. PROTON TRANSFER REACTIONS IN 7-AZAINDOLE HYDROGEN BONDED SYSTEMS Introduction-Absorption Spectra The spectra of aromatic nitrogen heterocyclic compounds show a general resemblance to the spectra of corresponding aromatic hydrocarbons. Systems whose spectra are being compared must have the same number and geometrical 01 arrangement of r electrons ; thus, an imino nitrogen in a heterocyclic molecule is equivalent to -CH?CH- group in an aromatic hydrocarbon. The spectra of aromatic hydrocarbons have been compared and classified by 302 , , , . Platt ; his notation is used for electronic tranSitions in heterocyclic systems, also. An aza-substitution does not change the spectrum of an aromatic hy- drocarbon very much; the most general difference seems to be an increase in the intensity of the lLb band. This feature is demonstrated very 164 clearly in the case of quinoline as compared to naphthalene . Similar effects are noted in the comparison of other heterocyclic and hydrocarbon systems. In Figures 32 and 33, we display the vapor phase absorption spectra of indole and 7-azaindole. The lowest energy transition of . . l . . indole has been aSSigned as A-——}Lb, and the second tranSition was assigned to be bA-——;Da. The lLb is more intense in indole than in 1 naphthalene spectrum. In 7-azaindole the .A-——-1La transition is approxi- mately short axis polarized and so the aza-substitution at the 7 position . 1 . . 1 . is expected to perturb the Lh tranSition more than the Lb cauSing the two bands to be closer together in 7-azaindole than in indole. Further, the aza-substitution might be expected to increase further the intensity 1 of the Lb band. Both phenomena can be seen in the displayed spectra. A more detailed analysis of the vapor spectra of these compounds has been performed by R. W. wagner303. wagner also performed Pariser-Parr-Pople calculations on these mole- cules, and in Figures 34 and 35, we reproduce his results on indole and 7-azaindole showing the r electron charge densities in the ground, 11b 1 we also give the charge density differences with plus and La states. sign as increase and minus sign as decrease. These results will be useful in interpreting the excited state acid-base properties of these molecules. Indole Absorption spectra have been recorded in the gas phase, hydrocarbon O and 7' (1L ) are a (3MP), diethyl ether, alcohol and water. The data 96 9 max shown in Table 15. Considering hydrocarbon (3MP) as a reference, there is a red shift in ether and ethanol solution but a blue shift in water. The interaction of 165 .Esuuowom Cowuauomc< uoam> oHonCH .Nm munwwm A53 reexamine new emu _ b ALISNHINI NOIldHOSHV 166 . _ .Bsuuoodm coHuaqucd uoom> oaoucflmumcn .mm ouzmam 25 5023522 com com omN OmN omN omN osm p . i» lb 1. b b w as m J Tm I no "N I "N l H ”N m z a m ‘J—d Figure 34. 34a. 34b. 167 Calculated Charge Densities for Indole Numbers at each atomic position denote n ~electron charge densities in electron units for ground state (top number), 1La state (middle number), and lLb state (bottom number). Numbers at each atomic position denote n -eleCtr0“ charge density differences from ground state (in electron units) for 1L state (top number), and state (bottom number) with plus sign indicating an increase in electron density and minus sign indi- cating a decrease in electron density. 1.046 0.962 1.093 1.012 1.175 0.970 ~0.084 +0.047 +0.163 -0.042 168 0.988 1.133 1.194 1.156 1.003 1.096 1.175 1.101 1.118 1" 1.264 1.480 1.232 1.210 1.382 Figure 34a +0.145 -0.191 +0.168 -0.059 +0.079 +0.005 ‘-‘-‘~l\l +0. 146 H +0.114 '0-270 -0.098 Figure 34b Figure 35. 35a. 35b. 169 Calculated Charge Densities for 7-Azaindole. Numbers at each atomic position denote n ~e1ectron charge densities in electron units for ground state (tOP number), lLb state (middle number), and La state (bottom number). Numbers at each atomic position denote n -e1ectron charge density differences from ground state (in electron units) for lLb state (top number), 30d La state (bottom number) with plus sign indicating an ‘ increase in electron density and minus sign indicatlng a decrease in electron density. 170 0.978 1-180 1.198 1.150 0.951 1.186 1.050 1.022 1.036 1.095 1.192 1.077 0.973 1.128 1.187 1.473 1.356 1.105 1.293 1.440 Figure 353 +0.202 +0.172 -0.247 ~0.012 -0.028 W -0.014 +0.097 -0.018 +0.155 -0.021 N +0.169 -0.368 +0.106 -0.033 :2 Figure 35b 171 Table 15. Absorption Spectral Shifts for Indole in Different Media ( cm—l) Solvent 90,0(1Lb) ‘gmax(1La) 3M? 34843 37594 Ether 34722 _ 36832 EtOH 34782 36832 H20 34965 37105 solvent with the solute in BM? since indole is polar (1L= 2.3D) is mainly of the dispersion type plus dipole-induced dipole interaction. Both interactions should cause red shift compared with vapor phase. In ether besides the dispersion and dipole-induced dipole interaction, dipole-dipole interaction is important because ether has a dipole moment of 1.3D. A further shift arises from hydrogen bonding of the pyrrolic hydrogen with ether's oxygen. In ethanol and water, an additional hydro- gen bond may be formed with the pi-electron of pyrrolic nitrogen. If this nitrogen is less basic in the excited state, a blue shift will be seen. Table 16 shows the shifts (in cm'l) of the absorption maxima of 1 1 the A-—-Lb and A——--La transitions with respect to the hydrocarbon solution. Table 16. Absorption Spectral Shifts (cm-1) for Indole in Different Media Relative to 3MP. State vapor ether EtOH H20 CH2012 111; +501 ~121 - 61 +122 -122 1La +897 ~762 -762 -489 -694 172 Wagner's calculations show that the electronic density on position 1 (pyrrolic nitrogen) decreases in the excited state and that the decrease is larger in die 1La state than in the lLb state. The absorption spectra (blue shifts in water) support this conclusion and give further evidence for the increased acidity of the pyrrolic hydrogen in the excited state. 7-Azaindole Table 17 shows the absorption maxima of 7-azaindole in several sol- vents. In hydrocarbon (3MP) the spectrum shows a red shift for both 1La and 11b compared to the vapor which is a manifestation of the dispersion forces. In ether we observe a further red shift which appears to be mainly due to the hydrogen bonding with the pyrrolic hydrogen. The red shift seems to indicate that this hydrogen is more acidic in the excited state. In ethanol there is further red shift which may be due partly to increased dipole-dipole interactions and also as Wagner's calculations suggest due to hydrogen bonding with the pyridine nitrogen. In water we observe a blue shift compared to ethanol although one would expect a red shift Table 17. Absorption Spectral Shifts for 7-Azaindole in Different Media(cm_1). N 1 ~ 1 Solvent yo 0( Lb) VEax( 1;) Vapor 34495 36258 3MP 34013 34843 Et20 34722 34482 34602 173 at least because of the higher dielectric constant of water. The effect seems to be associated with the higher acidity of water this causes a blue shift because of the solvent hydrogen interaction with the pi-electron charge density at the pyrrolic nitrogen. Excited-State Acid-Base Properties As we already pointed out in the introduction Forster247 and Weller253 were the first to show that electronic excitation may change drastically the acid-base properties of a molecule. If equilibrium is estalished during the excited state lifetime then the pKa* and pr* can be determined in a way analogous to the spectro- photometric determination of the ground state pK's, by fluorometric ti- tration. For this reason we have studied the fluorescence intensity as a func- tion of pH (corrections were made for changes in the absorption spectra) for indole and 7-azaindole. Figures 36 and 37 show the results. Indole As we see both at low and high pH the indole fluorescence is quenched. In the high pH region the quenching is ascribed to the ionization of the pyrrolic hydrogen according to the reaction +0H' *9 +Ho <1.) mm . From the midpoint of the titration curve one gets a pK£* = 12.3. Exactly 304 the same number was obtained by E. Vander Donckt . The excited state 305 PKb* ‘ 12.3 should be compared to the ground state pr = 16.97 which shows that indeed the pyrrolic hydrogen acidity has considerably increased in the excited state. 174 .oomm um oHoon 1H O 3:: O 0H JJ :3 H O U) m :3 8 i" :1 at 1H O a) 0 G d) 0 CD a) H O .‘3 H In a) .C U c: 0 £751 a. (H O U U 0 1H ‘H W re ,— (suun MoquV) Mgsuaiul aouaosaionH 4 AU nmvcoeok c_o_ovc_ #J ~O \D m a) H 5 00 0H L‘n 175 .oomm um oHomnwmnH0 a ox .v< kumuuwpu< mama maocwfim one .oaeamm woumuuamoaoo mfiu How can: mwz cowuwu H m 33qu use; .a: mwm :uwamampma coaumuaoxm .Anuvz noaxoé was AIJ 29.3on um m2m cw oHovchn¢un mo Aunwwmv moaoomouosam can Aumoqv moguauomn<§ loom oom one 00¢ com com \ ‘. /// \\ cu / x C / x N c E / c C / Nu. \ S e E. // \ R / nu u ,, \ L / F / \\ w. /<\ .H A L E R .wm madman (Q-Ol x) NOLLONLLXB HV'IOW ———— 180 m \N \ N H 3 z:— Kl\ l Figure 39. Biprotonic Phototautomerism in 7-Azaindole Hydrogen-Bonded Dimers. l8l )- t: (D Z LU .— 2;. [U U 2 LL! U (0 L11 (I C) D .1 h. l J l J 1 3.4 260 ?80 300 320 WAVELENGTH (NM) Figure 40° Corrected Fluorescence Excitation Spectra of 7-Azaindole in 3MP at 25°C. The Fluorescence wavelengths were 325 nm (Solid Curve, F1) and 475 nm (Dashed Curve, F2). 182 a result of a primary kinetic isotope effect. The sample used was a lO-ZM.solution of 7-azaindole (Nl-D) in 3MP cooled slowly to 77°K. Under these conditions, virtually all 7-azaindole molecules are present as dimers since4AH and equilibrium constant K.for dimerization at room tem- perature are -9.6 kcal/mole and 1.8 x 103M“1 respectively. By studying the kinetics of the excited state reaction we hope to be able to comment on the mechanism of the thermal reaction vs tunneling and on the shape of the potential well in which the protons are located. Excitation Delocalization in 7AI Dimer Before attempting the kinetic analysis of the double proton (deuteron) transfer we have to resolve an important question. Namely if during the time scale of the proton transfer the excitation is localized on one 7-azaindole molecule or is delocalized (spread) over the pair comprising the dimer. The significance of this question is that in the case that the excitation is localized, the two proton transfers are inequivalent while if the proton transfer happens from a stationary state involving both 7-azaindoles then both protons are facing the same energy barrier. To answer this question let us consider two identical molecules, a and b, in proximity in a rigid amorphous solution. If molecule a alone is excited at t = O, the system is then in a non-stationary state where the excitation will oscillate between the moieties. The probability of finding the excitation on b at time t is: ;;b* = Sinzza.sin2(w/2o4u/ht‘l/sinZa) where 2a = arctan(2u/5); die the difference (if any) in energy of the 0,0 bands of the two identical molecules. This difference may arise because of the possibility of interactions with different microenvironments. For the kind of environments used in our experiments (hydrocarbon matrices) 183 the interactions of the solute with the solvent is weak and mainly of the dispersion force type. In such a case 6 is expected to be very small at least compared with the interaction u = (¢*b*) between molecules a a a a and b. If 6 = 0 then we Obtain the familiar expression: p * = sin2(7r/2-4u/ht) ab where the prdbability of finding the excitation on molecule b oscillates between zero and one with a half period of t = h/4u. Although we realize that the transfer is oscillatory and, therefore, different from a kinetic first order process, we may consider the reciprocal value of that time as the "transfer rate" of this process: k = 4|uI/h (3'1) (119) As we have already discussed the usual spectroscopic practice is to express the intermolecular interaction operator Vab in a multipole series and keep the first nonrvanishing term in this case the dipole term. The interaction then is considered as arising from the coupling of the two transition di- polesM, u = I<|M|2/R3 x is a numerical factor which depends on the mutual orientation of both molecules, K= cos gab - 3cos 0aocost9b where gab is the angle between the transition moment vectors of both molecules while 0a and 0b are the angles between these reSpective vectors and the direction a——+b. If the lowest energy state of 7-azaindole in the dimer is the lLb state then k31 -1 while if there is a state inter- change as suggested by Taylor, El-Bayoumi and Kasha208 than k:: -2. Al- though in the point dipole approximation R is not well defined an approxima- tion value of R = 5A seems appropriate. In this way u can be approximated to be of the order of 103 cm-l. Using equation 119, we calculate a "transfer rate" of 1014 s”1 . Even in the case that u has been overestimated by say 184 an order of magnitude still the state of the dimer is a stationary state involving both 7-azaindole moieties. Kinetic Analysis let us now consider the kinetic analysis of the double proton trans- fer. The kinetic scheme considered is the following: D* :2 (2T)* /\ /\ kFT kIn? J \ D + hVD 2T + th 2T In the scheme we have pictured the two tautomers arising from the double deuteron transfer as sharing the excitation and constituting a single entity. Although this may be true by the same arguments used for the dimer it is not necessary for the following kinetic treatment (this applies to dimer too) since the reverse deuteron transfer is independent of concen- tration since the two tautomers are trapped in the same cavity of the rigid medium. So the two macroscopic rate constants kTD and kDT are first order rate constants. The kinetic equations after a.6-pulse excitation will be completely analogous to the ones describing the intramolecular excimer interaction in 1,3DNP. d[D*]/dt = -x [13*] + km. [(2%)] (120) d [(2T*)1/dt = -y [(2%)] + kTD[D*1 (121) where kD = kFD + km = km: + kIT x = 1(1) + kTD y = 1‘1: + km: Solving equations 120 and 121 with the initial condition that [(213%)] = 0 at t = 0, we obtain 185 art u... e: 3 :53 mo 538% 822.5 833% as: .55 4 .He manage 2313 Ill , , 2.212 ...... , a: n- O kusuaiul aauesaJonM 186 [D*]=|bflO/(x2-A1)[(x2 _ X)e-A1t +|(x _ k1)e-A2tl K2T)*] = [D*]OkTD/(x2 - 11)-(e'xlt - e-xzt) " Z = + + - + 2 where Al,2 J5(x y {(y x) 4kTDkDT) (l 2) and as usual changing to intensities we obtain _ _ 'Klt 'Kzt IDCt) - kFD[P*l/[D*10 - kFD(x2-x)(e +Ae )/(A2-A1) H where A (x - A1)/(A2 - x) -A a. [<2T>*1/ W1. = was From equation 122 we get that x1 + x2 = kD + kT + kDT + kTD (123) . = + + Al A2 kTDkT kaDT kaT (124) It -A 1T0» -e 25/(12-11) Rate Constants for Proton Transfer The dimer decay is found to be described (as expected) as the sum of two exponentials with unequal amplitudes. The values of)(1 and x2 found after deconvolution of the dimer decay are: 11 = 3.33 x 108 s’1 8 - x2 = 1.11 x 10 s 1 The decay part of the tautomer fluorescence is described by a single exponential RT = 1.0 x 108 8.1. The only unknown quantity that we should obtain in order to be able to use equations 123 and 124 and thus obtain the proton transfer rates is kD, the dimer decay constant in the absence of proton transfer. Since we can not prohibit the proton transferwfirom happening after the dimer is excited we have to use a model. As such we chose Nemethyl-7-azaindole which can not form dimers, and therefore no proton transfer can take place. The Observed lifetime of Nemethyl-7-azaindole of 3.8 ns does not correspond directly to the lifetime of the dimer. The reason for this lies in the fact that the excitation is spread’ to both members of the 187 dimer. Since the Einstein A coefficient for a transition between states u and l is given by _ o: 3 Aul - l/T 8nhV ngUI/gu where Bu is the Einstein B coefficient for induced emission and g1 and 1 gu are the degeneracies of the lower and upper states respectively and 1° the radiative lifetime, then T°/(kFD(kDT+kT>) h d d d h h 3° (F2/F1)H/(F2/F1)D = kTD/kTD.(kDT + k'1:)/(k1)1: + RT) (128) 191 l l ‘1 l I T I". ~ I \ {\ “ ' ‘1 \ : 1, x _. , \ - I \ r: I \ a 1 \ E "" l \ '1 5 I \ :3 .1 ' \ o - l ' \ .. 5 I 1' \ 8 1 \ 32‘ I ‘ O P- -\ b \ .- 3 / \ \ u. I, \ \ \ .. \ \ \ \ ... \ \ \ l l I 350 400 450 500 550 WAVELENGTH (NM) Figure 43. Effe t of Deuterium Substitution on the Fluorescence of 7AI (10' M) in 3MP at 77°K. The Degassed Solutions were Excited at 285 nm.Using Front Surface Excitation. Solid Curve 7AI and Dashed Curve 7AI-d1. 192 Since the isotope substitution is not expected to change the radiative lifetimes. Further since kT + k‘DT CikT and since kT = k’E‘T + knr’ k}; is expected to be equal to k; except in the improbable case of a very large and specific effect of one H/D atom on the non-radiative rates. Under these assumptions equation 128 can be simplified to h d (Fz/F1)H/(F2/F1)D = kTD/kTD (129) The experimental isotope effect measured as in equation 129 as the ratio of FZ/Fl for protonated and deuterated 7-azaindole in 3MP matrix at 77°K is 2.9 so (kh /kd ) 2.9 TD TD 77°K = h - Using our result for kd we obtain a kTD = 2.5 x 108 s 1. Further as we TD discussed in the introduction this kinetic isotope effect is connected to the zero-point energy difference so: h / d k = - ZPE RT kTD TD exp( / ) h d _ where kTD/kTD - 2.9 The ZPE thus obtained is around 60 cm’l. If the activation energy for the double proton transfer is 500 cm."1 then the activation energy for the double deuteron transfer is around 560 cm-1. This is in good agree- 1 ment with the estimated activation energy of 500-600 emf considering the double deuteron transfer at 77°K as a thermal process. It appears that a mechanism that involves motion of the deuterons over the potential barrier can explain the experimental data at 77°K‘because of very small energy barrier. Tunneling vs Proton Transfer Ingham and El-Bayoumi311 have reported that FZ/Fl does not change significantly even at 4°K (the authors cite various difficulties preventing accurate evaluation of the ratio). under the assumption that the radiative and non-radiative rates do not change from 77°K to 4°K, which is reasonable 193 since any vibronic effects reaponsible for these changes will be the same at 77°K.and 4°K, one gets that (kTD)77°K== (kTD)4°K° For the reaction to be a thermal one at this temperature (4°K) the activation energy should be around 30 cm-1 a result which points strongly to quantum mechanical tunneling. In the following we will try to build a simple model for the case that the tautomerism is following a tunneling mechanism. In the introduction we discussed briefly the phenomenon of proton tunneling and we derived a formula for the transmission coefficient through a parabolic barrier. This derivation implied a free particle striking the barrier from the left. In the case of the hydrogen bonded dimer the hydrogen (deuterium) is initially oscillating in the left-hand well between its two turning points with a characteristic frequency )’, that is now we are dealing with tunneling of a bound particle. If the oscil- lation frequency is V then the particle hits the barrier from the left %)/times per second and each time it hits the barrier the probability of getting through is g, the transmission coefficient. In this way we can define a tunneling rate constant as k = % g(s"1) and a characteristic tun time, tunneling time, as = l/ktun. One important question is if the Ttun vibrational levels of the well are occupied in a way dictated by Boltzmann statistics or tunneling occurs from a non-thermal distribution of states prepared by the action of light. In our case the tunneling time is of the order of 10”7 3 while the vibrational relaxation times are of the order of 10712 3, Obviously we have a Boltzmann distribution and more over because of the low temperature (77°K) we can safely assume that only the zeroth vibrational levels are populated. Let us now try to estimate the "hit frequency" . If one assumes that the bottom of the well is approximately parabolic one can make an estimate of V. A parabolic well is characterized 194 by a potential of the form V =G(x - x')2/2 A quantumwmechanical wave packet is then going to oscillate with the same frequency as a classical particle. The Hamiltonian may be expressed in the form H = p2/2m + %a(x - x')2 Consequently, if 2b is the distance between the two classical turning points, one has the relation E = 35ab2 = %hv and since ¢1= 4a2mV2 v= h/(4a2mb2) For b in A and for deuterons v = 5 x 1011/1»2 sec-1 0 For hydrogen bonded systems a value of 2b = 0.6 A seems reasonable and gives a frequency of )/= 5.6 x 1012 s"1 In the case of the 7-azaindole dimer both deuterons must change positions in order that tautomerization happens. From the previous dis- cussion on the delocalization of the excitation we can say that the two deuterons are moving in two identical potential wells presumably indepen- dently. If both hit the barrier on the correct side the probability that both tunnel through is g2. Since we assumed that the motions of the two deuterons are not coupled the phase difference of their vibrational motions will be random. It is reasonable to propose that only particles that are both moving toward the correct side of the barrier can give rise to tau- tomers. Since there are four possibilities the tunneling rate constant will be k = fiVg As we show in the introduction g can be written as g = exp(-fizxa0(2mV0)%/h) At 77°K,:22: .9024:ka— o_.o mod 8.0 v0.0 .2; oozonuomnd .2170.“ x 0..” one» .oomm um xoamfioo Iwmoaom 0:”. mo down—wade“; No.0 . a . . as «new; 30~vssossv / Downiaj 203 solvents which Forster and ROkos regard as evidence for a direct relation between the mechanism.responsible for the red shift and that responsible for the quantum-yield change. A similar correlation has been reported by Viktorova315 for several organic fluors in both single and bicomponent solvents, the quantum yield passing through a maximum value with increasing Stokes shift in the emission. The author explained this phenomenon in terms of discrete centers in mixed solutions corresponding to molecules of the fluor associated with different numbers of solvent molecules. Ei- singer and Navon300 have suggested that the predominant non-radiative re- laxation process for indole is a tunneling process between the potential surface corresponding to the fluorescent state and that corresponding to the ground state, followed by rapid vibronic relaxation. The authors give an isot0pe effect OECDDZO dole, which they attribute to the internal conversion process. A1%20 = 1.30 for indole and 1.19 for l-methyl-in- walker et al.316 suggested the important non-radiative path.may be electron transfer to the solvent. This possibility was suggested by the results of Grossweiner and Joschek317. walker et a1. ascribed the isot0pe effect to "differences in the salvation state of electrons in D20 and H20". Ricci318 gives aGDD OKQ‘ZO = 1.49 for indole and = 1.19 for l-methyl- indole. The fact that tie ratio is not unity for l-methyl-indole was at- tributed by the author as due to a tunneling type interaction in which the highly energetic carbon to hydrogen stretching vibrations transfer energy via coupling to the 0H and OD stretching vibrations of the solvent molecules. In D20 the lower frequency 0D stretching vibrations cannot dissipate the vibrational energy as effectively as a fluorescence quencher as Héo resulting in 19 per cent more fluorescence in D20 than H20. Ricci also found that the isotope effect increases as the methylation of the indole ring increases 204 reaching a value of 3.46 in the case of 2,3,6-trimethy1 indole. The author suggests that as the number of methyl groups on the indole ring are in- creased the importance of non-radiative decay through the CH stretching vibrations decreases resulting in a proportional increase in the importance of the isotopically dependent deprotonation mode of non-radiative decay. He considers two possibilities for the deprotonation step: (a) rate limi- ting proton transfer followed by a fast non-radiative decay of the anion ./ \* _H+ \* \ «Moo» N H _ or (b) proton loss as part of the internal conversion process: \\ a Q N 4 N\ + H+ + energy H Several other authors support the one or the other explanation (see for 319 320 example E. P. Busel et a1. ; M. S. walker et a1. and E. P. Kirby 321 et al. ),but no clear answer has been given yet. we have found that 7AI's fluorescence is quenched considerably in water. For example the fluorescence quantum yield of 7AI in a dilute solution in 3MP (no F2) is 0.24 while the fluorescence quantum yield in water is 0.03, an eightfold decrease. we also found that 7AI's fluorescence is greatly enhanced in D 0 being 3.63 times higher than in H20. Some results are 2 summarized in Table 19. By looking at the results in Table 19, we imme- diately see that the isotope effect is connected to some non-radiative deactivation path. The larger the quenching, the larger the isotope effect. So (DFUAI, H20)< (1%,(indole, H20) and QZO/QZO)7AI) Q20%0)indole' Second we see that when we replace the pyrrolic hydrogen with a methyl 205 group the quantum yield increases with a simultaneous decrease of the isotope effect and that actually CD N-methyl- 7-azaindole id) N-methyl- indole and ((11320, O)NM7AI 2'- (%20/qg20)m1 The above result shows very clearly the importance of the pyrrolic hydro- gen for the quenching path. In the case of 7AI the isotope effect differ- ence between 7AI and NM7AI is very large. In the case of indole vs 1- methyl-indole the difference is small and many authors have argued (Eisinger and Navon primarily) that the differences between indole and NMI may be just due to different coupling with the medium. This does not seem likely at least for 7AI. So at this point we can conclude that the N—H group participates in the quenching. Further, by comparing indole and 7AI quantum yields we can say that the quenching function of the pyridinic nitrogen can only be exemplified in the presence of the pyrrolic hydrogen or in other words the deactivation mechanism involves interaction with both sites. we also studied the fluorescence intensity of 7AI in D20 as a function Table 19. Fluorescence Quantum Yields of Indole, 7-Azaindole and Methyl Derivatives in water and Deuterium Oxide. Compound cDF(H20) (DF (D20)IHom osu mo aowumfiuo> Do to In V— mp up —F Op 0 m n o n v a . n u _ a — d a d 1.1 «.eeEoEA .we seamen othfi/ozotl> “2 'v C? v3 n6 A90 no 209 the well known Stern-Volmer equation IO/I = 1 + kQ [01 where I0 is the fluorescence intensity at neutral pH, kQ is the quenching rate constant («emole-1 8-1),‘718 the lifetime at neutral pH and Q (quencher), in this case is the GET. Figure«49 shows a plot of log(Io/I) vs -10ga0He. The plot has two linear sections intersecting at approximately the titrimetric pK£*. The sudden change in slope may mean a change in quenching mechanism. If in the high pH range the quenching is due to acid dissociation then this result may be considered as supporting the idea that at "neutral" pH the quenching involves a more complicated mechanism involving both the )N-H and 3N: sites. Bowen325 was the first to show the existence of activated quenching. That is: knr = Aexp(-EA/RT) (131) The quantum yield of fluorescence is given.by CDF = kF/(IH‘ + knr) or CDFm1 = 1 + kleknr or using equation 131 for knr HIDE"1 - 1) = %-1Aexp(-EA/RT) (132) Equation 132 provides a means for determining the acrivation energy EA by plotting (q%-1 - 1) vs l/T. we have studied the temperature dependence of 7AI's fluorescence in H 0 and D20 solutions (neutral pH). Absolute fluorescence quantum.yields at different temperatures were calculated by comparing the fluorescence intensity at a particular temperature with that at room.temperature((P% = 0.031). The results are displayed in Figure EN" The activation energy for both H20 and D20 is about the same 6 keel/mole while the preexponential factors 210 7' Azaindole in H20 - Log (0H) Figure 49. Quenching of 7AI's Fluorescence in Water by Hydroxyl Ions. lO/I vs -log(OH-) Where IO is the Intensity at Neutral pH. 70- 60.. 50»- 40.. 20— 'l ‘1’, -1 A 01 o~\loo~oo l Figure 50 . 211 0 Water A Deuterium Oxide Determination of the Activation Energies for the "Activated Quenching" of 7AI's Fluorescence in Water. Circles: Water, Triangles: Deuterium Oxide. 212 ‘AHQO and ADZO are different. Eisinger and Navon300 performed exactly the same analysis for tryptophan and found that again the activation energies (7 kcal/mole) were the same in both H20 and D20 and that the isotope effects can be traced to the frequency factors of the acrivated quenching. In this way a better expression for the isotope effect is: -l _ -l _ = = provided of course that the radiative lifetime does not change from H20 to D20 which has been shown to be valid for indole. The fact that the activation energies appear the same in H20 and D20 should not be taken as absolute because minor, but significant, changes may be masked by ex- -1 perimental error. Since kF is of the order 107 s then the preexponential factors are of the order 1012. If the activated quenching is due to an 326 activated intersystem crossing , one may expect a spin prohibition factor of about 10"4-10”6 which would result in producing a preexponential factor 7--109 (A for unimolecular processes, is usually 1013). The Observed of 10 preexponential factors seem.to disfavor an intersystem crossing process while solvent induced internal conversion process as suggested by Eisinger and Navon can not be excluded. Kirby and Steiner321 and Busel et al.319 have also studied the acti- vated-quenching of indole and indole derivatives in water. In their treat- ment they included a temperature independent deactivation mechanism so that equation 132 was generalized to (CDF‘I - 1) =-' kF-lAexp(-EA/RT) + kF'l‘knr' where knr' does not depent on temperature. Their analysis showed that the activation energies and the temperature independent quenching kér were the same in H 0 and D20. The temperature 2 independent process has been attributed to an Sl—---)T1 process while the temperature dependent process was Busel et al. to excited state acid-base 213 reactions while Kirby and Steiner do not give a specific assignment for the process although they disfavor the electron ejection idea. As we have seen the problem.is very complex and probably it is not just one quenching mechanism.as it is usually assumed but more than one operating simultaneously. In the case of 7AI in water solutions of dif- ferent pH.we can visualize the following kind of complexes 1‘” \ / LES: //U\:/i:> ‘\\ l \\ :4 v '.' I'k ,oe .0 H Ii H ' II III Complexes II and III predominating at medium and high pH respectively can revert upon light absorption to either the anion or a tautomeric species IV. Internal conversion from the anion and the tautomer may be efficient in water solutions. The isotope effect may be due to both kinetic isotope effects on the deprotonation of N-H and protonation of >N: (probably more important for structure III) and to suppression of the internal con- version process in the deuterated solvent (effect of 0-D vibrations). At low pH, structure I would quench in a different way than the ones considered above. While studying the quenching of 7AI fluorescence by acids we assumed that the quenching was due to protonation of N7. Although this might be true, we have noticed that the quenching is not followed by the appearance of the cationic fluorescence except at very IOW'pH. A very 327 similar case has been observed by Forster , in his study of 2-naphthyl- amine in water-sulfuric acid mixtures. He postulated a third species (be- sides the -NH2 and -NH3+) where the proton is shared between the amine and the solvent. Weller have suggested that a ring-protonated molecule (carbon 214 protonation) can play the role of the third species. In our case protona- tion in position 3 is possible. 328 Schulman and Liedke have formulated this third species as a stoichiometric complex (exciplex) between the H amino group and the hydronium ion -N"'H-qy+. Their explanation for the H quenching resulting from such a complex formation is that secondary hydrogen bonding between the hydronium ion and the aqueous solvent could provide an efficient quenching mechanism, the electronic excitation energy being taken up by the vibrations of the solvent structure. It appears quite likely that quenching by such a complex is important in 7AI at moderately low pH. We have studied some other aza-aromatics (see structures below) in H20 and D20 solutions. NOne of them showed any significant isotope effect supporting the idea of the cooperativity of the two sites in 7-azaindole. CHAPTER 5 EXPERIMENTAL (I) Experimentally StudieggMoleggles 1. 1,3-Bis-(g-naphthyl)propane_(l,3DNP). 1,3DNP was provided by Prof. M. Szwarc. It was synthesized according to the procedure described by Chandross and Dempster56. The product was analysed with Mass Spectra, NMR, Stoichiometric analysis. Finally we compared the absorption and emission spectra of the received sample with those reported by Chandross and Dempster and found no difference. 2. 1,3-Diphenv1prgggne (1,3DPP). 1,3DPP was synthesized by Prof. D. Earnum through reduction of 1,3-diphenylpropanone-2. The product was extensively purified by fractional vacuum distillation and was tested by gas chroma- tography and comparison.with the reported emission spectra of 1,3DPP by F . Hirayama49 . 3. 41(3-thenlpropyl)-pvrigine (97%)_(1,3PyPP), and 4-Benzylpyridine (99+%) (PygflzP). 1,3PyPP and PyCHéP were obtained from Aldrich Chemical Company and were further purified by repeated fractional vacuum distillations. The lower absorption band in 3MP was essentially that of v-picoline. 4. Phegylgoetic,»PhenylpropionicgggdrIndgn-lfiggrbggylic Acids were obtained from Aldrich Chemical Company. They were further purified by repeated recrystallizations and vacuum sublimation. 5. Indan was also obtained from Aldrich and was distilled under reduced pressure. 6. Indan-2~Carboxy1ic Acid was synthesized according to the procedure of 215 ¥ 216 Perkin and Revay329. The compound was identified by absorption spec- troscopy, NMR and mass spectroscopy. The purification involvent several recrystalizations and sublimations. 7. Indole was purchased from Calbiochem and was recrystallized once from an alcohol-water mixture then vacuum sublimed slowly for one day. 8. l-Methylindole was purchased from Eastman Organic Chemicals and was purified by fractional distillation under reduced pressure. 9. 7-Azaindole was purchased from Aldrich and was recrystallized from cyclohexane three times. 10. N]-d~7-Azaindole was prepared by refluxing 7-azaindole in alkaline D20 for one hour. Mass spectral analysis gave an isotopic purity of 91%. 11. NJ-Methyl-7-Azaindole and 7-Methyl-7H-pyrrolo(2,3-blpyridine (N7- Methyl-Tautomer were prepared according to the method of RObison and 330 RObison . NM7AI was purified by vacuum distillation and NMT was purified by paper chromatography. (II) Solvents ,flgggg. Only doubly-distilled water was used. Ethanol. 200 proof ethanol was distilled through a 1 meter vacuum jacket column. The distillation rate was adjusted such that a very slow rate (about 5 drops per minute) was maintained. Distillation continued until the benzene-alcohol azeotrope was no longer present as determined by an absorption spectrum of the distilled alcohol in a 10 cm cell. That is, the characteristic benzene UV absorption was no longer apparent. Ethanol was then distilled and used as needed. 3-Methylpentane S3MP). A modified version of the purification method of 331 Potts was used. Phillips Pure Grade 3-methylpentane (3MP) was shaken 217 for 30 minutes with a 50:50 mixture of concentrated sulfuric acid and concentrated nitric acid. It was then shaken 3 times for 30 minutes each with concentrated sulfuric acid. This was followed with several shakings using sodium carbonate sOlutions until the 002 production ceased. The 3MP was then shaken several times with water until the water remained clear compared to the initial yellow color it attained with the first shaking. After storing the 3MP overnight over anhydrous calcium sulfate it was placed in a flask and sodium ribbon was added. It was refluxed through a vacuum jacketed 1m column and distilled for use as needed. Passing the distillated through a 1m column of activated Silica Gel did not alter its absorption characteristics so this step was not required in the purification process. Meth 1 C clohexane 2-Meth 1-Butane Iso entane Dichlorqmethppe, Gl - cerol, and p-Dioxane were spectroquality solvents by Matheson Coleman and Bell (MC/B). Methyl Sulfoxide. Spectrophotometric Grade by Aldrich Chemical Company. Deuterium Oxide. 99.81 atom %, 0.2 uc Tritium/m1. International Chemical and Nuclear Coorporation (ICN). Ethyl Alcohol-d. 99% Aldrich Chemical Company. (III) Spectral Measurements Absorption Spectr . All reported absorption spectra were obtained with a Cary 15 spectrophotometer. Eggssigp_§pectra. Most of the fluorescence spectra were Obtained with an Aminco~Keirs spectrophosphorimeter equipped with a high pressure xenon arc lamp and a EMI 9781 R photomultiplier tube. All the phosphorescence spectra were obtained with the above instrument equipped with a rotating can phosphoroscope. 0.5 mm slits were used. 218 The emission spectra of 1,3DPP were obtained with a component system for higher resolution. Excitation wavelengths were selected by a B & L 10 cm grating blazed at 3000 A in a B & L 500 mm monochromator. The excitation illumination was focused on the sample whose emission was detected at a right angle relative to excitation. The emission mono- chromator was a Spex 1700-II which utilized a 10 cm B & L grating blazed at 5000 A. The emission spectrum was detected with a EMI 9558 QA photo- multiplier tube. The tube's high voltage was maintained by a Fluke 412B power supply which was normally operated at 1100 V. Signals from the detected emission were fed to a PAR HR-8 Lock-in Amplifier whose reference was provided by a light chopper. Finally the amplified emission signal was displayed on a strip-chart recorder. Some of the solution samples were degassed prior to their use in luminescence studies. This procedure consisted of attaching the sample tube with closed stopcock to the vacuum line, freezing the tube with liquid nitrogen, opening the stopcock to evacuate the tube above the frozen sample to a pressure of 10'.6 Torr, closing the stopcock, allowing the sample to thaw, refreezing, evacuating the tube and continuing this freeze-thaw cycling until the vacuum line ionization gauge did not quiver when the stopcock was opened after a freeze. Quantum Yields. Luminescence quantum yields for the phenyl carboxylic acids and indan derivatives were determined by comparison with toluene's fluorescence. The fluorescence quantum yield of toluene at 300°K in cyclo- hexane is 0.14 using diphenylanthracene with a quantum yield of 0.84 as l 6 a standard 7 . If two solutions absorbing the same number of photons in the same solvent emit N1 and N2 photons, respectively, the ratio of their luminescence quantum yields is equal to the ratio of the number of photons emitted, which are proportional to the areas under the fluorescence 219 bands, plotted on an energy scale: ¢§/¢1 = 112/N1 To obtain the same number of exciting photons absorbed by the two solutions, we adjust the concentrations so that they have the same optical density at the exciting wavelength. The wavelength of the excitation light was -———}B2u absorption band. No corrections for 13 the Aminco-Keirs response are necessary, since the emission spectra are taken at 260 nm, in the ;A in the same wavelength range. If quantum yields are to be compared in different solvents, correction must be made for refraction in the solvents. In a solvent with a refractive index n, there will be a broadening of the angle of aperture of the excited light beam by n2. Hence if n1 and n2 are the indices of refraction of the solvents corresponding to the two measurements, the ratio of the quantum yields Will be 2 cl’2/"’1 = Nz/N1(“2/“1) Quantum yields measurements at 77°anere performed by comparing the areas of the emission curves AF(77°K)/AF(room temperature) and correcting for the contraction and the change of the refractive index of the solvent as the result of cooling. The refractive indices at 77°waere deduced from those at room temperature by consideration of the specific refractivity (n-l)/d which is temperature invariant (where d is the density). Quantumyields for indole derivatives in various solvents at room temperature were obtained by utilizing a double beam quantum yield instru- ment interfaced with a PDP ll computer332. This instrument measures both emission and absorption and corrects for both monochromator and phototube response. It is programmed to correct for inner-filter effects for solutions with optical densities up to l. 220 The quantum yield standard used was 10-6M quinine sulfate in 1N H2304 for which the quantum yield is accepted as 0.54. Temperature Variation Studies. A quartz dewar with a flat quartz excita- tion window and a 1 cm square suprasil cuvette were used. The temperature of the sample was controlled by boiling liquid nitrogen using a power resistor and allowing the N2 gas to flow into the sample dewar. The temperature of the sample was monitored through a thermocouple (copper, constantan) attached to the outside of the cuvette immediately above the point of excitation. Comparing readings of thermocouple on the outside and inside of a cuvette containing solvent, one finds not more than 1° difference between thermocouples over wide ranges of temperature. Fluorescence Decay and Time-Resolved Spectral Studies Sipgle Photon limp Correlation Spectrometer The essential parts of the apparatus are a nanosecond flash lamp, special photomultiplier where individual photons are detected. A time- to amplitude converter (TAC) where the fluorescence photons are timed re- lative to the flash lamp, and a multichannel analyser (MCA) where the data are stored. A block diagram of the instrument is shown in Figure 51. The nanosecond lamp is essentially a relaxation oscillator where the voltage across two electrodes is increasing through charging a capacitor till the breakdown voltage of the gas filling the lamp (H2 was used) is reached where upon the lamp discharges and the cycle is repeated. Such a free-running lamp was used for the 1,3DNP experiment (ORTEC 9352). The width at half maximum (whm) was«’4 ns. In the proton transfer study a gated lamp was used (Photochemical Assodiates, University of western Ontario) which.employs a Thyratron which acts as an ultrafast switch and allows the lamp to discharge "’28 DISC L.___.____I _ [AMP H.V. CPU SAMPLE 56 DUVP BASE I. I r E 1' IM E ET A PPARATUS ‘— 7;: l (3‘1: 35.. I on” START STOP L111] M c; TEL! L coupua MOWER Block Diagram for Time Resolved Spectrophotometer. 222 when the thyratron grid is triggered with a pulse. A typical pulsing rate is 50 KHz. The whm was~I2 ns and the lamp intensity is~107-108 photons/ flash. The lamp flash is detected with the IP28 phototube and the resulting analog pulse, after discrimination against low-level noise (ORTEC Model 436) is used to start the TAC. The IP28 anode pulse times the occurrence of the lamp flash in less than 100 picoseconds. Fluorescence photons are detected by a fast photomultiplier 56DUVP. (Time resolution 800 picoseconds). This tube is capable of amplifying a single photon into an electrical pulse of several volts. This pulse is used to stop the TAC. The time resolution of the apparatus depends upon the uncertainty of timing the detection of single-photon pulses relative to the flash- lamp pulse. The principal time jitter arises from the fact that the amplitude of single-photon pulses is not constant but covers a broad range in pulse heights. The jitter associated with timing by single- level crossing was eliminated by using a "constant fraction timing" discriminator (ORTEC Model 463). In this type of instrument the input pulse is split into two channels; one channel is inverted and delayed and the other channel is simply attenuated. The channels are added such that the resultant bipolar pulse has the desired fractional trigger thresh- old associated with its zero crossing point. This effectively eliminates the time jitter. The heart of the instrument is the TAC (ORTEC Model 457), the opera- tion of which is illustrated in Figure 52. Each time the lamp flashes, a Synchronization pulse is sent to the TAC and initiates its time sweep (start). If a stop pulse is received from the photomultiplier during the time sweep, a TAC output pulse is generated with amplitude proportional to OUTPUT PULSE AM PTITUDE—-—> 223 \ __._.__._.______.____A \ Output pulse \ °P Reset Figure 52. W" Start \ \ \ If: F W{{: TIME—9 The Basic Idea of Time-To-Amplitude Conversion. Pulse Amplitude is Proportional to t8 - t The Output t0p start’ 224 the time (t t stop' start)’ The calibration of the TAC is performed with the ORTEC 462 Time Calibrator. The time base is a precision lOO-MHz crystal-controlled oscil- lator that is calibrated against WWV, the National Bureau of Standards frequency, and is temperature-compensated for accuracy throughout the normal operating range of 0 to 50°C. The TAC output is fed to the NBA. The MBA (NUCLEAR DATA Model 1100) takes this voltage V and converts it to an appropriate channel number C. The final result is a count stored in channel number C which is a record of the time at which the photon was observed by the photomultiplier. The analyzer channels now represent increments in time, and the counts in each channel are proportional to the probability for fluorescence emission from the sample between t and t +'At, where t is measured from an arbitrary but fixed point each flash. Note that each pulse of the lamp can produce only one count since the TAC only seeks one photon after it is started. By collecting many of these single photon events we eventually build in the analyzer the decay curve of the emitting species. The contents of the memory of MBA can be displayed on an oscilloscope (HEWLETT-PACKARD Model 130 BR) on an XY plotter or can be fed to a tele- type equipped with a tape punch. Through use of an accustic coupler, the teletype is connected with the CDC 6500 computer where the data are processed. The processed data are plotted through a digital plotter (HEWLETT-PACKARD 7200 A). The lamp pulsing rate and the photon counting rates are monitored through two digital counters (MONSANTO Mbdel 150A). For this method to produce true decay curves, the rate of detecting photons must be less than 10% of the lamp-repetition rate. Faster collection 225 of fluorescence photons results in what is called "pulse pile-up",'biasing of the decay curve in favor of early events. Typically the lamp repetition rate was 50 KHz and the count rate 100 counts/sec. The count collection was continued until typically 104-105 counts were collected in the channel corresponding to the maximum in the curve. The standard deviation of each point is given by Poisson statistics; i.e.,(r= (N)!5 where N is the number of counts in the channel in question. It can be shown by Laplace transform techniques that the instrument output is given by the convolution integral obsd t t) = G t')F(t-t')dt' IF< I0< or for digital data with time divided into channels bd t. I: 8 (ti) = 101 G(t')F(ti-t')dt' where G(t) is the time response of the fluorescence system when excited with the delta pulse of light, and F(t) is the time dependence of the lamp distorted by the detection system. To obtain the desired function G(t), one must solve this integral equatidn a process called deconvolution. The various deconvolution techniques have been reviewed by ware333. 62 we have used the deconvolution procedure suggested by ware et al. . According to this procedure G(t) is expressed as n G(t) =figlakexp(rt/7k) where yk have fixed values selected to span certain range. least-squares techniques are used to obtain the best weighting factors for each exponen- tial in order to minimize the sum of the squares of the deviations between the Observed decay curve and that calculated from the lamp curve and this empirical function: calcd ti ' ' (ti) = so .E.akexp(-ti/)k)F(ti-t )dt 226 th th The residual pi for the i data point (i channel) is given by pa = Igb5d(ti) _ IEaICd(ti) The deconvolution problem then reduces to obtaining the coefficients ak such that Zipiapi/Oaj = 0 j = 1,2,...,n TigepResglved Spectga. Time resolved spectra were obtained by measuring the emission spectrum through a time window. In practice, the instrument is first operated in the single-photon lifetime mode, and the upper and lower discriminator levels of the multichannel pulse-height analyser are adjusted in order to obtain the desired time window, that is only pulses between E and E +-AE are allowed to pass and so events occurring only between t and t +-At after excitation are recorded. This can be seen visually on the oscilloscope during the collection of the decay curve, since the upper and lower level discriminators will block out the leading and trailing edges of the decay curve and can be used to create a window with any time size. The instrument is then switched to multichannel scalling and a spectrum is obtained through the time window created by the analyser discriminators. BIBLIOGRAPHY BIBLIOGRAPHY l. T. H. Lyman, Astrophs. J., g9, I (1924). 2. L. A. Sommer, Proc. Natl. Acad. Sci. (U. S.), 13, 213 (1927). 3. Lord Rayleigh, Proc. Roy. Soc., A125,l (1929). 4. S. Mrozowski, Z. Physik, 106, 458 (1937). 5. Th. Forster, and K. Kasper, Z. Physik. Chem. (Frankfurt),_l, 275 (1954). 6. J. B. Birks, Acta Phys. Polon., 2Q, 367 (1964). 7. H. Leonhardt and A. weller, Ber. Bunsenges. Phys. Chem., 61, 791 (1963). 8. 0. Cheshnovsky, A. Gedanken, B. Raz, and J. Jortner, Chem. Phys. Lett., 22, 23 (1973). 9. B. Stevens, Spectrochim.Acta, 18, 439 (1962). 10. J. Tanaka, Bull. Chem. Soc. Japan,.3§, 1237 (1963). 11. P. F. Jones and M; Nicol, J. Chem. Phys.,,4§, 5440 (1968). 12. M. T. Vala, J. Haebig, and S. A. Rice, J. Chem. Phys., 43, 886 (1965). 13. Th. Forster, Angew. Chem. Intern. Ed., 8, 333 (1969). 14. J. B. Birks, Progress in Reaction Kinetics, 2, 181 (1970). 15. B. Stevens, Advances in Photochemistry, 8, 161 (1971). 16. A. Dalgarno and J. T. Lewis, Proc. Phys. Soc., (London) A62, 57 (1956). 17. F. London, Trans. Faraday Soc., 33, 8 (1937). 18. B. Pullman, P. Claverie, and J. Caillet, Proc. Natl. Acad. Sci. (U.S.), ,§§, 904 (1966). 19. R. J. W. LeFevre and K. M. S. Sundaram, J. Chem. Soc., 4442 (1963). 20. J. Frenkel, Phys. Rev., 31, 17 (1931); ibid, 31, 1276 (1931). 21. A. S. Davydov, J. Exptl. Theor. Phys. (U. S. S. R.), 18, 210 (1948). 227 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. '39. 40. 41. 42. 430 44. 228 A. S, Davydov, Theory of Molecular Excitons, Translation by M. Kasha and K. Oppenheimer, McGraw-Hill, N. Y., 1962. M. R. Philpott, Adv. in Chem. Phys., I. Prigogine and S. A. Rice editors, Jehn Wiley, 1973, p. 227. H. Margenau and N. R. Kestner, Theory of Intermolecular Forces, Second Edition, Pergamon Press 1969, p. 318. A. w. Overhauser, Phys. Rev., 101, 1702 (1956). P. Avakian and R. E. Merrifield, M01. Crystals, 5, 37 (1968). R. E. Merrifield, Ace. of Chem. Research, 1, 129 (1968). R. S. Mnlliken, J. Amer. Chem. Soc., 14, 811 (1952). s. P. McGlynn, Chem. Rev., _5_§, 1113 (1958). G. Briegleb, E1ektronen-Donator-Acceptor Kbmplexe Springer-Verlag, Berlin-Gottingen-Heidelberg, 1961. J. N. Murrel, Quart. Rev., 15, 191 (1961). R. S. Mulliken and w. B, Person, Ann. Rev. Phys. Chem., 13, 107 (1962). R. S. Becker and E. Chen, J. Chem. Phys., 35, 2403 (1966). R. S. Mulliken and w, B, Person, Molecular Complexes: A Lecture and Reprint Volume. Wiley N. Y., 1969. Th. Forster, Pure Appl. Chem., 4, 121 (1962). M. A, Slifkin, Nature,_2_9_(_), 766 (1963). R. M; Hedges and F. A. Matsen, J; Chem. Phys., 28, 950 (1958). J. R. Hoyland and L. Goodman, J. Chem. Phys., 36, 12, (1962). ibid,‘§§, 21 (1962). T. Azumi and S. P. MbGlynn, J. Chem. Phys., 41, 3131 (1964). E. Doller and Th. Forster, Z. Physik. Chem. N. F., 31, 274 (1962). E. Konijnenber, Thesis, University of Amsterdam, 1963. J. N. Murrel and J. Tanaka, Mbl. Phys., 1, 363 (1964). T. Azumi, A. '1'. Amstrong, and S. P. McGlynn, J. Chem. Phys., 3;, 3839 (1964). R. J. Buenker and S. D. Peyerimhoff, Chem. Phys. Lett., 3, 37 (1969). W. von Niessen, J. Chem. Phys., 22, 1948 (1971). fi—fi W. von Niessen, Theor. Chim. Acta (Berlin), 31, 111 (1973). ibid.,~3g, 13 (1973). 229 45. T. Azumi and H..Azumi, Bull. Chem. Soc. Japanxgg, 279 (1967). 46. A. K. Chandra and E. C. Lim, J. Chem. Phys., 48, 2589 (1968). 47. F. Hirayama, "Energy Transfer and Quenching in Plastic Scintillators", Thesis, University of Michigan. 48. S. S. Yanari, F. A. Bovey, and R. Lumry, Nature, 200, 242 (1963). 49. F. Hirayama, J. Chem. Phys., 4;, 3163 (1965). 50. D. H. Phillips and J. C. Schug, J. Chem. Phys., 59, 3297 (1969). 51. J. Eisinger, M. Gueron, R. G. Shulman and T. Yamane, Proc. Not. Acad. Sci., 5;, 1015 (1966). 52. N. J. Leonard, J. Iwamura, and J. Eisinger, Proc. Nat. Acad, Sci., fig, 352 (1969). 53. B. Stevens, Advances in Photochemistry, Vol. 8, 1971, pp. 216-217. 54. J. Ferguson, J. Chem. Phys., pg, 765 (1958). 55. J. B. Birks, Nature, 214, 1187 (1967). 56. E. A. Chandross, and C. J. Dempster, J} Am. Chem. Soc., 22, 3586 (1970). 57. H. J. Pownall and L. C. Smith, J. Am. Chem. Soc.,-25, 3136 (1973). 58. H. J. Calls and E. Sackmann, Biochim. Biophys. Acta, 339, 103 (1974). 59. J. B. Birks, T. A. King and I. H. Munro, Proc. Phys. Soc., 89, 355 (1962). 60. I. Isenberg and R. D. Dyson, Biophys. J., 2, 1337 (1969). 61. R. Schuyler and I. Isenberg, Rev. Sci. Instrum.x42, 813 (1971). 62. W. Ware, L. J. Doemeny and T. L. Nemzek, J. Phys. Chem.,.ZZ, 2083 (1973). 63. J. B. Birks, Photophysics of Aromatic Molecules, Wiley-Interscience, 1970, p. 70. 64. P. Claverie, in "Molecular Associations in Biology" Pullman editor, Academic Press, 1968. 65. A. I. R. Rae and R. Mason, Proc. Roy. Soc., A304, 487 (1968). 66. M. T. Vala, J. H. Hillier, S. A, Rice and J. Jertner, J; Chem. Phys., 3g, 23 (1966). 67. A. I. Kitaygorodsky, Tetrahedron 14, 230 (1961). 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 230 G. Favini and M. Simonetta, Theor. Chim. Acta,l, 294 (1963). M. J. Mantione, Theor. Chim. Acta, _1_5_, 141 (1969). 0. K. Rice and H. C. Ramsperger, J. Amer. Chem. Soc., _l_+__9_, 1617 (1927)° L. S. Kassel, J. Phys. Chem., 32, 225 (1928). ibid, 32, 1065 (1928). S. w. Benson, in "Foundations of Chemical Kinetics" McGraw-Hill 1960, pp. 245-246. F. W. Schuler and G. W. Murphy, J. Amer. Chem. Soc., 12, 3155 (1950). J. B. Aladekomo and J. B. Birks, Proc. Roy. Soc., A284, 551 (1965). B. K. Selinger, Austral. J. Chem., 12, 825 (1969). J. B. Birks, M. D. Lumb and I. H. Munro, Proc. Roy. Soc., A289, 289 (1964). R. B. Cundall and D. A. Robinson, Chem. Phys. Lett., 1.3, 257 (1972). R. L. Barnes and J. B. Birks, Proc. Roy. Soc. A291, 570 (1966). E. Doller and Th. Forster, Z. Phys. Chem. (N.F.‘) _4_8_, 58 (1965). B. Stevens and M. I. Ban, Trans. Faraday Soc., _6_(_)_, 1515 (1964). Th. Forster and H. P. Seidel, Z. Phys. Chem. (N. F. ), £58, 58 (1965). E. A. Moelwyn-Hughes, J. Chem. Soc., 850 (1940). J. B. Birks, M. D. Lumb andl. H. Munro, Acta Phys. Polon., g_6_, 379 (1964). C. Lewis and w. R. Ware, Mol. Photochem., _§, 261 (1973). F. Hirayama and s. Lipsky, J. Chem. Phys., 51, 1939 (1969). G. J. Hoijtink, Z. Elektrochem., _6_4, 156 (1960). J. B. Birks, M. D. Lumb and I. H. Munro, Proc. Roy. Soc., (London), A280, 289 (1964). R. L.Bsrnes and J. B. Birks, Proc. Roy. Soc. (London), A291, 570 (1966). T. Azumi and H. Azumi, Bull. Chem. Soc. Japan, 32, 1829 (1966) ibid, p2, 2317 (1966). N. Mataga, M. Tomura, and H. Nishimura, M01. Phys., _9_, 367 (1965). R. B. Cundall, L. C. Pereira and D. A. Robinson, Chem. Phys. Lett., _1_3_, 253 (1972). E. A. Chandross. J. Chem. Phys., 43, 4175 (1965). 231 93. J. R. Greenleaf, M. D. Lumb and J. B. Birks, J. Phys. B,.1, 1157 (1968). 94. M. Mataga, Y. Torihashi and Y. Ota, Chem. Phys. Lett., 1 .....S 385 (1967). 95. J. B. Birks, D. J. Dyson, and I. H. Munro, Proc. Roy. Soc., A275, 575 (1963). 96-. N. Mataga, Chem. Phys. Lett., _1_, 385 (1967). 97. W. H- Melhuish and W. S. Metcalf, J. Chem. Soc., 480 (1958). 98. P. Debye, Trans. Electrochem. Soc., 82, 205 (1942). 99. R. M. Noyes, Prog. Reaction Kinetics, 1, 131 (1961). 100. E. Fisher, Z. Elektrochem. 23, 16 (1949). 101. A. Gieser, A. Spernal and K. Wirtz, Z. Naturforsch. 8A, 522 (1953) ibid, pg, 532 (1953). 102. K. Kawaoka, A. U. Khan and D. R. Kerns, J. Chem. Phys., 46, 1842 (1967). 103. F. W. RObinson and R. P.Frosch, J. Chem. Phys., 31, 1962 (1962). 104. G. J. Heytink, Accounts Chem. Res., 2, 114 (1969). 105. R. w. Nicholls, P. A. Fraser, W. R. Jermain and R. P. McEachran, Astrophys. J.,‘13, 399 (1960). 106. M. D. Lumb and D. A. weyl, J. Mol. Spectroscy., 23, 365 (1967). 107. J. B. Aladekomo and J. B. Birks, Proc. Roy. Soc.,A284, 551 (1965). 108. J. Langelaar, R. P. H. Rettschnick, A. M. F. Lambooy and F. J. Hoytink, Chem. Phys. Lett., 1, 609 (1968). 109. J. Langelaar, Thesis, University of Amsterdam (1969). 110. R. D. Nelson, D. R. Lide, A. A. Maryott, in "Selected Values of Electric Dipole Mements for Molecules in the Gas Phase", National Bureau of Standards (NSRDS-NBS 10). 111. Handbook of Chemistry and Physics, The Chemical Rubber Company. 112. M. Itoh, T. Mimura, H. Usui, and T. Okamoto, J. Am. Chem. Soc., 25, 4388, (1973). 113. J. Kirkwood, J. Chem. Phys., _2_, 351 (1934). 114. R. G. Pearson. J. Chem. Phys., 29, 1478 (1952). 115. J. Timmermans, in "Physico-Chemical Constants of Pure Organic Compounds", Elsevier Publishing Company 1950. 116. M. B. Ledger and P. Suppan, Spectrochim. Acta, 23A, 3007 (1967). 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 232 J. B. Birks, in ”Photophysics of Aromatic Molecules" pp. 439-441. Mbhyi-Eldin M. Abu-Zeid, Prodeeding of the 6th International Sym- posium on Atomic, Melecular and Solid-State Theory and Quantum Biology, Interscience Publishers 1972. J. Beens, H. Knibbe, and A. weller, J. Chem. Phys., 91, 1183 (1967). D. J. Witt and C. R. Estee, Proc. S. Dakota Acad. Sic., 11, 229 (1952). . Hanack, in "Conformation Theory”, Academic Press 1965. . B. Person and G. C. Pimentel, J. Am. Chem. Soc., 19, 5321 (1953). K. S. Pitzer, J. Chem. Phys., 9, 711 (1940). 0. S. Khalil, R. H. Hofeldt and S. P. MbGlynn, J. Luminescence, 9, 229 (1973). G. Castro and R. M. Hochstrasser, J. Chem. Phys., 42, 4352 (1966). E. C. Lim and S. K. Chakrabarti, M01. Phys., 11, 293 (1962). I. H. Hillier, L. Glass and S. A. Rice, J. Chem. Phys., 9;, 3015 (1966). L. G. Christophorou, M. E. M. Abu-Zeid and J. G. Carter, J. Chem. Phys., 49, 3775 (1968). R. B. Cundall and A. J. R. Voss, Chem. Commun., 116 (1969). J. Langelaar, G. Jansen, R. P. H. Rettschnick and G. J. Hoytink, Chem. Phys. Lett., 11, 86 (1971). B. A. Baldwin, J. Chem. Phys., 99, 1039 (1969). W. Siebrand, J. Chem. Phys., 29, 1040 (1969). Chandross, J; Am. Chem. Soc., 93, 704 (1970). R. M. Hochstrasser, Rad. Research, 19, 107 (1963). R. M. Hochstrasser and T. F. Hunter, J. Chem. Phys., 39, 2739 (1964). P. Avakian and R. E. Merrifield, Phys. Rev. Lett., 19, 541 (1964). V. Ern, P. Avakian and R. E. Merrifield, Phys. Rev., 1&9, 862 (1966). R. E. Merrifield, Ace. of Chem. Res., 1, 129 (1968). M. Kasha, Discuss. Faraday Soc., 9, 14 (1950). H. Sponer and J. H. Rush, J. Chem. Phys., 19, 1847 (1952). L. E. Orgel, J. Chem. Soc., 121 (1955). 233 142. L. Salem, in “Molecular Orbital Theory of Conjugated Systems" w. A. ‘ Benjamin, Inc., N. Y. 1966 p. 462. 143. M. A. El-Sayed, J. Chem. Phys., 99, 552 (1962). 144. M. Kasha, in "Light and Life" W. D. McElroy and B. Glass, eds. The Johns Hopkins Press, Baltimore, 1961 p. 31. 145. H. P. Stephenson, J. Chem. Phys., 11, 1077 (1954). 146. E. Clementi, J. Chem. Phys., $9, 4731 (1967). 147. R. D. Brown and M. L. Heffernan, Austral. J. Chem._11, 554 (1959). 148. D. S. McClure, Electronic Spectra of Molecules and Ions in Crystals, Solid State Physics, Academic Press, Vol. 8-9 1959. 149. D. J. Cohen and L. Goodman, J. Chem. Phys., 49, 713 (1967). 150. R. B. Cundall, F. J. Fletcher and D. G. Milne, Trans. Faraday Soc., .99, 1146 (1964). 151. D. F. Evans, J. Chem. Soc., 3885 (1957). 152. Th. Forster, Fluoreszenz Organisher Verbindungen, Vandenhoeck and Ruprecht, Gottingen, 1951, p. 105. 153. E. M. Kosower and P. E. Klinedinst Jr., J. Am. Chem. Soc., 19, 3493 (1956). 154. G. Briegleb, J. Trencseni and w. Herre, Chem. Phys. Lett., 3 .....S 146 (1969). 155. R. M. Hechstrasser and J. W. Michaluk, J. Chem. Phys., 9;, 668 (1971). 156. R. D. Nelson Jr., D. R. Lide Jr., and A. A. Maryott, Natl. Std. Ref. Data Ser., Natl. Bur. Std. (US) 19, 29 (1967). 157. P. Cremaschi, A. Gamba, and M. Simonetta, Theor. Chim..Acta (Berlin) .11, 155 (1973). 158. S. Mataga and N. Mataga, Z. Physik. Chem. N. F-.129 231 (1959). 159. J. Del Bene and H. H. Jaffe, J. Chem. Phys., 49, 1221 (1968). 160. V. S. Krishna and L. Goodman, J..Am. Chem. Soc., 9;, 2042 (1961). 161. V. Zanker, Z. Physik. Chem. (Frankfurt) 1, 52 (1954). 162. N. GinSburg, W. W. Robertson, and F. A. Matsen, J. Chem. Phys., 13, 511 (1946). 163. E. M. Kosower, D. Hofmann, and K. Wallenfels, J. Am. Chem. Soc., 94, 2755 (1962). 164. 3. Shifrin, Biochim. Biophys. Acta, 96, 173 (1965). 165. J. w. Verhoeven, I. P. Dirkx and Th. J. de Boer, Tetrahedron Lett., 4399 (1966). 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. '188. 189. 234 A. Pople, D. L. Beveridge, and P. A. Dobosh, J. Chem. Phys.,.gz, 2026 (1967). Clementi, J. Chem. Phys.,,flé, 4731 (1967). .Adam, A. Grimison, R. Hbffimann, and C. Z. de Ortiz, J. Am. Chem. Soc., 29, 1509 (1968). Saito and K. Nukada, Tetrahedron Lett., 111 (1965). Chandross, J. Chem. Phys., 45, 397 (1966). 0. Loutfy and R, W. Yip, Can. J. Chem.,‘gl, 1881 (1973). Hirayama, J. Chem. Phys., 3g, 3163 (1965). B. Birks, in "Photophysics of Aromatic Molecules", Wiley Interscience 170, p. 424. N. Nezmaiko, I. E. ObYknovennaya and A. S. Cherkasov, Opt. Spectry., .21, 23 (1966). ibid, 2.__1_, 285 (1966). G. Wéber and J. R. Lakowicz, Chem. Phys. lett._22, 419 (1973). H. watson and M. Ashraf El-Bayoumi, J. Chem. Phys., 22, 5464 (1971). M. Rosenberg and E. C. Eimutis, J. Phys. Chem., 19, 3494 (1966). leonhardt and A. Wéller, Ber. Bunseges. Physik. Chem. 61, 791 (1963). B. Birks, in "Photophysics of Aromatic Mblecules" pp. 72 and 460. weber and M. Shinitzky, Proc. Natl. Acad. Sci., U.S._§§, 823 (1970). Itoh and T. Azumi, Chem. Phys. lett.,.22, 395 (1973). S. Bayliss and E. G. MbRae, J. Phys. Chem. 58, 1002 (1954). Reid, J. Chem. Phys.,.gg, 1212 (1952). 1M.'Moodie and C. Reid, J. Chem. Phys.,‘gg, 1510 (1952). Iwata, J. Tanaka and S. Nagakura, J. Chem. Phys., 41, 2203 (1967). Christodouleas and S. P. MbGlynn, J. Chem. Phys.,.ég, 166 (1964). P. MhGlynn and J. D. Boggus, J. Am. Chem. Soc.,.§Q, 5096 (1958). P. MbGlynn, J. D. Boggus and E. Elder, J. Chem. Phys., 32, 357 (1966). Badger, B. Brocklehurst and R. D. Russel, Chem. Phys. lett.,‘1, 122 (1967). 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 2010 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. J. L. J. U.S., 63, 253 (1969). 235 Badger and B. Brochlehurst, Trans. Faraday Soc., 65, 2576 (1969). Badger and B. Brocklehurst, Trans. Faraday Soc., 6;, 2582 (1969). Edlund, P. 0. Kinell, A. Land and A. Shimizu, J. Chem. Phys., 46, 3679 (1969)., W. McLafferty, Mass Spectrometry of Organic Ions, Academic Press, New York, 1963. R. Goldschmidt, R. Potashnik, and'M. Ottolenghi, J. Phys. Chem., 15, 1025 (1971). J. Land, J. T. Richards, and J. K. Thomas, J. Phys. Chem. 16, 1025 (1971). Orbach, R. Potashnik, and M. Ottolenghi, J. Phys. Chem., 16, 1133 (1972). P. McGlynn, T. Azumi, and M. Kinoshita, McIecular Spectroscopy of the Triplet State, Prentice-Hall Englewood Cliffs, N. J., 1969, Pp. 284-325. White, Biochem. J. 11, 217 (1959). W. Cowgill, Arch. Biochem. Biophys., 199, 36 (1963). Feitelson, J. Phys. Chem.,.§§, 391 (1964). Bishai, E. Kuntz, and L. Augenstein, Biochim. Biophys. Acta, 140, 381 (1967). Tournon and M. A. El-Bayoumi, J. Am.Chem. Soc., 23, 6396 (1971). J. Mflttal, J. P. Mittal and E. Hayon, J..Am. Chem. Soc., 25, 6203 (1973). Tournon and'M. A. El-Bayoumi, J. Chem. Phys., 26, 5128 (1972). .A. Chandross and A. T. Thomas, Chem. Phys. Lett., 9, 393 (1971). Tournon, E. Knntz, and M. A. El-Bayoumi, Photochem. Photobiol. 1g, 425 (1972). watanabe, J. Chem. Phys., 26, S42 (1957). .A. Taylor, M. A. El-Bayoumi, and M. Kasha, Proc. Nat. Acad. Sci. Briegleb, Z. Physik. Chem., (leipzig) B51, 9 (1941). S. Rowlinson, Trans. Faraday Soc., A], 120 (1951). Lennard-Jones, Proc. Roy. Soc., A14, 198 (1949). Lennard-Jones and J. A. Pople, Proc. Roy. Soc., A205, 155 (1951). 236 213. C. A. Coulson, Trans. Faraday Soc., §§, 433 (1942). 214. C. A. Coulson, Proc. Roy. Soc., A291, 63 (1951). 215. C. A. Coulson, Research (London).19, 149 (1957). 216. C. A. Coulson and V, Danielsen, Arkiv. Fysik. 8, 205 (1955). 217. N. D. Sokolov, Zh. Eksperim, i Teor. Fiz. 23, 315 (1952). 218. H. Tsubomura, Bull. Chem. Soc., Japan, 21, 445 (1954). 219. S. Bratoz, Symp. Forces Intermoleculaires, Bordeaux, 1965. 220. G. Bessis and S. Bratoz, J. Chim. Phys., 51, 769 (1960). 221. P. A. Kollman and L. C. Allen, J. Am. Chem. Soc., 22, 6101 (1970). 222. P. A. Kollman and L. C. Allen, Chem. Rev., 12, 283 (1972). 223. J. E. Del Bene, J. Chem. Phys., 58, 3139 (1973). 224. E. Clementi, J. Mahl, and W. von Niesen, J. Chem. Phys., 54, 508 (1971). 225. P. 0. Lowdin, Rev. Mod. Phys., 35, 724 (1963). 226. C. L. Bell and G. M. Barrow, J. Chem. Phys., 31, 1158 (1959). 227. G. M. Barrow, Spectrochim. Acta 16, 799 (1960). 228. Krueger, Can. J. Chem., 42, 201 (1964). 229. G. E. Bacon, in D. Hadzi (editor), Hydrogen Bonding, Pergamon London, 1959 p. 23. 230. S. W. Peterson and H. A. Leby, Acta Cryst., 19, 70 (1957). 231. A. Piekara, J. Chem. Phys., éé, 2145 (1962). 232. A. R. Ubbelohde and K. J. Gallagher, Acta Cryst.,.§, 71 (1955). 233. Y. Imry, I. Pelah and E. Wiener, J. Chem. Phys., 43, 2332 (1965). 234. S. Ganesan, Nature, 185, 757 (1960). 235. A. Grimison, J. Phys. Chem.,.él, 962 (1963). 236. C. C. Costain and G. P. Srivastava, J. Chem. Phys., 41, 1620 (1964). 237. R. Blinc and M. Pintar, J. Chem. Phys., 35, 1140 (1961). 238. J. L. Wbod, J. M01. Structure, 13, 141 (1972). 239. R. Blinc, D. Hadzi and A. Novak, Z. Elektrochem.,léfi, 567 (1960). 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 237 R. Blinc and D. Hadzi, Spectrochim. Acta,.1§, 852 (1960). D. Hadzi, Pure and Appl. Chem..11, 435 (1965). 2M. Kasha, Disc. Faraday Soc.,,g, 14 (1950). G. C. Pimentel, J. Am. Chem. Soc., 12, 3323 (1957). G. M. Barrow, J. Am. Chem. Soc., _7__8_, 5802 (1956). H. Baba, A. Matsuyama, and H. Kokubum, J. Chem. Phys., 41, 895 (1964). K. Wéber, Z. Physik. Chem. (Leipzig) 31;, 18 (1931). Th. Forster, Z. Elektrochem., 54, 42 (1950). N. Mataga and Y. Kaifu, M01. Phys., 1, 137 (1963). A. Weller, Naturwiss.,.42, 175 (1955); Z. Elektrochem., 69, 1144 (1956). K. z. Hirota, Physik. Chem. (Frankfurt) §_5_, 222 (1962). Y. V. Noboikin, B. A. Zadorozhnyi, and E. N. Pavlova, Opt. Spectry. .2, 347 (1960). Y. V. Noboikin, B. A. Zadorozhnyi, and E. B. Pavlova, Opt. Spectry. g, 312 (1959); ibid,‘§, 231 (1959). A. Wéller, Z. Elektrochem., 56, 662 (1952). A. weller, Z. Physik. Chem., 18, 163 (1958). A. Wéller, in "Progress in Reaction Kinetics", Vol. 1, Pergamon Press, London, 1961, Chap. 7. G. Jackson and G. Porter, Proc. Roy. Soc., (London), A260, 13 (1961). M. R. Loken, J. W. Hayes, J. R. Gohlke, and L. Brand, Biochem., 11 4779 (1972). , W. R. ware, P. R. Shukla, P. J. Sullivan, and R. V, Bremphis, J. Chem. Phys.,.ég, 4048 (1971). F. Hund, z. Physik, 3;, 805 (1927). E. P. Wigner, Z. Phys. Chem. (Leipzig), £12, 203 (1932). D. M. Dennison and G. E. Uhlenbeck, Phys. Rev., 41 313 (1932). E. Grunwald, in "Progress in Physical Organic Chemistry". Cohen, Streitwieser, and Taft, eds. Wiley, N. Y. 1965 Vol 3. 238 264. H. S. Johnston, Adv. Chem. Phys., 3, 131 (1961). 265. E. F. Caldin, Chem. Rev., Q_9_, 135 (1969). 266. C. Haas and D. F. Horning, J. Chem. Phys., 32, 1763 (1960). 267. n. Hadzi, J. Chem. Phys., 34, 1445 (1961). 268. P. O. Lowdin, Adv. Quantum Chem. _2_, 213 (1965). 269. J. Bigeleisen and M. Wolfsberg, Adv. Chem. Phys., _1, 15 (1958). 270. J. Bigeleisen, J. Chem. Phys., 11, 675 (1949). 271. J. Bigeleisen and M. Goeppert-Mayer, J. Chem. Phys., 15, 261 (1947). 272. L. Melander, Arkiv. Kemi, 2, 211 (1950). 273. F. H. Westheimer, Chem. Rev., §_1_, 265 (1961). 274. P. M. Laughton and R. E. Robertson, Can. J. Chem. _3_4_, 1714 (1956). 275. R. E. Robertson, Can. J. Chem. 32, 613 (1957). 276. R. E. Robertson and P. M. Laughton, Can J. Chem., 32, 1319 (1957). 277. F. D, Rossini, J. W. Knowlton and H. L. Johnston, J. Res. Natl. Bur. Stand., _2_4, 369 (1940). 278. W. M. Jones, J. Chem. Phys., _4_8, 207 (1968). 279. C. G. Swain and R. F. W. Bader, Tetrahedron _1_0_, 182 (1960). 280. C. G. Swain, A. D. Ketley and R. F. W. Bader, J. Am. Chem. Soc., _2_1_8, 2353 (1959). 281. R. B. Henry, M. Kasha, Ann. Rev. Phys. Chem., _1_9_, 161 (1968). 282. G. W. Robinson and R. P. Frosch, J. Chem. Phys., 31, 1962 (1962); ibid, §_8_, 1187 (1963). 283. J. Jortner, R. 8. Berry and R. M. Hochstrasser, Adv. Photochem., _Z, 149 (1969). 284. W. Siebrand and D. F. Williams, J. Chem. Phys., 46, 403 (1967). 285. E, C. Lim and J. D. Laposa, J. Chem. Phys., 4;, 3257 (1964). 286. J. D. Laposa, E. c. Lim and R. E. Kellogg, J. Chem. Phys., 41;, 3025 (1965). 287. R. Watts and S. J. Strickler, J. Chem.Phys., _4_2, 3867 (1968). 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 239 T. Gierke, R. J. watts and S. J. Strickler, J. Chem. Phys., 59, 5425 (1969). B. R. Henry and J. L. Charlton, J. Am. Chem. Soc., 22, 2782 (1973). N. Hirota and C. A. Hutchison, J. Chem. Phys., 46, 1561 (1966). J. L. Kropp and M.W. Winsdor, J. Chem. Phys., 4;, 1599 (1965). R. Borkowski, H. Forest and D. Grafstein, J. Chem. Phys.,_42, 2974 (1965). - J. Kropp and W. Dawson, J. Chem. Phys., 45, 2419 (1966). V. L. Ermolaev and E. B. Sveshnikova, Chem. Phys. Lett., 23, 349 (1973). Th. Forster, Ann. Physik,.g, 55 (1948). D. L. Dexter, J. Chem. Phys., 21, 836 (1953). V. L. Ermolaev and E. B. Sveshnikova, Opt. 1 Spektroskopiya 39, 379 (1971). L. Stryer, J..Am. Chem. Soc., 88, 5708 (1966). Th. Forster and K. Rokos, Chem. Phys. Lett. 1, 279 (1967). J. Eisinger and G. Navon, J. Chem. Phys. 59, 2069 (1969). J. R. Platt, J. Chem. Phys.,.lg, 101 (1951). IL B. Klevens and 3. R. Platt, J. Chem. Phys.,-11, 470 (1949). R. W.‘Wagner, Ph. D. Thesis, Michigan State University, 1971. E. Vander Donckt, Bull. Soc. Chim. Belges,.Z§, 69 (1969). G. Yagil, J. Phys. Chem., 1;, 1034 (1962). J. W. Bridges and R. T. Williams, Biochem. J..191, 225 (1968). J. Longworth, R. Rahn, and R. Shulman, J. Chem. Phys., 42, 2930 (1966). T. K. Adler and A. Albert, J. Chem. Soc., 1794 (1960). 1961. IM. A. El-Bayoumi, Ph. D. Thesis, Florida State University, K.C. Ingham, M. Abu-Elgheit and M. A. El-Bayoumi, J. Am. Chem. Soc., .23, 5023 (1971). K. C. Ingham and M. A. El-Bayoumi, J. Am. Chem. Soc., _9_6_, 1674 (1974). H. A. Benesi and J. H. Hildebrand, J. Am. Chem. Soc., 11, 2703 (1949). 313. 314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 324. 325. 326. 327. 328. 329. 330. 331. 332. ' 333. 334. 240 L. Scott, Rec. Trav. Chim., Pays-Bas, _7__5_, 787 (1956). Bowen and D. Seaman, Luminescence of Organic and Inorganic Materials, Kaleman and Spruch Editors, Wiley N. Y. 1962. V. Viktorova, Opt. Spectry. 22, 206 (1967). S. Walker, T. W. Bednar, and R. Lumry, in "Molecular Luminescence", E. C. Lim editor, Benjamin, N. Y. 1969. I. Grossweiner and H. I. Joschek, Adv. Chem. Ser., 50, 279 (1965). W. Ricci, Photochem. Photdbiol.,.lg, 67 (1970). P. Busel, T. L. Bushaeva, and E. A. Burshein, Optics and Spectry., 2._9_, 268 (1970). S. Walker, T. W. Bednar, R. Lumry and F. Humphries, Photochem. Photo- biol., _l_4_, 147 (1971). P. Kirby and R. F. Steiner, J. Phys. Chem., 14, 4480 (1970). K. Glasoe and F. A. Long, J. Am. Chem Soc., 64, 188 (1960). B. Martin, Science, 132, 1198 (1963). L. Wehry and L. B. Rogers, J. Am.Chem. Soc., 88, 351 (1966). J. Bowen, Disc. Faraday Soc.,.gz, 40 (1959). C. Lflm, J. D. Laposa, and J. M. H. Yu, J. M01. Spectry., 12, 412 (1966). Th. Forster, Chem. Phys. Lett., _1__7_, 309 (1972). S. ‘W M. G. Schulman and P. Liedke, Z. Physik. Chem., N. F..§4, 317 (1973). H. Perkin and G. Revay, J. Chem. Soc., 228 (1894). M. Robinson and B. L. Robinson, J. Am. Chem. Soc.,-11, 6554 (1955). S. Potts, J. Chem. Phys.,.gQ, 809 (1952). F. Holland, R. E. Teets, and A. Timnick, Anal. Chem., 45, 145 (1973). R. ware,in "Creation and Detection of the Excited States"'Vol. 1A, Lamola Ed. Marcel Dekka, N. Y. 1971. Melinger, and F. Wilkinson, Trans. Faraday Soc., 62, 1785 (1966). MICHIGAN STATE UNIVERS ITY LIBRARIES IIIIIIIIIII II 2 5073 308 HIHIII I III 1293 0 I Il .clauv ...._........ .71.... . x 3.2:. Kw...im..z. V.. u. ...: a? 1.57 .32.: , n 5 . as (a 3%... . hear... .... «figug T ...! if . n... . I. .«3 a as». 3 5L. 1. x .3. 155...... , . .... .... . .33.... («swung 1- .701 .—